Carbon-based materials have emerged as one of the most prevalent supports for SACs due to their exceptional properties, including high specific surface area, superior electrical conductivity, remarkable chemical stability, and tunable surface chemistry. Commonly employed carbon-based supports encompass graphene, carbon nanotubes, ordered mesoporous carbon, and carbon fibers [25–27]. The synthesis of SACs on these supports can be achieved through either bottom-up or top-down approaches. The bottom-up strategy involves directly introducing single-atom catalytic sites onto pre-existing carbon supports via methods such as impregnation. Conversely, the top-down approach entails the initial preparation of polymeric precursors through self-assembly, followed by carbonization and acid etching to successfully anchor SACs onto the carbon supports [28,29]. These methodologies offer the flexibility to precisely modulate the surface architecture of the carbon supports, thereby inducing alterations in the electronic structure of the metal active sites through the creation of intrinsic defects. This, in turn, leads to significant variations in catalytic activity and selectivity. Furthermore, these supports facilitate the formation of stable coordination structures with metal single atoms]through defect sites (e.g., vacancies, edge sites) and heteroatom doping (e.g., N, S, P), thereby effectively regulating the catalytic performance [30]. Additionally, the π-conjugated system of carbon materials facilitates electron transfer, which can modulate the electronic structure of the single atoms and enhance their catalytic activity [31]. In recent years, carbon-based SACs have demonstrated remarkable performance in ORR, CO₂RR and NRR [32–36]. It has been demonstrated that the active centers of SACs typically consist of isolated metal atoms forming stable coordination structures with coordinating atoms (e.g., N, O, S) on the support. This configuration not only effectively stabilizes the metal single atoms but also optimizes the adsorption behavior of reaction intermediates by modulating the electronic structure, thereby significantly enhancing catalytic performance. For instance, manganese (Mn)-based SACs with Mn-O sites exhibit exceptional catalytic activity in the NRR due to the appropriate energy and symmetry of their 3d orbitals, which facilitate the adsorption and activation of N₂ molecules. Our team successfully constructed Mn-O₃N₁ sites anchored on porous carbon (Mn-O₃N₁/PC) through precise regulation of Mn-O bonding conditions. This catalyst achieved an NH₃ yield of 66.41 µg h^‒1 mg_cat^‒1 at −0.35 V (vs. RHE), approximately four times higher than that of the Mn-N₄/PC control catalyst (Fig. 2a). This performance enhancement is attributed to the unique geometry and electronic structure of the Mn-O₃N₁ site, which not only promotes N₂ adsorption and activation but also significantly reduces the free energy change in the potential-determining step [37]. Similarly, Tran et al. [38] discovered that positively charged Co single atoms supported on N-doped carbon nanofibers serve as highly efficient active sites for catalyzing the reduction of N₂ to NH₃, achieving an NH₃ yield of 67.6 µg h^‒1 mg^‒1 at 0.1 V with a maximum Faradaic efficiency (FE) of 56.9%. Density functional theory (DFT) calculations revealed that charge transfer between the Co single atom and the support modulates the local electronic structure, inducing the formation of Lewis acid pairs and enabling the Co site to exhibit high intrinsic activity for NRR (Fig. 2b). Additionally, porous carbon materials are considered excellent supports for SACs due to their gas storage capacity and hydrophobicity, which enhance the N₂ concentration near the single-atom catalytic sites [39]. Building on this, Liao et al. [40] immobilized Mo on the quadruple hollow sites of phosphotungstic acid embedded in carbon nanotubes (CNTs) to obtain the Mo-PTA@CNT electrocatalyst. Electrochemical testing revealed that this catalyst achieved an NH₃ yield of 51 µg h^‒1 mg^‒1 at −0.1 V with an FE of 83%. Similarly, niobium (Nb)-based SACs have also demonstrated remarkable performance in NO reduction reactions (NORR). For example, Nb SACs supported on B and N co-doped carbon nanotubes (Nb-SA/BNC) achieved NH₃ yield of 8.2 × 10^‒8 mol cm^‒2 s^‒1 in room-temperature electrochemical denitrification. The introduction of B and N heteroatoms not only improved the electrical conductivity of the catalyst but also provided abundant anchor sites to stabilize the Nb single atoms and optimize their electronic structure. Field emission scanning electron microscopy (FESEM) and transmission electron microscopy (TEM) characterizations revealed that Nb-SA/BNC possesses a nanotube array structure with a specific surface area of 875.5 m²/g and hierarchical micro/mesoporous features, facilitating mass transfer and reactant diffusion, thereby accelerating the electrocatalytic process. Theoretical calculations further indicated that the Nb single-atom sites promote NO adsorption and lower the energy barrier of the potential-determining step [41]. Considering the ordered porous carbon matrix facilitates the exposure of active sites and the bulk transport of electrolytes during electrolysis. A N, S co-doped ordered porous carbon-supported Pd SACs (PdSA/N, S-OPC) was prepared using a template-assisted wet chemistry method. This catalyst exhibited exceptional hydrogen oxidation reaction (HOR) activity and stability under alkaline conditions, outperforming commercial Pt/C catalysts and ranking among the highest reported performances. Experimental and theoretical analyses suggested that this superior HOR performance stems from the synergistic effect of the Pd-N₄ coordination structure and N/S doping, which optimizes the adsorption behavior of H_ad species, significantly enhancing catalytic performance (Fig. 2c) [30]. Fan et al. [42] synthesized mesoporous Fe-N-C materials through a self-assembly and pyrolytic carbonization strategy, which demonstrated high activity in the electrocatalytic reduction of nitrate to N₂ (Fig. 2d). In the context of CO₂RR and ORR bifunctional catalysts, Nb SACs supported on nitrogen-doped ordered mesoporous carbon (Nb-N-C) achieved a FE_CO of up to 90% and an ORR half-wave potential of 0.84 V. When applied in zinc-air batteries (ZABs), the peak power density reached 115.6 mW/cm² (Fig. 2e). DFT calculations revealed that the formation of Nb-N coordination bonds effectively lowered the energy barriers for the generation of key intermediates (*COOH and *O) in CO₂RR and ORR, thereby enhancing catalytic performance [34]. In addition, Fe-based SACs have shown excellent performance in NRR. For example, N, S co-doped graphene (NSDG)-loaded Fe SACs (Fe-SAS/NSDG) significantly enhanced the NRR activity through the Fe-N₃S active center. The introduction of S enhanced the affinity of the material for N₂ and optimized the adsorption and activation process of N₂ molecules. As shown in Figs. 2f-i, the lowest adsorption energy is found when the N₂ tip is adsorbed on Fe-N₃S [43].

Similarly, the Fe-Zn diatomic catalyst (Fe-Zn-SA/NC) anchored on a porous N-doped carbon skeleton exhibited excellent ORR performance in the full pH range (Fig. 3a) [44]. B and N co-doped carbon supports loaded with single-atom Mn sites exhibit excellent HER and hydrazine oxidation reactions (HzOR) bifunctional catalytic performance in alkaline electrolytes [32]. As confirmed by XPS analyses, N-doped carbon black-supported Ag SACs (SA-Ag/NC) exhibited enriched pyridine N species on their surface. These sites facilitated the stable dispersion of single atoms through strong coordination with Ag atoms, resulting in exceptional NH₃ yield and FE in electrocatalytic reactions [35]. In addition, N-doped hollow carbon sphere-supported Ce SACs demonstrated remarkable performance in the NORR in acidic solutions, achieving a FE_NH3 of 91% at −0.7 V (vs. RHE) with a yield of 1023 µg h^‒1 mg_cat^‒1. Characterization analyses confirmed the existence of the Ce-N₄ coordination structure (Figs. 3b-g). Theoretical simulations indicated (Figs. 3h-k) that the Ce-N₄ active site not only activated NO molecules through strong electronic interactions but also lowered the free energy barrier for the transition of *NO to *NOH intermediates, enhancing NORR kinetics and suppressing competitive hydrogen evolution [12]. In NRR, Fe supported on N-doped frameworks exhibited exceptionally high catalytic activity (ISAS-Fe/NC). The introduction of N atoms altered the electron cloud distribution of the carbon material, creating partially positively charged neighboring carbon atoms that facilitated the adsorption and immobilization of Fe single atoms. The formation of the Fe-N₄ active center optimized the adsorption and activation process of reactant molecules, significantly improving NRR catalytic activity. Additionally, the high conductivity of carbon-based materials promoted rapid electron transfer between the catalyst and reactants, reducing the reaction overpotential and further enhancing catalytic efficiency [45]. Xia et al. [46] achieved highly selective CO₂-to-CO conversion by constructing an asymmetric ligand-structured Ni-N₃-S atomic site catalyst (Ni₁—NSC), which exhibited a FE_CO exceeding 99% at a high current density of −320 mA/cm² (Fig. 3l). Spherical aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) revealed that Ni atoms were uniformly dispersed on the carbon support surface. In situ attenuated total reflection surface-enhanced infrared absorption spectroscopy (ATR-SEIRAS) and differential electrochemical mass spectrometry (DEMS) analyses demonstrated that the Ni-N₃-S sites exhibited weaker binding strength to *CO and lower kinetic overpotential for CO₂-to-CO conversion compared to conventional Ni-N₄ sites. DFT calculations further confirmed that the enhanced activity and selectivity of the Ni₁—NSC catalyst were attributed to the lower Gibbs free energy of the *COOH intermediate on the Ni-N₃-S sites.

Despite the significant advantages of carbon-based supports in SACs, their stability under high temperatures or strongly oxidizing environments remains a challenge. Carbon corrosion can lead to the agglomeration and loss of single atoms, as well as limited interaction strength with highly reactive metals. Therefore, the development of novel support materials and the optimization of single-atom anchoring strategies are critical directions for future research, aiming to achieve more efficient and stable SACs designs.

Metal oxides have garnered significant attention as ideal supports for SACs due to their structurally diverse crystalline frameworks, customizable surface characteristics, and adaptable electronic properties [47]. These materials demonstrate exceptional versatility in catalytic applications, particularly through their three primary categories of support materials: Titanium-based oxides (Ti-based oxides), tungsten-based oxides (W-based oxides), and molybdenum-based oxides (Mo-based oxides). A critical feature of these oxide supports lies in their inherent defect engineering capabilities, which manifest as oxygen vacancies (Ov), metal vacancies (Mv), crystallographic edge irregularities, and atomic-scale step discontinuities. These structural imperfections serve as favorable anchoring sites that stabilize isolated metal atoms through strong coordination bonds, effectively preventing atomic aggregation during catalytic processes. The functional superiority of metal oxide-supported SACs stems from dynamic interfacial interactions between the support matrix and immobilized metal species. Through synergetic electron transfer mechanisms, the oxide substrate can either donate or accept electron density from anchored single atoms, thereby inducing charge redistribution at the atomic interface. This electronic modulation significantly alters the d-band center position of metal active sites, a critical parameter governing catalytic behavior [23]. For instance, electron-deficient metal centers may exhibit optimized binding energies for specific reaction intermediates, while electron-rich configurations could facilitate the desorption of products. Amorphous α-MoO_x supported on nitrogen-doped carbon (Pt-SA/α-MoO_x) demonstrates exceptional catalytic performance in the HER (Fig. 4a). This catalyst achieves an ultra-high mass activity of 52.0 A/mg_pt at an overpotential of 50 mV, primarily attributed to the synergistic effect between the amorphous α-MoO_x support and Pt single atoms. Electron energy loss spectroscopy (EELS) analysis confirms the presence of abundant oxygen vacancies in the α-MoO_x support, a finding corroborated by electron paramagnetic resonance (EPR) results. These oxygen vacancies form strong interactions with Pt single atoms, optimizing the adsorption and desorption processes of hydrogen atoms (H*), thereby significantly enhancing HER activity. Furthermore, the reduction of Mo⁶⁺ to lower valence states (Mo⁵⁺, Mo⁴⁺, Mo³⁺, and Mo²⁺) in α-MoO_x induced lattice-electron coupling distortions, further boosting electrochemical activity. The introduction of nitrogen-doped carbon not only improved the conductivity of the support but also facilitated charge transfer during electrocatalysis, underpinning the high performance of Pt-SA/α-MoO_x. The significance of amorphous α-MoO_x lay in its high defect concentration, tunable electronic structure, and strong synergistic effects with single atoms. These unique properties not only stabilized single-atom active sites but also significantly enhanced catalytic performance by optimizing the local electronic environment [48]. In another study, Gao et al. [49] successfully reduced the bandgap and enhanced the conductivity of TiO₂ by introducing Nb single atoms. Characterization results revealed that Nb doping significantly improved the adsorption capacity of N₂ on the Ti surface, while the Nb-Ti-O structure suppressed the HER and accelerated electron transfer in the NRR. At a potential of −0.5 V (vs. RHE), the catalyst achieved an NH₃ production rate of 213 µg h^‒1 mg^‒1 with a FE of 9.2%. Moreover, researchers synthesized MgO nanosheets and their supported single-atom Pt catalyst (Pt-SA/MgO) for toluene oxidation. Compared to pure MgO, Pt-SA/MgO demonstrated significantly enhanced catalytic activity. Studies showed that the surface of Pt-SA/MgO readily generates oxygen vacancies, which facilitate the activation of molecular oxygen and the formation of reactive oxygen species. Based on experimental data and theoretical calculations, a mechanism for reactive oxygen-promoted toluene oxidation was proposed: In the presence of H₂O, molecular oxygen more readily dissociates on the oxygen vacancies of Pt-SA/MgO to form *OH radicals, which are the primary reactive oxygen species [50]. Currently, two main approaches are employed to enhance the activity of SACs supported on metal oxides. Constructing oxygen vacancies to increase the anchoring capacity of the support for single atoms, thereby improving single-atom loading, and introducing dopants to further enhance the electronic regulation of the support on single atoms. For example, Liu et al. [51] loaded La, which has unsaturated orbitals, onto oxygen-deficient MoO_3−x. The strong interaction between La and two coordinated lattice oxygens in MoO_3−x (O₂-La-O_x coordination) effectively prevented the detachment or aggregation of single atoms, avoiding catalyst deactivation (Figs. 4b-k). Tang et al. [52] presented a protonated iridium oxide (H_3.8Ir_1−xRu_xO₄) encapsulating monotonically dispersed ruthenium atoms, which exhibits excellent activity and stability in acidic water oxidation. Single Ru doping facilitates the induction of localized oxygen vacancies in the Ir-O lattice, which synergistically enhances the adsorption of OOH* intermediates and improves the OER activity. In addition, the preferential oxidation of Ru and the electronegativity of the oxygen vacancies significantly stabilized the Ir-O active site and improved the OER stability. As a result, the H_3.8Ir_1−xRu_xO₄ catalyst has an overpotential of 255 mV at 10 mA/cm² and exhibits excellent catalytic durability in acidic electrolytes for more than 1100 h, representing a significant increase in stability by an order of magnitude compared to pristine H_3.8IrO₄. The proton exchange membrane electrolyzer using H_3.8Ir_1−xRu_xO₄ catalyst as the anode exhibited stable performance for more than 1280 h at a high current density of 2 A/cm². Introducing dopants to further enhance the electronic regulation of the support on single atoms. Based on DFT calculations, Lin et al. [53] proposed that introducing heteroatoms (e.g., Fe, Co, Ni, Cu, Zn, Ga and Ge) into Keggin-type polyoxometalate supports can adjust the coordination environment of Mo single atoms, promoting the electrocatalytic NRR. The charge distribution of O atoms at the 4-H site was closely related to the type of central heteroatom. When the central heteroatom was Co or As, more electrons were transferred from W atoms to O atoms. Charge density difference analysis indicated that Mo atoms on CoTA and ASTA exhibit stronger electronegativity, allowing more electrons to flow from the supported Mo atoms to the N≡N antibonding orbital, effectively activating the N≡N bond (Figs. 4l and m). Partial crystal orbital Hamiltonian population (COHP) and PDOS analysis of the N—N bond demonstrate that electron transfer from Mo to N₂ antibonding orbitals activates the N₂ molecule and extends the N=N bond length (Figs. 4n and o). Therefore, the Mo atoms supported by CoTA and AsTA have a more significant activation effect on N₂ than those supported by GaTA and ZnTA.

In summary, metal oxide supports offer high design flexibility. By constructing oxygen vacancies or introducing heteroatoms, their electronic structures can be significantly improved, optimizing the strength of their interactions with single atoms. However, their loadings are generally low. In practical applications, the choice of oxide support, dopant elements, and single-atom species can be tailored to specific needs, enabling precise control and optimization of catalytic performance.

MXenes were first synthesized in 2011 by Prof. Yury Gogotsi's group through selective etching of Ti₃AlC₂ MAX phase precursors using hydrofluoric acid (HF), yielding the novel 2D layered material Ti₃C₂T_x, later named "MXenes" to denote two-dimensional transition metal carbides, nitrides, or carbonitrides [54]. Their general formula is M_n+1X_nT_x, where M represents transition metals (e.g., Sc, Ti, V), X denotes carbon/nitrogen, and T_x corresponds to surface terminations (e.g., -O, -OH, -F). Both the M and X layers in MXenes are chemically tunable, while the diverse surface terminations enable the creation of a vast MXenes family with distinct properties. As 2D materials, MXenes possess atomic-level thickness and structural diversity, while maintaining stability in harsh environments (acidic, alkaline, and organic solvents). At the same time, MXenes have excellent electrical properties, similar to their precursor MAX, MXenes without surface group saturation mostly exhibit metallic properties, and the d electronic state of the transition metal atoms has an energy band through the Fermi energy level, which can conduct electrons [55]. After being saturated by the surface group, the surface group forms a new chemical bond with the transition metal atoms, and the d-orbitals of the transition metal atoms are partially occupied by the electrons of the surface group. The d-orbitals of the transition metal atoms hybridize with the valence electron orbitals of the surface group, and the energy of the hybrid orbitals decreases. MXenes without surface group saturation will become semiconductors after surface saturation, while some MXenes retain their metallic properties [56]. These attributes confer MXenes with unique catalytic advantages. Specifically, their ultrahigh surface area, tunable bandgap, mechanical robustness, and abundant surface defects establish MXenes as ideal platforms for SACs. The formation of stable coordination bonds between surface terminations and metal single atoms enhances both the dispersion and stability of active sites. Simultaneously, charge redistribution at the metal-support interface modulates the electronic structure of catalytic centers, enabling optimized adsorption energies for reaction intermediates and superior electrocatalytic performance [57,58]. A highly efficient and stable electrochemical carbon monoxide reduction (COR) catalyst was developed by anchoring copper single atoms on Ti₃C₂T_x MXenes nanosheets, achieving an unprecedented selectivity of 98% for multicarbon products. In COR, Cu-SA/Ti₃C₂T_x exhibited superior C₂ product selectivity (98%) compared to Cu-NP/Ti₃C₂T_x. Notably, at −0.7 V (vs. RHE), the FE for C₂H₄ reached 71%, one of the highest values reported for Cu-based COR catalysts. Furthermore, Cu-SA/Ti₃C₂T_x demonstrated remarkable electrochemical stability over 68 h. X-ray absorption fine structure (XAFS) analysis confirmed the stable anchoring of Cu single atoms via O coordination on the Ti₃C₂T_x substrate (Figs. 5a-c). Theoretical simulations revealed (Figs. 5d-i) that atomically dispersed Cu-O₃ sites significantly promote C—C coupling of CO molecules, generating the critical *CO—CHO intermediate and reducing the free energy barrier of the potential-determining step, thereby explaining the catalyst's high activity and selectivity [59]. A novel electrocatalyst comprising atomically dispersed Pt anchored on MXenes through Pt-O and Pt-Ti coordination bonds (denoted as Pt-SA/MXenes) was developed for efficient HER. The asymmetric coordination environment induces local electric field polarization, endowing the Pt-SA/MXenes with exceptional alkaline HER performance. Specifically, the catalyst achieves a current density of 10 mA/cm² at a mere overpotential of 33 mV and demonstrates remarkable stability over 27 h of continuous operation. Notably, the mass activity reaches 23.5 A/mg_Pt at 100 mV overpotential, surpassing that of commercial Pt/C by a factor of 29.4. Theoretical calculations based on DFT suggested that the polarized electric field effectively modulates the electronic structure of Pt-SA/MXenes, thereby lowering the energy barrier for H* intermediate adsorption/desorption and enhancing the overall HER catalytic efficiency [60]. Additionally, we proposed a novel self-driven dual hydrogen production system based on oxygen-functionalized Ti₃C₂O_x MXenes-supported single Rh atoms (Rh-SA/Ti₃C₂O_x). This bifunctional catalyst exhibited outstanding performance in both pH-universal HER and HzOR. DFT calculations indicated that the introduction of single Rh atoms optimized the electronic structure of Ti₃C₂O_x, facilitating hydrogen adsorption in HER and dehydrogenation in HzOR (Figs. 5j and k) [61]. Recent advancements have led to the development of MXene-supported metal catalysts, such as Ru/Mo₂CT_x [62], Ti₃C₂O_x-PtSA [63] and Pt-on-Pd/Ti₃C₂O_x [64], which have been successfully applied in electrochemical reactions. For instance, Du et al. [65] synthesized V₄C₃T_x-supported Ni nanoparticle (Ni@MX) composites via a two-step freeze-drying and hydrogen thermal reduction process. These catalysts demonstrated exceptional activity for electrocatalytic NRR in 0.1 mol/L KOH, achieving an NH₃ yield of 21.29 µg h^‒1 mg^‒1 and an FE of 14.86%. DFT calculations revealed that the high activity stemmed from synergistic effects between Ni sites and oxygen vacancies on V₄C₃T_x, which enhanced nitrogen adsorption and reduced Gibbs free energy barriers. Moreover, the V₄C₃T_x MXenes support suppressed the competing HER, further boosting NRR activity. Single-atom Ru sites anchored on Ti₃C₂T_x MXenes nanosheets were discovered to function as trifunctional electrocatalysts for acidic HER, OER, and ORR. Theoretical calculations showed that isolated Ru-O₂ sites optimize reactant/intermediate adsorption and lower the energy barrier of the potential-determining step, accelerating HER, OER, and ORR kinetics [66]. Although MXenes have been synthesized by liquid-phase and molten-salt etching methods, they still suffer from slow reaction kinetics for the removal of A species from the MAX phase as well as long production times (5–48 h). Yang's team [58] developed a minute-scale production method to produce MXenes by selectively etching the MAX phase (Ti₂AlC) under metal chloride (ZnCl₂) vapor (Ti₂CCl_x) (Fig. 5l). By this method, the etching time is less than one-sixtieth of the liquid phase etching and molten salt etching methods. The resulting MXenes can be directly used as an efficient platform for immobilizing single-atom Pt, with good HER electrocatalytic performance, a Tafel slope as low as 31 mV/dec, and a cycling stability of up to 5000 cycles. Although MXenes are often used as substrates for SACs, the knowledge of the interactions between single atoms and MXenes is still relatively scarce. Based on the abundant intrinsic defects on the surface of MXenes and the tunable functional groups, Prof Rogach's team [67] innovatively constructed a variety of single-atom Co-based catalysts (Co@V₂CT_x, Co@Nb₂CT_x, and Co@Ti₃C₂T_x). The HER and OER overpotentials of Co@V₂CT_x are 35 mV and 242 mV at a current density of 10 mA/cm², respectively. Through synchrotron radiation spectroscopy, DFT calculations, and comparative analysis of the electronic structures and oxidation states of Co active sites on different MXenes substrates, the study revealed strong electron-orbital coupling and enhanced charge transfer dynamics in Co@V₂CT_x, which synergistically improved both HER and OER activities (Fig. 5m). This work systematically elucidates the interaction mechanisms between single atoms and MXenes substrates, offering critical insights for the rational design of SACs.

Despite their high conductivity, tunable surface chemistry, and mechanical stability, MXenes face challenges in synthesis and processing, and their long-term stability under electrochemical conditions requires further investigation.

MOFs, characterized by their porous crystalline structure formed through the self-assembly of metal ions or clusters with organic ligands, offer an exceptional platform for SACs. MOFs were first reported by Yaghi's team in 1995 [68]. Since then, a vast number of MOFs have been successfully synthesized, with several classical MOFs becoming research hotspots in materials science. To date, researchers have developed over 20,000 distinct MOFs structures with varied morphologies through the combination of diverse metal ions and organic ligands. Their unique architectural features include ultra-high specific surface area, tunable pore size, and versatile coordination environments [17]. By immobilizing metal single atoms within the MOFs framework via coordination bonds, high dispersion and stability of active sites can be achieved, making MOFs particularly advantageous for catalytic reactions requiring precise control over the coordination environment. Furthermore, catalytic performance is intrinsically governed by the microenvironment surrounding active sites, making the rational design of advanced SACs pivotal for achieving high-efficiency catalytic systems. For instance, Jiang et al. [69] designed a Pd NPs@MOFs composite by modifying the ligand branches of MOFs, integrating metal active sites with tailored microenvironments to optimize interactions between Pd nanoparticles (NPs) and substrates (Fig. 6a). This composite demonstrated superior catalytic activity and selectivity in the dehydrocoupling of organosilanes and dehydrogenation of halogenated nitrobenzenes, outperforming conventional Pd NPs supported on surfactants or other supports. The morphology of MOFs significantly influences their catalytic properties, as different structures expose distinct lattice planes, thereby affecting reactivity [70,71]. The diversity in MOF morphology arises from the coordination dynamics between metal ions, ligands, and solvent molecules [71]. Solvent properties, such as polarity, intermolecular forces, and boiling point, further modulate crystal growth [72]. For example, Bai et al. [73] synthesized Fe-MOFs with controllable morphologies via a one-step solvothermal method by regulating competitive coordination in the metal-solvent system (Fig. 6b). Upon pyrolysis, these Fe-MOFs transformed into Fe₂O₃ with varied morphologies and exposed active crystalline planes, leading to differentiated electrocatalytic performances. Although MOFs exhibit significant potential in the field of catalysis, their limited chemical stability, mechanical stability, and electrical conductivity have restricted their widespread application. Through pyrolysis treatment, MOFs can be transformed into MOFs-derived carbon materials, which not only retain the porous structure and high specific surface area of the MOFs precursors but also significantly enhance chemical stability, mechanical stability, and electrical conductivity. Additionally, the porous structure of MOFs can be utilized to encapsulate metal nanoparticles, which, after pyrolysis, are uniformly dispersed on carbon or oxide supports. By precisely tuning the structure of the MOFs precursors, the morphology and electronic structure of the derived materials can be meticulously designed, thereby optimizing metal-support interactions and further improving catalytic performance. This strategy provides a novel direction for the development of highly efficient catalysts [74,75]. For instance, Jiang et al. developed a nitrogen-doped graphene-coated PdNiCo alloy catalyst derived from MOFs, where trace palladium atoms enhanced electron transfer efficiency, and the interface between the alloy and nitrogen-doped carbon synergistically modulated hydrogen adsorption energy, significantly boosting HER activity. This catalyst exhibited performance comparable to commercial Pt/C [76]. In addition, considering that MOFs can effectively transfer their advantages to carbon-based materials during pyrolysis, Wen et al. [77] developed a microwave-assisted rapid pyrolysis method to prepare nitrogen-doped carbon-supported single-atom nickel catalysts (Ni_1–N-C) from Ni-ZIF-8 precursors within 3 min (Fig. 6c), achieving a FE_CO of 96% in CO₂RR. Similarly, our team synthesized Fe-NP/MNCF, a bifunctional catalyst for ORR and CO₂RR, by pyrolyzing Fe-doped ZIF-8 modified with phosphomolybdic acid hydrate (PMo). Fe-NP/MNCF demonstrated exceptional ORR activity, surpassing commercial Pt/C, and remarkable stability in CO₂ electrolysis driven by ZABs. Raman spectroscopy revealed that PMo introduction increased defect density, enhancing electrocatalytic activity (Fig. 6d), while BET analysis confirmed a high specific surface area of 674.85 m²/g, facilitating reactant mass transfer (Fig. 6e). XPS analysis identified graphite-N, pyridine-N, and metal-N species, with the latter forming Fe-N-C structures that optimized ORR performance by promoting 4-electron pathways and inhibiting H₂O₂ formation (Fig. 6f). In addition, no metal-O peaks were detected in Fe-NP/MNCF (Fig. 6g), suggesting that it effectively avoids the loss of active sites due to Fe oxidation. It was shown that polyacid doping not only significantly increased the surface charge distribution and defect density of the catalyst, but also optimized the multilevel pore structure of ZIF-8, providing more channels for electron transfer. This innovative MOF modification strategy opens a new way to design efficient ORR and CO₂RR catalysts [75]. Chen et al. utilized Zn/Mo-MOFs to anchor Mo single atoms and ultra-small MoC clusters on nitrogen-doped porous carbon, demonstrating strong N₂ adsorption capabilities for NRR [78]. However, challenges such as high costs and metal aggregation during carbonization remain. Alternative approaches, such as coupling MIL-53(Fe) with TiO₂, have achieved notable NH₃ yields without carbonization [79], while Liu et al. [80] prepared Fe-anchored B- and N-co-doped carbon substrates from ferrocenyl boronic acid-coated ZIF-8, exhibiting exceptional activity in CO₂RR and Zn-CO₂ batteries (Figs. 6h-k).

A comparative analysis of the four support types reveals distinct advantages and limitations in different electrocatalytic reactions. Carbon-based materials excel in ORR and HER due to their high conductivity and tunable surface chemistry, while metal oxides and MOFs show superior performance in OER and CO₂RR, respectively. MXenes, with their unique combination of conductivity and surface functionality, demonstrate potential across multiple reactions but require further development. The catalytic activity of SACs is influenced by several key factors, including the dispersion and stability of single atoms, the electronic interaction between supports and active sites, and the mass transfer properties of the support. Optimizing these factors through innovative support design and synthesis strategies is essential for enhancing SACs performance.

Through precise regulation of active sites via support engineering, SACs exhibit significant potential in renewable energy conversion and storage. Their unique structures and exceptionally high atomic utilization rates make them pivotal materials in various energy technologies, offering innovative solutions for achieving efficient and cost-effective energy conversion and storage. In fuel cells, high-performance single-atom ORR catalysts can substantially enhance the efficiency and durability of fuel cells while reducing costs, thereby accelerating the commercialization of fuel cell technology [81–83]. In water electrolysis for hydrogen production, SACs with superior HER and OER catalytic properties can significantly improve electrolysis efficiency, reduce energy consumption, and facilitate large-scale green hydrogen production, laying a solid foundation for the development of the hydrogen energy industry [84–86]. Furthermore, in energy storage systems such as metal-air batteries and lithium-ion batteries, SACs are expected to play a critical role by enhancing battery performance and cycle life, providing new approaches to addressing energy storage challenges [34,87–89].

In summary, SACs with precisely regulated active sites through support engineering demonstrate vast application potential in fields such as fuel cells, water electrolysis, metal-air batteries, and lithium-ion batteries. Their remarkable characteristics: high efficiency, low cost, and controllability, provide essential technical support for the efficient conversion and storage of renewable energy. These advancements are likely to drive the rapid development and widespread adoption of clean energy technologies. Looking ahead, as research into the regulation mechanisms of SACs active sites deepens and breakthroughs in large-scale preparation technologies emerge, SACs are poised to play an even more critical role in the energy sector, offering robust technical solutions for achieving carbon neutrality goals.

In the field of environmental remediation, SACs, with their exceptional activity, selectivity, and atomic utilization efficiency, demonstrate immense potential, particularly in wastewater treatment and exhaust gas purification. By precisely regulating active sites through support engineering, SACs can efficiently catalyze the conversion of environmental pollutants, providing innovative solutions for achieving green and sustainable environmental remediation. For instance, in electrocatalytic carbon dioxide reduction, SACs engineered for CO₂RR can convert carbon dioxide into valuable chemicals or fuels. This not only enables the resource utilization of carbon dioxide but also effectively mitigates greenhouse gas emissions, contributing significantly to global climate change mitigation [90–93]. In wastewater treatment, SACs can efficiently degrade persistent organic pollutants such as dyes, pesticides, and pharmaceutical residues into harmless small molecules like CO₂ and H₂O through processes such as the electrochemical oxidation reaction, thereby achieving water purification and offering a powerful tool for water resource protection and recycling [94–96]. Additionally, due to their excellent atomic efficiency and tunable electronic structure, SACs exhibit unusual molecular adsorption and electron transfer capabilities, which are also valuable in the detection and degradation of polluted gases [97–100].

In summary, SACs with precisely regulated active sites through support engineering present broad application prospects in environmental remediation. Their outstanding performance in carbon dioxide resource utilization, wastewater treatment, and exhaust gas purification provides critical technical support for achieving sustainable environmental protection [101–104]. In the future, as research into the regulation mechanisms of SACs active sites advances and large-scale preparation technologies mature, SACs are expected to play an even more pivotal role in environmental remediation, offering robust technical guarantees for building a clean and healthy ecological environment.