Emerging single-atom iron catalysts for advanced catalytic systems

Abstract

Due to the elusive structure–function relationship, traditional nanocatalysts always yield limited catalytic activity and selectivity, making them practically difficult to replace natural enzymes in wide industrial and biomedical applications. Accordingly, single-atom catalysts (SACs), defined as catalysts containing atomically dispersed active sites on a support material, strikingly show the highest atomic utilization and drastically boosted catalytic performances to functionally mimic or even outperform natural enzymes. The molecular characteristics of SACs (e.g., unique metal–support interactions and precisely located metal sites), especially single-atom iron catalysts (Fe-SACs) that have a similar catalytic structure to the catalytically active center of metalloprotease, enable the accurate identification of active centers in catalytic reactions, which afford ample opportunity for unraveling the structure–function relationship of Fe-SACs. In this review, we present an overview of the recent advances of support materials for anchoring an atomic dispersion of Fe. Subsequently, we highlight the structural designability of support materials as two sides of the same coin. Moreover, the applications described herein illustrate the utility of Fe-SACs in a broad scope of industrially and biologically important reactions. Finally, we present an outlook of the major challenges and opportunities remaining for the successful combination of single Fe atoms and catalysts.

---

1. Introduction

In the production of chemicals, catalysts play an indispensable role. For example, the industrial production of ammonia requires the use of Fe catalysts to boost the reaction rates. In recent years, catalysts have undergone rapid developments from bulk to nano levels, and as the size of the catalyst decreases, their catalytic performance in terms of activity and selectivity substantially improve. In this context, what happens when the size of the catalyst reaches the atomic level is a highly fascinating topic although technologically very challenging.

In 1996, Xu and colleagues prepared a five-coordination titanium single-site heterogeneous catalyst for the catalytic epoxidation of cycloolefins.¹ At this time, single-atom catalysts (SACs) did not attract the attention of researchers. It was not until 2011 that Zhang et al.² proposed the concept of SACs and synthesized SACs with a single platinum (Pt) atom anchored on the surface of FeO_x nanocrystals for the oxidation of CO. Since then, owing to the superior catalytic activity and selectivity with utmost atomic usage efficiency, SACs have arguably become the most active frontier the heterogeneous catalysis.

SACs, which are characterized by atomically dispersed active sites, interestingly show precisely located metal centers, unique metal–support interactions and an adjustable coordination environment.^3,4 Hence, SACs inherit the advantages of homogeneous and heterogeneous catalysts.⁵ To date, many SACs based on various metals, such as Ru, Pt, Er, Ag, Au and Pd, have been developed and display excellent catalytic performances.^6–8 However, despite their potential, the current examples highlighting the further practical application of noble metal catalysts have been limited to some extent. Two of the main obstacles are related to their (1) relatively high cost and (2) relatively low abundance in the Earth.⁹ Interestingly, transition metals hold vast potential to replace precious metal catalysts. Researchers have developed non-precious metal-based SACs (Fe, Co, Ni, Cu, Mo, etc.), resulting in a booming of the library of SACs.^10–12 Among the non-precious metals, Fe is abundant in the Earth and displays excellent catalytic properties with good stability and recyclability. Besides, the facile scalable route for the synthesis of Fe-SACs ensures their large-scale application in industrial production.^13–16 Today, Fe-SACs are extensively used in a wide range of research fields (Fig. 1). For example, in the energy storage and conversion fields, Fe-SACs have impressive electrochemical catalytic properties in fuel cells and metal–air cells, which can highly reduce the energy barriers of reaction pathways including but not limited to the nitrogen reduction reaction (NRR), hydrogen evolution reaction (HER), oxygen evolution reaction (OER), oxygen reduction reaction (ORR) and carbon dioxide reduction reaction (CO₂RR). Importantly, Fe-SACs are demonstrated to have natural enzyme-like catalytic activities, which can functionally mimic or even outperform some metalloproteases, and hence they have been recently explored in the biological catalysis field for biosensing, wound disinfection and cancer treatment. In this review, initially, we present an overview of different support materials, focusing on how to achieve precise control of high Fe loading and single-atomic dispersion on the support surface. Subsequently, the typical characterization tools and applications of Fe-SACs in biocatalysis, energy storage and conversion are introduced. To appreciate this progress,the summary and future opportunities in the rapidly growing studies of Fe-SACs are finally discussed.

Fig. 1 Schematic illustration of single-atom iron catalysts for a broad scope of critical applications from biocatalysis to energy storage and conversion.

2. Synthesis of Fe-SACs using different support materials

Fe-SACs possess superior catalytic activities owing to their abundant exposed sites and unsaturated coordination environment. Notably, single atoms anchored on the surface of a support have a high surface energy. Thus, during their preparation and catalysis processes, these atoms tend to agglomerate and form clusters or even nanoparticles, which will deteriorate their catalytic effect or fail. Accordingly, strong interactions between the single metal atoms and the substrate not only can prevent the agglomeration of single Fe atoms but also adjust the geometry and electronic structures of Fe-SACs, improving their catalytic performance.¹⁷ Therefore, the rational design of appropriate support materials and suitable synthesis processes are particularly important for the catalytic performance of Fe-SACs.

2.1. Metal–organic frameworks (MOF) and their derivative-supported Fe-SACs

The unique advantage of MOFs is their designability, e.g., adjustable chemical compositions, structure, pore, particle size and shape, which can provide a high surface area, porosity (about 90% of the free volume), more exposed active sites and abundant pore positions.^18–20 As anticipated, MOFs and their derivatives have become ideal support materials for anchoring an atomic dispersion of Fe and other single metal atoms.

MOFs are composed of evenly distributed metal nodes and organic linkers.²¹ Among the MOF materials, the zeolite imidazole framework (ZIF) is most commonly utilized to construct SACs.²² One typical ZIF material is ZIF-8, which is composed of Zn atoms as metal nodes. Owing to the low boiling point of Zn atoms, ZIF-8 is extensively used as a support material for the synthesis of Fe-SACs via the spatial confinement strategy.^23–25 Fe atoms are confined in a certain space to prevent their agglomeration or migration. This strategy requires that the size of the metal precursor be small enough to enter the “cage”, but not be too small, where its size must be larger than half of the cavity in the MOF material. This can guarantee that only one molecule of metal precursor enters one “cage”, preventing its agglomeration during the next pyrolysis treatment.¹⁸ By classifying the relevant literature, several representative size parameters can be derived, providing a critical guide for the design of Fe-SAC:ZIF-8 with a cavity diameter of ∼11.6 Å and the metal precursors of Fe(acac)₃ and ferrocene with the molecular size of 9.7 and 6.4 Å, respectively. Li et al.²⁶ used a cage-encapsulated-precursor pyrolysis strategy to prepare a highly reactive isolated single-atom Fe/N-doped porous carbon catalyst (Fe/ZIF-SAC) with an Fe loading of 2.16 wt% (Fig. 2a). During the synthesis process, molecular-scale cages were formed together with the assembly of Zn²⁺ and 2-methylimidazole. Importantly, one Fe(acac)₃ molecule, in principle, was trapped in one cage, resulting from the size-matching effect. After pyrolysis at 900 °C under an Ar atmosphere, ZIF-8 was transformed into an N-doped porous carbon framework. Meanwhile, the Fe(acac)₃ molecule in the cage was reduced by carbonization of the organic linker, leading to the production of isolated single Fe atoms anchored on nitrogen species. To verify the above-mentioned size-matching effect, it should be investigated what will happen if an Fe precursor with a larger molecular size than the ZIF-8 cavity diameter is selected to synthesize Fe-SACs. Recently, Fe(II) phthalocyanine (FePc) has been identified to be an excellent metal precursor with a relatively large molecular size of about 14.6 Å.¹⁴ Accumulating evidence indicates that FePc will destroy the microcavity of ZIF-8, causing it to rupture.^27,28 In the subsequent pyrolysis process, ZIF-8 was converted into N-doped carbon and FePc was simultaneously reduced in situ to produce edge-dispersed Fe–N sites.²⁶ However, a drawback inherent to the nature of this control experiment was that excessive FePc could form Fe₂O₃ nanoparticles. Thus, acid etching treatment was required to remove impurities, but clearly this process caused the loss of Fe atoms.

Fig. 2 Synthesis of Fe-SACs using different support materials. (a) Schematic illustration of the synthesis of isolated single iron atoms anchored on N-doped porous carbon. Reprinted with permission from ref. 26. Copyright 2017, John Wiley and Sons, Inc. (b) Schematic illustration of the synthesis of atomically dispersed iron-nitrogen active sites in porphyrinic triazine-based frameworks. Reprinted with permission from ref. 42. Copyright 2018, the American Chemical Society. (c) Schematic illustration of the synthesis of atomically dispersed iron-nitrogen species. Reprinted with permission from ref. 52. Copyright 2017, John Wiley & Sons, Inc. (d) Schematic illustration of the synthesis of porous dual-metal doped g-C₃N₄ patches attached to carbon nanotube bundles. Reprinted with permission from ref. 61. Copyright 2018, The Royal Society of Chemistry. (e) Schematic illustration of synthesis of SACs from metal foam (metal = Fe, Co, Ni, and Cu) by strong bonding from the dangling bonds of GO support. Reprinted with permission from ref. 4. Copyright 2019, John Wiley & Sons, Inc.

In the case of MOFs and their derivatives as a support, another method used for the preparation of Fe-SACs is the coordination design strategy.²⁹ This strategy requires that the organic linking group in the MOF material has one or more atoms (e.g., N, O and S) containing lone pairs of electrons. These atoms can serve as a “claw” to trap Fe atoms, thereby anchoring the dispersed Fe atoms on the surface of MOF and their derivatives.^30,31

As a support material, we reiterate that the biggest advantage of MOFs and their derivatives is their designability, but this is just two sides of the same coin. MOFs and their derivatives support can be beneficial or detrimental for the preparation of SACs. The low chemical stability and weak electrical conductivity of MOFs and their derivatives make it difficult to transfer their charge for electrocatalytic applications.²⁰ Considering this, we need to further study an alternative approach to improve the conductivity of SACs using MOFs and their derivative materials as a support, for example, combining them with conductive agents (graphene, nickel foam, etc.) to form a composite nanomaterial.

2.2. Covalent organic frameworks (COF) and their derivative-supported Fe-SACs

The organic building blocks of covalent organic frameworks (COFs) are connected via covalent bonds to construct a porous framework with a periodic structure.^32,33 Similar to MOF materials, COFs have a high surface area, low density, tunable pore size, skeleton, composition and porosity and a variety of structural topologies.^34–37 These features make them attractive in wide areas, such as gas storage, catalysts, optoelectronics, separation, and proton conduction.^38,39

Actually, COFs and their derivatives under pyrolysis have been found to act as ideal precursors for the production of porous carbon materials. When COFs and their derivatives are employed as a support, Fe^2+/3+ can facilely diffuse into their cavity owing to the strong coordination interaction between Fe^2+/3+ and N position in COFs and their derivatives. The periodic structures of the separation unit in COFs and their derivatives will remarkably inhibit the agglomeration of Fe atoms during the pyrolysis process. Based on the connectivity and geometry of their optimized molecular building blocks, COFs can be separated into two-dimensional (2D) layered structures and three-dimensional (3D) network structures.³⁴ In general, 2D layered structures are widely used in the field of catalysts. 2D layered COFs are stacked face-to-face with the largest π orbital overlap, thereby promoting the charge migration along the framework and improving their conductivity. In addition, covalent bonds with extended π conjugation are connected to the skeleton, which can endow COFs with better physical and chemical stability.^34,40

Atom-dispersed Fe metal sites can be introduced in COFs and their derivatives through two different paths. The first path is the direct use of Fe metal macrocyclic clusters as building blocks, such as Fe-based porphyrins and Fe phthalocyanine.⁴¹ Recently, Cao et al.⁴² prepared a highly stable atom-dispersed Fe–N₄ species on porous porphyrin triazine-based frameworks (Fe/PTF-SACs) via a direct ionothermal method (Fig. 2b). Under the catalysis of molten ZnCl₂, 5,10,15,20-tetrakis(4-cyanophenyl)porphyrinato-Fe(III) chloride (Fe-TPPCN) was transformed to Fe/PTF-SACs through a trimerization reaction in an evacuated ampule, followed by removal of ZnCl₂ using diluted HCl. Taking a closer look at above synthetic processes, the contribution of molten ZnCl₂ was the most prominent. It not only could catalyze the trimerization reaction of the Lewis acid, but also act as a pore-forming agent to produce mesopores. Consequently, the atomically dispersed Fe–N₄ active sites were more exposed to the highly dispersed reactive species, which was conducive for the transport of substances and maximized the atom utilization. Meanwhile, materials under high-temperature pyrolysis will inevitably undergo partial graphitization, hence accelerating the electron transfer and improving the electrocatalytic activity of Fe-SACs.

Thus far, more than 200 2D layered COF materials have been developed. However, these COF materials usually exist in the form of insoluble, cross-linked powders or films, which highly limit their applications. Hence, it is particularly critical to develop water-soluble COFs as support materials.^43,44 Since the crosslinking event is mainly stopped under swelling conditions rather than dissolution conditions, grafting hydrophilic groups is not feasible. Peng et al.⁴⁵ designed and constructed iron quasi-phthalocyanine into conjugated a 2D network with a rigid structure and heterogeneously charged the coordinated single-atom centers to synthesize 2D COF materials. The strong interactions between the negative hydroxyl groups and positively charged quasi-phthalocyanine centers enabled the rigid structures of the COF to be easily solvated. Thus, the robust conjugation systems and Fe–N coordination centers of the COF materials caused them to display outstanding ORR properties. The factors affecting the catalytic performance of Fe-SACs included the Fe loading and the role of the support. A high N content in the support can anchor more Fe atoms, thereby increasing the metal loading, while a high specific surface area in the support can improve the mass transfer process in the catalytic process. Therefore, when using COF derivatives as a support, increasing their specific surface area becomes the main problem. Wang et al.⁴⁶ synthesized an imine COF thin film at room temperature through an electrochemical exfoliation synthesis method. The COF thin film possessed high crystallinity and a hierarchical porous structure, which made it display relatively excellent mass transfer performances. Recently, Xiong et al.⁴⁷ synthesized a hollow COF with high crystallinity and uniform morphology. These COF materials were demonstrated to exhibit a large specific surface area, radially oriented nanopore channels, uniform morphology and tunable particle size. Cooper and colleagues prepared a highly crystalline and stable COF material with good CO₂ adsorption ability via the pre-assembly of reversible covalent bonds and post-synthesis framework reconstruction into irreversible covalent bonds.⁴⁸ The above-mentioned COF materials all displayed a high specific surface area, and thus their catalytic activity will be greatly improved when they are used to prepare Fe-SACs.

2.3. Carbon nanotube (CNT)-supported Fe-SACs

CNTs are mainly composed of carbon atoms arranged in a hexagonal shape to form a coaxial circular tube with single to dozens of layers.³⁶ The carbon atoms in CNTs are dominated by sp² hybridization. Simultaneously, the hexagonal grid structure has a certain degree of bending to form a spatial topology, yielding a certain sp³ hybrid bond. Thus, the chemical bond formed has a mixture of sp² and sp³ hybrid states, and these p orbitals will overlap outside the graphene sheet of CNTs, showing a highly delocalized large π bond. Consequently, CNTs have good electrical conductivity. At high temperatures, because the Fe atoms are flowing, they tend to migrate and aggregate to produce inactive nanoparticles, especially at a high content of Fe precursors. Thus, to improve the activity of the catalyst, the selection of the appropriate support to anchor Fe atoms is an effective route to increase the single-atom Fe loading for the preparation of Fe-SACs.⁴⁹ CNTs provide abundant exposed Fe–N_x active sites, good chemical stability, high mechanical properties and unique porosity. Moreover, the excellent electrical conductivity of CNTs accelerates the electron transfer efficiency.^9,36,50,51 Thus, all these features enable the use of CNTs as a support with vast potential for anchoring Fe atoms.

According to whether the CNTs are simultaneously formed in the process of synthesizing Fe-SACs, their synthesis routes can be divided into two categories. One is the direct use of commercially available CNTs as a support. This synthesis method generally involves three steps, i.e., adsorption of Fe³⁺ cation (sometimes Na⁺ is also required), pyrolysis in an inert gas (sometimes NH₃ is needed) to obtain atomically dispersed Fe–N_x active sites, and use of the acid leaching strategy to remove the nanoparticles or iron oxide generated by Fe agglomeration during pyrolysis. Accordingly, single-atom Fe anchored on CNT catalysts with high catalytic performances will be achieved.^52–54 For example, Chen et al.⁵² proposed an effective method involving S to achieve atomically dispersed Fe–N_x species anchored on N, S co-decorated CNTs (Fig. 2c), generating new carbon-based bifunctional electrocatalysts. FeCl₃ as the iron source and KSCN as the sulfur source were added to synthesize the N,S co-doped single atomic catalysts with a unique hierarchical structure and rich atom dispersed Fe–N_x species. The obtained Fe/CNT-SACs had a higher specific surface area (241.5 m² g⁻¹) compared with the SACs without S salt (173.9 m² g⁻¹). The increased surface area and pore volume of the catalyst facilitated the acid leaching. In addition, the addition of S salts can prevent the formation of iron-based carbides, which are not easily removed by acid leaching.

The other type of synthesis relies on the application of different carbon sources to synthesize CNTs during the preparation of Fe-SACs. Likewise, this synthetic route roughly involves three steps. The first is assembly polymerization, in which various Fe and carbon sources are introduced. Secondly, high-temperature pyrolysis is carried out in an inert atmosphere such as N₂ or NH₃. Finally, acid leaching is performed to remove other unwanted metal species,^9,55 which will lead to the loss of Fe–N_x active sites, thereby reducing its catalytic performance.

By comparing the last step for the synthesis of CNT-supported Fe-SACs, acid etching treatment is usually needed and essential, regardless of the synthetic methods, to eliminate the nanoparticles or iron oxides generated from the agglomeration of Fe during pyrolysis. We advocate that the acid leaching strategy plays an increasingly crucial role in accelerating the exposure of Fe–N_x active sites, but it has to be delicately considered in future studies. Clearly, excessive acid leaching will also be irreversible to remove atomically dispersed single-atom Fe species, which possibly make the as-synthesized Fe-SACs fail in satisfying catalytic goals. Recently, the metal (Zn) atomic isolation strategy and silicon protective layer-assisted strategy have emerged as exciting alternatives to avoid Fe agglomeration in the synthesis of CNT-supported Fe-SACs, entering the development spotlight.^54,56

2.4. Graphitic carbon nitride (g-C₃N₄)-supported Fe-SACs

Graphitic carbon nitride (g-C₃N₄) is a graphene analogue consisting of periodically linked tris-s-triazine units, which is produced via the thermal condensation of organic monomers, including urea, melamine and dicyandiamide.^57–60 Considering its molecular structure, g-C₃N₄ can be regarded as N-doped graphene. Furthermore, the g-C₃N₄ framework contains rich and uniformly distributed pyridine N components with abundant lone pairs of electrons and large N coordination cavity, making it an impressive support for anchoring Fe atoms for the synthesis of Fe-SACs.⁴¹

There are usually two routes for the fabrication of Fe-SACs using g-C₃N₄ as a support. The first is a direct in situ bottom-up synthetic method, which involves the direct heating and polymerization of mixtures of Fe salt and organic monomers to obtain Fe-SACs with high catalytic activities.^41,57 For example, Sang et al.⁵⁸ synthesized Fe-coordinated g-C₃N₄via the thermal condensation of dicyandiamide and FeCl₂ in an N₂ atmosphere. Notably, when g-C₃N₄ alone was selected as the supporting material, the catalytic effect of the as-prepared Fe-SACs was not very good due to its poor conductivity. Thus, to solve this issue, CNTs or crystalline carbon structure serving as a conductive substrate can be introduced in the g-C₃N₄ framework. Hybrid compositions not only facilitate charge transfer, but also suppress the strong accumulation of g-C₃N₄ layers, hence enabling a high Fe loading.^41,57 For example, Song et al.⁶¹ adopted a high-temperature polymerization method to develop a structure composed of g-C₃N₄ nanopatch-enveloped CNTs, where isolated Fe atoms were embedded in the tri-s-triazine unit by forming a metal–N_x structure (Fig. 2d). The scheme showed that NH₄Cl served as a dynamic gas template, which could generate ammonia gas during the pyrolysis process, and consequently adjust the porous structure of g-C₃N₄. The loosely packed g-C₃N₄ segment adhered to the CNT framework was conducive for electron transfer and increasing the exposed active sites.

Significantly, the majority of g-C₃N₄ materials are prepared via the thermal polymerization of melamine, cyanamide or dicyandiamide. However, these precursors are expensive and toxic. Accordingly, for large-scale industrial production in the future, we still need to select other precursors (e.g., urea) to produce g-C₃N₄-supported Fe-SACs.^58,62–64

Another popular strategy to synthesize Fe-SACs using g-C₃N₄ as a support is the post-synthesis method (also called wet chemical method).⁵⁷ This method involves impregnating the target metal salt in the liquid phase on the surface or edge defects of g-C₃N₄, and then reducing the metal ions using a reducing agent such as NaBH₄ or low temperature heat treatment.⁴¹ The main advantage of this method is that the reaction conditions are relatively mild, which will not destroy the original basic structure of the support g-C₃N₄ during the synthesis of Fe-SACs. Moreover, the diffusion of the target metal ions will be more complete. There is a large number of coordination sites on g-C₃N₄, facilitating the uniform distribution of Fe ions on the surface of the g-C₃N₄ support.^41,57 For example, Cao et al.⁶⁵ proposed a surfactant-assisted wet chemical method to synthesize single-atom Fe supported on N-doped g-C₃N₄ (denoted as SA-Fe/NG), creating unique Fe-pyrrolic-N active species with an Fe atom and eight C atom active sites next to pyrrolic N.

2.5. Graphene-supported Fe-SACs

Graphene is a hexagonal 2D material composed of sp² hybridized carbon.^66,67 It is common in nature, but difficult to peel off the single-layer structure from the graphite widely used in daily life. The arrangement of carbon inside graphene is consistent with a monoatomic layer of graphite bonded by sp² hybrid orbitals, in which each carbon atom has an unbonded electron in the p orbital. The p orbital of adjacent carbons is perpendicular to the plane direction formed by the δ bond and forms π-type bonds with a semi-full state, endowing graphene with excellent electrical conductivity and heat conduction. Compared with CNTs, both graphene and graphene nanosheets (GN) exhibit a high specific surface area and superior mechanical property, which are crucial requirements of support materials to anchor Fe atoms.^66,68,69

However, the interaction between graphene and iron atoms is relatively weak. The single Fe atoms are prone to aggregate, which deteriorates the catalytic activities of Fe-SACs. Thus, to address this issue, N doping and vacancy defects are two commonly employed modification methods.⁷⁰ Defective graphene with vacancies that can firmly anchor metal atoms, and hence prevent their agglomeration, is often used as a support for the synthesis of Fe-SACs. Generally, to form graphene vacancies (monovacancy and divacancy graphene), a specific area is irradiated with a focused electron beam. The formed vacancies play multiple roles in the adsorption of single Fe atoms and prevent their agglomeration, thereby leading to highly dispersed single-atom active sites.^71–73

Graphene oxide (GO), a derivative of graphene, is a single layer of graphite peeled from graphite oxide.⁶⁷ Compared with graphene, which is composed of sp² hybridized carbon, GO contains unoxidized aromatic regions (sp² carbon atoms) and fatty six-membered ring regions (sp³ carbon atom) formed by oxidation damage to the lattice. Importantly, its surface and edges have more O-containing groups, e.g., –CH–(O)–CH–, –OH, –COOH and –C=O, which can complex with metal ions, especially those with multivalence.⁷⁴ Moreover, due to its high specific surface energy, good hydrophilicity and mechanical property, GO has become an ideal support for the preparation of SACs.⁶⁷ Qu et al.⁴ modified GO with a large amount of O-containing dangling bonds, which can provide anchoring sites for Fe atoms (Fig. 2e). In the synthetic process, the slurry of GO was mixed with the metal foam of iron. At this time, a large amount of iron and the suspended oxygen groups on the surface of graphene oxide were in close contact, and the iron transferred electrons to the oxygen groups. Consequently, a large number of Fe^δ+ (0 < δ < 3) species was formed. Besides, owing to the coordination, each iron atom and the four oxygen dangling bonds on the surface of GO formed Fe–O bonds. Finally, the iron atoms were pulled out of the metal foam under the action of ultrasound to yield their natural isolation, which prevented the iron atoms from agglomerating on the surface of GO.

Notably, the use of GO as a support for anchoring single metal atoms has certain drawbacks that cannot be circumvented. For example, under pyrolysis treatment, GO is prone to undergo irreversible accumulation. This phenomenon, which is an inherent nature of GO, causes a series of destructive consequences, including deterioration of the accessible surface area, blocking of catalytically active centers, and obstruction of mass diffusion/transfer, ultimately having a negative effect on the electrocatalytic performance. Thus, the design and fabrication of GO-supported Fe-SACs remain a grand challenge. Only recently, Tang et al.⁶⁹ proposed a sophisticated and universal template-engaged method to synthesize 3D GO hollow nanospheres, which can exhibit improved mechanical strength, accelerated reaction kinetics and enhanced catalytic performance when employed as a catalyst support. From a positive point of view, 3D GO materials are expected to be an ideal platform to immobilize isolated single metal atoms.

2.6. Non-carbon-based material-supported Fe-SACs

The derivatives of MOFs and COFs, g-C₃N₄, CNTs and graphene are implicated as carbon-based supports, while non-carbon-based materials (MoS₂, MoSe₂, WS₂, WSe₂ and other transition metal sulfides) also deserve to be studied to support single Fe atoms. The basal plane of the crystal surface of transition metal chalcogenides is controlled by introducing multiple defect sites, enabling single atoms to be anchored to form high-loading SACs. In a typical experiment, Zheng et al.⁷⁵ employed thiourea-coordinated Fe atoms as a precursor to synthesize atomically dispersed Fe supported on ultra-thin 2D transition metal chalcogenides. Notably, this method was not only suitable for metallic Fe, but could also be applied to anchor other metals, such as Co, Ni, Cu, Pt, Pd and Ru. Impressively, the single metal loading reached an astonishing 10%. Taking the MoS₂-supported Fe as an example, n-BuLi was initially used to treat transition metal chalcogenides such as MoS₂, and then they were peeled into ultra-thin monolayers. There were abundant S vacancies in these monolayers, which self-assembled with Fe–thiourea complexes to yield nuclei under hydrothermal environments. Finally, a single atomic Fe catalyst with MoS₂ as a carrier was formed through H₂ reduction at high temperature. At 2022, Li et al.⁷⁶ constructed Fe single-atom catalytically active sites coordinated with S on mesoporous TiO₂via a lattice confinement strategy. The spectroscopy analysis indicated that Fe atoms were anchored in the TiO₂ lattice in the form of FeS₂O₂ coordination structures, which served as the active centers for the electrocatalytic nitrogen reduction reaction. It should be noted that the lattice confinement strategy is versatile and can be employed to synthesize single-atom catalysts, including Ni₁S@TiO₂, Co₁S_x@TiO₂, and Mo₁S_x@TiO₂. In a typical experiment, Zhang et al.⁷⁷ synthesized 3D single-atom Fe-doped macroporous/mesoporous TiO₂–SiO₂ photocatalysts via an evaporation-driven self-assembly route. In this work, polystyrene spheres were arranged to produce polystyrene opal as a hard template for macropores and P123 micelles were used as a soft template for mesopores. Hydrolysis and calcination were performed to remove polystyrene and P123, leading to the formation of macroporous/mesoporous structures with the presence of massive single Fe atoms.

Besides, the surface of transition metals contains a large amount of hydroxide groups, hence providing suitable anchoring sites for the preparation of SACs to coordinate with metal atoms via M–O(OH) interactions. Recently, He et al.⁷⁸ constructed single Pt atoms anchored on a CuO support by utilizing the electron transfer between Pt atoms and CuO. Zhang et al.⁷⁹ synthesized atomically dispersed Pt on RuO₂via an impregnation-adsorption method, which exhibited excellent catalytic activity for methanol oxidation under alkaline conditions. The above-mentioned examples are associated with single Pt atoms supported by transition metal oxides because electron transfer could easily occur between Pt and transition metal oxides. Likewise, it needs to be further investigated if it also possible to synthesize Fe-SACs through interactions between single Fe atoms and other transition metal oxides.

3. Characterization of SACs

The information of active species (e.g., atomic structures, oxidation states and coordination chemistry) is closely related to the catalytic activity of SACs. With the development of advanced characterization techniques, an atomically precise understanding of SACs becomes possible, which can have quantifiable benefits for catalytic performance.

3.1 Electron microscopy

Regarding high-resolution transmission electron microscopy (HRTEM), its detection limit is at the nanometer or sub-nanometer levels. Thus, it is impossible to clearly observe the distribution of single atoms on a support by HRTEM. Fortunately, sub-angstrom-resolution aberration-corrected scanning transmission electron microscopy (AC-STEM) can image individual atoms dispersed on different support materials.¹⁷ To outline the whole picture of SACs, the combination of HRTEM and AC-STEM is a widely accepted and indispensable characterization method. Firstly, the use of the HRTEM technique can observe the entire support with a nanoscale imaging resolution (Fig. 3a and b). For example, Jiang et al.¹⁴ utilized ZIF-8 and FePc precursors to synthesize atom-dispersed Fe–N₄ species anchored on N-doped carbon, exhibiting abundant mesopores with a dodecahedron shape (Fig. 3b). Impressively, the imaging resolution will reach the atomic scale under AC-STEM characterization. The red circle in Fig. 3c represents a single scattered Fe atom. Meanwhile, auxiliary characterizations, e.g., energy-dispersive X-ray spectroscopy (EDS) and electron energy loss spectroscopy (EELS), together with electron microscopy can be employed to analyze the structure and elemental distribution of single metal atoms.

Fig. 3 Characterization of SACs. (a) TEM image of FePc molecules encapsulated in the cavities of ZIF. (b) TEM image and (c) HAADF-STEM image of hierarchically porous Fe-SACs. (d) Fe K-edge XANES (inset was the magnified image) and (e) FT k₃-weighted EXAFS spectra of Fe-SACs and the reference samples. (f) Corresponding FT-EXAFS fitting curves of Fe-SACs. Reprinted with permission from ref. 14. Copyright 2018, the American Chemical Society.

3.2 Spectroscopy analysis

In addition to the use of the aforementioned electron microscopy technology to characterize the distribution of individual metal atoms on the support surface, it is also necessary to identify the catalytic property-associated coordination information such as electronic and atomic structures of SACs. The most commonly utilized spectroscopy technology is X-ray absorption fine structure spectroscopy (XAFS). According to the mechanism and peak shapes, XAFS can be divided into two categories,⁸⁰i.e., X-ray absorption near edge spectroscopy (XANES) and extended X-ray absorption fine structure spectroscopy (EXAFS).

The XANES technique provides the valence state, charge transfer and other information of the single metal atoms in SACs. Interestingly, due to the presence of the “fingerprint effect”,^81,82 XANES has the capability of establishing the type of element. It is important to note that considering the definition of SACs, there should be no metal–metal bond (e.g., Fe–Fe bond) in the active sites on the surface of support. EXAFS analysis can clarify this atomically structural information, a decisive criterion in evaluating the successful synthesis of SACs. If EXAFS analysis shows that that there is no information on metal–metal bonds, it is basically inferred that only a single metal atom exists in SACs. Besides, some other characterizations of the active centers can be elucidated by EXAFS technology, including the type of atom coordinated with the central atom, the coordination number and the distance between the coordinated atoms and central atoms.¹⁷ As shown in Fig. 3d, taking Fe foil, FePc, FeO and Fe₂O₃ as references, the Fe K-edge XANES results indicated that the oxidation state of the Fe atoms in Fe-SACs were between the Fe(II) and Fe(III) states.¹⁴ According to the Fourier transform curve of EXAFS (Fig. 3e), it can be seen that there was only one main peak at 1.5 Å, corresponding to the Fe–N species, with the absence of Fe–O or Fe-Fe bonds, together revealing that a single Fe atom was coordinated with the N atoms on the support. After linearly fitting the Fourier transform curve of EXAFS, the number of N atoms coordinated with Fe atoms was calculated to be about 4 (Fig. 3f), indicating that the active sites of the Fe-SACs were atomically dispersed Fe–N₄ species.¹⁴

Here, we stress that although the above-described modern characterization techniques have become powerful tools for studies at the atomic scale, the common spectroscopy tools can also offer key parameters of SACs in detail, e.g., mutual interactions between isolated metal and supports and atomically local structural information. This analysis data may allow common spectroscopy tools to complement or extend current advanced approaches for atomic catalysis research.⁶² In fact, as early as the 1970s, Yates et al.⁸³ used infrared spectroscopy (IR) to confirm the presence of a single Rh atom in a catalyst. The underlying principle is monitoring the interaction between the adsorbed molecules (molecular probe) and the support of SACs.⁸⁴ The variation in vibration frequency and intensity of the probe could be utilized to extract the characteristics of the active centers by calibration. Likewise, nuclear magnetic resonance (NMR) spectroscopy also contributes to clarifying the single-atom species and binding ligands in SACs, e.g., monoatomic Pt catalysts with a very low loading.⁸⁴ However, thus far, NMR spectroscopy has not been applied for the characterization of Fe-SACs.⁸⁵

4. Applications of Fe-SACs

Due to the excellent catalytic properties of Fe-SACs, their application fields are impressively wide, especially in energy storage and biocatalysis. In the case of energy storage, Fe-SACs play key roles in electrocatalytic, thermocatalytic and photocatalytic processes. In the biocatalysis field, Fe-SACs show excellent multi-enzymatic activities, making them suitable therapeutics in biosensing, wound disinfection, oxidative stress cell protection and cancer treatment.^86–89

4.1. Fe-SACs for energy storage

To satisfy the large energy demand, the development of highly efficient catalysts is particularly critical for sustainable energy technologies.⁹⁰ From a practical point of view, endowing catalysts with the capability to reduce the relatively large energy barriers of certain reactions has not yet been fully achieved owing to the inadequate design strategies and techniques available for the preparation of highly active catalysts.^5,91 Consequently, Fe-SACs have currently entered the development spotlight for energy storage. According to the type of energy supply, catalysis can be divided into electrocatalysis, photocatalysis and thermal catalysis.⁵⁷ Therefore, below we introduce the application of Fe-SACs in the above-mentioned three fields.

4.1.1. Fe-SACs in electrocatalysis. In the energy storage and transfer fields, electrocatalysis is the most widely studied. Therefore, the development of high-performance electrocatalysts is crucial to achieve fast energy conversion, which helps to lower the energy barrier and facilitate the kinetics of electrochemical processes.⁹² In the field of energy storage, Fe-SACs play a crucial roles in electrocatalytic reactions including the ORR, HER, OER, NRR, CO₂RR and H₂O₂ production.^93–100
4.1.1.1. ORR process. Metal–air batteries and fuel cells are ideal clean energy conversion equipment owing to their advantages of high energy density, low cost and zero emission of pollution gas.^52,55 However, despite this promise, because of the slow kinetics of the ORR, there are several challenges (e.g., low energy conversion efficiency, poor stability in corrosive electrolytes and unsatisfactory lifespan) in their progression for energy storage.^101,102 The ORR process has multi-step reaction pathways, which can be roughly divided into three categories (Fig. 4). One is the 4e⁻ reaction pathway (formulas (1)–(4)), as follows: O₂ + 2H₂O + 4e⁻ → 4OH⁻. The second is the 2e⁻ reaction pathway (formulas (5) and (6)), as follows: O₂ + H₂O + 2e⁻ → HO₂⁻ + OH⁻. The third is a series of reactions, where the 2e⁻ and the 4e⁻ pathways coexist.

Fig. 4 4e⁻ reaction pathway and 2e⁻ reaction pathway on single-atom iron catalysts.

The 4e⁻ reaction pathway is as follows:

O₂ + H₂O + e⁻ + * → OOH* + OH⁻. (1)

OOH* + e⁻ → O* + OH⁻ (2)

O* + H₂O + e⁻ → OH* + OH⁻ (3)

OH* + e⁻ → * + OH⁻ (4)

The 2e⁻ reaction pathway is as follows:

O₂ + H₂O + e⁻ + * → OOH* + OH⁻. (5)

OOH* + e⁻ → OOH⁻ (6)

The Fenton reaction is as follows:

Fe²⁺ + H₂O₂ → Fe³⁺ + OH⁻ + ˙OH (7)

where * refers to the surface site.

In the ORR process, many reaction intermediates (O*, OOH*, OH*, etc.) are produced and the adsorption of these intermediates on the surface of the electrode is vital to the ORR kinetics. Thus, an ideal ORR catalyst should show a suitable binding energy with the reaction intermediates.¹⁰³ When Fe-SACs catalyze the ORR process following the 2e⁻ reaction pathway, H₂O₂ will be generated, and then Fe reacts with the H₂O₂ to yield the strong oxidizing ˙OH (formula (7)). Notably, ˙OH results in the electrical oxidation of the carbon surface and the proliferation of micropores, which together oxidize the active sites to inactivate them.¹⁰⁴ Consequently, the stability of the battery system will be deteriorated. Besides, because the energy transfer efficiency of the 2e⁻ pathway is relatively low, it should be prevented as much as possible. As a comparison, the 4e⁻ pathway is the optimal electron pathway to improve the ORR activity, which is required by the battery cathode.⁵⁵

In fact, not all the Fe sites in Fe-SACs have a catalytic effect, such as the Fe atoms inside the support. Only the exposed active sites contribute to the catalytic properties. In this context, it is an efficient and direct way to improve the ORR activity by increasing the utilization of the metal atoms. Multiple lines of evidence reveal that the ORR process mainly occurs at the “catalyst–electrolyte–oxygen” three-phase interface. Guided by this discovery, recent studies attempted to increase the density of the Fe active sites exposed on the three-phase interface by adjusting the morphology of the catalysts.^105–109 The reactants and products are transported in large quantities between the catalyst layers, and the utilization rate of the Fe active sites is improved. In a typical experiment, Hou et al.¹¹⁰ fabricated Fe-SACs with an overhang-eave structure, enabling more stretched edges, which can act as the three-phase interface of the ORR reaction. Also, the mass transfer of the ORR reaction substances between the catalyst layers was remarkably promoted, and the Fe active sites were exposed to the maximum extent, resulting in excellent ORR activity. In an alkaline solution, the detected onset potential and half-wave potential were 1.0 and 0.85 V, respectively, which are comparable to that of the commercial Pt/C catalysts (Fig. 5a–c). Designing and controlling the local structure of the metal sites are key to achieving efficient catalysis by single-atom metal catalyst sites. Generally, in the synthesized Fe-SACs, the plane of the active site Fe–N₄ is straight. If the plane of the active site is adjusted, will it affect its catalytic activity? In a typical experiment, Wu et al.¹¹¹ tuned the catalytic activity of Fe–N sites of surface single atoms through a stress-modulating method and observed enhanced ORR activities. In the synthetic process, the self-assembled structure of chiral surfactant molecules was used as a template, and then pyrolysis was performed to generate polypyrrole with a helical structure. Consequently, the active sites of Fe–N₄ were rich in compressive stress. A large number of isolated non-planar curved Fe–N₄ right-handed single Fe atoms was distributed on the helical hollow carbon fibers during the pyrolysis process. The electrochemical test results showed that the Fe-SACs had a higher onset potential, half-slope potential and kinetic current density compared with other types of catalysts in both acidic and basic conditions. DFT calculation was further utilized to determine the planar and curved Fe–N₄ structures. The charge density difference diagrams demonstrated that the charges of the Fe atoms can transfer to the adjacent N atoms, irrespective of a flat or curved Fe–N₄ structure (Fig. 6a and b). The Bader charge of Fe in the curved Fe–N₄ structure was larger than that of the planar one, indicating that the curved structure facilitated the charge transfer. Moreover, compared with the planar Fe–N₄ structure, the d-band center of the curved Fe–N₄ structure shifted down. Consequently, the adsorption energy with oxygen-containing intermediates was reduced, thereby improving the ORR activity (Fig. 6c and d). The Gibbs free energy diagram results revealed that for the planar Fe–N₄ structure, the potential determination step was *OH + 3OH⁻ + e⁻ → * + 4OH⁻, and the onset potential was determined to be about −0.15 V. The potential determination step of the curved Fe–N₄ structure was *O₂ + 2H₂O + 4e⁻ → *OOH + OH⁻ + H₂O + 3e⁻, and the onset potential was −0.10 V (Fig. 6e and f). The above-mentioned data indicated that the curved Fe–N₄ structure was more active than the planar Fe–N₄ structure.

Fig. 5 Applications of Fe-SACs in energy storage. (a) ORR polarization curves and (b) corresponding E_1/2 and J_L values of Fe-SACs. (c) LSV curves of Fe-SACs at different rotation rates. Reprinted with permission from ref. 110. Copyright 2020, John Wiley & Sons, Inc. (d) N₂ adsorption/desorption isotherms of meso/micro-Fe/Co-SACs and micro-Fe/Co-SACs. (e) Tafel plot and (f) RDE polarization curves of meso/micro-Fe/Co-SACs and reference samples for the OER. Reprinted with permission from ref. 161. Copyright 2018, John Wiley & Sons, Inc. (g) HER polarization curve and (h) Tafel plot of catalyst Fe/Pt-SACs and reference samples in 0.5 M H₂SO₄. (i) DFT calculation of free energy ΔG_H* of H* adsorption for the indicated catalysts in 1 M KOH. Reprinted with permission from ref. 104. Copyright 2018, John Wiley & Sons, Inc.

Fig. 6 (a and b) Charge density difference diagrams (yellow: electron accumulation and cyan: electron depletion), (c and d) projected density of states, (e and f) ORR Gibbs free energy changes of planar Fe–N₄ (k, m, and o) and curved Fe–N₄ (l, n, and p). Reprinted with permission from ref. 111. Copyright 2021, John Wiley & Sons, Inc.

The electrolyte pH also affects the ORR activities of Fe-SACs. In acidic media, the ORR activity of Fe-SACs will be seriously damaged because of the acid leaching, causing a large overpotential (40–400 mV) and extremely poor stability.¹¹² Previous studies demonstrated that pyridine N, which plays an decisive contribution in regulating the ORR activity, will be coupled with H⁺ in acidic media.⁵⁵ This complexation reduces the charge density of carbon atoms adjacent to the Fe–N₄ active sites, largely inhibiting the ORR activity. Besides, a second and likely more fundamental reason for the slow ORR kinetics of Fe-SACs is that the binding energy between the Fe–N₄ active sites and O₂ or the reaction intermediates is not optimal. This guideline helps efforts to focus on identifying a way to optimize the adsorption energy between Fe and O₂ and reduce the bond energy of the Fe–O bond. The difference in the coordination environment of atom-dispersed Fe active centers results in a variation in charge density around them, which will in principle affect the adsorption energy and the final ORR activity.³⁶ Doping with oxygen atoms during the synthesis of Fe-SACs by forming specific local structures (e.g., carbonyl groups) can facilitate the production of OOH* and the reduction of OH* species. This synergistic effect drastically decreases the overpotential of Fe-SACs, and thus boost their ORR activity.¹¹³ N doping is the most widely used route in the design and synthesis of Fe-SACs. In the pyrolysis process, the formation of carbon defects and an increase in the polarity of the surface of carbon atoms adjacent to N will occur, thereby promoting the rapid transfer of electrolyte and electrons and increasing the ORR activity.^114–116 N has a higher electronegativity (χ = 3.07) than that of C (χ = 2.55), and thus for N-doped carbon formed by pyrolysis, the N atoms exhibit strong ability to attract electrons.^36,41 Density functional theory (DFT) indicated that the active site of Fe–N₄ formed by co-doping of Fe and N is very similar to Pt, which can absorb oxygen, and subsequently break the O=O bonds during the ORR reaction.¹¹⁷ Taking a closer look at Fe-SACs, N-doped carbon has three structures, namely pyrrole N, pyridine N (affording coordination sites for Fe atoms in the form of Fe–N_x) and graphite N (mainly affecting the geometry and electronic structure of carbon skeleton in Fe-SACs). Studies have shown that pyridine N and graphite N are advantageous species for ORR activities. Specifically, pyridine N can improve the onset potential of the ORR, while graphite N increases the kinetic current and the limiting diffusion current density of Fe-SACs, and thus an increase in both N contents in the catalyst is beneficial for the ORR activity.^117–119

In addition to O and N doping, S, P, C, B and F doping is also possible.¹²⁰ In a typical experiment, Hu et al. synthesized a single atom catalyst (FeN₄Cl₁/NC) with the FeN₄Cl₁ configuration with a metal triazole salt (MET) skeleton as the precursor and chlorine secondary ligand 4,5-dichloroimidazole (dcIm) as the raw material. The catalyst had a high iron content (2.78 wt%) and a relatively excellent void/volume ratio of 0.92. In FeN₄Cl₁/NC, because the Cl atom had a lower electronegativity, the d-band delocalization will be enhanced. Consequently, the electronic structure of the atomic iron changed, showing an optimal free energy of adsorption with the reaction intermediate *OH during the ORR process. Under alkaline conditions (0.1 M KOH), the measured half slope potential was 0.91 V, which was greater than that of the FeN₄/NC catalyst (E_1/2 = 0.89V) without Cl doping. Also, the Tafel slope (35.8 mV dec⁻¹) was smaller than that of FeN₄/NC (49.3 mV dec⁻¹) and Pt/C (66.5 mV dec⁻¹) (Fig. 7a–c). The results of the RRDE test revealed that the average electron transfer number of FeN₄Cl₁/NC in the ORR process was 3.97, and the yield of H₂O₂ was lower than 2%. The data suggested that the ORR process followed a four-electron transfer route (Fig. 7d). Importantly, the content of iron had a direct influence on the ORR performance. When the content of iron was 0.35%, FeN₄Cl₁/NC had the largest E_1/2 and j_k (Fig. 7e). Simultaneously, FeN₄Cl₁/NC exhibited longer stability. After 5000 consecutive potential cycles, E_1/2 showed a slight negative shift (Fig. 7f). Under acidic conditions (0.1 M HClO₄), its half slope potential was 0.79 V, which was greater than that of FeN₄/NC (E_1/2 = 0.76 V), and it had a lower Tafel slope (47.4 mV dec⁻¹) (Fig. 7g). The DFT calculation results are shown in Fig. 7h. The two FeN₄Cl₁/NC and FeN₄/NC catalysts followed the four-electron transfer path during the ORR process. For the FeN₄/NC catalyst, the rate-determining step was the last step, and it can be seen that the FeN₄ active site had a strong adsorption of *OH, which made the final desorption process extremely difficult, thereby inhibiting the ORR rate. For the FeN₄Cl₁/NC catalyst, all the electron transfer steps of the adsorption free energy ΔG were negative, indicating that due to the coordination of Cl atoms, *OH was easy to desorb from the FeN₄ active site in the final process. The result of the electron density difference when the *OH was adsorbed on the active site showed that the axial electron transfer between the iron atom and the chlorine atom will affect the charge of the iron atom, thereby reducing the free energy of adsorption between the iron atom and *OH, thus contributing to the ORR process (Fig. 7i and j).¹²¹

Fig. 7 Electrocatalytic ORR performance of FeN₄Cl₁/NC. ORR polarization curves (a) and comparison of J_k at 0.85 V and E_1/2 (b) and Tafel plots (c) of various catalysts. (d) Electron transfer number (n, top) and H₂O₂ yield (bottom) versus potential for various catalysts. (e) Comparison of J_k at 0.85 V and E_1/2 of FeN₄Cl₁/NC samples with different Fe loadings. (f) ORR polarization curves of FeN₄Cl₁/NC before and after 5000 cycles. (g) LSV polarization curves for various catalysts. (h) Free energy diagram of the ORR on FeN₄/NC and FeN₄Cl₁/NC (U = 0). Calculated electron density difference for OH* adsorbed on (i) FeN₄Cl₁/NC and (j) FeN₄/NC. Reprinted with permission from ref. 121. Copyright 2021, John Wiley & Sons, Inc.

The S atom has strong ability to attract electrons, with an electronegativity of ∼2.58. Thus, the co-doping of S and N dramatically alters the electron density around the Fe active sites of Fe-SACs. Consequently, the C atoms appended to the Fe active sites display more a positive charge, which is conducive to the adsorption of O₂, and thus improve the ORR activity.⁵² Generally, the application scope of S doping is relatively narrow and it is preferentially doped at the edges and defects of graphene, whereas P atoms can be randomly doped in carbon-based materials.¹²² Recently, Yu et al.¹²³ proposed a new conceptual strategy to reset the local coordination environment of Fe–N_x active sites by controlling the dynamic transfer of Fe–S bonds in Fe-SACs catalysts. The spectral and theoretical results showed that the selective cleavage of Fe–S bonds caused the incorporation of electron-withdrawing sulfur oxide on the Fe–N_x active site, thereby promoting the desorption of oxygen-containing intermediates from the active site. Cumulative evidence showed that the doping of P atoms increased the contents of pyridine N in Fe-SACs, adjusting the electronic structure of the Fe active site.¹²⁴ Accordingly, the adsorption of OH* intermediates was weakened, increasing the ORR activity. The doping of B atoms not only can provide electron-deficient sites (improving the electron transfer at the Fe–N_x–C sites), but also enhances the interactions of oxygen-containing species, thereby enhancing the ORR activity.^118,124 Doping with F atoms can increase the adsorption capacity for O₂ and enhance the ORR activity due to the synergistic effect between nitrogen and fluorine.¹²⁰

In summary, the coordination environment of Fe–N_x active sites is determined by the doping of heteroatoms and the number of N atoms coordinated with Fe atoms. It can be found that for the majority of Fe-SACs, the number of N atoms coordinated to the central Fe atom is 4, and other coordination numbers of 2, 3, and 5 were also reported.^125–127 According to DFT analysis, the coordination states of the Fe active sites, especially the number of coordination atoms, will have a huge impact on the catalytic activity by affecting the binding energy between the O₂/active intermediates (OH*, OOH*, O*, etc.) and the central Fe atom.¹²⁸ In a typical experiment, Lai et al.¹²⁵ synthesized an Fe–N/C electrocatalyst with five-coordinated Fe–N_x active sites through the host–guest chemistry strategy. According to DFT calculations, the five-coordinated Fe–N_x active site can increase the ORR reaction rate in acidic medium by reducing the reaction barrier and the adsorption energy of the OH* reaction intermediate. Importantly, reports validated that the introduction of other transition metal elements M (Co, Mn, Cu, Zn, Ni, etc.) in Fe-SACs to generate a bimetallic site catalyst with N₃Fe–MN₃ structures is beneficial for strengthening the binding capability with O₂.¹²⁹ To a certain extent, the bimetallic SACs will increase the bond length of the O=O double bonds, and accordingly weaken the energy of these double bonds, thus facilitating their cleavage. Similarly, single-atom catalysts with isolated diatomic metal sites can also enhance their catalytic effects by adjusting the electronic states of adjacent metal atoms. For example, Han et al. synthesized atomically dispersed platinum with iron anchored on a nitrogen-doped carbon support (Fe–N₄/Pt–N₄@NC), where the Pt–N₄ molecule was adjacent to the Fe–N₄ active site. The DFT calculation and electrochemical analysis results showed that Pt–N₄ molecules can activate the O₂ adsorbed on the Fe–N₄ active sites, and also re-hybridize the 3d orbitals of Fe and 2p orbitals of O, thereby optimizing the adsorption of oxygen intermediates and accelerating the entire ORR dynamics process. Simultaneously, as for comparison, Co–N₄/Pt–N₄@NC and Mn–N₄/Pt–N₄@NC catalysts were also prepared. The results showed that the existence of Pt–N₄ molecules had no effect on the ORR process of Co–N₄/Pt–N₄@NC, but inhibited the ORR process of the Mn–N₄/Pt–N₄@NC catalyst. The main reason for this was due to the existence of Pt–N₄ molecules. There was a large free energy of adsorption between the active site of Mn–N₄ and *OH, which made it difficult to desorb from the active site and reduced the ORR catalytic performance.¹³⁰ These characteristics enabled the bimetallic SACs to follow the 4e⁻ pathways during the ORR reaction, representing an enhanced catalytic performance (Table 1).^112,131,132

Table 1 Summary of various Fe-SACs used in ORR catalysis

Support	Active site	E 1/2 (V)	E onset (V)	Ref.
The symbol “—” indicates no results were reported in the study.
N-Doped carbon	Fe–Nx	0.84	0.96	119
N,S co-doped carbon	Fe–S4N2	0.90	—	15
N,S co-doped carbon	Fe–Nx	0.97	—	336
N,P co-doped carbon	Fe–N4	0.86	—	122
N,S,P co-doped carbon	Fe–N4	0.91	—	124
N,P co-doped carbon	Fe–N3P	0.87	0.94	337
N,Cl co-doped carbon	Fe–Cl1N4	0.92	—	338
N,B co-doped carbon	Fe–Nx	0.84	0.97	118
N,F co-doped carbon	Fe–Nx	0.82	0.90	120
N-Doped carbon	OFeN4–O–FeN4O	0.83	0.93	113
N-Doped carbon	Fe–N4	0.85	—	117
N-Doped carbon	Fe–N5	0.74	0.86	125
Multiwalled CNTs	Fe–N5	0.92	—	126
N,F co-doped porous carbon	Fe–N5	0.88	0.99	127
N-Doped carbon	Fe–N2	0.86	1.05	339
N doped carbon	Fe–N2	0.93	0.99	340
N,S co-doped carbon	Fe–N3S	0.84	0.96	341
N-Doped porous carbon	N3Fe–CoN3	0.86	1.06	112
N-Doped carbon	Co–N4, Fe–N4	0.86	0.96	114
N-Doped carbon	N3Fe–CoN3	0.95	0.95	132
N-Doped carbon	Fe–N4	0.89	1.01	342
N-Doped carbon	Fe–N4, Mn–N4	0.90	—	343
N,S co-doped carbon	Fe–N4, Mn–N2S2	0.90	—	344
N-Doped carbon	Fe–N4, Co–N4	—	0.93	345
N-Doped carbon	Fe–N4	0.86	—	346
N-Doped carbon	Fe–N4	0.91	—	14
N-Doped carbon	Fe–N4	0.87	1.01	52
N-Doped carbon	Fe–N4	0.91	1.05	347
N-Doped carbon	Fe–N4	0.88	—	65
N-Doped carbon	Fe–Nx	0.82	—	54
N-Doped carbon	Fe–Nx	0.87	1.05	69

---

Due to the paramagnetic properties of ground-state oxygen in oxygen-catalyzed reactions, its production, electron transport during reduction reactions, and orbital interactions between intermediates and catalysts are closely related to the spin state of the catalyst materials, and thus the kinetics and thermodynamics of the reaction are also very sensitive to the spin configuration. In a typical experiment, Wu et al.¹³³ found that applying a 2% shrinkage stress to FeN₄ could move the 3d orbital of Fe in the center of the Fe–N bond shrinkage, promote the adsorption capacity of FeN₄ for O₂ molecules, and then improve the ORR reaction (Fig. 8a and b). Using the local spin configuration to analyze the ORR performance on FeN₄, it can be seen from the DOS diagram that after applying stress, the self-selected density increased and the Fe–N bond contracted. Besides, the iron ion changed from an intermediate spin state to a high spin state (Fig. 8c), resulting in a wider spin channel (Fig. 8d) and a larger adsorption bond order for O₂ (Fig. 8e). All these factors contributed to both the kinetic and thermodynamic aspects of the ORR.¹³⁴ Kang et al.¹³⁵ found that the magnetic moment of FeN₄ and N-rich FeN₄ materials had a linear relationship with the ORR reaction descriptor (ΔG_*OH), and the magnetic moment of 0.25 μ_B corresponded to the optimal ΔG_*OH, while the magnetic moment value of the original FeN₄ site was close to 2 μ_B, corresponding to a too low value of ΔG_*OH. Consequently, its ORR performance was relatively inferior (Fig. 9a–c). From the perspective of spintronics, the reduction of the magnetic moment revealed that the spin state of the iron site changed from a high-spin state to a low-spin state. After analyzing the spin–orbit interaction between the iron site and *OH, it was found that the FeN₄ material with a smaller magnetic moment will show a smaller bond level for the adsorption of *OH, and thus the adsorption energy of *OH will also be weaker. The value of ΔG_*OH will also increase, which corresponds to a better ORR performance (Fig. 9d and e).¹³⁴

Fig. 8 Spin-related charge transfer and orbital interactions of the ORR on FeN₄ with and without 2% Fe–N bond contraction. (a) Density of states (DOS) of the 3d orbitals of the central Fe cation in the two FeN₄ sites. 0% and −2% denote pure FeN₄ site and FeN₄ site with 2% Fe–N bond contraction, respectively. (b) Adsorption configuration of O₂ adsorption on two FeN₄ sites. (c) Possible Fe spin configurations in FeN₄. (d) Spin channel of Fe site in FeN₄. (e) Orbital interactions between Fe sites and the triplet O₂ molecule. (a and b) Reprinted with permission from ref. 132. Copyright 2019, John Wiley & Sons, Inc. (c–e) Reprinted with permission from ref. 133. Copyright 2020, John Wiley & Sons, Inc.

Fig. 9 Spin-related orbital interactions on pristine FeN₄ and spin-engineered FeN₄. (a) Representative configurations of FeN₄ with different amounts of surrounding graphitic N atoms. (b) Overpotential (η_ORR) as a function of *OH adsorption free energy (ΔG_*OH). (c) Linear relationship between ΔG_*OH and on-site magnetic moments (namely, the spin) of the Fe center in FeN₄ sites. (d) Possible spin state of the Fe center with high and low magnetic moment. (e) Orbital interactions between Fe center and *OH intermediate. (a–c) Reprinted with permission from ref. 135. Copyright 2019, John Wiley & Sons, Inc. (d and e) Reprinted with permission from ref. 134. Copyright 2020, John Wiley & Sons, Inc.

Transition metals (M = Mn, Fe, Co, Ni, etc.) possess 3d unoccupied orbitals, which can accommodate foreign electrons and reduce the bonding strength between OOH* and O*/OH* intermediates.¹³⁶ In the ORR, the activity of the catalyst is mainly affected by its electronic structure. The catalyst boosted the formation of M^(m+1)+–O₂²⁻ bonds on the its surface, ensuring the rapid replacement of O²⁻/OH⁻ and the regeneration of OH⁻.¹³⁷ Based on the coordination environment, Fe(III) possessed multiple spin states, including low-spin t⁵_2g e⁰_g, medium-spin t⁴_2g e¹_g, and high-spin t³_2g e²_g.¹³⁸ The low-spin electron configuration was d²_xy d²_yz d¹_xz. No electrons occupied the antibonding orbitals, which resulted in a strong M^m+/O² interaction and stable M^(m+1)+–O₂²⁻ bond, making it difficult to realize the M^(m+1)+–O₂²⁻/M^m+–OOH transition.¹³⁹ The electron configuration of high spin was d¹_xy d¹_yz d¹_xz, and the high e_g filling caused poor adsorption capacity and poor performance.¹⁴⁰ The mid-spin electron configuration was d_xy² d¹_yz d¹_xz and the single d_z² electron in the mid-spin state could easily penetrate the anti-bonding π orbital of oxygen, showing high ORR activity.¹⁴¹ Therefore, by improving the surrounding chemical environment of transition metals and modulating their electron spin states, the electronic structure of the catalytic site can be precisely regulated, and high-performance ORR catalysts with tunable activity are expected to be obtained. For example, Zhang et al.¹⁴² prepared Fe, Mn/N–C electrocatalysts with dispersed Fe, Mn bimetallic atoms for the ORR process. Magnetic measurements showed that the introduction of Mn–N structures led to the delocalization of Fe^III electrons and the spin state transition of Fe^III from low spin (t⁵_2g e⁰_g) to medium spin (t⁴_2g e¹_g), which can facilitate the penetration of the antibonding π orbital of oxygen. DFT calculation revealed that the Mn–N groups dispersed by adjacent atoms can effectively activate the Fe(III) site through the spin state transition and electronic regulation. Consequently, the Fe, Mn/N–C catalyst could moderately interact with oxygen with an optimized bond length and adsorption energy, which promoted the kinetic process of the ORR reaction (Fig. 10).

Fig. 10 Regulating Fe-spin state by atomically dispersed Mn–N in Fe-SACs with high oxygen reduction activity. Optimized structure of (a) Fe/N–C and (c) Fe, Mn/N–C. Optimized atomic structures for the main process of the ORR: (b) Fe/N–C and (d) Fe, Mn/N–C. (e) Pathways for Fe, Mn/N–C are summarized at U = 0 V, 0.72 V, and 1.23 V, respectively. (f) pH-corrected free energy diagram of Fe, Mn/N₆⁻¹. Reprinted with permission from ref. 142. Copyright 2021, Springer Nature Limited.

Fe-SACs have been widely used for the ORR due to their abundant atomically dispersed active sites (Fe–N₄).^143,144 However, although they exhibit high ORR activity, they are prone to demetallization, surface carbon oxidation and carbon corrosion caused by Fenton reactions under acidic conditions, which reduce their stability.^145–148 Therefore, developing ORR catalysts with high activity and stability in acidic media remains a significant challenge.

4.1.1.2. OER process. In contrast to the ORR process, the OER process is its reverse reaction.¹⁴⁹ It mainly occurs in water splitting or the rechargeable process of metal–air batteries. Given that the OER process involves 4e⁻ transfer steps, its kinetics are sluggish, where a large overpotential is generally required, resulting in a low energy utilization rate.¹⁵⁰

For the OER process, which electrochemical pathway will occur depends on the medium conditions. In a neutral and alkaline media, the overall reaction equation is as follows: 4OH⁻ → O₂ + 2H₂O + 4e⁻ (formulas (8)–(11)). In acidic medium, the reaction equation is 2H₂O → O₂ + 4H⁺ + 4e⁻ (formulas (12)–(15)).¹⁵¹ The following steps involve multiple proton-coupled electron transfers. In general, the OER process requires an overpotential of 1.23 V.¹⁷

In a neutral or alkaline medium (Fig. 11):

OH⁻ + * → OH* + e⁻. (8)

OH* + OH⁻ → O* + H₂O + e⁻ (9)

O* + OH⁻ → OOH* + e⁻ (10)

OOH* + OH⁻ → * + O₂ + H₂O + e⁻ (11)

In an acidic medium (Fig. 12):

H₂O + * → OH* + H⁺ + e⁻. (12)

OH* → O* + H⁺ + e⁻ (13)

O* + H₂O → OOH* + H⁺ + e⁻ (14)

OOH* → * + O₂ + H⁺ + e⁻ (15)

Recent studies have shown that the catalytic performance of OER active catalysts is related to the adsorption free energy of the active intermediates (OH*, O* and OOH*). For example, Man et al.¹⁵² plotted the negative value (−η) of the theoretical overpotential of various metal oxide catalysts in the OER process and the standard free energy of the (ΔG_O* − ΔG_OH*) step to obtain a volcano plot (Fig. 13). The interactions between the active intermediates and the catalytic sites play an important role in the catalytic performance. Importantly, the strong or weak interaction will be detrimental to the catalytic performance of the catalyst. It can be seen from the volcano plot that during the OER process, the theoretical overpotentials (η) of various metal oxide catalysts and (ΔG_O* − ΔG_OH*) show good agreement with the electrocatalytic activity of water oxidation. The strong or weak interaction between the metal oxides (SnO₂, VO₂ and CrO₂) at the bottom of the volcano plot and the reaction intermediates (OH*/O*/OOH*) result in a high overpotential for water oxidation, while the metal oxides (RuO₂, Co₃O₄, and PtO₂) at the apex of the volcano plot have the most suitable interaction with the reaction intermediates. When the metal at the apex is used as an OER catalyst, the OER overpotential is the lowest and close to zero. In this regard, according to the volcano plot, the most favorable catalysts for the OER process are IrO₂, RuO₂, and Co₃O₄. Thus far, the most widely used OER catalysts are the oxides of precious metals Ru and Ir. Among them, RuO₂ has high OER activity, but RuO₂ easily forms a high oxidation state and dissolves in the electrolyte under the high potential of the anode, leading to a low stability performance.¹⁵³ Moreover, because the precious metals Ru and Ir have low Earth reserves, high cost, and poor stability, there is an urgent need for non-precious metal OER electrocatalysts to replace RuO₂ catalysts. Accordingly, considering the abundant reserves of Fe in the Earth, it has been reported that Fe-SACs can be used as OER electrocatalysts to replace precious metals due to their high electrical conductivity, superior chemical stability, and ability to react with oxygen.¹⁵⁴

Fig. 11 4e⁻ reaction pathways of the OER processes on single-atom iron catalysts in neutral or alkaline medium.

Fig. 12 4e⁻ reaction pathways of the OER processes on the single-atom iron catalysts in acidic medium.

Fig. 13 Activity trends for the OER on various metal oxide catalysts. The negative value of the theoretical overpotential is plotted against the descriptor for the OER (the standard free energy of HO* oxidation). Reprinted with permission from ref. 152. Copyright 2012, The Royal Society of Chemistry.

Under alkaline conditions, there are generally two variables for the OER test of the catalyst. The first is the overpotential. The smaller the overpotential of a catalyst is, the higher its OER performance. The other is the Tafel slope. When studying the OER dynamics, the Tafel plots is obtained from the corresponding polarization curve at low overpotential, and then the Tafel equation can be obtained by linear fitting.⁶¹ Studies have shown that the RDS can be determined based on the Tafel slope. If the Tafel slopes of two catalysts are approximately the same, they have a similar RDS in the OER process. When the Tafel slope is less than about 80 mV dec⁻¹, this indicates that in the specific path of the OER, the step after the first electron transfer is the RDS, which is OH* + OH⁻ → O* + H₂O + e⁻. The OH surface species are rearranged by surface reaction. When the Tafel slope is greater than 100 mV dec⁻¹, the RDS is the formation of the OH* intermediate in the first electron transfer step.^61,155,156

Research results showed that when Fe, Co, and Ni are used as catalysts, the OER catalytic effect follows the order of Ni > Co > Fe. According to DFT calculation, it was found that N doping will cause the redistribution of the charge of the carbon material.¹⁵⁷ In the N-doped carbon catalyst, the C atom close to the N atom became the preferential binding site of the oxygen intermediate in the OER process, rather than the doped N atom itself. Therefore, for SACs, the M–N_x metal center and the C atom close to the N atom may act as active sites. Oxygen intermediates such as OH*, O*, OOH* and O₂* will be generated in the OER process. Different intermediates may have different adsorption positions in the catalyst. Some of them are adsorbed on the M–N_x metal active center, while others will be adsorbed on the C atoms close to the N atom. This adsorption ability is mainly related to the difference in the adsorption energy between the oxygen intermediate and the two positions. In this context, whether the C atom will become the adsorption site of the oxygen intermediates in the OER process largely depends on the number of d electrons of the metal. For the Fe and Co transition metals, their 3d electron numbers are 6 and 7, respectively. Therefore, all the oxygen intermediates are more strongly bound to the M–N_x active site in the metal center than the C site. In the case of SACs synthesized from Fe and Co metals, the two intermediates (O* and OH*) preferentially combine with the C atom close to the N atom, and the OOH* intermediate preferentially adsorbs on the metal centers (Fig. 14a and b). The limit barrier of the OER reaction is a critical parameter to evaluate catalytic activity (Fig. 14c). According to the DFT calculation results, for Fe-SACs, the RDS is O* + OH⁻ → OOH* + e⁻, with a limit barrier of 0.97 eV. The large limit barrier indicates that the chemical adsorption between the oxygen intermediate and the active site is strong, thereby increasing the activation barrier during the reaction. When the current density was 10 mA cm⁻², the overpotentials of the Fe, Co, and Ni catalysts were 488 mV, 402 mV, and 331 mV, respectively (Fig. 14d). The Tafel slope can be using the Tafel plot. Because the overpotential and Tafel slope of metal catalysts both show the trend of Fe > Co > Ni, the OER activity of the single-atom catalysts follow the trend of Ni > Co > Fe (Fig. 14e). Fe will yield a semiconducting Fe–OOH* species with the O₂ intermediate of OOH* in the OER reaction, hence reducing the electron transfer rate. Moreover, the bonding strength between the Fe active sites and OOH* causes the deactivation of these active sites, which directly inhibit the OER activity.¹⁵⁸ Therefore, to improve the catalytic performance of Fe-SACs, the most effective method is the doping of Ni or Co atoms. It is widely acknowledged that the doping of Co atoms is beneficial to improve the conductivity of catalysts and can also activate the inactive metal atoms. Thus, the OER activity is increases obviously.^{149,159–161} Typically, Thomas’ group synthesized hierarchically mesoporous/microporous carbon nanosheets with highly dispersed Fe/Co bimetallic centers by heteroatom doping (Fe/Co–N–C).¹⁶¹ This synthesis method employed FeCl₃·6H₂O as the active salt and SiO₂ as a template to adjust the porosity and thickness of the carbon nanosheets. A large number of mesoporous/microporous structures was generated with a specific surface area of 1321 m² g⁻¹ (Fig. 5d), which was beneficial for the diffusion of electrolyte and gas during the ORR/OER process, thereby enhancing the catalytic activity. The authors measured the ORR and OER activity of the catalyst meso/micro-Fe/Co–N–C by rotating disk electrode technology in alkaline media. It had a high onset potential (E_onset = 0.954 V) and half-wave potential (E_1/2 = 0.886 V). Moreover, due to the rich mesoporous/microporous structure, this catalyst showed a high limit current density, reaching 6.3 mA cm⁻², which was higher than that of the commercial Pt/C catalyst (5.6 mA cm⁻²). It can be concluded from the OER polarization curve that meso/micro-Fe/Co–N–C had a lower potential of 1.67 V when the current density was 10 mA cm⁻². According to the Tafel plot corresponding to the OER diagram, the catalyst was demonstrated to show a lower voltage gap value (E_gap = 0.78 V) at a current density of 10 mA cm⁻², indicating the good ORR/OER bifunctional activity of meso/micro-Fe/Co–N–C (Fig. 5e and f). As is known, if a catalyst has high OER activity, it will have a moderate affinity with the OH intermediate and can improve the adsorption of hydroxide and the equilibrium adsorption rate of the catalyst surface. Consequently, the last step of the OER path generation of oxygen becomes a rate-determining step, which ultimately reduces the Tafel slope. Obviously, in addition to the introduction of Co atoms in Fe-SACs, Ni atoms can also be co-doped, which is conducive to the formation, association and dissociation of OOH*.¹⁶²

Fig. 14 (a) Possible reaction schemes for M–N₄ active site during the OER process. (b) Possible reaction scheme of the dual-center mechanism for the catalyst with MN₄ active site in the OER process. (c) Free energy diagrams of Fe-NHGF, Co-NHGF and Ni-NHGF and other single-atom catalysts in the OER process. (d) Evaluation diagram of OER activity of NHGF, Fe-NHGF, Co-NHGF and Ni-NHGF in 1 M KOH electrolyte. (e) Tafel plots of the corresponding catalysts shown in d. (f) TOF value of Ni-NHGF catalyst and other metal catalysts. Reprinted with permission from ref. 157. Copyright 2018, Springer Nature Limited.

In metal–air batteries or fuel cells, dual-functional electrocatalysts are usually required to drive multiple electrochemical reactions simultaneously, but a single atom active center with only a single metal component is difficult to satisfy high selectivity and activity for different electrocatalytic reactions. As can be seen from the above discussion, among the SACs of Fe, Co, and Ni, the Fe-SACs (Fe > Co > Ni) are the most catalytically effective for the ORR process, and the Ni-SACs (Ni > Fe > Co) are the best for the OER process. Therefore, it is a direct way to develop dual-functional electrocatalysts by integrating two single-atom catalysts on a support. In principle, the catalysts of the respective atoms correspond to their respective catalytic processes, making them not only have excellent ORR catalytic effect, but also have excellent OER catalytic effect. Recently, Chen et al.¹⁶³ embedded Ni atoms and Fe atoms in the inner and outer walls of graphene hollow nanospheres through a step-by-step self-assembly strategy. Given that the two active centers of Ni–N₄ and Fe–N₄ were distributed on different sides of the graphene hollow nanospheres in the catalyst, the obtained SACs could balance the competition between the ORR and OER rate-limiting steps by reducing the mutual interference of ORR and OER active centers, showing higher OER and ORR catalytic activities. According to electrochemical tests and density functional calculations, the outer Fe–N₄ active center had a higher catalytic performance for the ORR process, while the inner Ni–N₄ active center could boost the OER reactions. During the OER process, the RDS was the first discharge step, corresponding to the formation of the OH* intermediate (Fig. 15a). Moreover, the E_gap was measured to be about 0.79 V, suggesting the excellent activity and electrode reversibility as a bifunctional oxygen catalyst. According to the d-band center theory, the center value of the d-band farther from the Fermi level indicated that the adsorption capacity was weaker.¹⁶⁴ By measuring the projected density of states (PDOS) of the single atom sites of Ni and Fe, the binding between the Ni–N₄ active center and oxygen was much weaker than that with the Fe–N₄ active center, which resulted in the weaker ORR activity of the Ni–N₄ active center, but its OER activity was better than that of the Fe–N₄ active center (Fig. 15b and c). After the catalyst was assembled into a zinc–air battery, it showed excellent electrode efficiency and cycle stability.

Fig. 15 (a) DFT-optimized adsorption configurations of the reaction adsorbates. (b) DFT-calculated reaction free energies of oxygen electrocatalytic reactions (U = 0 V) and (c) PDOS of the single-atom sites of Fe–N₄ and Ni–N₄. The black dashed line in (c) indicated the Fermi level. The red and blue vertical bars represent the calculated d-band centers of Fe and Ni, respectively. Reprinted with permission from ref. 161. Copyright (2020), John Wiley & Sons, Inc.

Due to the excessively high spin state of iron oxides (filling amount is about 2.0), the binding strength of iron oxides to the reaction intermediates is inhibited, resulting in unsatisfactory OER catalytic activity. Thus, to solve this issue, Shen et al.¹⁴⁰ anchored Fe(III) atoms on ultrathin TiO₂ nanoribbons through an adsorption-oxidation strategy (the surface of ultrathin TiO₂ nanoribbons was rich in –OH as sites to anchor Fe and prevent aggregation), thereby synergistically reducing the spin state of Fe (1.08). The formed σ bond with oxygen-containing intermediates was stronger during the catalytic OER process, which could enhance the adsorption of oxygen-containing intermediates (Fig. 16a–c). Meanwhile, orbital hybridization of Fe 3d and Ti 3d with electron delocalization boosted the electrical conductivity of the nanoribbons, thereby reducing the ohmic loss. All these factors together resulted in excellent OER activity (Fig. 16d–f).

Fig. 16 DFT calculations. (a–c) Side view model for DFT calculation of Fe-TiO₂ (a), Fe₂O₃ (b) and TiO₂ (c) models after structural optimization. (d) TDOS and projected DOS, and the dashed lines represent the d-band centers of Ti 3d and Fe 3d. (e and f) Projected Fe 3d (e) and Ti 3d (f) DOS of selected atoms in the models. Reprinted with permission from ref. 140. Copyright 2020, John Wiley & Sons, Inc.

4.1.1.3. HER process. Because hydrogen energy (H₂) can provide energy during the combustion process and emit zero CO₂, it has become a substitute for fossil fuels to reduce the greenhouse effect and solve the energy crisis.^165,166 Electrochemical water splitting is a clean and sustainable method for the preparation of hydrogen, which is the combination of two semi-reactions (Fig. 17), that is, the HER at the cathode and OER at the anode.^17,41,167 Under standard conditions, the Gibbs free energy of water splitting is 273.2 kJ mol⁻¹ with an equilibrium potential of 1.23 V.¹⁶⁸ However, in practical applications, due to some kinetic obstacles, the equilibrium potential required for electrochemical water splitting is greater than 1.23 V, forming an overpotential. Thus, to accomplish water splitting, the use of catalysts is required to overcome this large equilibrium potential.¹⁶⁶

Fig. 17 Possible e⁻ reaction pathways for the HER processes on single-atom iron catalysts.

The HER is a 2e⁻ transfer reaction, which takes place on the electrode surface. The first step is proton-coupled electron transfer on the catalyst surface and form an H* intermediate at the active centers (Volmer–Heyrovsky or discharge reaction, formula (16)). The second step is divided into two pathways according to the amount of active H* intermediate on the electrode surface. If the number of H* is small, a single H* will combine with H⁺ and electrons to form H₂ (formula (17)). Otherwise, if there are massive amounts of active H* intermediates on the electrode surface, two adjacent H* will recombine to form H₂ (formula (18)).^169,170

The Volmer–Heyrovsky reaction mechanism is as follows:

H⁺ + e⁻ + * → H* (16)

H* + H⁺ + e⁻ + * → H₂ + * (17)

The Volmer–Tafel reaction mechanism is as follows:

H⁺ + e⁻ + * → H* (16)

2H* + H₂ + 2* (18)

Regardless of the reaction pathway, H* is the only chemical reaction intermediate. In this case, the free energy of H adsorption (ΔG_H*) can be utilized as a scalar to evaluate the HER performances.^17,41 Generally, catalysts with ΔG_H* infinitely close to zero have the highest HER activity.^165,171 To date, it is very difficult to clarify the specific working principles of different HER electrocatalysts. However, the Tafel slope can represent the rate-determining step (RDS), providing valuable insights into possible reaction pathways. Therefore, when measuring the HER performance of a catalyst, the Tafel slope is an indispensable parameter. Based on the Tafel slope, the reaction RDS can be determined, inferring the reaction mechanism of the catalyst used in HER catalysis. According to the Butler–Volmer kinetics, the RDS can be divided into three categories. The first type is that when the second step of the HER process, corresponding to the Tafel step (2H* → H₂ + 2*), which is the RDS, the Volmer reaction is very fast. The resulting Tafel slope is 30 mV dec⁻¹ at 25 °C. The second type is that when the Heyrovsky step (H* + H⁺ + e⁻ + * → H₂ + *) is the RDS, the reaction rate of the first step is also very fast, and its Tafel slope is 39 mV dec⁻¹ at 25 °C. The third category is the Volmer reaction (H⁺ + e⁻ + * → H*) as the RDS. At this time, the discharge of H⁺ is very slow, and hydrogen will be released through the recombination reaction step or the Tafel step. The resulting Tafel slope is 118 mV dec⁻¹ at 25 °C.¹⁷²

Multiple lines of evidence support that Fe-SACs have high catalytic activity for the ORR (refer to Section 4.1.1), but their catalytic performance for the HER process is still relatively mediocre.¹⁶⁶ Thus, it is necessary to dope Fe-SACs with other transition metals to manipulate the electronic structures of the Fe active sites, thereby improving their catalytic activity for the HER process. Zeng et al.¹⁰⁴ grafted a single Pt atom on the active site of Fe–N₄ by bridging O₂ to form a single atom catalyst with the structure of Pt₁–O₂–Fe₁–N₄ (Fig. 5g–i). Fe/Pt-SACs exhibited outstanding HER catalytic activity with an onset overpotential of about 15 mV, which was obviously less than that of the commercial Pt/C catalyst. Moreover, the Tafel slope of Fe/Pt-SACs was determined to be 42 mV dec⁻¹, which is very close to that of the Pt/C catalyst (38 mV dec⁻¹), revealing that HER process of Fe/Pt-SACs followed the Volmer–Heyrovsky reaction mechanism. Interestingly, Fe/Pt-SACs had a fast electron transfer rate based on the higher exchange current density of 0.039 mA cm⁻² (Fig. 5h), boosting the HER kinetics. Grafting Pt₁–O₂ to the Fe–N₄ active center afforded Fe with a spatial and electronic effect, which can facilitate the dissociation of water, thereby enhancing the HER dynamics. According to DFT calculation (Fig. 5i), the ΔG_H* of Fe/Pt-SACs was about 0.16 eV, which is equivalent to that of the commercial Pt/C catalyst (∼0.09 eV).

When preparing Fe-SACs, the morphology and structure of the catalyst should be optimized. If the synthesized catalysts have a high specific surface area and density of active centers, more active catalytic centers will be accessible. Furthermore, the HER catalytic activity will be greatly improved by considering the following aspects: (1) suitable porosity between adjacent structures of the catalysts to achieve charge transport and electrolyte penetration into the deep active center of the catalyst surface and (2) higher crystallinity and appropriate topological defects to increase the electron density and conductivity of the material.

4.1.1.4. CO₂RR process. The amount of CO₂ in the air is increasing gradually, resulting from the massive consumption of fossil fuels, and the greenhouse effect-associated global warming has triggered a series of global issues. Luckily, CO₂ can be reduced to various C₁ products, including HCOOH, HCHO, CH₃OH, CH₄ and CO, under the catalysis of SACs.^173,174

Experiments and theories show that the reaction mechanism of CO₂ to form CO through the reduction process is as follows: CO₂ produces a COOH* intermediate by a proton-coupled single electron transfer reaction (formula (19)), which, in turn is transformed to the CO* intermediate and H₂O (formula (20)). The last step is the desorption of CO at the active sites (formula (21)).^41,175 A key debate regarding the CO₂RR mechanism is that the generation of the COOH* intermediate requires two steps, namely, the proton-uncoupled electron transfer mechanism.^91,176

The proton-coupled electron transfer mechanism (Fig. 18) is as follows:

CO₂ + H⁺ + e⁻ + * → COOH* (19)

COOH* + H⁺ + e⁻ → CO* + H₂O (20)

CO* → CO + * (21)

Fig. 18 Proton-coupled electron transfer mechanism of CO₂RR processes on single-atom iron catalysts.

Recent evidence demonstrated that the Fe in Fe-SACs carries a large positive charge with a valence state between 0 and +3, making it beneficial to effectively convert CO₂ into CO during the CO₂RR process. For example, Gu et al. used ZIF-8 as a support to synthesize Fe-SACs. The K-edge XANES spectra indicated that the Fe in the catalysts showed a +3 valence state. The CO₂RR experiment was carried out in 0.5 M KHCO₃ saturated with CO₂. When the overpotential was 0.34 V, the partial current density reached 94 mA cm⁻², and the Faraday efficiency of the cathode CO was calculated to be higher than 90% (Fig. 19a). Due to the presence of pyrrole N ligands, in the process of electrocatalysis, the Fe in the catalysts maintained the oxidation state of +3, and thus these Fe-SACs outperformed their counterparts in terms of catalytic activity and stability. To validate the influence of the oxidation state of Fe on the CO₂RR catalytic effect, the Tafel slopes of Fe²⁺–N–C and Fe³⁺–N–C catalysts were determined to be about 117 and 71 mV dec⁻¹, respectively (Fig. 19b), suggesting that the Fe³⁺ site had a faster CO₂ adsorption rate than the common Fe²⁺ site. Moreover, when some deionized water or tap water was used to replace KHCO₃, its Faraday efficiency and j_CO almost remained constant, lasting at least 12 h (Fig. 19c).¹⁷⁶ Furthermore, the type and number of N coordinated with the Fe atom have been shown to affect the catalytic activity of Fe-SACs.^177,178

Fig. 19 Applications of Fe-SACs in energy storage. (a) j_CO and (b) Tafel plots of Fe³⁺–N–C in an H-cell (red) and on a gas diffusion electrode (blue), and of Fe²⁺–N–C in an H-cell (black), respectively. (c) Chronoamperometry curve and faradaic efficiency of CO production (dots) by Fe³⁺–N–C in H-cell at −0.37 V versus RHE. Reprinted with permission from ref. 176. Copyright (2019), the American Association for the Advancement of Science. (d) Mechanism diagrams of SACs with Fe–N₄ active sites in the NRR process. Reprinted with permission from ref. 193. Copyright (2016), the American Chemical Society. (e) N₂ adsorbed on the active site in an end-on and side-on structure and corresponding Bader charge distribution of *N₂. (f) Gibbs free energy diagram of Fe–(O–C₂)₄ catalyst for the NRR process through distant and alternating pathways and (g) enzymatic and consecutive pathways. Reprinted with permission from ref. 95. Copyright (2020), John Wiley & Sons, Inc.

Not only the oxidation state of iron in Fe-SACs will affect the CO₂RR process, but also the reaction temperature in the electrochemical reduction of CO₂ will affect the catalytic effect of Fe-SACs. In a typical experiment, Lin et al.¹⁷⁹ separately synthesized atom-dispersed iron and nickel anchored on nitrogen-doped carbon supports and studied the temperature in the range of 303 K and 343 K. At a higher overpotential, increasing the reaction temperature increased the CO Faraday efficiency and current density of the Fe-SACs. When the temperature increased from 303 K to 323 K, the CO current density of the Fe-SACs increased rapidly from 114.4 mA cm⁻² to 179.7 mA cm⁻². With an increase in temperature to 343 K, the CO current density in this process increased slowly, and then plateaued. The final CO current density was 185.8 mA cm⁻². In contrast, increasing the temperature inhibited the CO Faraday efficiency of Ni–N–C. When the temperature increased from 323 K to 343 K, the CO current density showed a downward trend from 252.2 mA cm⁻² to 202.9 mA cm⁻². The DFT calculation results indicated that an increase in temperature accelerated the reaction kinetics. Moreover, the difference in the CO₂RR process of these two catalysts, i.e., Fe-SACs and Ni-SACs, was mainly due to the difference in the adsorption strength between the key reaction intermediates and the active sites. Therefore, the reaction temperature was also one of the main factors affecting the catalytic performance of Fe-SACs in the CO₂RR process. However, Sun et al.¹⁸⁰ introduced a single P atom in the form of a P–C bond in a nitrogen-doped carbon-supported single iron atom catalyst for the CO₂RR process. Due to the addition of P atoms, the electron density of Fe in the active center increased. Accordingly, the formation of the *COOH intermediate was promoted, making it have a superior CO₂RR catalytic performance under a lower overpotential. Under a low overpotential at 320 mV, the CO Faraday efficiency of Fe-SAC/NPC was maintained at 97% within 24 h. The current density of CO was stable at 5 mA cm⁻² and it had a low Tafel slope (59 mV dec⁻¹).

A good CO₂RR catalyst should not only be able to balance the adsorption energy of the *COOH and *CO intermediate products, but also be beneficial for the desorption of CO. The coordination environment around the active center of Fe–N_x, e.g., the type of N coordinated with the Fe atom and the number of N atoms coordinated with the Fe atom, affected its catalytic activity. Studies have shown that the Fe–N₃ active center had a high CO₂RR catalytic activity and can effectively convert CO₂ to CO. In a typical experiment, Wang et al.¹⁸¹ synthesized a single Fe atom uniformly dispersed N-doped porous carbon support via a gas diffusion strategy (Fe₁NC/S_X–Y, X and Y represented the particle size of ZIF-8 precursor and pyrolysis temperature, respectively). When the particle size of ZIF-8 was 100 nm and the temperature was 1000 °C, the prepared Fe₁NC/S₁-1000 catalyst had the best CO₂RR catalytic activity (Fig. 20a). At −0.6 V, the partial current density of CO of Fe₁NC/S₁-1000 was 6.8 mA cm⁻², which was about 1.2, 2.0 and 86 times that of Fe₁NC/S₁-800, Fe₁NC/S₁-900 and NC/S₁-1000, respectively (Fig. 20b). At −0.65 V, the calculated turnover frequency was 2225 h⁻¹, and its catalytic effect was comparable to that of Fe³⁺–N–C catalysts (Fig. 20c). To study the coordination environment of Fe atoms (type, content and number of coordinated N atoms), the content of various N and the corresponding FE of CO at different temperatures were measured. When the temperature was set as 800 °C, 900 °C and 1000 °C, the coordination atoms of the prepared catalysts were 3.8, 3.3 and 3.0, respectively. Moreover, the content of graphite N was 0.55, 1.11 and 1.32 wt%, respectively. Notably, the FE of CO was 81.8%, 90% and 95.6%, respectively. Simultaneously, the authors also constructed three structural models of FeN₄, FeN₃ (embedded in the graphene plane, and completely encapsulated by C atoms), and FeN₃V (V represents vacancies) for density functional theory calculations (Fig. 20d and e). In the CO₂RR process of the three models, only the process of converting *COOH to *CO was exothermic. Consequently, the catalytic activity of each model could be evaluated by calculating the two steps of CO₂ conversion to *COOH and CO desorption from the active sites. Among the three models, the FeN₃V model had the best catalytic activity and could balance the adsorption energy of the *COOH and *CO active intermediates, which was beneficial for the formation of CO. The above-mentioned experimental results and theories demonstrated that the synergy between graphite N and Fe–N₃ species improved the CO₂RR catalytic activity of the catalyst. The graphite N can redistribute the charge around the Fe atoms and promote the transport of electrons.

Fig. 20 (a) FE_CO and (b) J_CO of Fe₁NC/S₁-Y series and NC/S₁-1000. (c) TOFs of Fe₁NC/S₁-1000 compared to that of other reported nonprecious metal/N-doped carbon-based SACs: Fe-SACs, Fe³⁺–N–C, Fe–N₅, Fe_0.5d, Co–N₂, A-Ni-NSG, AD–Sn/N–C, and Co–N₅/HNPCSs. (d) Top view of the optimized atomic structures for FeN₄, FeN₃, and FeN₃V models embedded on graphene layer. The yellow, blue, and gray balls represent Fe, N, and C atoms, respectively. (e) Calculated free-energy diagrams for the CO₂ER pathway of FeN₄, FeN₃, and FeN₃V moieties. Reprinted with permission from ref. 181. Copyright 2020, John Wiley & Sons, Inc.

Both experiments and theory show that in metal SACs, their catalytic performance for the CO₂RR follows the order of Ni > Fe > Co, which is mainly related to the binding energy between each metal active site and the active intermediates during the CO₂RR process.^41,182 The onset potential in the CO₂RR processes of the catalyst with Fe/Co–N₄ active sites is relatively low, whereas it is difficult for the CO* active intermediate to desorb from metal atoms because of the high binding energy between them. The fatal consequence is the protonation of the CO* intermediates (or catalyst poisoning) and generation of other products, causing the Fe/Co–N₄ catalysts to exhibit low CO selectivity. Considering SACs with Ni–N₄ active centers, they have a higher current density, but when generating COOH* intermediates, their kinetics is slower, resulting in a decrease in the reaction rate.^183–185 For metallic nickel, in the CO₂RR process, its catalytic effect is greater than that of metallic iron. Therefore, the introduction of monoatomic nickel into a single atomic iron-based catalysts can improve their CO₂RR catalytic activity. Recently, Jiao et al. adopted a zinc-assisted atomization strategy to dope single iron and zinc atoms in MOF-derived nitrogen-doped carbon (Fe₁–Ni₁–N–C), where the Fe–N₄ and Ni–N₄ active centers were adjacent to each other. Due to the interactions between the adjacent sites, the CO₂RR catalytic activity was much larger than Fe₁–N–C and Ni₁–N–C of nitrogen-doped carbon supports modified with Fe and Ni single atoms. The CO₂ electrochemical reduction test results showed that Fe₁–Ni₁–N–C had a higher current density and exhibited the largest Faraday efficiency (96.2%) at −0.5 V, which was better than that of Fe₁–N–C and Ni₁–N–C. Meanwhile, the CO current density of Fe₁–Ni₁–NC (j_CO = 2.4 mA cm⁻²) was also greater than that of Fe₁–N–C (2.1 mA cm⁻²), Ni₁–N–C (0.1 mA cm⁻²) and N–C (0.7 mA cm⁻²). Fe₁–Ni₁–N–C had faster kinetics, and its Tafel slope (83 mV dec⁻¹) was smaller than that of Fe₁–N–C and Ni₁–N–C. In the −0.5 V state, Fe₁–Ni₁–N–C could run continuously for 10 h and the Faraday efficiency and current density of CO remained unchanged, indicating that it had high stability. DFT calculation demonstrated that the coupling of adjacent Ni and Fe atoms in Fe₁–Ni₁–N–C was beneficial for the activation of CO₂, thereby reducing the barrier for the formation of COOH* intermediates, ultimately improving the electrochemical reduction performance of Fe₁–Ni₁–N–C for CO₂.¹⁸⁶ Overall, designing bimetallic SACs may successfully negotiate the above-mentioned intrinsic barriers.¹⁷⁵ Various electrochemical pathways can be performed on the optimal active centers and the catalytic activity of CO₂RR will be improved by the synergistic effect.

4.1.1.5. NRR process. On the surface of the Earth, the content (∼78%, volume fraction) of N₂ is extremely rich, but nitrogen basically exists in the form of gaseous dimolecular N₂, which cannot be directly used by organisms or employed as industrial raw materials. Thus, nitrogen fixation technology has been rapidly developed to convert nitrogen to ammonia (biological way) and hydrazine (industrial way) or to oxidize nitrogen to form nitrogen oxide. Generally, the nitrogen fixation reaction refers to the reduction of N₂ to ammonia with high chemical reaction activity. To date, the main method for industrial synthesis of ammonia is the Haber–Bosch reaction.¹⁸⁷ However, the environmental impact (e.g., huge energy consumption and CO₂ emission) associated with this reaction is not consistent with the concept of green chemistry and sustainable development.^188,189 The electrochemical synthesis of NH₃ under mild conditions was developed, in which renewable electricity and H₂O as a hydrogen source are employed to directly convert N₂ into NH₃ by accepting electrons at the cathode (namely the NRR process, formula (22)), accompanying the OER reactions at the anode (formula (23)).

Cathode: N₂ + 6H⁺ + e⁻ → 2NH₃ (22)

Node: 3H₂O−6e⁻ → 6H⁺ + [/]O₂ (23)

There are three reaction mechanisms that are related to Fe as an electrocatalyst in the NRR reactions, namely, the distal mechanism, alternating mechanism and enzyme mechanism.^190–192Fig. 19d shows the reaction pathways when Fe acts as the active center of the NRR process.¹⁹³ Similar to the Fe–N complexes, which are effective NRR electrocatalysts, Fe-SACs also display great potential as NRR catalysts.^194,195 In 2020, Zhang et al. designed Fe-SACs with an Fe–(O–C₂)₄ structure, in which an Fe atom was coordinated with 4 O atoms instead of common N atoms. An electrochemical test was performed in alkaline solution. When the applied potential was −0.1 V, the Faraday efficiency of NH₃ reached the maximum of 29.3% with a yield of 32.1 μg h⁻¹ mg_cat.⁻¹, showing outstanding NRR catalytic activity. N₂ (bond length of 1.12 Å) adsorbs on the active center in two configurations, namely “side-on” (1.188 Å) and “end-on” (1.146 Å) (Fig. 19e). The various configurations will affect the amounts of electrons transferred from the Fe–(O–C₂)₄ structure to the adsorbed N₂, leading to different catalytic properties. As shown in Fig. 19f and g, the “end-on” adsorption occurred via the remote and alternate reaction mechanism, and the “side-on” adsorption preferred to undergo the continuous hydrogenation reaction mechanism. It is vital to note that multiple reaction mechanisms possibly participate in the NRR process due to the acceptable energy input.⁹⁵ Recently, Li et al.¹⁹⁶ synthesized a new type of NRR single-atom catalyst. They used SiO₂ nanospheres as a sacrificial template to anchor a single iron atom on N,O co-doped inverse opal carbon via high-temperature thermal addition. The structure had a high specific surface area (S_BET = 361 m² g⁻¹), which was conducive to mass transfer and electron transportation. Among them, a single iron atom was coordinated with two nitrogen atoms and four oxygen atoms. The charge transfer between oxygen atoms and iron atoms not only can promote the transportation of nitrogen, but also effectively reduce the binding energy between the active sites and *N₂. All these factors together can facilitate the desorption of NH_x, increase the electron transfer rate, and effectively accelerate the rate determination step (*N₂ → *NNH).

During the NRR reaction, a competitive reaction occurs simultaneously, i.e., the HER. The NRR is a 6-proton-coupled electron transfer process, while the HER is a 2-proton-coupled electron transfer process. Moreover, only a small overpotential is required to initiate the HER process, and thus it is the most common competitive side reaction for the NRR process. Accordingly, the redesign of electrocatalysts exhibiting a weak binding energy with hydrogen is needed to break this inevitable barrier.¹⁹⁷ Therefore, to achieve a higher binding strength between Fe-SACs and N₂ than that of hydrogen, systematic material-design methods have been recently developed for the fabrication of Fe-SACs, such as doping with heteroatoms (B, O, S, P, etc.), varying the coordination number of the Fe center atom, carbon support defect, S or N vacancies, active sites of bimetallic atoms, and others.^198,199 Recently, Wang et al.²⁰⁰ investigated the effect of the electronic state of the Fe-SAC active center on the NRR process. The NRR process occurred at the active site of FeN₄, instead of MoN₄. The MoN₄ adjacent to the FeN₄ active site can tune the spin state of the Fe atom from a high spin state to an intermediate spin state . The unpaired electrons of Fe were reduced, which was favorable for the reverse contribution of π and promoted bonding with the unpaired electrons of *N₂. Thereby, it could further accelerate the first-step hydrogenation process (Fig. 21a–c). Simultaneously, the change in spin state caused an increase in the charge density of the Fe sites, which promoted the adsorption of N₂. Meanwhile, the H intermediates adsorbed on the Mo sites can assist N hydrogenation via hydrogen flooding. The separated d-electrons and empty d-orbitals were more favorable for the overlapping of the Fe 3d and N 2p orbitals, which significantly lowered the energy barrier for the rate-determining step (Fig. 21d–g). In the NRR process, most scientists employ N₂ as a raw material to convert it into ammonia. A few of them use nitrate as a raw material to synthesize ammonia, mainly because this process involves 8 electron transfer steps and many reaction pathways produce more by-products (NO₂, NO₂⁻, NO, N₂O, N₂, NH₂OH, NH₃, and NH₂NH₂), and reducing NO₃⁻ to N₂ involves an N–N coupling step. This step generally requires two adjacent active sites for catalysis. Impressively, single-atom catalysts possess these characteristics. Wu et al. synthesized a single-atom iron catalyst for the synthesis of ammonia using NO₃⁻ as the raw material, and proved that the FeN₄ active sites have excellent activity and selectivity in the process of converting NO₃⁻ to NH₃.²⁰¹

Fig. 21 Magnetic susceptibility and ⁵⁷Fe Mössbauer spectra of the catalysts. (a) Simplified schematic of N₂ bonding to transition metals. (b) Magnetic susceptibility of FePPc. (c) Room-temperature ⁵⁷Fe Mössbauer spectrum of FeMoPPc. The data (scattered dots) are shown with the best fit (black dash line) and the deconvolution of two components, as indicated by the legend. DFT calculations of the NRR activity on FeMoPPc catalysts. (d) Different free-energy diagrams for the NRR on Fe atoms of FeMoPPc and Mo atoms of FeMoPPc. (e) Different free-energy diagrams for the HER on Fe atoms of FeMoPPc and Mo atoms of FeMoPPc. (f) Charge density difference calculations of N₂ bonded to Fe atoms of FePPc and FeMoPPc (the electron excess area and electron deficiency area are represented by yellow and blue-green, respectively). (g) Charge density difference calculations of the N₂ bonded to Mo atoms of MoPPc and FeMoPPc. Reprinted with permission from ref. 200. Copyright 2021, John Wiley & Sons, Inc.

Similar to the CO₂RR, the NRR is a reaction involving multiple reaction intermediates (including N₂ adsorption/activation and NH₃ desorption), and the HER acts as a major competing reaction, making selectivity a formidable challenge.^202–204 In addition, the relatively poor catalytic activity due to the slow adsorption of nitrogen and the high cracking energy of the N–N triple bond (940.95 kJ mol⁻¹) is also an urgent problem to be solved.²⁰⁵ Therefore, improving the catalytic activity and selectivity of Fe-SACs during the NRR process has become the main research topic. Studies have shown that metal nitrides have higher activity for the NRR at low onset potentials, while inhibiting the HER, mainly because they have a stronger binding energy for nitrogen than hydrogen.^206,207 Therefore, the combined application of metal nitrides and Fe-SACs may be a good approach to catalyze the NRR. Meanwhile, understanding of the details of the interface mechanism will also provide the necessary guidance for the implementation of more targeted strategies to design better catalysts.^208,209

4.1.1.6. H₂O₂ production. As a type of reactive oxygen species (ROS), H₂O₂ is widely used in various industries (food, biology, chemistry, environmental science, pharmaceutical preparation, papermaking, etc.) due to its strong sterilization effect and non-toxic and harmless decomposition products.^210–212 Especially in the food production process, H₂O₂ is the widely used because its use can make food beautiful, whiten it, and increase its shelf life.²¹³

Currently in industry, the most common method for the synthesis of H₂O₂ is the hydrogenation of anthraquinone with hydrogen, followed by oxidation in an organic medium.^214,215 However, this method has several disadvantages. The first is that a large amount of reaction by-products and waste will be produced. Also, a large amount of energy will be consumed. Secondly, the generated H₂O₂ has a high concentration, which is easy to decompose.²¹⁶ Therefore, the electrochemical synthesis of H₂O₂ has become an emerging technological method to replace the traditional method to prepare H₂O₂.^217,218 The electrochemical synthesis of H₂O₂ mainly has the following advantages: (1) mild reaction conditions, less requirements for temperature and environmental pressure, (2) electricity as an energy source does not require the combustion of fossil fuels and (3) use of water and air as raw materials and a green source. The electrochemical production of H₂O₂ mainly relies on the 2e⁻ process of the ORR, with the reaction equations shown in eqn (24) and (25).²¹⁹ H₂O₂ acts as a by-product in the 4e⁻ reaction process of the ORR, and thus during the reaction process, the 4e⁻ process needs to be inhibited. Therefore, it is very important to choose a suitable catalyst.

In alkaline medium: O₂ + 2H₂O + 2e⁻ → H₂O₂ + 2OH⁻ (24)

Acidic medium: O₂ + 2H⁺ + 2e⁻ → H₂O₂ (25)

An ideal electrochemical catalyst should be able to strongly adsorb O₂ to generate the oxygen-containing *OOH intermediate in the electrochemical synthesis of H₂O₂, and then the weak adsorption of *OOH on the surface of the catalyst can easily desorb and generate H₂O₂.²²⁰ However, if the adsorption of *OOH on the surface of the catalyst is relatively strong, it will stay on the surface of the catalyst for a long time, resulting in dissociation, generating *O and *OH, and finally being reduced to H₂O. Therefore, the best catalyst should have a large adsorption energy between O₂ and the smallest *OOH adsorption energy, and simultaneously have the largest kinetic barrier for the reduction/dissociation of O* and OH* by *OOH, thus making it the most selective for the electrochemical production of H₂O₂.^221,222 While Fe-SACs have excellent ORR properties, they can also produce H₂O₂via the 2e⁻ pathway.²²³ In a typical experiment, Jiang et al.²²⁴ synthesized a series of single-atom catalysts in which transition metal atoms (Fe, Pd, Co, and Mn) coordinated with C, O, and N on carbon nanotube vacancies. Among the transition metal-catalyzed production of H₂O₂, Fe-SACs exhibited the best activity and selectivity. In 0.1 M KOH, when the onset potential was 0.822 V vs. RHE, the generation current of H₂O₂ reached 0.1 mA cm⁻², and the H₂O₂ generation rate was about 43 mA cm⁻². More interestingly, under both neutral and alkaline conditions, Fe-SACs had 95% H₂O₂ selectivity. The results showed that changing the coordination structure of the active site of Fe-SACs can improve the catalytic efficiency of the 2e⁻ step of the ORR to produce H₂O₂. When the N atom was replaced by an O atom, the binding energy between the *OOH reaction intermediate and the C site on Fe-SACs was weakened, which was easily decomposed into *O and *OH. The 4e⁻ pathway was carried out to generate the H₂O product. On the contrary, when the structure of the active site was Fe–C–O, the binding energy between *OOH and C site was enhanced, which resulted in the 2e⁻ pathway to produce H₂O₂.

4.1.2. Fe-SACs in thermocatalysis. In the field of energy storage, in addition to abovementioned electrocatalysis reactions, there are thermocatalysis and photocatalysis reactions.⁵⁷ Generally, most heterogeneous catalysis belongs to the category of thermocatalysis.²²⁵ Thermal catalysis includes many types of reactions, e.g., C–H bond activation, CO catalytic oxidation, selective hydrogenation, methane activation, NO oxidation and N₂O decomposition.^73,226–228 Given that Fe-SACs have been demonstrated to show the unique advantages of heterogeneous and homogeneous catalysts, they can be used as thermal catalysts.
4.1.2.1. Oxidation of CO. To address the increasingly serious environmental pollution, the conversion of CO in automobile exhaust and industrial waste gas into non-toxic gas at room temperature or low temperature has become an urgent problem.²²⁹ The catalytic oxidation of CO is one of the main and vital prototype reactions in the heterogeneous catalysis process.²³⁰ Li et al.²³¹ investigated the mechanism of the catalytic oxidation of CO by atomically dispersed Fe embedded on graphene with first-principle computation. By carefully comparing the reaction between the adsorbed O₂ and CO by the Langmuir–Hinshelwood (LH) and Eley–Rideal (ER) mechanisms, the results showed that the Fe-SACs had better catalytic activity for the catalytic oxidation of CO through the ER mechanism (Fig. 22a). Firstly, CO combines with an O₂ molecule adsorbed on the surface of Fe-SACs to form the CO₃* intermediate state (CO + O₂* → CO₃*). The intermediate CO₃* will combine with another CO molecule to release two CO₂ molecules (CO + CO₃* → 2CO₂) (Fig. 22b). Similarly, Liu et al.²²⁵ also used first-principles calculations to calculate the catalytic mechanism of single-atom Fe-doped graphene nanosheets in the catalytic oxidation of CO. The N atoms coordinating with Fe atoms made single Fe atoms have more positive charges, thereby adsorbing more reactive gases. The introduction of Fe can adjust the electronic structure and magnetic properties of the graphene system, thereby improving its catalytic activity. As illustrated in Fig. 23, the LH and ER mechanisms were used to study the catalytic oxidation of CO. The results revealed that the LH mechanism as the starting state was more advantageous than the ER mechanism during the reaction process because the formation of reaction intermediates had a smaller energy barrier, and then the reaction intermediates generated CO₂ through the ER reaction.

Fig. 22 (a) Atomic configurations of the initial state (IS), transition state (TS), intermediate state (MS), and final state (FS) for CO oxidation on Fe-embedded graphene. (b) Energy profiles for CO oxidation over Fe on defective graphene. Reprinted with permission from ref. 231. Copyright 2010, the American Chemical Society.

Fig. 23 Minimum energy profiles and configurations of the different states for CO oxidation on Fe–GN₄ by (a) LH mechanism and (b) ER mechanism. Black, blue, red and green balls represent the C, N, O and Fe atoms, respectively. Reprinted with permission from ref. 225. Copyright 2017, The Royal Society of Chemistry.

As a typical magnetic transition metal, the spin state of the Fe single-atom active center undoubtedly plays an important role in regulating the catalytic performance of various chemical processes. Therefore, theoretical calculations and multiscale simulations are also effective means to determine the working mechanism of Fe-SACs. The underlying mechanisms of various Fe-SACs spin degrees of freedom with high performance can be resolved through theoretical calculations and simulations. In a typical experiment, Zhang et al.²³² proposed that the 2D ferromagnetic organometallic framework Mn₂C₁₈H₁₂ can be used as an ideal platform to efficiently realize the activation of the spin triplet O₂ and the oxidation of CO based on the basic principles of quantum mechanics. The physical mechanism of the two-dimensional ferromagnetic organometallic framework single-atom catalyst system to efficiently excite O₂ was a new “charge-spin dual cooperative mechanism” (Fig. 24a). O₂ achieved a transition from a spin triplet state to a spin singlet state when the O₂ molecule was adsorbed on a certain Mn single-atom active site on the ferromagnetic 2D MOF structure. Importantly, the single-atom active site of the catalyst and its neighboring single atoms had a very different balance of labor and cooperation in terms of charge and spin degrees of freedom. During O adsorption, the charge transfer was mainly provided by the adsorption active site, and its adjacent Mn atoms are weakly related to the active site in terms of charge degrees of freedom. However, according to the Wigner spin conservation selection rule, the adjacent Mn atoms have a significant synergistic response in the spin degrees of freedom, revealing that these single atoms accommodate most of the spin magnetic moment transferred from the O₂ molecule. Consequently, the efficient excitation of O₂ from the spin triplet state to the spin singlet state can be realized. The quantum synergistic mechanism effectively reduced the rate-determining step barrier of the reaction process, thereby promoting the catalytic oxidation of CO.^233,234 This novel quantum effect of “charge-spin synergistic catalysis” has been further extended to various magnetic X₂C₁₈H₁₂ (X = Mn, Fe, Co, and Ni) systems. Interestingly, a clear linear scaling relationship was found between the degree of chemical activity of these M-SACs for O₂ excitation and the spin excitation energy (Fig. 24b–e).

Fig. 24 (a) Schematic diagram of the “charge-spin synergistic catalysis” mechanism of O₂ activation by magnetic single-atom sites. (b) Adsorption energy E_ads, (c) charge transfer Δρ and (d) O–O stretching vibrational frequency ν of the O₂ molecule on the X-I sites with respect to the quantity of spin accommodated on X-II site ΔM(X-II) of 2D X₂C₁₈H₁₂ MOF structures (X = Mn, Fe, Co, and Ni). (e) ΔM(X-II) as a function of spin excitation energy (SEE). Reprinted with permission from ref. 232. Copyright 2022, the American Chemical Society.

4.1.2.2. Activation of C–H bond. Given that the C–H bond (409 kJ mol⁻¹) in the benzene molecule is relatively stable, it is very difficult to oxidize benzene to phenol.²³⁵ To date, a widely employed route is the use of transition metals as catalysts for oxidation catalysis at high temperature (50–140 °C); however, this method generally consumes significant energy and shows poor catalytic property.²³⁶ Also, the catalytic oxidation of benzene to phenol at room temperature is an important barrier to be addressed. Deng et al.²³⁷ used FePc as the iron source and GNs as the carbon source to anchor atomically dispersed Fe–N₄ species. In the presence of H₂O₂ oxidant, the FeN₄/GN-x catalyst (x represents the Fe content in the catalyst) could catalytically oxidize benzene to phenol at room temperature. It can be seen from Fig. 25a that when the iron content was 2.7%, the corresponding catalyst had the highest catalytic activity. The conversion rate of benzene reached 23.4%, showing a yield of phenol of over 18% (Fig. 25b). DFT was used to calculate the oxidation process of benzene (Fig. 25c–e). FeN₄/GN-2.7 used as a catalyst to catalyze the oxidation of benzene, which involved five steps. Firstly, H₂O₂ is decomposed into H₂O₂ and O atoms, while O atoms are adsorbed on the active site Fe–N₄ to form Fe=O species, and then a second H₂O₂ molecule is dissociated again. The dissociated O atoms combine with Fe on the other side of the Fe=O species to form O=Fe=O species. The O atom in the active center of O=Fe=O will adsorb benzene molecules to form a C–H bond, with an energy barrier of 0.59 eV. The Fe–O bond will break to form phenol and Fe=O species, and then cyclically perform oxidation reactions. Meanwhile, the Bader charge analysis results showed that FeN₄/GN-2.7 had 0.14 more electrons than FePc, which made it easier for FeN₄/GN-2.7 to combine with O to form an Fe–O bond, thereby enhancing the catalytic oxidation activity of benzene. Moreover, Liu et al.²³⁸ used nano-MgO as a sacrificial template and synthesized monoatomic Fe-based catalysts containing different amounts of FeN_x (x = 4–6) species by adjusting different pyrolysis temperatures. When the pyrolysis temperature was 700 °C, the catalytic effect was the best. The obtained Fe-SACs contained 28.3% high-spin FeN₆, 53.8% low-spin FeN₆ and 17.9% medium-spin FeN₅ species. The most important species for C–H bond oxidation was FeN₅ species, and its activity was about 10 times higher than that of the high/low-spin FeN₆ species and about 3 times higher than that of the FeN₄ species. The experiment results showed that the Fe-SACs could selectively oxidize C–H bonds and catalyze the oxidation of various substrates such as aromatic hydrocarbons, heterocyclic alkanes and aliphatic alkanes, with a selectivity of up to 99%.

Fig. 25 (a) Performance of the direct oxidation of benzene to phenol by FeN₄/GN samples compared with GF, GN, and FePc. (b) Phenol yield of FeN₄/GN-2.7 for the direct oxidation of benzene to phenol with different reaction times. (c) Formation energies of FeN₄/GN and Fe/GN structures. The formation energy was calculated as follows: E_Fe-embedded − E_Fe-bulk − E_(N)GN, where E_Fe-embedded and E_Fe-bulk are the total energies of FeN₄/GN and the Fe/GN structure and an Fe atom in Fe bulk, respectively, and E_(N)GN was the total energy of the optimized structure of FeN₄/GN or Fe/GN with the Fe atom removed from the system. (d) Free energy diagram of the oxidation of benzene to phenol on FeN₄/NG. The gray, blue, light blue, red, and white balls represent C, N, Fe, O, and H atoms, respectively. Proposed mechanism for C₂H₆ activation on the O–FeN₄–O active site. (e) Scheme for the reaction mechanism of the oxidation of benzene to phenol on FeN₄/NG. Reprinted with permission from ref. 237. Copyright 2015, the American Association for the Advancement of Science.

The conversion of ethane into high value-added chemical products (such as C₂H₅OH, CH₂=CH₂ and CH₃COOH is also a critical step in industrial production.^239,240 However, the difficulty of the cleavage of the C–H bond makes this conversion process more difficult and usually requires a high temperature (>400 °C), but its selectivity is poor. Meanwhile, some side reactions will occur, which may cause the product to over-oxidize or result in catalyst poisoning.²⁴¹ Interestingly, Fe-SACs can convert ethane into C₂ products through the catalytic oxidation of H₂O₂ at room temperature. In a typical experiment, Wang et al.²⁴² synthesized atomic-grade Fe–N₄ species uniformly dispersed on N-doped carbon. When H₂O₂ was used as an oxidant, the FeN₄/GN catalyst could effectively convert ethane into CH₃COOH, CH₃CH₂OOH, CH₃CH(OH)₂, CH₂CH₂ and other high value-added C₂ products. As shown in Fig. 26a, consistent with the oxidation of benzene mentioned above, the activation of ethane was carried out according to the free radical mechanism. When ethylene is adsorbed on the O–Fe–O species, the C–H bond will be stretched to form a type of ˙CH₂CH₃ radical, and then form ethanol with the OH group or destroy another C–H bond to form CH₂CH₂. Notably, the produced ethanol can be used again as the reactant for oxidation catalysis. The C–H bond on the same carbon atom is activated, and finally CH₃CHO is produced. The energy barriers of these three reactions are relatively low (Fig. 26b). Likewise, Cui et al.²⁴³ used Fe-SACs to convert methane into some high value-added C₁ products (CH₃OH, CH₃OOH, HOCH₂OOH, HCOOH, etc.) at room temperature. The conversion mechanism was roughly the same as that of ethane. It also first formed the O–Fe–O active intermediate, then activated methane through the free radical mechanism, and finally yielded the C₁ product.

Fig. 26 (a) Reaction cycle with the reactants (black), products (black, blue and pink), radical intermediates (red) and atomic structures of the active site and transition states TS1-TS3. (b) Reaction pathway with atomic structure of intermediates and the corresponding activation barrier. Reprinted with permission from ref. 242. Copyright 2019, Elsevier.

4.1.2.3. Selective hydrogenation. Selective hydrogenation means that when the substrate contains two or more unsaturated functional groups, the catalyst selectively hydrogenates a specific functional group.^244–246 Selective hydrogenation not only can improve the yield of the target product and reduce the yield of the by-products, but also can be performed in a mild manner, thereby reducing the energy consumption.²⁴⁶ It plays a pivotal role in the fine chemical, pharmaceutical, agrochemical and printing and dyeing industries.^247–250 Due to the excellent catalytic activity of Fe-SACs, they can also be used to catalyze selective hydrogenation.

As a type of heterogeneous catalysis, the selective hydrogenation of unsaturated functional groups has become a focus of chemists, especially for the hydrogenation of nitro compounds. The mechanism is complex, and thus noble metal catalysts with excellent selectivity are required.^251–255 Therefore, Yun et al.²⁵⁶ developed a new material of Fe single atoms and Fe₂O₃ clusters uniformly dispersed on N-doped carbon for the selective hydrogenation of nitro compounds (Fig. 27a). Using nitrobenzene as a model and hydrazine hydrate as a hydrogen source, the catalytic performance of Fe-SACs was tested. At room temperature, nitrobenzene could be completely hydrogenated to aniline within 12 min with a conversion rate of over 99%, which was greater than that of the FeN₄ catalyst (50%) and iron nanoparticles (25%). Its TOF (1923 h⁻¹) was higher than that of the FeN₄ catalyst (748 h⁻¹) and iron nanoparticles (75 h⁻¹) alone. The DFT calculation results showed that the FeN₄ catalyst was more active for the decomposition of hydrazine hydrate to produce hydrogen. The decomposition process of hydrazine hydrate followed the order of N₂H₄ → N₂H₄* → N₂H₃* → N₂H₂* → N₂H* → N₂* → N₂ (Fig. 27b and c). The Fe₂O₃ cluster surface was more active for the reduction of aniline by nitrobenzene. The process of reduction and hydrogenation of nitrobenzene followed the sequence of PhNO₂ →PhNO₂* → PhNOOH* → PhN(OH)₂* → PhNOH* → PhNHOH* → PhNH* → PhNH₂*→ PhNH₂ (Fig. 27d).²⁵⁷ Therefore, the results of the DFT calculations indicated that FeN₄ can catalyze the decomposition of hydrazine hydrate to generate hydrogen.

Fig. 27 (a) Schematic diagram of the selective hydrogenation of nitrobenzene catalyzed by FeSAs/Fe₂O₃ACs/NPC. Gibbs free energy profiles for (b) nitrobenzene reduction processes on the FeN₄ site embedded in graphene (lines in blue) and the Fe₂O₃(012) surface (lines in red), respectively. The dashed arrows indicate the overall barriers in the processes. Optimized structures of intermediates in hydrazine dehydrogenation and nitrobenzene reduction processes on the FeN₄ site embedded in graphene (c) and Fe₂O₃(012) surface (d), respectively. Light gray: C in graphene; dark gray: C in molecules; white: H; red: O in Fe₂O₃; pink: O in molecules; blue: N; and purple: Fe. Reprinted with permission from ref. 256. Copyright 2020, the American Chemical Society.

Similarly, Tian et al.²⁵⁸ synthesized atomically dispersed Fe metal sites supported on nitrogen-doped carbon for the selective hydrogenation and transfer hydrogenation of nitrobenzene. The catalytic performance for the selective hydrogenation and transfer hydrogenation of nitrobenzene to aniline was investigated using H₂ and isopropanol as the hydrogen sources. When the temperature was set at 160 °C, the conversion and selectivity of nitrobenzene were higher than 99% with H₂ as the hydrogen source. When the temperature was 220 °C, using isopropanol as the hydrogen source, the conversion and selectivity of nitrobenzene transfer hydrogenation both reached 99%. DFT calculations further revealed that the Fe-SACs exhibited an excellent catalytic performance for selective hydrogenation at lower temperature mainly owing to their efficient activation of reactants and intermediates. The increase in temperature can make Fe-SACs easily overcome the reaction energy barrier of isopropanol dehydrogenation, thereby improving the catalytic efficiency of its transfer hydrogenation.

4.1.3. Fe-SACs in photocatalysis. To make full use of solar energy, developing excellent photocatalysts to convert solar energy into the energy has become a potential development strategy.^259–261 However, due to the slow separation of electron–hole pairs and limited surface active sites, the catalytic performance of conventional photocatalysts, depending on their energy band structure and surface structure, should be further improved.^262,263 In principle, the isolated reaction centers in single-atom photocatalysts not only can provide more active sites for photocatalytic reactions, but also expand the light-harvesting range and improve the charge separation/transfer efficiency, thereby becoming an excellent substitute for conventional photocatalysts.^264–266 Moreover, due to the extremely simple structure of single-atom catalysts, it is possible to more accurately understand the reaction mechanism of photocatalytic reactions by clarifying the relationship between structure and performance, which in turn boost the design of reasonable photocatalysts.²⁵⁹ In recent years, single-atom photocatalysts have developed rapidly and are widely used in the field of photocatalysis for organic synthesis, pollutant degradation, H₂O₂ production and CH₄ conversion.^267–272

The generation of H₂ fuel by photocatalytic water splitting is a competitive and sustainable energy route.^273,274 Generally, this process requires extremely active and stable light-harvesting semiconductors to meet the stringent requirements of suitable band positions and fast interfacial charge transfer processes. Accordingly, Zhang et al.²⁶⁷ synthesized atomically dispersed Fe integrated in porous coiled graphitic carbonitride (Fe@g-C₃N₄) for inducing ultra-high solar photon-driven hydrogen production. Fe@g-C₃N₄ had an excellent photocatalytic hydrogen production rate (3390 μmol h⁻¹ g⁻¹) and a good apparent quantum efficiency (6.89%) at 420 nm. The DFT calculation suggested that the presence of Fe atoms can optimize the electron/band structure of the g-C₃N₄ semiconductor, which helped to expand the light absorption range and shift the conduction band up to enhance the reducibility. Consequently, the separation efficiency of photoexcited electron–hole pairs improved. Simultaneously, g-C₃N₄ can shift its d orbital position to the Fermi level to affect the electronic structure of Fe, which not only can afford active sites, but also accelerate the interfacial charge transfer process, thereby improving its catalytic efficiency.

In addition to the above-mentioned nitrogen fixation methods, there are other ways to reduce N₂ to NH₃, such as N₂ photo-fixation. Fu et al.²⁷⁵ reported the use of a combination of triphenylphosphine (PPh₃) and sodium iodide (NaI) as photosensitizers for the catalysis of small organic molecules in the alkylation of silyl enol ethers. Inspired by this catalytic process, Hou et al.²⁷⁶ attempted to adsorb PPh₃ on activated carbon and coupled it with a single-atom iron catalyst to form Fe₁/C-PPh₃/NaI for the heterogeneous catalysis of N₂ photoimmobilization. Fe₁/C-PPh₃/NaI exhibited excellent catalytic activity in the heterogeneous catalysis process of N₂ photoimmobilization, and the yield of NH₃ was as high as 98 μmol (g_cat h)⁻¹, which was 1.5 times that of NH₃ when [Ru(bpy)₃]Cl₂ was used as the photosensitizer.²⁷⁷ It can be found that under light conditions, PPh₃/NaI as a photosensitizer can transfer hot electrons to Fe atoms through activated carbon, and then generated PPh₃–I˙ radicals and oxidize H₂O to O₂. Meanwhile, hot electrons will reduce N₂ to NH₃ on Fe atoms (Fig. 28). NO₃⁻ as one of the most common pollutants can be used as a nitrogen source to synthesize NH₃, which not only can treat environmental pollution, but also produce NH₃, the most basic chemical product.²⁷⁸ It should be noted that the conversion of NO₃⁻ to NH₃ involved multiple electronic reactions and reaction pathways, and thus understanding the reaction mechanism was extremely important for the selection of suitable catalysts.²⁷⁹ The porphyrin group of chlorophyll exhibits unique photosensitivity, and the Fe atom in nitrogenase can be used as a binding site to combine with N₂ through an anti-bond. Inspired by this, Shang et al.²⁸⁰ synthesized a porphyrin-based metal–organic framework-based single-atom Fe catalyst for the photocatalytic synthesis of NH₃. The NH₃ yield (635 μg g_cat.⁻¹) and production rate (127 μg h⁻¹ g_cat.⁻¹) increased by 82% and 50%, respectively. Alternatively, Wu et al.²⁸¹ achieved the photocatalytic immobilization of nitrogen by coordinating photocatalytic N₂ hydrogenation and water oxidation through a doping strategy. Both the experimental results and DFT results indicated that photoinduced hole-trapping polarons were formed on the Fe dopants, resulting in a high content of Fe(IV) species and promoting N₂ hydrogenation on the adjacent oxygen vacancies.

Fig. 28 (a) Optimized reaction path for the dissociation of N₂ to NH₃ over Fe₁O₄/C. *Represents the adsorbed sites. (b) Schematic diagram of the process of N₂ photofixation over Fe₁/C-PPh₃/NaI. Reprinted with permission from ref. 276. Copyright 2020, the American Chemical Society.

Cumulative studies support that Fe-SACs can also be used for other photocatalytic reactions, such as organic synthesis, activation of C–H bonds and degradation of organic materials.^282–285 In a typical experiment, Wen et al.²⁸³ synthesized a biomimetic photocatalyst with a single-atom iron site for organic synthesis by coupling carbon nitride with hemin via a chemical cross-linking method. Under the irradiation of visible light, it could sulfonate olefins to form β-ketosulfones (94% yield) under normal temperature and pressure. More interestingly, under near-infrared light, it can be used for the sulfonation reaction of androstenone. This was mainly because carbon nitride can enhance the separation ability of photogenerated electron pairs and holes, thereby improving the activity and stability of photocatalysts. Xiao et al.²⁸² synthesized a single Fe atom aggregated on carbon nitride for the photocatalytic activation of O₂ to generate the superoxide anion (O₂˙⁻) to promote the activation of α-C–H bond of ethylbenzene to form acetophenone with higher conversion and selectivity than 99%. DFT calculations showed that a single Fe atom coordinates with 4 N atoms to form a low-spin Fe–N₄ active center, which provides an adsorption site for O₂ and allows the electrons generated in carbon nitride to transfer to the adsorbed O₂, producing O₂˙⁻. Zhao et al.²⁸⁵ synthesized a single-atom photocatalyst with Fe–N₄ active sites and used it for the catalytic activation of peroxomonosulfate and sulfamethoxazole degradation. Under visible light, 98.7% of sulfamethoxazole with an initial concentration of 10 mg L⁻¹ was degraded using 50 mg mL⁻¹ of catalyst. The photogenerated electrons accelerated the cycle of triple bond Fe(II)/triple bond Fe(III), and the photogenerated holes further boosted the generation of ¹O₂.

4.2. Applications of Fe-SACs in biocatalysis

Enzymatic reactions are an indispensable process in metabolism, where enzymes play a crucial role as the catalyst.²⁸⁶ Essentially, enzymes possess the advantages of fast reaction rate under mild conditions, unique specificity and adaptability, making them widely used in disease diagnosis and treatment.²⁸⁷ Unfortunately, natural enzymes also have some inherent drawbacks, e.g., high cost, poor stability, easy to inactivate, difficulty in recovery and storage, and poor selectivity.^89,288 In 2007, Perrett et al. first discovered the peroxidase (POD)-like activity of Fe₃O₄ nanoparticles, and since then a series of nanomaterials that can functionally mimic the catalytic properties of natural enzymes (called nanozymes) have been significantly developed (Fig. 29).²⁸⁹ Nanozymes have the characteristics of natural enzymes and nanomaterials including low cost, high stability, facile storage, adjustable activity and easy mass production, which enable them to display impressive application prospect.^290,291 In fact, compared with natural enzymes, the selectivity and activity of nanozymes are not perfect, while their catalytic activity is not as high as expected owing to the low density of active sites.²⁹² The uneven surface and geometric structures of nanozymes led to a complicated catalytic mechanism. It is difficult to establish the relationship between actual catalytically active sites and catalytic activity, further hindering the application of nanozymes.

Fig. 29 Single-atom iron catalysts functionally mimicking several different antioxidant enzymes.

To bridge the huge gap between the catalytic performances natural enzymes and nanozymes, single-atom nanozymes (SAzymes) have been recently developed, showing very impressive catalytic activity and selectivity.^293,294 More interestingly, the uniformity of the active sites and geometric structures are conducive to the accurate identification and characterization of active sites in catalytic reactions, and also helps to understand the catalytic mechanism of SAzymes.²⁸⁷ Experiments and theory suggest the similar active sites (M–N_x) of SAzymes to natural enzymes, especially in terms of geometric, chemical and electronic structures.²⁹² For example, the active site of horseradish peroxidase (HRP) has a single heme b cofactor with a proximal ligand.²⁹⁵ Notably, this molecular structure is consistent with the structure of the Fe–N₄ active centers in SAzymes. By clarifying the relationship between the structures and activity of natural enzymes, continued interdisciplinary efforts have been devoted to designing SAzymes with the same activity or even better activity than natural enzymes for large-scale production, showing a great application prospect in the field of biocatalysis.^89,296,297

4.2.1. Tumor treatment. Nowadays, chemotherapy is still the most common and effective tumor treatment method, but patients generally suffer from low drug bioavailability and unsatisfactory therapy specificity, leading to serious side effects.²⁹⁸ The physical therapy methods, e.g., radiotherapy, ultrasound therapy and photothermal therapy, can achieve localized treatment; however, these methods may cause severe damage to the surrounding normal tissues and even induce undesirable tumor metastasis.²⁹⁹ The design of intrinsically biocompatible therapeutic systems, which do not have any traditional or newly discovered cytotoxic drugs, is a very exciting goal towards practical tumor treatments. In this context, chemodynamic therapy has been proposed based on the difference between the microenvironment of tumors and that of normal tissues.^298,300,301

The working strategy underlying chemodynamic therapy is to initiate a Fenton reaction to convert H₂O₂ into a strong oxidizing reactive oxygen radical (˙OH), which then induces cell tumor apoptosis.^302,303 Compared with general nanozymes, the higher atom utilization and selectivity of Fe-SACs enable them to more easily catalyze the oxidation of H₂O₂ to generate strong oxidative ˙OH, which is widely used in tumor therapy. In a typical experiment, He et al. studied the therapeutic effect of Fe-SACs on the activity of lung cancer cells. As shown in Fig. 30a, in the presence of oxygen, Fe-SACs could significantly reduce the number of lung cancer cells, which was observed by fluorescent imaging. The killing rate of Fe-SACs on lung cancer cells reached 88% (Fig. 30b and c). These results all proved that Fe-SACs can generate active oxygen by activating oxygen to induce cancer cell apoptosis. In the presence of oxygen, ³O₂ can be adsorbed on the active Fe–N₄ sites, thereby changing its electronic structure by activating ³O₂ to produce ¹O₂. Then, the protons and electrons from the carrier surface were adsorbed to form HO₂^·−, and finally produce ROS to induce cancer cell apoptosis.⁸⁶ Moreover, Shi's group applied Fe-SACs with POD-like activities to catalyze H₂O₂ to produce strong oxidizing ˙OH species under acidic condition. The in vivo and in vitro study demonstrated that Fe-SACs could not only kill tumor cells, but also induced the accumulation of lipid peroxides, hence causing tumor cell ferroptosis. The synergy of these two effects can play a key role in tumor treatment.³⁰⁴ Notably, Fe-SACs can also serve as catalase (CAT)-like enzymes and photosensitizers to yield reactive oxygen radicals under light irradiation, inducing the apoptosis of cancer cell.²⁹⁶ In meantime, the favorable biodegradability and biocompatibility of Fe-SACs guarantee their desirable biosafety both in vivo and in vitro. NO_x is a transmembrane complex enzyme with 6 subunits, which is mainly expressed in immune cells. The intracellular subunit of the complex enzyme can convert NADPH to NADP⁺, transfer electrons to the gp91^phox subunit of the cell membrane through the transmembrane region and use its heme active site to convert O₂ molecules into superoxide free radicals. A lack of NO_x may lead to immunodeficiency and cause chronic granulomatous disease (CGD). Recently, Wu et al. designed and synthesized Fe–N-doped two-dimensional graphene materials based on the structural characteristics of NO_x active sites and electron transport (FeNGR). Their studies confirmed that FeNGR can efficiently convert NADPH to NADP⁺, with the generation of oxygen free radicals simultaneously. The conversion rate was as high as 93%. Compared with protease, FeNGR was more stable and could maintain 85% catalytic activity in extreme reaction systems (pH = 1–12, temperature 4–90 °C, and 0–100% organic solvent). In addition, FeNGR could replace the biological function of NO_x and catalyze the generation of superoxide radicals in NO_x-deficient cells, further activating the downstream immune signaling molecules.³⁰⁵

Fig. 30 Applications of Fe-SACs in biocatalysis. (a) Fluorescence microscopy images of lung cancer cells treated with and without carbon black and Fe-SACs for 24 h. (b) Proliferation rate and (c) MTT assay of lung cell viability incubated with and without carbon black and Fe-SACs. Reprinted with permission from ref. 86. Copyright (2021), the American Chemical Society. (d) UV-vis absorption spectra of TMB oxidized by Fe-SACs with different concentrations of H₂O₂ and (e) corresponding calibration curve. (f) Reproducibility and (g) selectivity of H₂O₂ detection by Fe-SACs. Reprinted with permission from ref. 308. Copyright (2019), the American Chemical Society. (h) Confocal laser scanning microscopy image of HeLa cell with Fe-SACs. (i) Cytotoxicity of Fe-SACs. (j) Intracellular ROS scavenging by Fe-SACs in β-Lap-treated cancer cells using flow cytometry. (k) Effects of different concentrations of Fe-SACs on the viability of β-Lap-treated cancer cells. Reprinted with permission from ref. 286. Copyright (2019), The Royal Society of Chemistry.

4.2.2. Biosensing. Among the biosensing technologies, the most commonly studied one is the colorimetric sensing route, which is mainly attributed to its convenient operation and facile detection by the naked eyes.²⁹² Colorimetric sensing is technology based on the color variation in chemical reactions. When Fe-SACs are used as nanozymes to analyze certain substances, the popular color substrate is 3,3′,5,5′-tetramethylbenzidine (TMB) or o-phenylenediamine (OPD). Since Fe-SACs have a variety of enzyme activities, they hold great potential for the detection and qualitative and quantitative analysis of substances. ROS, especially H₂O₂, are produced by oxygen metabolism in cells and play a key role in a myriad of biological processes including cell proliferation and differentiation.³⁰⁶ When the concentration of H₂O₂ in the cell is high, it will cause oxidative stress, and further initiate a series of diseases, such as Alzheimer's and Parkinson's disease.^213,307 Studies revealed that the concentration of H₂O₂ in cancer cells (100 μM) is much higher than that in normal cells (20 nM).³⁰¹ In this regard, the concentration of H₂O₂ can be used as an important indicator of the response to diseases.²¹³

Recently, Fe-SACs have been reported to show POD-like activity comparable to that of HRP, which can catalyze H₂O₂ to yield ˙OH, and then oxidize the colorless substrate TMB to the blue oxidation product oxTMB.^308,309 By elevating the H₂O₂ concentration, the absorbance of the product at 652 nm also gradually increased with a linear relationship (Fig. 30d and e). Interestingly, Fe-SACs also have excellent durability and anti-interference properties (Fig. 30f and g). Taking advantage of the POD-like catalytic activity of Fe-SACs, colorimetric sensing can be further extended to detect ascorbic acid, butyrylcholinesterase and acetylcholine.^53,310,311 Of course, the oxidase-like activity of Fe-SACs is another working principle for colorimetric sensing. For example, Wu et al.³¹² confirmed the feasibility of Fe-SACs with oxidase-like activity for the indirect detection of acetylcholinesterase. Chen et al.³¹³ employed the oxidase-like activity of Fe-SACs to evaluate the contents of alkaline phosphatase.

Besides the aforementioned colorimetric sensing by the enzyme-like activity of Fe-SACs, the detection of substances can be performed by using electrochemical, electrochemiluminescence and fluorescence methods.³¹⁴ For example, Gu et al.³¹⁵ introduced as-prepared Fe-SACs in an electrochemiluminescence system to improve the luminescence signal. Wang et al.³¹⁶ used the fluorescence method to detect acetylcholinesterase by Fe-SACs with POD-like activity.

4.2.3. Oxidative stress cytoprotection. ROS (O₂˙⁻, H₂O₂, ˙OH, etc.) are intermediate products of cell metabolism, which are mainly produced in the mitochondria of the cell.^86,317 Studies have shown that cardiovascular diseases, bronchitis, and otitis are associated with excessive ROS.^318–320 Thus, it is particularly important to maintain the balance of ROS production and elimination in the body and regulate the normal level of ROS in cells. Generally, the cellular antioxidant enzymes consist of CAT, POD, SOD and glutathione peroxidase (GPx). Cells rely on these natural enzymes to maintain the physiological level of ROS. When the ROS content in cells is high, the activity of antioxidant enzymes in the body will be broken. At this time, an external nanozyme (e.g., SAzymes) with antioxidant activity is needed to control the level of ROS in the body.

CAT catalyzes the decomposition of H₂O₂ to produce O₂, hence eliminating ROS, and is used as an oxidative stress relief agent to protect normal tissues and cells. SOD can eliminate ROS by catalyzing the disproportionation of O₂˙⁻ to H₂O₂ and O₂. Ma et al. reported that Fe-SACs showed both CAT- and SOD-like antioxidant activities, and further evaluated their ability to remove ROS in cells. HeLa cells exhibited the efficient cellular uptake and accumulation of Fe-SACs in the cytosol, as evidenced by the black punctate of Fe-SACs inside HeLa cells (Fig. 30h). Cell viability studies indicated the minimal cytotoxicity of Fe-SACs (Fig. 30i). To verify the ROS removal ability of Fe-SACs, β-lapachone (β-Lap), which can be bioactivated by enzymes overexpressed in HeLa cells to produce massive ROS in short time, was added to the cell culture medium. 2′,7′-Dichlorofluorescein diacetate (DCFDA) was utilized to fluorescently stain the in situ formed ROS. Compared with the cells without β-Lap treatment, the fluorescence intensity of the cells treated by β-Lap increased 6.8 times (Fig. 30j). Surprisingly, when HeLa cells were pretreated in Fe-SACs, and then treated with β-Lap, no obvious fluorescence was observed, indicating that the ROS produced by β-Lap were eliminated by Fe-SACs. The above-mentioned results all supported that Fe-SACs had the ability to eliminate ROS, thereby protecting cells from oxidative stress and damage (Fig. 30k).²⁸⁶ Moreover, Fe-SACs were also found to show CAT-like and GPx-like activity, which can effectively eliminate ROS and control the H₂O₂ level in cells.⁸⁹ In summary, Fe-SACs with multiple antioxidant enzyme activities are more effective than a single antioxidant enzyme in reducing the oxidative stress response of cells.³²¹

4.2.4. Wound disinfection. Regardless of the type of inflammation (infectious or non-infectious inflammation), if proper treatments are not taken, it will cause unpredictable consequences to human health.^322,323 However, clinically, due to the complexity of inflammation, it is difficult to distinguish the types of inflammation, and thus antibiotics are utilized to treat inflammation. The abuse of antibiotics will make the treatment of non-infectious inflammation less effective and increase the resistance of bacteria to antibiotics.^324,325 The new frontier of drug therapy to replace the existing antibiotics should show a good therapeutic effect on both infectious and non-infectious inflammation. Considering two reactions (namely POD could catalyze H₂O₂ to yield ˙OH under acidic condition and oxidase (OXD) catalyzes O₂ to generate H₂O₂ or O₂˙⁻), the products of these enzymatic reactions are ROS that can be used to kill bacteria.^88,291 It is the basic working strategy for treating inflammation caused by bacterial infection. Meanwhile, superoxide dismutase (SOD) catalyzes O₂^·− to produce H₂O₂ and CAT can catalyze H₂O₂ to yield O₂ and H₂O, respectively. Clearly, these two enzymes can eliminate ROS and treat non-infectious inflammation induced by excessive ROS. To prepare a drug that can treat both infectious inflammation and non-infectious inflammation, it needs to possess multi-enzymatic activities. Impressively, Fe-SACs were found to show the above-mentioned four types of enzyme-like activities.⁸⁹

Typically, in 2020, Fan et al. designed hollow porous carbon spheres co-doped with N and Fe (Fe/N-HCNs), which exerted multienzyme mimicking activities, such as POD-, OXD-, CAT- and SOD-like catalytic activities. To determine the therapeutic effect of Fe/N-HCNs on infectious inflammation, animal experiments were performed for wound healing of animal skin infections. On the 5th day, the scars of skin infection wound in the Fe/N-HCNs + H₂O₂ group were obviously reduced (Fig. 31a and b). On the 11th day, the scars disappeared and the wounds healed completely. Due to the OXD-like catalytic activity of Fe/N-HCNs, the animal wound was protected from bacterial infection and inflammation. Fig. 31c indicates that Fe/N-HCNs had good biocompatibility. To further validate the therapeutic effect of Fe/N-HCNs on animal skin infection wounds, hematoxylin–eosin (H&E) staining of the wounds 11th day later was conducted (Fig. 31d). A large number of inflammatory cells appeared in the control group. Importantly, when Fe/N-HCNs + H₂O₂ synergistic treatment was performed, optimal treatment effects were observed. The inflammatory cells disappeared quickly, which is associated with the healed skin infection wound. The above-mentioned data shows that Fe/N-HCNs can serve as a good anti-infective inflammation drug, which effectively treat inflammation caused by bacterial infection and promote wound healing. Furthermore, based on the SOD-like activity of Fe/N-HCNs, the therapeutic effect on non-infectious inflammation was monitored. The experimental results of colitis induced by dextran sodium sulfate (DSS) as a model are shown in Fig. 31e and f. Colon shortening is the most important feature of colitis. The colon after DSS treatment was significantly shortened, but after treatment using Fe/NHCNs, the colon became longer, revealing that Fe/NHCNs had an excellent therapeutic effect on colitis. After DSS treatment, compared with the control group, the tissue mucosal structure of the mice was severely damaged. As anticipated, Fe/NHCNs could restore the tissue mucosal structures of the mice to a certain extent and inflammatory cells had also been highly reduced. Together, the prominent multi-enzyme activities of Fe/N-HCNs served as an anti-inflammatory alternative for both infectious and noninfectious inflammation.³²⁶ Recently, Li et al. synthesized a single-atom nanoenzyme with FeN₃P as the active center. The precise coordination of N and P affects the electronic structure of the Fe atom in the active center, and thus the as-obtained SACs exhibited excellent-POD-like enzyme activity for tumor treatment. To identify the excellent enzyme activity of the FeN₃P-SAzyme, the authors also prepared an FeN₄-SAzyme with only N coordination and Fe₃O₄ nanoparticles for comparison. The measured specific activity showed that the enzyme activity of FeN₃P-SAzyme (316 U mg⁻¹) was 10 times that of FeN₄-SAzyme (33.8 U mg⁻¹) and 30 times that of the Fe₃O₄ nanoparticles (9.12 U mg⁻¹). Furthermore, the inhibitory effects of FeN₃P-SAzyme and FeN₄-SAzyme on cells showed concentration and time dependence (Fig. 32a and b), respectively, while the Fe₃O₄ nanoparticles basically had no inhibitory effect on the cells. After FeN₃P-SAzyme was endocytosed by the cells, it finally entered into the lysosome (Fig. 32c and d). H₂DCFDA was used as a fluorescent probe to detect ROS in HepG₂ liver cancer cells treated by FeN₃P-SAzyme. After the addition of GSH, the intracellular ROS was reduced, indicating that FeN₃P-SAzyme had POD-like activity and can be used to inhibit tumor cells (Fig. 32e). The mouse experiment is shown in Fig. 32f and g. After using FeN₃P-SAzyme to treat mouse tumors, the tumor growth of the mice was significantly inhibited.³²⁷

Fig. 31 Applications of Fe-SACs for infected-wound healing. (a) Photographic images of the extent of wound healing. (b and c) Changes in wound size in the mice in each group on days 0, 1, 3, 5, 8, and 11. (d) Changes in body weight of the mice during wound healing in each group. (e) Hematoxylin and Eosin (H&E) staining wound sections from each group (n = 4). (f) H&E staining sections of colons. (g) Photo of the colons in each group. Reprinted with permission from ref. 326. Copyright (2020), the American Chemical Society.

Fig. 32 Applications of Fe-SACs for infected-wound healing. (a) Cell viability of HepG2 hepatoma cells after incubation with different concentrations of FeN₃P-SAzyme, FeN₄-SAzyme and Fe₃O₄. (b) SAzymes induced time-dependent cell death. (c) Visualization of the internalization of FeN₃P-SAzyme in HepG2 tumour cells using TEM. (d) Confocal images of the intracellular uptake of FeN₃P-SAzyme in lysosomes. (e) Intracellular oxidative species in HepG2 cells treated with FeN₃P-SAzyme in the absence or presence of H₂O₂ or antioxidant glutathione (GSH). (f and g) Anti-tumour activity (f) and toxicity (g) of FeN₃P-SAzyme. Reprinted with permission from ref. 327. Copyright (2020), Springer Nature Limited.

4.2.5. Applications of Fe-SACs in other applications. In addition to the above-mentioned electrochemical catalysis and biocatalysis, Fe-SACs hold vast potential for other applications including but not limited to the catalytic oxidation of benzene under mild conditions, methane/ethane conversion and degradation of organic matter, epoxidation and C−H sulfonation.^{237,238,283,328,329}

The organic compounds in industrial wastewater pose a serious threat to humans. Hence, it is necessary to develop catalysts to degrade organic compounds. Among them, the most effective strategy is to use strong oxidants to remove difficult-to-degrade organics. ˙OH is an excellent strong oxidant, owing to its high oxidation potential of ∼2.8 V.³⁰³ Since Fe-SACs have been found to show POD-like activity, which catalyze H₂O₂ to generate a large amount of ˙OH, Fe-SACs in principle efficiently remove organic pollutants. For example, the Fe-SACs designed by An et al. in the presence of H₂O₂ had removal rates of 96%, 98% and 98% for methylene blue, rhodamine B and phenol, respectively.⁸⁷ Fe-SACs also catalyzed the dehydrogenation of diphenylhydrazine to azobenzene with a selectivity of 100% and yield of 75.3%.⁸⁶ Meanwhile, the POD-like activity of Fe-SACs enabled them to initiate Fenton reaction in the presence of H₂O₂, therefore effectively degrading phenol and other phenol aqueous solutions with a removal rate of ∼83%.³²⁹

In the production of fine chemicals, epoxidation occupies an important position, and selecting a suitable catalyst (e.g., single-atom catalysts) is particularly important. In a typical experiment, Tian et al.³³⁰ synthesized a highly dispersed Fe2 cluster via a wet-chemical strategy and used it for trans-stilbene epoxidation with a high conversion rate of 91% and selectivity of 93%. Wen et al.²⁸³ designed biomimetic photocatalysts with single-atom Fe active sites by chemical cross-linking. Under the irradiation of visible light or near-infrared light, the catalyst could promote the sulfonation of olefins to form β-ketosulfones with a yield of up to 94%. Obviously, Fe-SACs have other applications, such as air filtration systems, preventing the emission of volatile gases and electrochemical organic conversion.³³¹ In the future development, Fe-SACs can be more widely applied in the fields mentioned above.

5. Conclusion and outlook

Since Zhang et al. proposed the concept of SACs in 2011, SACs have attracted considerable interest as industrial catalysts and therapeutic nanozymes.² In recent years, the development of SACs, especially Fe-SACs, promises to inherit the main advantages of homogeneous catalysts (isolated active site), heterogeneous catalysts (good stability and facile to separate) and natural enzymes (excellent activity and selectivity). These advantages make Fe-SACs shine in the fields of electrochemical catalysis, in which the larger energy barriers for electrocatalytic reactions (e.g., ORR, OER, HER, CO₂RR and NRR) will be clearly reduced. Interestingly, the application of Fe-SACs in catalytic biomedicine has emerged. The active center (e.g., Fe–N_x species) of Fe-SACs shows a similar structure to the catalytically active center of metalloprotease, thus offering huge potential for achieving superior catalytic activity and selectivity to functionally mimic or even outperform natural enzymes. Multiple lines of evidence support the multi-enzymatic (POD-, OXD-, CAT- and SOD-like) activities of Fe-SACs, which have a greater application prospect in the biomedical domains of biosensing, wound disinfection, cancer treatment and oxidative stress cytoprotection.

However, despite the significant progress achieved, the majority of synthetic routes and application of Fe-SACs are still in the laboratory stage. In our opinion, there is still plenty of work to be done to overcome the major challenges in the design, synthesis, characterization and application of Fe-SACs (Fig. 33). These challenges are as follows:

Fig. 33 Prospects and future challenges of Fe-SACs, ranging from design, synthesis, and characterization to applications.

(1) One inevitable technological challenge to engineer SACs towards practical application is developing cost-effective and green synthetic processes for SACs. For the preparation of Fe-SACs, the energy consumption is always large. Thus far, most Fe-SACs are ultimately prepared via high-temperature (700–1000 °C) pyrolysis treatment. This high-temperature procedure requires a high energy input, which makes the use of Fe-SACs difficult in large-scale industrial production and hinders their further applications.

(2) Engineering of the composition, crystal phase (if applicable) and architecture of support materials and manipulating the metal–support interactions are essential to optimize the catalytic properties of Fe-SACs. However, integrating all these parameters into a limited nano/micro-scale space of support materials has not yet to be fully achieved due to the currently inadequate design strategies and techniques available for the synthesis of Fe-SACs. Considering the accumulating evidence, in addition to the chemical compositions of support materials, their uniform cavity and ordered structure significantly contribute to the spatial distribution of single Fe atoms. Firstly, an ideal support material should possess a large specific surface area, making it more convenient to obtain charge transport and electrolyte penetration in the deep active centers of the catalyst surface. Secondly, the tailored topological defects and appropriate crystallinity of support materials will be greatly beneficial for increasing the electron density and conductivity of SACs, as well as further activating the inactive single Fe atoms. This is where natural organisms provide a rich source of inspiration for the innovative synthesis methods of support materials. Nature abounds with intricate hierarchical textures consisting of amorphous and crystalline hybrid phases, which synergistically intertwine to offer extraordinary functions, e.g., enhanced surface area, superior mechanical integrity and high transport rate.^332,333 Notably, these features are an integrated embodiment for high-performance electrochemical catalytic pathway. Other examples of natural evolution-inspired strategies are the use of pre-formed solid intermediates as starting phases, which can enable the bottom-up creation of smart materials with robust functionality compared to the substances found in nature. To date, a representative example is that the magnetosome of magnetotactic bacteria always exploits solid, poorly crystalline ferrihydrite intermediates as precursors in the biomineralization of stoichiometric and structurally pure Fe₃O₄ crystals.³³⁴ The synthesis of bioinspired materials via the solid phase is an alternative approach to manufacture a variety of sophisticated support materials with precise control over their chemical composition, crystal structure, size and morphology by means of biomolecular templates.

(3) Functionally similar to natural enzymes, the design and synthesis of Fe-SACs enabling precise control over a high metal loading and single-atomic dispersion are a challenging yet highly rewarding endeavor, possibly resulting in long-awaited industrial breakthrough in utilizing catalysts. The catalytic activity of Fe-SACs is directly related to the loading amounts of single-atom Fe metals on the support material. However, Fe-SACs usually suffer from a low-loading of Fe atoms (generally less than 10%), which is inherent to the classical material-design dilemma for Fe-SACs, namely, the high Fe loading and single-atomic dispersion (the two typical characters) tend to be mutually exclusive. Because single metal atoms inherently display a huge specific surface energy, they easily undergo migration, agglomeration and sintering during their preparation and catalytic reactions, hence deteriorating the overall catalytic activities of Fe-SACs. However, Wang et al.³³⁵ recently synthesized a series of single-atom catalysts with a high loading by calcining coordination polymers at high temperature, and the synthesis reached the gram scale. Interestingly, this method was suitable for a variety of metal atoms, including Fe (30.0 wt%), Ni (21.6 wt%), Cu (22.4 wt%), Zn (21.1 wt%), Ru (13.5 wt%), Rh (3.5 wt%), Pd (3.8 wt%), Pt (3.2 wt%) and Ir (4.4 wt%). Therefore, this high loading is expected to shine in the field of industrial catalysis.

(4) From the point of view of practical applications, how to avoid the poisoning of Fe-SAC catalysts remains a significant challenge. Unlike common catalysts, SACs will be completely deactivated if the intermediates or by-products are firmly combined with SACs. Generally, it is difficult for active intermediates to desorb from single metal atoms, owing to the high binding energy between them. Therefore, it is worth further studying to regenerate the isolated metal active sites for the recycling use of Fe-SACs.

(5) The identification of the relationship between the internal structure and catalytic activity of Fe-SACs is particularly important for the rational synthesis of catalysts, which requires more advanced characterization tools to observe the atomic active sites. Beyond the modern characterization techniques that have been utilized, new-generation convergent probes with a quicker detection rate and better temporal and spatial resolution should be developed to record instantaneous variations in the atomic active centers on the carrier. Furthermore, advanced instruments in other research fields including cryogenic transmission electron microscopy, which has evolved into an indispensable method for the characterization of structural biology, can be introduced in SAC-related fields. These new techniques are pivotal to prevent the collapse of the support skeleton and original single Fe atom aggregation during the characterization of Fe-SACs.

Considering aforementioned fundamental and technical hurdles to be settled, Fe-SACs have recently emerged as an interdisciplinary research field from chemistry, physics, catalytic science and biomedicine. Notably, the research on Fe-SACs is still in its infancy and further development of this field calls for more efforts. We expect that significant and promising directions would continue to emerge well beyond the outlook described herein.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (51803161), the Natural Science Foundation of Hubei Province, China (2021CFB344), the Young Top-notch Talent Cultivation Program of Hubei Province, China, and the Shanghai Municipal Science and Technology Major Project.

References

1. M. A. Roberts, G. Sankar, J. M. Thomas, R. H. Jones, H. Du, J. Chen, W. Pang and R. Xu, Nature, 1996, 381, 401–404.
2. 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.
3. F. Li, G.-F. Han, H.-J. Noh, S.-J. Kim, Y. Lu, H. Y. Jeong, Z. Fu and J.-B. Baek, Energy Environ. Sci., 2018, 11, 2263–2269.
4. Y. Qu, L. Wang, Z. Li, P. Li, Q. Zhang, Y. Lin, F. Zhou, H. Wang, Z. Yang, Y. Hu, M. Zhu, X. Zhao, X. Han, C. Wang, Q. Xu, L. Gu, J. Luo, L. Zheng and Y. Wu, Adv. Mater., 2019, 31, 1904496.
5. J. Li, H. Liu, M. Wang, C. Lin, W. Yang, J. Meng, Y. Xu, K. A. Owusu, B. Jiang, C. Chen, D. Fan, L. Zhou and L. Mai, Chem. Commun., 2019, 55, 334–337.
6. X. Ge, P. Zhou, Q. Zhang, Z. Xia, S. Chen, P. Gao, Z. Zhang, L. Gu and S. Guo, Angew. Chem., Int. Ed., 2019, 59, 232–236.
7. L. Kuai, Z. Chen, S. Liu, E. Kan, N. Yu, Y. Ren, C. Fang, X. Li, Y. Li and B. Geng, Nat. Commun., 2020, 11, 48.
8. Q. Liu, J. Wang, J. Zhang, Y. Yan, X. Qiu, S. Wei and Y. Tang, Catal. Sci. Technol., 2020, 10, 450–457.
9. C. Zhu, S. Fu, J. Song, Q. Shi, D. Su, M. H. Engelhard, X. Li, D. Xiao, D. Li, L. Estevez, D. Du and Y. Lin, Small, 2017, 13, 1603407.
10. W. Chen, J. Pei, C.-T. He, J. Wan, H. Ren, Y. Zhu, Y. Wang, J. Dong, S. Tian, W.-C. Cheong, S. Lu, L. Zheng, X. Zheng, W. Yan, Z. Zhuang, C. Chen, Q. Peng, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2017, 56, 16086–16090.
11. C. Ling, X. Bai, Y. Ouyang, A. Du and J. Wang, J. Phys. Chem. C, 2018, 122, 16842–16847.
12. L. Han, X. Liu, J. Chen, R. Lin, H. Liu, F. Lue, 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.
13. Y. Cheng, S. He, S. Lu, J.-P. Veder, B. Johannessen, L. Thomsen, M. Saunders, T. Becker, R. De Marco, Q. Li, S.-Z. Yang and S. P. Jiang, Adv. Sci., 2019, 6, 1802066.
14. R. Jiang, L. Li, T. Sheng, G. Hu, Y. Chen and L. Wang, J. Am. Chem. Soc., 2018, 140, 11594–11598.
15. Q. Li, W. Chen, H. Xiao, Y. Gong, Z. Li, L. Zheng, X. Zheng, W. Yan, W.-C. Cheong, R. Shen, N. Fu, L. Gu, Z. Zhuang, C. Chen, D. Wang, Q. Peng, J. Li and Y. Li, Adv. Mater., 2018, 30, 1800588.
16. F. Xiao, G.-L. Xu, C.-J. Sun, M. Xu, W. Wen, Q. Wang, M. Gu, S. Zhu, Y. Li, Z. Wei, X. Pan, J. Wang, K. Amine and M. Shao, Nano Energy, 2019, 61, 60–68.
17. J. Guo, J. Huo, Y. Liu, W. Wu, Y. Wang, M. Wu, H. Liu and G. Wang, Small Methods, 2019, 3, 1900159.
18. A. Han, B. Wang, A. Kumar, Y. Qin, J. Jin, X. Wang, C. Yang, B. Dong, Y. Jia, J. Liu and X. Sun, Small Methods, 2019, 3, 1800471.
19. Q.-L. Zhu, W. Xia, L.-R. Zheng, R. Zou, Z. Liu and Q. Xu, ACS Energy Lett., 2017, 2, 504–511.
20. J. Meng, X. Liu, C. Niu, Q. Pang, J. Li, F. Liu, Z. Liu and L. Mai, Chem. Soc. Rev., 2020, 49, 3142–3186.
21. W.-H. Li, W.-H. Deng, G.-E. Wang and G. Xu, EnergyChem, 2020, 2, 100029.
22. J. Wang, G. Han, L. Wang, L. Du, G. Chen, Y. Gao, Y. Ma, C. Du, X. Cheng, P. Zuo and G. Yin, Small, 2018, 14, 1704282.
23. H. Huang, K. Shen, F. Chen and Y. Li, ACS Catal., 2020, 10, 6579–6586.
24. X. Chen, N. Wang, K. Shen, Y. Xie, Y. Tan and Y. Li, ACS Appl. Mater. Interfaces, 2019, 11, 25976–25985.
25. X. Wang, H. Zhu, C. Yang, J. Lu, L. Zheng and H.-P. Liang, Carbon, 2022, 191, 393–402.
26. 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.
27. C. Zhao, X. Dai, T. Yao, W. Chen, X. Wang, J. Wang, J. Yang, S. Wei, Y. Wu and Y. Li, J. Am. Chem. Soc., 2017, 139, 8078–8081.
28. V. Armel, S. Hindocha, F. Salles, S. Bennett, D. Jones and F. Jaouen, J. Am. Chem. Soc., 2017, 139, 453–464.
29. Y. Hou, M. Qiu, M. G. Kim, P. Liu, G. Nam, T. Zhang, X. Zhuang, B. Yang, J. Cho and M. Chen, Nat. Commun., 2019, 10, 1392.
30. L. Jiao, G. Wan, R. Zhang, H. Zhou, S.-H. Yu and H.-L. Jiang, Angew. Chem., Int. Ed., 2018, 57, 8525–8529.
31. D.-L. Meng, C.-H. Chen, J.-D. Yi, Q. Wu, J. Liang, Y.-B. Huang and R. Cao, Research, 2019, 2019, 1768595.
32. M. S. Lohse and T. Bein, Adv. Funct. Mater., 2018, 28, 1705553.
33. H. Wang, Z. Zeng, P. Xu, L. Li, G. Zeng, R. Xiao, Z. Tang, D. Huang, L. Tang, C. Lai, D. Jiang, Y. Liu, H. Yi, L. Qin, S. Ye, X. Ren and W. Tang, Chem. Soc. Rev., 2019, 48, 488–516.
34. G. Lin, H. Ding, D. Yuan, B. Wang and C. Wang, J. Am. Chem. Soc., 2016, 138, 3302–3305.
35. X. Ao, W. Zhang, Z. Li, J.-G. Li, L. Soule, X. Huang, W.-H. Chiang, H. M. Chen, C. Wang, M. Liu and X. C. Zeng, ACS Nano, 2019, 13, 11853–11862.
36. J.-C. Li, P.-X. Hou and C. Liu, Small, 2017, 13, 1702002.
37. S. M. J. Rogge, A. Bavykina, J. Hajek, H. Garcia, A. I. Olivos-Suarez, A. Sepulveda-Escribano, A. Vimont, G. Clet, P. Bazin, F. Kapteijn, M. Daturi, E. V. Ramos-Fernandez, F. X. Llabres i Xamena, V. Van Speybroeck and J. Gascon, Chem. Soc. Rev., 2017, 46, 3134–3184.
38. X. Yan, H. Liu, Y. Li, W. Chen, T. Zhang, Z. Zhao, G. Xing and L. Chen, Macromolecules, 2019, 52, 7977–7983.
39. Y. Zeng, R. Zou, Z. Luo, H. Zhang, X. Yao, X. Ma, R. Zou and Y. Zhao, J. Am. Chem. Soc., 2015, 137, 1020–1023.
40. N. Huang, L. Zhai, D. E. Coupry, M. A. Addicoat, K. Okushita, K. Nishimura, T. Heine and D. Jiang, Nat. Commun., 2016, 7, 12325.
41. Y. Zhu, W. Peng, Y. Li, G. Zhang, F. Zhang and X. Fan, Small Methods, 2019, 3, 1800438.
42. J.-D. Yi, R. Xu, Q. Wu, T. Zhang, K.-T. Zang, J. Luo, Y.-L. Liang, Y.-B. Huang and R. Cao, ACS Energy Lett., 2018, 3, 883–889.
43. A. M. Evans, L. R. Parent, N. C. Flanders, R. P. Bisbey, E. Vitaku, M. S. Kirschner, R. D. Schaller, L. X. Chen, N. C. Gianneschi and W. R. Dichtel, Science, 2018, 361, 52–57.
44. T. Hasell and A. I. Cooper, Nat. Rev. Mater., 2016, 1, 16053.
45. P. Peng, L. Shi, F. Huo, S. Zhang, C. Mi, Y. Cheng and Z. Xiang, ACS Nano, 2019, 13, 878–884.
46. L. Wang, C. Xu, W. Zhang, Q. Zhang, M. Zhao, C. Zeng, Q. Jiang, C. Gu and Y. Ma, J. Am. Chem. Soc., 2022, 144, 8961–8968.
47. Z. Xiong, B. Sun, H. Zou, R. Wang, Q. Fang, Z. Zhang and S. Qiu, J. Am. Chem. Soc., 2022, 144, 6583–6593.
48. W. Zhang, L. Chen, S. Dai, C. Zhao, C. Ma, L. Wei, M. Zhu, S. Y. Chong, H. Yang, L. Liu, Y. Bai, M. Yu, Y. Xu, X.-W. Zhu, Q. Zhu, S. An, R. S. Sprick, M. A. Little, X. Wu, S. Jiang, Y. Wu, Y.-B. Zhang, H. Tian, W.-H. Zhu and A. I. Cooper, Nature, 2022, 604, 72–79.
49. J.-C. Li, M. Cheng, T. Li, L. Ma, X. Ruan, D. Liu, H.-M. Cheng, C. Liu, D. Du, Z. Wei, Y. Lin and M. Shao, J. Mater. Chem. A, 2019, 7, 14478–14482.
50. Y. Li, W. Zhou, H. Wang, L. Xie, Y. Liang, F. Wei, J.-C. Idrobo, S. J. Pennycook and H. Dai, Nat. Nanotechnol., 2012, 7, 394–400.
51. Y. Ma, L. Sun, W. Huang, L. Zhang, J. Zhao, Q. Fan and W. Huang, J. Phys. Chem. C, 2011, 115, 24592–24597.
52. P. Chen, T. Zhou, L. Xing, K. Xu, Y. Tong, H. Xie, L. Zhang, W. Yan, W. Chu, C. Wu and Y. Xie, Angew. Chem., Int. Ed., 2017, 56, 610–614.
53. N. Cheng, J.-C. Li, D. Liu, Y. Lin and D. Du, Small, 2019, 15, 1901485.
54. J.-C. Li, Z.-Q. Yang, D.-M. Tang, L. Zhang, P.-X. Hou, S.-Y. Zhao, C. Liu, M. Cheng, G.-X. Li, F. Zhang and H.-M. Cheng, NPG Asia Mater., 2018, 10, e461–e461.
55. D. Liu, J.-C. Li, Q. Shi, S. Feng, Z. Lyu, S. Ding, L. Hao, Q. Zhang, C. Wang, M. Xu, T. Li, E. Sarnello, D. Du and Y. Lin, ACS Appl. Mater. Interfaces, 2019, 11, 39820–39826.
56. Y. J. Sa, D.-J. Seo, J. Woo, J. T. Lim, J. Y. Cheon, S. Y. Yang, J. M. Lee, D. Kang, T. J. Shin, H. S. Shin, H. Y. Jeong, C. S. Kim, M. G. Kim, T.-Y. Kim and S. H. Joo, J. Am. Chem. Soc., 2016, 138, 15046–15056.
57. Y. Wang, J. Mao, X. Meng, L. Yu, D. Deng and X. Bao, Chem. Rev., 2019, 119, 1806–1854.
58. Y. Oh, J. O. Hwang, E.-S. Lee, M. Yoon, V.-D. Le, Y.-H. Kim, D. H. Kim and S. O. Kim, ACS Appl. Mater. Interfaces, 2016, 8, 25438–25443.
59. X. Wang, X. Chen, A. Thomas, X. Fu and M. Antonietti, Adv. Mater., 2009, 21, 1609.
60. Y. Zheng, Y. Jiao, Y. Zhu, Q. Cai, A. Vasileff, L. H. Li, Y. Han, Y. Chen and S.-Z. Qiao, J. Am. Chem. Soc., 2017, 139, 3336–3339.
61. D. Liu, S. Ding, C. Wu, W. Gan, C. Wang, D. Cao, Z. U. Rehman, Y. Sang, S. Chen, X. Zheng, Y. Wang, B. Ge and L. Song, J. Mater. Chem. A, 2018, 6, 6840–6846.
62. Y. Peng, X. Yu, W. Yin, W. Dong, J. Peng and T. Wang, ACS Appl. Bio Mater., 2020, 3, 59–67.
63. W.-J. Ong, L.-L. Tan, S.-P. Chai, S.-T. Yong and A. R. Mohamed, Nano Energy, 2015, 13, 757–770.
64. R. B. N. Baig, S. Verma, R. S. Varma and M. N. Nadagouda, ACS Sustainable Chem. Eng., 2016, 4, 1661–1664.
65. L. Yang, D. Cheng, X. Zeng, X. Wan, J. Shui, Z. Xiang and D. Cao, Proc. Natl. Acad. Sci. U. S. A., 2018, 115, 6626–6631.
66. Y. Tang, J. Zhou, Z. Shen, W. Chen, C. Li and X. Dai, RSC Adv., 2016, 6, 93985–93996.
67. H. Amani, E. Mostafavi, H. Arzaghi, S. Davaran, A. Akbarzadeh, O. Akhavan, H. Pazoki-Toroudi and T. J. Webster, ACS Biomater. Sci. Eng., 2019, 5, 193–214.
68. 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.
69. X. Qiu, X. Yan, H. Pang, J. Wang, D. Sun, S. Wei, L. Xu and Y. Tang, Adv. Sci., 2019, 6, 1801103.
70. Z.-y Gao, W.-j Yang, X.-l Ding, G. Lv and W.-p Yan, Phys. Chem. Chem. Phys., 2018, 20, 7333–7341.
71. A. W. Robertson, B. Montanari, K. He, J. Kim, C. S. Allen, Y. A. Wu, J. Olivier, J. Neethling, N. Harrison, A. I. Kirkland and J. H. Warner, Nano Lett., 2013, 13, 1468–1475.
72. X. Zhao, X. Liu, B. Huang, P. Wang and Y. Pei, J. Mater. Chem. A, 2019, 7, 24583–24593.
73. W. Yang, Z. Gao, X. Liu, X. Li, X. Ding and W. Yan, Catal. Sci. Technol., 2018, 8, 4159–4168.
74. 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.
75. J. Zheng, K. Lebedev, S. Wu, C. Huang and S. Tsang, J. Am. Chem. Soc., 2021, 143, 7979–7990.
76. J. Chen, Y. Kang, W. Zhang, Z. Zhang, Y. Chen, Y. Yang, L. Duan, Y. Li and W. Li, Angew. Chem., Int. Ed., 2022, 61, e202203022.
77. S. Wu, Z. Chen, W. Yue, S. Mine, T. Toyao, M. Matsuoka, X. Xi, L. Wang and J. Zhang, ACS Catal., 2021, 11, 4362–4371.
78. Z. Jiang, M. Tian, M. Jing, S. Chai, Y. Jian, C. Chen, M. Douthwaite, L. Zheng, M. Ma, W. Song, J. Liu, J. Yu and C. He, Angew. Chem., Int. Ed, 2022, 61, e202200763.
79. Z. Zhang, J. Liu, J. Wang, Q. Wang, Y. Wang, K. Wang, Z. Wang, M. Gu, Z. Tang, J. Lim, T. Zhao and F. Ciucci, Nat. Commun., 2021, 12, 5235.
80. R. Nietuby, E. Sobczak and K. E. Attenkofer, J. Alloys Compd., 2001, 328, 126–131.
81. J. Yang, B. Chen, X. Liu, W. Liu, Z. Li, J. Dong, W. Chen, W. Yan, T. Yao, X. Duan, Y. Wu and Y. Li, Angew. Chem., Int. Ed., 2018, 57, 9495–9500.
82. X. Li, W. Bi, L. Zhang, S. Tao and Y. Xie, Adv. Mater., 2016, 28, 2427–2431.
83. J. T. Yates, T. M. Duncan, S. D. Worley and R. W. Vaughan, J. Chem. Phys., 1979, 70, 1219–1224.
84. J. Liu, ACS Catal, 2017, 7, 34–59.
85. J. H. Kwak, J. Hu, D. Mei, C. W. Yi, D. H. Kim, C. H. Peden, L. F. Allard and J. Szanyi, Science, 2009, 325, 1670–1673.
86. F. He, L. Mi, Y. Shen, T. Mori, S. Liu and Y. Zhang, ACS Appl. Mater. Interfaces, 2018, 10, 35327–35333.
87. S. An, G. Zhang, J. Liu, K. Li, G. Wan, Y. Liang, D. Ji, J. T. Miller, C. Song, W. Liu, Z. Liu and X. Guo, Chinese J. Catal., 2020, 41, 1198–1207.
88. M. Huo, L. Wang, H. Zhang, L. Zhang, Y. Chen and J. Shi, Small, 2019, 15, 1901834.
89. M. Lu, C. Wang, Y. Ding, M. Peng, W. Zhang, K. Li, W. Wei and Y. Lin, Chem. Commun., 2019, 55, 14534–14537.
90. B. Lu, Q. Liu and S. Chen, ACS Catal., 2020, 10, 7584–7618.
91. A. S. Varela, M. Kroschel, N. D. Leonard, W. Ju, J. Steinberg, A. Bagger, J. Rossmeisl and P. Strasser, ACS Energy Lett., 2018, 3, 812–817.
92. Q. Zhang and J. Guan, Adv. Funct. Mater., 2020, 30, 2000768.
93. X. Wang, Y. Jia, X. Mao, D. Liu, W. He, J. Li, J. Liu, X. Yan, J. Chen, L. Song, A. Du and X. Yao, Adv. Mater., 2020, 32, 2000966.
94. L. Bai, Z. Duan, X. Wen and J. Guan, J. Catal., 2019, 378, 353–362.
95. S. Zhang, M. Jin, T. Shi, M. Han, Q. Sun, Y. Lin, Z. Ding, L. R. Zheng, G. Wang, Y. Zhang, H. Zhang and H. Zhao, Angew. Chem., Int. Ed., 2020, 59, 13423–13429.
96. Y. Wu, J. Cai, Y. Xie, S. Niu, Y. Zang, S. Wu, Y. Liu, Z. Lu, Y. Fang, Y. Guan, X. Zheng, J. Zhu, X. Liu, G. Wang and Y. Qian, Adv. Mater., 2020, 32, 1904346.
97. T. N. Huan, N. Ranjbar, G. Rousse, M. Sougrati, A. Zitolo, V. Mougel, F. Jaouen and M. Fontecave, ACS Catal, 2017, 7, 1520–1525.
98. J. He, T. Zheng, D. Wu, S. Zhang, M. Gu and Q. He, ACS Catal, 2022, 12, 1601–1613.
99. X. Li and Z. Xiang, Nat. Commun., 2022, 13, 57.
100. C. C. Wang, D. W. Luo, B. X. Hou, M. Gholizadeh, Z. X. Tan and X. J. Han, J. Renewable Mater., 2022, 10, 1337–1348.
101. Z. Yang, Y. Wang, M. Zhu, Z. Li, W. Chen, W. Wei, T. Yuan, Y. Qu, Q. Xu, C. Zhao, X. Wang, P. Li, Y. Li, Y. Wu and Y. Li, ACS Catal., 2019, 9, 2158–2163.
102. J.-C. Li, F. Xiao, H. Zhong, T. Li, M. Xu, L. Ma, M. Cheng, D. Liu, S. Feng, Q. Shi, H.-M. Cheng, C. Liu, D. Du, S. P. Beckman, X. Pan, Y. Lin and M. Shao, ACS Catal., 2019, 9, 5929–5934.
103. J. Zhang, Q.-A. Huang, J. Wang, J. Wang, J. Zhang and Y. Zhao, Chin. J. Catal., 2020, 41, 783–798.
104. X. Zeng, J. Shui, X. Liu, Q. Liu, Y. Li, J. Shang, L. Zheng and R. Yu, Adv. Energy Mater., 2018, 8, 1701345.
105. L. Jiao and H.-L. Jiang, Chem, 2019, 5, 786–804.
106. D.-S. Bin, Z.-X. Chi, Y. Li, K. Zhang, X. Yang, Y.-G. Sun, J.-Y. Piao, A.-M. Caoand and L.-J. Wan, J. Am. Chem. Soc., 2017, 139, 13492–13498.
107. Y.-G. Sun, J.-Y. Piao, L.-L. Hu, D.-S. Bin, X.-J. Lin, S.-Y. Duan, A.-M. Cao and L.-J. Wan, J. Am. Chem. Soc., 2018, 140, 9070–9073.
108. S.-N. Zhao, J.-K. Li, R. Wang, J. Cai and S.-Q. Zang, Adv. Mater., 2022, 34, 2107291.
109. K.-M. Zhao, S. Liu, Y.-Y. Li, X. Wei, G. Ye, W. Zhu, Y. Su, J. Wang, H. Liu, Z. He, Z.-Y. Zhou and S.-G. Sun, Adv. Energy Mater., 2022, 12, 2103588.
110. C.-C. Hou, L. Zou, L. Sun, K. Zhang, Z. Liu, Y. Li, C. Li, R. Zou, J. Yu and Q. Xu, Angew. Chem., Int. Ed., 2020, 59, 7384–7389.
111. J. Yang, Z. Wang, C.-X. Huang, Y. Zhang, Q. Zhang, C. Chen, J. Du, X. Zhou, Y. Zhang, H. Zhou, L. Wang, X. Zheng, L. Gu, L.-M. Yang and Y. Wu, Angew. Chem., Int. Ed., 2021, 60, 22722–22728.
112. J. Wang, Z. Q. Huang, W. Liu, C. R. Chang, H. L. Tang, Z. J. Li, W. X. Chen, C. J. Jia, T. Yao, S. Q. Wei, Y. Wu and Y. D. Lie, J. Am. Chem. Soc., 2017, 139, 17281–17284.
113. L. Gong, H. Zhang, Y. Wang, E. Luo, K. Li, L. Gao, Y. Wang, Z. Wu, Z. Jin, J. Ge, Z. Jiang, C. Liu and W. Xing, Angew. Chem., Int. Ed., 2020, 59, 13923–13928.
114. X. Jia, W. Gu, S. Zhang and E. Wang, ChemElectroChem, 2017, 4, 2005–2011.
115. T. Sharifi, E. Gracia-Espino, A. Chen, G. Hu and T. Wagberg, Adv. Energy Mater., 2020, 10, 1902084.
116. L. Zheng, Y. Zhao, H. Zhang, W. Xia and J. Tang, ChemSusChem, 2022, 15, e202102642.
117. H. Zhang, S. Hwang, M. Wang, Z. Feng, S. Karakalos, L. Luo, Z. Qiao, X. Xie, C. Wang, D. Su, Y. Shao and G. Wu, J. Am. Chem. Soc., 2017, 139, 14143–14149.
118. K. Yuan, S. Sfaelou, M. Qiu, D. Lützenkirchen-Hecht, X. Zhuang, Y. Chen, C. Yuan, X. Feng and U. Scherf, ACS Energy Lett., 2018, 3, 252–260.
119. Y. Li, T. Liu, W. Yang, Z. Zhu, Y. Zhai, W. Gu and C. Zhu, Nanoscale, 2019, 11, 19506–19511.
120. Y.-G. Lee and H.-J. Ahn, Appl. Surf. Sci., 2019, 487, 389–397.
121. L. Hu, C. Dai, L. Chen, Y. Zhu, Y. Hao, Q. Zhang, L. Gu, X. Feng, S. Yuan, L. Wang and B. Wang, Angew. Chem., Int. Ed, 2021, 60, 27324–27329.
122. J.-C. Li, H. Zhong, M. Xu, T. Li, L. Wang, Q. Shi, S. Feng, Z. Lyu, D. Liu, D. Du, S. P. Beckman, X. Pan, Y. Lin and M. Shao, Sci. China Mater., 2019, 63, 965–971.
123. L. Yu, Y. Li and Y. Ruan, Angew. Chem., Int. Ed., 2021, 60, 25296–25301.
124. Y. Chen, S. Ji, S. Zhao, W. Chen, J. Dong, W.-C. Cheong, R. Shen, X. Wen, L. Zheng, A. I. Rykov, S. Cai, H. Tang, Z. Zhuang, C. Chen, Q. Peng, D. Wang and Y. Li, Nat. Commun., 2018, 9, 5422.
125. Q. Lai, L. Zheng, Y. Liang, J. He, J. Zhao and J. Chen, ACS Catal., 2017, 7, 1655–1663.
126. P.-J. Wei, G.-Q. Yu, Y. Naruta and J.-G. Liu, Angew. Chem., Int. Ed., 2014, 53, 6659–6663.
127. Y.-M. Zhao, P.-C. Zhang, C. Xu, X.-Y. Zhou, L.-M. Liao, P.-J. Wei, E. Liu, H. Chen, Q. He and J.-G. Liu, ACS Appl. Mater. Interfaces, 2020, 12, 17334–17342.
128. A. Zitolo, V. Goellner, V. Armel, M.-T. Sougrati, T. Mineva, L. Stievano, E. Fonda and F. Jaouen, Nat. Mater., 2015, 14, 937.
129. S. Yang, T. Zhang, G. Li, L. Yang and J. Y. Lee, Energy Stor. Mater., 2017, 6, 140–148.
130. A. Han, X. Wang, K. Tang, Z. Zhang, C. Ye, K. Kong, H. Hu, L. Zheng, P. Jiang, C. Zhao, Q. Zhang, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2021, 60, 19262–19271.
131. Z. Lu, B. Wang, Y. Hu, W. Liu, Y. Zhao, R. Yang, Z. Li, J. Luo, B. Chi, Z. Jiang, M. Li, S. Mu, S. Liao, J. Zhang and X. Sun, Angew. Chem., Int. Ed., 2019, 58, 2622–2626.
132. S. Yang, X. Xue, J. Zhang, X. Liu, C. Dai, Q. Xu, J. Lian, Y. Zhao, G. Li, H. Li and S. Yuan, Chem. Eng. J., 2020, 395, 125064.
133. J. Li, H. Zhang, W. Samarakoon, W. Shan, D. A. Cullen, S. Karakalos, M. Chen, D. Gu, K. L. More, G. Wang, Z. Feng, Z. Wang and G. Wu, Angew. Chem., Int. Ed., 2019, 58, 18971–18980.
134. Y. Sun, S. Sun, H. Yang, S. Xi, J. Gracia and Z. J. Xu, Adv. Mater., 2020, 32, 2003297.
135. D. Xia, X. Yang, L. Xie, Y. Wei, W. Jiang, M. Dou, X. Li, J. Li, L. Gan and F. Kang, Adv. Funct. Mater., 2019, 29, 1906174.
136. H. Wang, R. Liu, Y. Li, X. Lü, Q. Wang, S. Zhao, K. Yuan, Z. Cui, X. Li, S. Xin, R. Zhang, M. Lei and Z. Lin, Joule, 2018, 2, 337–348.
137. C. Mu, J. Mao, J. Guo, Q. Guo, Z. Li, W. Qin, Z. Hu, K. Davey, T. Ling and S.-Z. Qiao, Adv. Mater., 2020, 32, 1907168.
138. Z. Li, Z. Zhuang, F. Lv, H. Zhu, L. Zhou, M. Luo, J. Zhu, Z. Lang, S. Feng, W. Chen, L. Mai and S. Guo, Adv. Mater., 2018, 30, 1803220.
139. X.-T. Wang, T. Ouyang, L. Wang, J.-H. Zhong, T. Ma and Z.-Q. Liu, Angew. Chem., Int. Ed., 2019, 58, 13291–13296.
140. G. Shen, R. Zhang, L. Pan, F. Hou, Y. Zhao, Z. Shen, W. Mi, C. Shi, Q. Wang, X. Zhang and J.-J. Zou, Angew. Chem., Int. Ed., 2020, 59, 2313–2317.
141. T. Ouyang, X.-T. Wang, X.-Q. Mai, A.-N. Chen, Z.-Y. Tang and Z.-Q. Liu, Angew. Chem., Int. Ed., 2020, 59, 11948–11957.
142. 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.
143. L. Gong, H. Zhang, Y. Wang, E. Luo, K. Li, L. Gao, Y. Wang, Z. Wu, Z. Jin, J. Ge, Z. Jiang, C. Liu and W. Xing, Angew. Chem., Int. Ed., 2020, 59, 13923–13928.
144. J. Han, X. Meng, L. Lu, J. Bian, Z. Li and C. Sun, Adv. Funct. Mater., 2019, 29, 1808872.
145. C. C. L. McCrory, S. Jung, I. M. Ferrer, S. M. Chatman, J. C. Peters and T. F. Jaramillo, J. Am. Chem. Soc., 2015, 137, 4347–4357.
146. X. Xie, C. He, B. Li, Y. He, D. A. Cullen, E. C. Wegener, A. J. Kropf, U. Martinez, Y. Cheng, M. H. Engelhard, M. E. Bowden, M. Song, T. Lemmon, X. S. Li, Z. Nie, J. Liu, D. J. Myers, P. Zelenay, G. Wang, G. Wu, V. Ramani and Y. Shao, Nat. Catal., 2020, 3, 1044–1054.
147. F. Zhou, P. Yu, F. Sun, G. Zhang, X. Liu and L. Wang, J. Mater. Chem. A, 2021, 9, 6831–6840.
148. J. Li, M. T. Sougrati, A. Zitolo, J. M. Ablett, I. C. Oğuz, T. Mineva, I. Matanovic, P. Atanassov, Y. Huang, I. Zenyuk, A. Di Cicco, K. Kumar, L. Dubau, F. Maillard, G. Dražić and F. Jaouen, Nat. Catal., 2021, 4, 10–19.
149. C.-Y. Su, H. Cheng, W. Li, Z.-Q. Liu, N. Li, Z. Hou, F.-Q. Bai, H.-X. Zhang and T.-Y. Ma, Adv. Energy Mater., 2017, 7, 1602420.
150. L. Bai, C.-S. Hsu, D. T. L. Alexander, H. M. Chen and X. Hu, J. Am. Chem. Soc., 2019, 141, 14190–14199.
151. J. Wu, L. Xiong, B. Zhao, M. Liu and L. Huang, Small Methods, 2020, 4, 1900540.
152. Á. Valdés, J. Brillet, M. Grätzel, H. Gudmundsdóttir, H. A. Hansen, H. Jónsson, P. Klüpfel, G.-J. Kroes, F. Le Formal, I. C. Man, R. S. Martins, J. K. Nørskov, J. Rossmeisl, K. Sivula, A. Vojvodic and M. Zäch, Phys. Chem. Chem. Phys., 2012, 14, 49–70.
153. P. Li, M. Wang, X. Duan, L. Zheng, X. Cheng, Y. Zhang, Y. Kuang, Y. Li, Q. Ma, Z. Feng, W. Liu and X. Sun, Nat. Commun., 2019, 10, 1711.
154. L. An, J. Feng, Y. Zhang, Y.-Q. Zhao, R. Si, G.-C. Wang, F. Cheng, P. Xi and S. Sun, Nano Energy, 2019, 57, 644–652.
155. W. Zhai, S. Huang, C. Lu, X. Tang, L. Li, B. Huang, T. Hu, K. Yuan, X. Zhuang and Y. Chen, Small, 2022, 18, 2107225.
156. X. Zhang, Q. Zhang, J. Cui, J. Yan, J. Liu and Y. Wu, Nanoscale, 2022, 14, 3212–3223.
157. H. Fei, J. Dong, Y. Feng, C. S. Allen, C. Wan, B. Volosskiy, M. Li, Z. Zhao, Y. Wang, H. Sun, P. An, W. Chen, Z. Guo, C. Lee, D. Chen, I. Shakir, M. Liu, T. Hu, Y. Li, A. I. Kirkland, X. Duan and Y. Huang, Nat. Catal., 2018, 1, 63–72.
158. S. L. Candelaria, N. M. Bedford, T. J. Woehl, N. S. Rentz, A. R. Showalter, S. Pylypenko, B. A. Bunker, S. Lee, B. Reinhart, Y. Ren, S. P. Ertem, E. B. Coughlin, N. A. Sather, J. L. Horan, A. M. Herring and L. F. Greenleette, ACS Catal, 2017, 7, 365–379.
159. M. Zeng, Y. Liu, F. Zhao, K. Nie, N. Han, X. Wang, W. Huang, X. Song, J. Zhong and Y. Li, Adv. Funct. Mater., 2016, 26, 4397–4404.
160. R. Gao, Y. Yin, F. Niu, A. Wang, S. Li, H. Dong and S. Yang, ChemElectroChem, 2019, 6, 1824–1830.
161. S. Li, C. Cheng, X. Zhao, J. Schmidt and A. Thomas, Angew. Chem., Int. Ed., 2018, 57, 1856–1862.
162. M. Ma, A. Kumar, D. Wang, Y. Wang, Y. Jia, Y. Zhang, G. Zhang, Z. Yan and X. Sun, Appl. Catal., B, 2020, 274, 119091.
163. J. Chen, H. Li, C. Fan, Q. Meng, Y. Tang, X. Qiu, G. Fu and T. Ma, Adv. Mater., 2020, 32, e2003134.
164. B. Hammer and J. K. Norskov, Nature, 1995, 376, 238–240.
165. Y. Yang, Z. Lun, G. Xia, F. Zheng, M. He and Q. Chen, Energy Environ. Sci., 2015, 8, 3563–3571.
166. T. He, C. Zhang and A. Du, Chem. Eng. Sci., 2019, 194, 58–63.
167. Q. Zhang and J. Guan, Adv. Funct. Mater., 2020, 30, 2000768.
168. J. Li and G. Zheng, Adv. Sci, 2017, 4, 1600380.
169. J. Wang, F. Xu, H. Jin, Y. Chen and Y. Wang, Adv. Mater., 2017, 29, 1605838.
170. X. Yu, J. Zhao, L.-R. Zheng, Y. Tong, M. Zhang, G. Xu, C. Li, J. Ma and G. Shi, ACS Energy Lett., 2018, 3, 237–244.
171. S. Sultan, J. N. Tiwari, A. N. Singh, S. Zhumagali, M. Ha, C. W. Myung, P. Thangavel and K. S. Kim, Adv. Energy Mater., 2019, 9, 1900624.
172. X. Yu, J. Zhao, L.-R. Zheng, Y. J. Tong, M. Zhang, G. Xu, C. Li, J. Ma and G. Shi, ACS Energy Lett., 2018, 3, 237–244.
173. 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.
174. J. Chen, T. Wang, X. Wang, B. Yang, X. Sang, S. Zheng, S. Yao, Z. Li, Q. Zhang, L. Lei, J. Xu, L. Dai and Y. Hou, Adv. Funct. Mater., 2022, 32, 2110174.
175. 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.
176. J. Gu, C.-S. Hsu, L. Bai, H. M. Chen and X. Hu, Science, 2019, 364, 1091.
177. T. Wang, X. Sang, W. Zheng, B. Yang, S. Yao, C. Lei, Z. Li, Q. He, J. Lu, L. Lei, L. Dai and Y. Hou, Adv. Mater., 2020, 32, 2002430.
178. 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.
179. L. Lin, H. Li, Y. Wang, H. Li, P. Wei, B. Nan, R. Si, G. Wang and X. Bao, Angew. Chem., Int. Ed, 2021, 133, 26786–26790.
180. X. Sun, Y. Tuo, C. Ye, C. Chen, Q. Lu, G. Li, P. Jiang, S. Chen, P. Zhu, M. Ma, J. Zhang, J. H. Bitter, D. Wang and Y. Li, Angew. Chem., Int. Ed., 2021, 60, 23614–23618.
181. T. Wang, X. Sang, W. Zheng, B. Yang, S. Yao, C. Lei, Z. Li, Q. He, J. Lu, L. Lei, L. Dai and Y. Hou, Adv. Mater., 2020, 32, 2002430.
182. F. Pan, H. Zhang, K. Liu, D. Cullen, K. More, M. Wang, Z. Feng, G. Wang, G. Wu and Y. Li, ACS Catal, 2018, 8, 3116–3122.
183. X. Qin, S. Zhu, F. Xiao, L. Zhang and M. Shao, ACS Energy Lett., 2019, 4, 1778–1783.
184. H. B. Yang, S.-F. Hung, S. Liu, K. Yuan, S. Miao, L. Zhang, X. Huang, H.-Y. Wang, W. Cai, R. Chen, J. Gao, X. Yang, W. Chen, Y. Huang, H. M. Chen, C. M. Li, T. Zhang and B. Liu, Nat. Energy, 2018, 3, 140–147.
185. W. Ju, A. Bagger, G.-P. Hao, A. S. Varela, I. Sinev, V. Bon, B. Roldan Cuenya, S. Kaskel, J. Rossmeisl and P. Strasser, Nat. Commun., 2017, 8, 944.
186. L. Jiao, J. Zhu, Y. Zhang, W. Yang, S. Zhou, A. Li, C. Xie, X. Zheng, W. Zhou, S.-H. Yu and H.-L. Jiang, J. Am. Chem. Soc., 2021, 143(46), 19417–19424.
187. Z. Wang, Z. Yu and J. Zhao, Phys. Chem. Chem. Phys., 2018, 20, 12835–12844.
188. Y. Huang, D. D. Babu, Z. Peng and Y. Wang, Adv. Sci, 2020, 7, 1902390.
189. J. G. Chen, R. M. Crooks, L. C. Seefeldt, K. L. Bren, R. M. Bullock, M. Y. Darensbourg, P. L. Holland, B. Hoffman, M. J. Janik, A. K. Jones, M. G. Kanatzidis, P. King, K. M. Lancaster, S. V. Lymar, P. Pfromm, W. F. Schneider and R. R. Schrock, Science, 2018, 360, eaar6611.
190. J. Chatt, J. R. Dilworth and R. L. Richards, Chem. Rev., 1978, 78, 589–625.
191. L. Zhang, W. Zhao, W. Zhang, J. Chen and Z. Hu, Nano Research, 2019, 12, 1181–1186.
192. J. S. Anderson, G. E. Cutsail III, J. Rittle, B. A. Connor, W. A. Gunderson, L. Zhang, B. M. Hoffman and J. C. Peters, J. Am. Chem. Soc., 2015, 24, 7803–7809.
193. X.-F. Li, Q.-K. Li, J. Cheng, L. Liu, Q. Yan, Y. Wu, X.-H. Zhang, Z.-Y. Wang, Q. Qiu and Y. Luo, J. Am. Chem. Soc., 2016, 138, 8706–8709.
194. J.-C. Liu, X.-L. Ma, Y. Li, Y.-G. Wang, H. Xiao and J. Li, Nat. Commun., 2018, 9, 1610.
195. J. S. Anderson, J. Rittle and J. C. Peters, Nature, 2013, 501, 84–87.
196. Y. Li, J. Li, J. Huang, J. Chen, Y. Kong, B. Yang, Z. Li, L. Lei, G. Chai, Z. Wen, L. Dai and Y. Hou, Angew. Chem., Int. Ed., 2021, 60, 9078–9085.
197. J. Greeley, T. F. Jaramillo, J. Bonde, I. Chorkendorff and J. K. N?Rskov, Nat. Mater., 2006, 5, 909–913.
198. M. Zafari, D. Kumar, M. Umer and K. S. Kim, J. Mater. Chem. A, 2020, 8, 5209–5216.
199. Y. T. Liu, L. Tang, J. Dai, J. Yu and B. Ding, Angew. Chem., Int. Ed., 2020, 59, 13623–13627.
200. 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.
201. Z.-Y. Wu, M. Karamad, X. Yong, Q. Huang, D. A. Cullen, P. Zhu, C. Xia, Q. Xiao, M. Shakouri, F.-Y. Chen, J. Y. Kim, Y. Xia, K. Heck, Y. Hu, M. S. Wong, Q. Li, I. Gates, S. Siahrostami and H. Wang, Nat. Commun., 2021, 12, 2870.
202. C. J. M. van der Ham, M. T. M. Koper and D. G. H. Hetterscheid, Chem. Soc. Rev., 2014, 43, 5183–5191.
203. K. Kugler, M. Luhn, J. A. Schramm, K. Rahimi and M. Wessling, Phys. Chem. Chem. Phys., 2015, 17, 3768–3782.
204. V. Kordali, G. Kyriacou and C. Lambrou, Chem. Commun., 2000, 1673–1674.
205. Y. Ma, T. Yang, H. Zou, W. Zang, Z. Kou, L. Mao, Y. Feng, L. Shen, S. J. Pennycook, L. Duan, X. Li and J. Wang, Adv. Mater., 2020, 32, 2002177.
206. Y. Abghoui, A. L. Garden, V. F. Hlynsson, S. Björgvinsdóttir, H. Ólafsdóttir and E. Skúlason, Phys. Chem. Chem. Phys., 2015, 17, 4909–4918.
207. E. Skúlason, T. Bligaard, S. Gudmundsdóttir, F. Studt, J. Rossmeisl, F. Abild-Pedersen, T. Vegge, H. Jónsson and J. K. Nørskov, Phys. Chem. Chem. Phys., 2012, 14, 1235–1245.
208. S. Licht, B. Cui, B. Wang, F. F. Li, J. Lau and S. Liu, Science, 2014, 345, 637–640.
209. Z. W. Seh, J. Kibsgaard, C. F. Dickens, I. Chorkendorff, J. K. Nørskov and T. F. Jaramillo, Science, 2017, 355, eaad4998.
210. N. Cheng, J.-C. Li, D. Liu, Y. Lin and D. Du, Small, 2019, 15, 1901485.
211. T. Zhang, Y. Xing, Y. Song, Y. Gu, X. Yan, N. Lu, H. Liu, Z. Xu, H. Xu, Z. Zhang and M. Yang, Anal. Chem., 2019, 91, 10589–10595.
212. L. Jiao, W. Xu, H. Yan, Y. Wu, C. Liu, D. Du, Y. Lin and C. Zhu, Anal. Chem., 2019, 91, 11994–11999.
213. Y. Peng, X. Yu, W. Yin, W. Dong, J. Peng and T. Wang, ACS Appl. Bio Mater., 2020, 3, 59–67.
214. K. Yuan, D. Lützenkirchen-Hecht, L. Li, L. Shuai, Y. Li, R. Cao, M. Qiu, X. Zhuang, M. K. H. Leung, Y. Chen and U. Scherf, J. Am. Chem. Soc., 2020, 142, 2404–2412.
215. R. Ciriminna, L. Albanese, F. Meneguzzo and M. Pagliaro, ChemSusChem, 2016, 9, 3374–3381.
216. H. Zhao and Z. Y. Yuan, ChemSusChem, 2021, 14, 1616–1633.
217. Y. Jiang, P. Ni, C. Chen, Y. Lu, P. Yang, B. Kong, A. Fisher and X. Wang, Adv. Energy Mater., 2018, 8, 1801909.
218. K. Jiang, S. Back, A. J. Akey, C. Xia, Y. Hu, W. Liang, D. Schaak, E. Stavitski, J. K. Nørskov, S. Siahrostami and H. Wang, Nat. Commun., 2019, 10, 3997.
219. W. Zhu and S. Chen, Electroanalysis, 2020, 32, 2591–2602.
220. X. Guo, S. Lin, J. Gu, S. Zhang, Z. Chen and S. Huang, ACS Catal, 2019, 9, 11042–11054.
221. 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.
222. Q. Gao, Ionics, 2020, 26, 2453–2465.
223. J. Zhang, H. Yang and B. Liu, Adv. Energy Mater., 2021, 11, 2002473.
224. K. Jiang, S. Back, A. J. Akey, C. Xia, Y. Hu, W. Liang, D. Schaak, E. Stavitski, J. K. Nørskov, S. Siahrostami and H. Wang, Nat. Commun., 2019, 10, 3997.
225. Z. Liu, T. He, K. Liu, W. Chen and Y. Tang, RSC Adv., 2017, 7, 7920–7928.
226. Y. Tang, J. Zhou, W. Chen, H. Chai, Y. Li, Z. Feng and X. Dai, Mol. Catal., 2019, 476, 110524.
227. E. H. Song, J. M. Yan, J. S. Lian and Q. Jiang, J. Phys. Chem. C, 2012, 116, 20342–20348.
228. H. Y. Zhuo, X. Zhang, J. X. Liang, Q. Yu, H. Xiao and J. Li, Chem. Rev., 2020, 120, 12315–12341.
229. C. J. Zhang and P. Hu, J. Am. Chem. Soc., 2001, 123, 1166–1172.
230. M. D. Ackermann, T. M. Pedersen, B. L. M. Hendriksen, O. Robach, S. C. Bobaru, I. Popa, C. Quiros, H. Kim, B. Hammer, S. Ferrer and J. W. M. Frenken, Phys. Rev. Lett., 2005, 95, 255505.
231. Y. Li, Z. Zhou, G. Yu, W. Chen and Z. Chen, J. Phys. Chem. C, 2010, 114, 6250–6254.
232. L. Zhang, X. Ren, X. Zhao, Y. Zhu, R. Pang, P. Cui, Y. Jia, S. Li and Z. Zhang, Nano Lett., 2022, 22, 3744–3750.
233. E. Wigner and E. E. Witmer, Z. Phys., 1928, 51, 859–886.
234. J. H. Moore, Phys. Rev. A, 1973, 8, 2359–2362.
235. G. Ding, W. Wang, T. Jiang, B. Han, H. Fan and G. Yang, ChemCatChem, 2013, 5, 192–200.
236. S.-i Niwa, et al. , Science, 2002, 295, 105–107.
237. D. Deng, X. Chen, L. Yu, X. Wu, Q. Liu, Y. Liu, H. Yang, H. Tian, Y. Hu and P. Du, Sci Adv., 2015, 1, e1500462–e1500462.
238. W. Liu, L. Zhang, X. Liu, X. Liu, X. Yang, S. Miao, W. Wang, A. Wang and T. Zhang, J. Am. Chem. Soc., 2017, 139, 10790–10798.
239. L. Kong, J. Li, Q. Liu, Z. Zhao, Q. Sun, J. Liu and Y. Wei, J. Energy Chem., 2016, 25, 577–586.
240. H. Zhu, D. C. Rosenfeld, D. H. Anjum, V. Caps and J.-M. Basset, ChemSusChem, 2015, 8, 1254–1263.
241. M. Peymani, S. M. Alavi and M. Rezaei, Int. J. Hydrogen Energy, 2016, 41, 19057–19069.
242. S. Wang, H. Li, M. He, X. Cui, L. Hua, H. Li, J. Xiao, L. Yu, N. P. Rajan, Z. Xie and D. Deng, J. Energy Chem., 2019, 36, 47–50.
243. X. Cui, H. Li, Y. Wang, Y. Hu, L. Hua, H. Li, X. Han, Q. Liu, F. Yang, L. He, X. Chen, Q. Li, J. Xiao, D. Deng and X. Bao, Chem, 2018, 4, 1902–1910.
244. M. Crespo-Quesada, F. Cárdenas-Lizana, A.-L. Dessimoz and L. Kiwi-Minsker, ACS Catal, 2012, 2, 1773–1786.
245. Q. Feng, S. Zhao, Q. Xu, W. Chen, S. Tian, Y. Wang, W. Yan, J. Luo, D. Wang and Y. Li, Adv. Mater., 2019, 31, 1901024.
246. Z. Sun, S. Wang and W. Chen, J. Mater. Chem. A, 2021, 9, 5296–5319.
247. T. He, C. Zhang, L. Zhang and A. Du, Nano Research, 2019, 12, 1817–1823.
248. E. M. Sulman, P. M. Valetsky, M. G. Sulman, L. M. Bronstein, A. I. Sidorov, V. Y. Doluda and V. G. Matveeva, Chem. Eng. Process, 2011, 50, 1041–1053.
249. S. R. More and G. D. Yadav, ACS Omega, 2018, 3, 7124–7132.
250. D. B. Bagal and B. M. Bhanage, Adv. Synth. Catal., 2015, 357, 883–900.
251. R. V. Jagadeesh, A. E. Surkus, H. Junge, M. M. Pohl, J. Radnik, J. Rabeah, H. Huan, V. Schünemann, A. Brückner and M. Beller, Science, 2013, 342, 1073–1076.
252. R. V. Jagadeesh, K. Murugesan, A. S. Alshammari, H. Neumann, M. M. Pohl, J. Radnik and M. Beller, Science, 2017, 358, 326–332.
253. Z. Zhang, C. Ma, Y. Tu, R. Si, J. Wei, S. Zhang, Z. Wang, J.-F. Li, Y. Wang and D. Deng, Nano Res., 2019, 12, 2313–2317.
254. L. Huang, Y. Lv, S. Liu, H. Cui, Z. Zhao, H. Zhao, P. Liu, W. Xiong, F. Hao and H. A. Luo, Ind. Eng. Chem. Res., 2020, 59, 1422–1435.
255. Y.-G. Ji, K. Wei, T. Liu, L. Wu and W.-H. Zhang, Adv. Synth. Catal., 2017, 359, 933–940.
256. R. Yun, F. Zhan, N. Li, B. Zhang, W. Ma, L. Hong, T. Sheng, L. Du, B. Zheng and S. Liu, ACS Appl. Mater. Interfaces, 2020, 12, 34122–34129.
257. T. Sheng, Y.-J. Qi, X. Lin, P. Hu, S.-G. Sun and W.-F. Lin, Chem. Eng. J, 2016, 293, 337–344.
258. S. Tian, M. Hu, Q. Xu, W. Gong, W. Chen, J. Yang, Y. Zhu, C. Chen, J. He, Q. Liu, H. Zhao, D. Wang and Y. Li, Sci. China Mater., 2021, 64, 642–650.
259. Z.-H. Xue, D. Luan, H. Zhang and X. W. Lou, Joule, 2022, 6, 92–133.
260. S. Cestellos-Blanco, H. Zhang, J. M. Kim, Y.-x Shen and P. Yang, Nat. Catal., 2020, 3, 245–255.
261. P. Zhang and X. W. Lou, Adv. Mater., 2019, 31, 1900281.
262. H. Tong, S. Ouyang, Y. Bi, N. Umezawa, M. Oshikiri and J. Ye, Adv. Mater., 2012, 24, 229–251.
263. B. Wang, H. Cai and S. Shen, Small Methods, 2019, 3, 1800447.
264. Y. Li, Z. Wang, T. Xia, H. Ju, K. Zhang, R. Long, Q. Xu, C. Wang, L. Song, J. Zhu, J. Jiang and Y. Xiong, Adv. Mater., 2016, 28, 6959–6965.
265. C. Gao, J. Low, R. Long, T. Kong, J. Zhu and Y. Xiong, Chem. Rev., 2020, 120, 12175–12216.
266. L. Zeng and C. Xue, Nano Res., 2021, 14, 934–944.
267. W. Zhang, Q. Peng, L. Shi, Q. Yao, X. Wang, A. Yu, Z. Chen and Y. Fu, Small, 2019, 15, 1905166.
268. J. Li, P. Liu, Y. Tang, H. Huang, H. Cui, D. Mei and C. Zhong, ACS Catal, 2020, 10, 2431–2442.
269. W. Wu, E. Cui, Y. Zhang, C. Zhang, F. Zhu, C.-H. Tung and Y. Wang, ACS Catal, 2019, 9, 6335–6341.
270. C. Wang, A. Li, C. Li, S. Zhang, H. Li, X. Zhou, L. Hu, Y. Feng, K. Wang, Z. Zhu, R. Shao, Y. Chen, P. Gao, S. Mao, J. Huang, Z. Zhang and X. Han, Adv. Mater., 2019, 31, 1903491.
271. Z. Chen, S. Wu, J. Ma, S. Mine, T. Toyao, M. Matsuoka, L. Wang and J. Zhang, Angew. Chem., Int. Ed., 2021, 60, 11901–11909.
272. X. Ge, P. Zhou, Q. Zhang, Z. Xia, S. Chen, P. Gao, Z. Zhang, L. Gu and S. Guo, Angew. Chem., Int. Ed., 2020, 59, 232–236.
273. T. Hisatomi and K. Domen, Nat. Catal, 2019, 2, 387–399.
274. J. Liu, H. Wu, F. Li, X. Feng, P. Zhang and L. Gao, Adv. Sustainable Syst., 2020, 4, 2000151.
275. M. C. Fu, R. Shang, B. Zhao, B. Wang and Y. Fu, Science, 2019, 363, 1429–1434.
276. T. Hou, H. Peng, Y. Xin, S. Wang, W. Zhu, L. Chen, Y. Yao, W. Zhang, S. Liang and L. Wang, ACS Catal, 2020, 10, 5502–5510.
277. C. Gao, S. Chen, Y. Wang, J. Wang, X. Zheng, J. Zhu, L. Song, W. Zhang and Y. Xiong, Adv. Mater., 2018, 30, e1704624.
278. Y. Wang, A. Xu, Z. Wang, L. Huang, J. Li, F. Li, J. Wicks, M. Luo, D.-H. Nam, C.-S. Tan, Y. Ding, J. Wu, Y. Lum, C.-T. Dinh, D. Sinton, G. Zheng and E. H. Sargent, J. Am. Chem. Soc., 2020, 142, 5702–5708.
279. G.-F. Chen, Y. Yuan, H. Jiang, S.-Y. Ren, L.-X. Ding, L. Ma, T. Wu, J. Lu and H. Wang, Nat. Energy, 2020, 5, 605–613.
280. S. Shang, W. Xiong, C. Yang, B. Johannessen, R. Liu, H.-Y. Hsu, Q. Gu, M. K. H. Leung and J. Shang, ACS Nano, 2021, 15, 9670–9678.
281. S. Wu, Z. Chen, W. Yue, S. Mine, T. Toyao, M. Matsuoka, X. Xi, L. Wang and J. Zhang, ACS Catal, 2021, 11, 4362–4371.
282. X. Xiao, Z. Ruan, Q. Li, L. Zhang, H. Meng, Q. Zhang, H. Bao, B. Jiang, J. Zhou, C. Guo, X. Wang and H. Fu, Adv. Mater., 2022, 34, 2200612.
283. J. Wen, X. Yang, Z. Sun, J. Yang, P. Han, Q. Liu, H. Dong, M. Gu, L. Huang and H. Wang, Green Chem., 2020, 22, 230–237.
284. Z. Zhao, W. Zhang, W. Liu, Y. Li, J. Ye, J. Liang and M. Tong, Chem. Eng. J, 2021, 407, 127167.
285. G. Zhao, W. Li, H. Zhang, W. Wang and Y. Ren, Chem. Eng. J., 2022, 430, 132937.
286. W. Ma, J. Mao, X. Yang, C. Pan, W. Chen, M. Wang, P. Yu, L. Mao and Y. Li, Chem. Commun., 2019, 55, 159–162.
287. L. Jiao, H. Yan, Y. Wu, W. Gu, C. Zhu, D. Du and Y. Lin, Angew. Chem., Int. Ed., 2020, 59, 2565–2576.
288. Y. Wang, K. Qi, S. Yu, G. Jia, Z. Cheng, L. Zheng, Q. Wu, Q. Bao, Q. Wang, J. Zhao, X. Cui and W. Zheng, Nanomicro Lett., 2019, 11, 102.
289. H. Wei and E. Wang, Chem. Soc. Rev., 2013, 42, 6060–6093.
290. J. Wu, X. Wang, Q. Wang, Z. Lou, S. Li, Y. Zhu, L. Qin and H. Wei, Chem. Soc. Rev., 2019, 48, 1004–1076.
291. L. Huang, J. Chen, L. Gan, J. Wang and S. Dong, Sci. Adv., 2019, 5, eaav5490.
292. X. Zhang, G. Li, G. Chen, D. Wu, X. Zhou and Y. Wu, Coord. Chem. Rev., 2020, 418, 213376.
293. H. Zhang, X. F. Lu, Z. Wu and X. Lou, ACS Cent. Sci., 2020, 6, 1288–1301.
294. R. Yan, S. Sun, J. Yang, W. Long, J. Wang, X. Mu, Q. Li, W. Hao, S. Zhang, H. Liu, Y. Gao, L. Ouyang, J. Chen, S. Liu, X.-D. Zhang and D. Ming, ACS Nano, 2019, 13, 11552–11560.
295. X. Huang and J. T. Groves, Chem. Rev., 2018, 118, 2491–2553.
296. L. Wang, X. Qu, Y. Zhao, Y. Weng and S. Zhou, ACS Appl. Mater. Interfaces, 2019, 11, 35228–35237.
297. H. Xiang, W. Feng and Y. Chen, Adv. Mater., 2020, 32, 1905994.
298. M. Huo, L. Wang, Y. Chen and J. Shi, Nat. Commun., 2017, 8, 357.
299. W. Zhang, J. Lu, X. Gao, P. Li, W. Zhang, Y. Ma, H. Wang and B. Tang, Angew. Chem., Int. Ed., 2018, 57, 4891–4896.
300. Z. Tang, Y. Liu, M. He and W. Bu, Angew. Chem., Int. Ed., 2019, 58, 946–956.
301. X. Chen, Y. Liu and W. Bu, Sci. Sin.: Chim., 2020, 50, 159–172.
302. C. Zhang, W. Bu, D. Ni, S. Zhang, Q. Li, Z. Yao, J. Zhang, H. Yao, Z. Wang and J. Shi, Angew. Chem., Int. Ed., 2016, 55, 2101–2106.
303. S. Zhai, X. Hu, Y. Hu, B. Wu and D. Xing, Biomaterials, 2017, 121, 41–54.
304. M. Huo, L. Wang, Y. Wang, Y. Chen and J. Shi, ACS Nano, 2019, 13, 2643–2653.
305. D. Wu, J. Li, S. Xu, Q. Xie, Y. Pan, X. Liu, R. Ma, H. Zheng, M. Gao, W. Wang, J. Li, X. Cai, F. Jaouen and R. Li, J. Am. Chem. Soc., 2020, 142, 19602–19610.
306. X. Jiang, H. Wang, R. Yuan and Y. Chai, Anal. Chem., 2018, 90, 8462–8469.
307. B. Dou, J. Yang, R. Yuan and Y. Xiang, Anal. Chem., 2018, 90, 5945–5950.
308. L. Jiao, W. Xu, H. Yan, Y. Wu, C. Liu, D. Du, Y. Lin and C. Zhu, Anal. Chem., 2019, 91, 11994–11999.
309. L. Shen, M. A. Khan, X. Wu, J. Cai, T. Lu, T. Ning, Z. Liu, W. Lu, D. Ye, H. Zhao and J. Zhang, Biosens. Bioelectron., 2022, 205, 114097.
310. X. Niu, Q. Shi, W. Zhu, D. Liu, H. Tian, S. Fu, N. Cheng, S. Li, J. N. Smith, D. Du and Y. Lin, Biosens. Bioelectron., 2019, 142, 111495.
311. M. S. Kim, J. Lee, H. S. Kim, A. Cho, K. H. Shim, T. N. Le, S. S. A. An, J. W. Han, M. I. Kim and J. Lee, Adv. Funct. Mater., 2020, 30, 1905410.
312. Y. Wu, L. Jiao, X. Luo, W. Xu, X. Wei, H. Wang, H. Yan, W. Gu, B. Z. Xu, D. Du, Y. Lin and C. Zhu, Small, 2019, 15, 1903108.
313. Q. Chen, S. Li, Y. Liu, X. Zhang, Y. Tang, H. Chai and Y. Huang, Sens. Actuators, B, 2020, 305, 127511.
314. L. Yao, S. Gao, S. Liu, Y. Bi, R. Wang, H. Qu, Y. Wu, Y. Mao and L. Zheng, ACS Appl. Mater. Interfaces, 2020, 12, 6268–6275.
315. W. Gu, H. Wang, L. Jiao, Y. Wu, Y. Chen, L. Hu, J. Gong, D. Du and C. Zhu, Angew. Chem., Int. Ed., 2020, 59, 3534–3538.
316. M. Wang, L. Liu, X. Xie, X. Zhou, Z. Lin and X. Su, Sens. Actuators, B, 2020, 313, 128023.
317. B. C. Dickinson and C. J. Chang, Nat. Chem. Biol., 2011, 7, 504–511.
318. J. Zhao, W. Gao, X. Cai, J. Xu, D. Zou, Z. Li, B. Hu and Y. Zheng, Theranostics, 2019, 9, 2843–2855.
319. Z. Jiulong, C. Xiaojun, G. Wei, Z. Linlin, Z. Duowu, Z. Yuanyi, L. Zhaoshen and C. Hangrong, ACS Appl. Mater. Interfaces, 2018, 10, 26108–26117.
320. J. Yao, Y. Cheng, M. Zhou, S. Zhao, S. Lin, X. Wang, J. Wu, S. Li and H. Wei, Chem. Sci., 2018, 9, 2927–2933.
321. Y. P. Lee, S. H. Kim, J. W. Bang, H. S. Lee, S. S. Kwak and S. Y. Kwon, Plant Cell Rep., 2007, 26, 591–598.
322. S. Reuter, S. C. Gupta, M. M. Chaturvedi and B. B. Aggarwal, Free Radic. Biol. Med., 2010, 49, 1603–1616.
323. X. Zhuobin, Q. Zhiyue, L. Qi, H. Yixin, L. Dandan, S. Xinggui, F. Kelong, X. Juqun, G. Yunhao and T. Yan, Nat. Commun., 2018, 9, 3713.
324. S. Levy, Nat. Med. Suppl., 2004, 10, S122.
325. H. Bai, H. Yuan, C. Nie, B. Wang, F. Lv, L. Liu and S. Wang, Angew. Chem., Int. Ed., 2015, 127, 13208–13213.
326. L. Fan, P. Sun, Y. Huang, Z. Xu, X. Lu, J. Xi, J. Han and R. Guo, ACS Appl. Bio Mater., 2020, 3, 1147–1157.
327. S. Ji, B. Jiang, H. Hao, Y. Chen, J. Dong, Y. Mao, Z. Zhang, R. Gao, W. Chen, R. Zhang, Q. Liang, H. Li, S. Liu, Y. Wang, Q. Zhang, L. Gu, D. Duan, M. Liang, D. Wang, X. Yan and Y. Li, Nat. Catal., 2021, 4, 407–417.
328. X. Cui, H. Li, Y. Wang, H. Yuanli, L. Hua, H. Li, X. Han, Q. Liu, F. Yang, L. He, X. Chen, Q. Li, J. Xiao, D. Deng and X. Bao, Chem, 2018, 4, 1902–1910.
329. C. Zhao, C. Xiong, X. Liu, M. Qiao, Z. Li, T. Yuan, J. Wang, Y. Qu, X. Wang, F. Zhou, Q. Xu, S. Wang, M. Chen, W. Wang, Y. Li, T. Yao, Y. Wu and Y. Li, Chem. Commun., 2019, 55, 2285–2288.
330. S. Tian, Q. Fu, W. Chen, Q. Feng, Z. Chen, J. Zhang, W.-C. Cheong, R. Yu, L. Gu, J. Dong, J. Luo, C. Chen, Q. Peng, C. Draxl, D. Wang and Y. Li, Nat. Commun., 2018, 9, 2353.
331. B. Singh, M. B. Gawande, A. D. Kute, R. S. Varma, P. Fornasiero, P. McNeice, R. V. Jagadeesh, M. Beller and R. Zbořil, Chem. Rev., 2021, 121, 13620–13697.
332. U. G. K. Wegst, H. Bai, E. Saiz, A. P. Tomsia and R. O. Ritchie, Nat. Mater., 2015, 14, 23–36.
333. Z. Chu and S. Seeger, Adv. Mater., 2015, 27, 7775–7781.
334. G. Mirabello, J. J. M. Lenders and N. A. J. M. Sommerdijk, Chem. Soc. Rev., 2016, 45, 5085–5106.
335. Y. Xiong, W. Sun, P. Xin, W. Chen, X. Zheng, W. Yan, L. Zheng, J. Dong, J. Zhang, D. Wang and Y. Li, Adv. Mater., 2020, 32, 2000896.
336. E. Luo, M. Xiao, Y. Wang, J. Ge, C. Liu and W. Xing, ChemCatChem, 2018, 10, 3653–3658.
337. K. Yuan, D. Luetzenkirchen-Hecht, L. Li, L. Shuai, Y. Li, R. Cao, M. Qiu, X. Zhuang, M. K. H. Leung, Y. Chen and U. Scherf, J. Am. Chem. Soc., 2020, 142, 2404–2412.
338. Y. Han, Y. Wang, R. Xu, W. Chen, L. Zheng, A. Han, Y. Zhu, J. Zhang, H. Zhang, J. Luo, C. Chen, Q. Peng, D. Wang and Y. Li, Energy Environ. Sci., 2018, 11, 2348–2352.
339. H. Shen, E. Gracia-Espino, J. Ma, H. Tang, X. Mamat, T. Wagberg, G. Hu and S. Guo, Nano Energy, 2017, 35, 9–16.
340. C. Zhu, Q. Shi, B. Z. Xu, S. Fu, G. Wan, C. Yang, S. Yao, J. Song, H. Zhou, D. Du, S. P. Beckman, D. Su and Y. Lin, Adv. Energy Mater., 2018, 8, 1801956.
341. 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.
342. Z. Wang, H. Jin, T. Meng, K. Liao, W. Meng, J. Yang, D. He, Y. Xiong and S. Mu, Adv. Funct. Mater., 2018, 28, 1802596.
343. S. Gong, C. Wang, P. Jiang, L. Hu, H. Lei and Q. Chen, J. Mater. Chem. A, 2018, 6, 13254–13262.
344. B. Wang, J. Zou, X. Shen, Y. Yang, G. Hu, W. Li, Z. Peng, D. Banham, A. Dong and D. Zhao, Nano Energy, 2019, 63, 103851.
345. T. Palaniselvam, V. Kashyap, S. N. Bhange, J.-B. Baek and S. Kurungot, Adv. Funct. Mater., 2016, 26, 2150–2162.
346. R. Ma, G. Lin, Q. Ju, W. Tang, G. Chen, Z. Chen, Q. Liu, M. Yang, Y. Lu and J. Wang, Appl. Catal., B, 2020, 265, 118593.
347. J. Han, X. Meng, L. Lu, J. Bian, Z. Li and C. Sun, Adv. Funct. Mater., 2019, 29, 1808872.

---