Enhancing Hydrogen Peroxide Synthesis through Coordination Engineering of Single-Atom Catalysts in the Oxygen Reduction Reaction: A Review

Hydrogen peroxide (H2O2) is an important chemical with a diverse array of applications. However, the existing scenario of centralized high-concentration production is in contrast with the demand for low-concentration decentralized production. In this context, the on-site green and efficient two-electron oxygen reduction reaction (ORR) for H2O2 production has developed into a promising synthetic approach. The development of low-cost, highly active, and durable advanced catalysts is the core requirement for realizing this approach. In recent years, single-atom catalysts (SACs) have become a research hotspot owing to their maximum atom utilization efficiency, tunable electronic structure, and exceptional catalytic performance. The coordination engineering of SACs is one of the key strategies to unlock their full potential for electrocatalytic H2O2 synthesis and holds significant research value. Despite considerable efforts, precisely controlling the electronic structure of active sites in SACs remains challenging. Therefore, this review summarizes the latest progress in coordination engineering strategies for SACs, aiming to elucidate the relevance between structure and performance. Our goal is to provide valuable guidance and insights to aid in the design and development of high-performance SACs for electrocatalytic H2O2 synthesis.


Introduction
Hydrogen peroxide (H 2 O 2 ) is a vital basic chemical widely used in various fields, including chemical synthesis, papermaking, textiles, healthcare, and environmental protection. It also holds potential as an energy carrier [1][2][3][4]. At present, industrial H 2 O 2 production primarily depends on the anthraquinone process [5]. However, this method is not only complex and energy demanding, but also generates a substantial amount of organic waste [6][7][8]. Additionally, to reduce storage and transportation costs, industrially produced H 2 O 2 is concentrated through distillation to high concentrations, posing potential explosion risks. Nevertheless, many real-world applications only require diluted H 2 O 2 . The contradiction between the centralized production of high concentrations and the demand for the dispersed production of low concentrations presents a challenging dilemma [9][10][11].
To address the above-mentioned challenges, a cleaner and safer synthesis approach for H 2 O 2 is via the direct electrocatalytic oxygen reduction reaction (2e − ORR) using clean electrical energy and O 2 (O 2 + 2e − + 2H + → H 2 O 2 ) [12][13][14][15]. This method offers a more environmentally friendly substitute to the conventional anthraquinone process, especially well-suited for on-site applications that require small amounts of low-concentration H 2 O 2 , such as water purification and clean disinfection [16][17][18]. However, the 2e − ORR process suffers from slow reaction kinetics and competition with the 4e − ORR process that leads to the generation of H 2 O. This competition results in low activity and selectivity, becoming the two main obstacles in electrocatalytic H 2 O 2 production [19][20][21]. Hence, the design and development of advanced catalysts boasting high activity, selectivity, and long-term stability for the 2e − ORR have become crucial issues that urgently need to be addressed in electrocatalytic H 2 O 2 synthesis.
Various catalysts have been employed for the electrocatalytic synthesis of H 2 O 2 to date, including noble metals and their alloys, non-noble metals and their oxides, carbides, non-metallic carbon-based catalysts, and single-atom catalysts [22][23][24][25][26][27][28]. Among the numerous catalysts reported, single-atom catalysts (SACs) with isolated metal atoms anchored on the surface of the support have combined the advantages of both homogeneous and heterogeneous catalysts, offering ultra-high atom utilization, high activity, and high selectivity. As a result, they have emerged as the ideal catalysts for the electrocatalytic H 2 O 2 synthesis [29][30][31]. Particularly, carbon-based SACs have attracted significant interest because of their low cost, wide availability, and relatively simple and tunable active site structures (M-L x -C, metal-ligand-carbon). These properties make them an excellent platform for studying the mechanism and optimizing their performance of electrocatalytic H 2 O 2 synthesis [32][33][34]. Therefore, carbon-based SACs are the primary focus of our subsequent discussion.
The coordination engineering of carbon-based SACs, as a key strategy to regulate the electronic structure of active sites and optimize ORR activity and selectivity, holds great research significance [35]. Despite considerable efforts in recent times, a comprehensive understanding of the structure-performance relationship in carbon-based SACs is still lacking, making their rational design and synthesis challenging [33,[35][36][37]. Therefore, this review summarizes the latest achievements in the coordination engineering of SACs for 2e − ORR, aiming to expand our understanding of the structure-performance correlation and provide valuable insights into the development of advanced SACs for electrocatalytic H 2 O 2 synthesis.

Mechanism of Electrocatalytic Oxygen Reduction for H 2 O 2 Synthesis
The electrocatalytic oxygen reduction reaction (ORR) is an intricate procedure encompassing multiple proton-coupled electron transfer (PCET) steps. In practical reaction systems, both the 4e − ORR and 2e − ORR pathways coexist and compete with each other (Figure 1a) [38,39]. As an essential reaction in the processes of the storage and conversion of renewable energy, 4e − ORR that thermodynamically favors the generation of H 2 O is mainly used in proton exchange membrane fuel cells (PEMFCs) [40,41] and metal-air batteries [42][43][44][45]. In contrast, the 2e − ORR pathway leading to the formation of H 2 O 2 can serve as a hopeful substitute to address the limitations of the anthraquinone process and enable the on-site decentralized production of H 2 O 2 [46][47][48].
The 4e − ORR involves three reaction intermediates, *OOH, *O, and *OH (* denotes the catalytic active sites; *OOH represents the state of the OOH intermediate adsorbed at the active sites, they convey the same meaning as those stated here in the following context), and all reaction steps in the acidic media are as follows: * + O 2 (g) + H + + e − → * OOH (1) In contrast, the 2e − ORR pathway involves only one reaction intermediate, *OOH, and its reaction steps are as follows: * + O 2 (g) + H + + e − → * OOH (5) The first step of ORR is the adsorption and activation of O 2 molecules on the catalyst surface, and its adsorption configuration significantly influences the reaction tendency [49]. There are three typical adsorption configurations for O 2 molecules on the catalyst surface ( Figure 1b) [50]: Pauling-type adsorption (end-on adsorption on a single atom), Griffithtype adsorption (side-on adsorption on a single atom), and Bridge-type adsorption (side-on adsorption on two atoms). For metal catalysts, O 2 molecules tend to adsorb in a side-on manner on surface atoms, causing a weakened O-O bond, which is more prone to breaking, favoring the 4e − ORR pathway for generating H 2 O. On the contrary, SACs have isolated metal atoms as catalytic centers, which preferentially adsorb O 2 molecules in an end-on manner, favoring the retention of the O-O bond and generating H 2 O 2 through the 2e − ORR pathway [39,51].
In addition to the adsorption configuration, according to the Sabatier principle [52], an ideal 2e − ORR catalyst ought to maintain a moderate binding strength with the *OOH intermediate to achieve a balance between adsorption and desorption, resulting in a volcanoshaped relationship (Figure 1c) [50,53]. The optimal catalyst located at the peak of the volcano curve would have both a low-reaction energy barriers, indicating high activity, and the ability to promptly desorb the generated *OOH, ensuring good selectivity. Furthermore, the binding strength of SACs with the intermediate *OOH is largely determined by the electronic structure of the active site. Therefore, it is imperative to apply appropriate strategies for regulating the electronic structure of SACs.
The first step of ORR is the adsorption and activation of O2 molecules on the catalyst surface, and its adsorption configuration significantly influences the reaction tendency [49]. There are three typical adsorption configurations for O2 molecules on the catalyst surface ( Figure 1b) [50]: Pauling-type adsorption (end-on adsorption on a single atom), Griffith-type adsorption (side-on adsorption on a single atom), and Bridge-type adsorption (side-on adsorption on two atoms). For metal catalysts, O2 molecules tend to adsorb in a side-on manner on surface atoms, causing a weakened O-O bond, which is more prone to breaking, favoring the 4e − ORR pathway for generating H2O. On the contrary, SACs have isolated metal atoms as catalytic centers, which preferentially adsorb O2 molecules in an end-on manner, favoring the retention of the O-O bond and generating H2O2 through the 2e − ORR pathway [39,51].
In addition to the adsorption configuration, according to the Sabatier principle [52], an ideal 2e − ORR catalyst ought to maintain a moderate binding strength with the *OOH intermediate to achieve a balance between adsorption and desorption, resulting in a volcano-shaped relationship ( Figure 1c) [50,53]. The optimal catalyst located at the peak of the volcano curve would have both a low-reaction energy barriers, indicating high activity, and the ability to promptly desorb the generated *OOH, ensuring good selectivity. Furthermore, the binding strength of SACs with the intermediate *OOH is largely determined by the electronic structure of the active site. Therefore, it is imperative to apply appropriate strategies for regulating the electronic structure of SACs.

Regulation of Coordination Environment
In carbon-based SACs, the isolated metal atoms that act as active centers are dispersed on the carbon support through coordination with non-metallic atoms. Compared to metal surfaces, the spatial separation of metal atoms in SACs on proper supports leads to their unique electronic structure [54]. As the research on carbon-based SACs expands, an increasing number of studies indicate that regulating the electronic distribution of the central metal atom's frontier d-orbitals is crucial for achieving ideal catalytic activity [53,55,56]. Since p-block elements are commonly used as coordinating atoms in the first coordination layer, the d-p orbital hybridization dominates the electronic structure of most reported active sites in carbon-based SACs [57,58,59,60]. Therefore, changes in the coordinating atoms within the first coordination layer directly impacts the electronic structure of the active sites, thereby affecting the activity and selectivity of ORRs. In addition, when heteroatoms are doped into the carbon support, long-range interactions can also occur in the second or further coordination layers ( Figure 2) [35,61,62]. Thus, we first discussed the direct coordination environments in the following

Regulation of Coordination Environment
In carbon-based SACs, the isolated metal atoms that act as active centers are dispersed on the carbon support through coordination with non-metallic atoms. Compared to metal surfaces, the spatial separation of metal atoms in SACs on proper supports leads to their unique electronic structure [54]. As the research on carbon-based SACs expands, an increasing number of studies indicate that regulating the electronic distribution of the central metal atom's frontier d-orbitals is crucial for achieving ideal catalytic activity [53,55,56]. Since p-block elements are commonly used as coordinating atoms in the first coordination layer, the d-p orbital hybridization dominates the electronic structure of most reported active sites in carbon-based SACs [57][58][59][60]. Therefore, changes in the coordinating atoms within the first coordination layer directly impacts the electronic structure of the active sites, thereby affecting the activity and selectivity of ORRs. In addition, when heteroatoms are doped into the carbon support, long-range interactions can also occur in the second or further coordination layers ( Figure 2) [35,61,62]. Thus, we first discussed the direct coordination environments in the following sections, including strategies to adjust coordinating atoms, coordination number, and axial coordination. Secondly, we reviewed the modulation of indirect coordination environments. Finally, the development prospects and challenges of carbon-based SACs' coordination engineering were summarized. sections, including strategies to adjust coordinating atoms, coordination number, and axial coordination. Secondly, we reviewed the modulation of indirect coordination environments. Finally, the development prospects and challenges of carbon-based SACs' coordination engineering were summarized.

Adjustment of the First Coordination Layer
In carbon-based SACs, the first-coordination-layer atoms anchor the metal center to the carbon substrate. Consequently, the characteristics of the atoms in the first coordination layer exert a direct impact on the geometric and electronic configurations of the active sites [63,64,65]. Different coordination atoms often exhibit variations in their atomic radius, electronegativity, and electronic structure, among other properties. To achieve a be er electrocatalytic performance for H2O2 synthesis, it is a straightforward and sensible strategy to choose appropriate first-coordination-layer atoms for specific metal centers, modify their electronic structure, and optimize their interactions with reaction intermediates.

The Types of Coordination Atoms
The most common coordinating atoms for anchoring metal centers in carbon-based SACs are nitrogen (N) atoms, which have similar atomic radii to carbon, but higher electronegativity. Therefore, the extensive research and documentation have focused on catalysts featuring M-N4-type sites [66,67]. However, even with the same N element, different types of N atoms (pyridinic N/pyrrolic N) can lead to discrepancies in the catalytic performance. Among an array of transition metal-based carbon SACs used for ORR, Co-SACs have been a controversial presence. Numerous experimental and computational evidence has indicated that Co-SACs exhibit the best d-band center and *OOH intermediate adsorption energy. Therefore, it is an excellent catalyst for the electrochemical synthesis of H2O2 [68,69,70]. However, reports of CoN4 sites showing high activity for 4e − ORR are not uncommon [71,72,73]. To unravel this mystery, Chen et al. revealed the "structure-performance" relationship of CoN4 catalysts with different N coordinations through theoretical simulations combined with experiments: pyrrolic-type CoN4 centers predominantly drive the 2e − ORR, whereas pyridinic-type CoN4 catalyzes 4e − ORR (Figure 3a) [74]. The difference in selectivity is a ributed to the fully paired dorbitals in pyridinic-type CoN4, while pyrrolic-type CoN4 exhibits high-spin states with unpaired electrons in the d-orbitals, resulting in weaker electron transfer with *OOH intermediates, making it more inclined towards 2e − ORR. Ultimately, Co-N SACDp with all-pyrrolic N coordination demonstrated outstanding H2O2 selectivity of up to 94% and an excellent mass activity of 14.4 A gcatalyst −1 at 0.5 V versus RHE in 0.1 M HClO4.
Nevertheless, stable M-N4-C structures do not always show the best activity and selectivity for the electrochemical synthesis of H2O2 [75]. Selecting suitable coordinating

Adjustment of the First Coordination Layer
In carbon-based SACs, the first-coordination-layer atoms anchor the metal center to the carbon substrate. Consequently, the characteristics of the atoms in the first coordination layer exert a direct impact on the geometric and electronic configurations of the active sites [63][64][65]. Different coordination atoms often exhibit variations in their atomic radius, electronegativity, and electronic structure, among other properties. To achieve a better electrocatalytic performance for H 2 O 2 synthesis, it is a straightforward and sensible strategy to choose appropriate first-coordination-layer atoms for specific metal centers, modify their electronic structure, and optimize their interactions with reaction intermediates.

The Types of Coordination Atoms
The most common coordinating atoms for anchoring metal centers in carbon-based SACs are nitrogen (N) atoms, which have similar atomic radii to carbon, but higher electronegativity. Therefore, the extensive research and documentation have focused on catalysts featuring M-N 4 -type sites [66,67]. However, even with the same N element, different types of N atoms (pyridinic N/pyrrolic N) can lead to discrepancies in the catalytic performance. Among an array of transition metal-based carbon SACs used for ORR, Co-SACs have been a controversial presence. Numerous experimental and computational evidence has indicated that Co-SACs exhibit the best d-band center and *OOH intermediate adsorption energy. Therefore, it is an excellent catalyst for the electrochemical synthesis of H 2 O 2 [68][69][70]. However, reports of CoN 4 sites showing high activity for 4e − ORR are not uncommon [71][72][73]. To unravel this mystery, Chen et al. revealed the "structure-performance" relationship of CoN 4 catalysts with different N coordinations through theoretical simulations combined with experiments: pyrrolic-type CoN 4 centers predominantly drive the 2e − ORR, whereas pyridinic-type CoN 4 catalyzes 4e − ORR (Figure 3a) [74]. The difference in selectivity is attributed to the fully paired d-orbitals in pyridinic-type CoN 4 , while pyrrolictype CoN 4 exhibits high-spin states with unpaired electrons in the d-orbitals, resulting in weaker electron transfer with *OOH intermediates, making it more inclined towards 2e − ORR. Ultimately, Co-N SAC Dp with all-pyrrolic N coordination demonstrated outstanding H 2 O 2 selectivity of up to 94% and an excellent mass activity of 14.4 A g catalyst −1 at 0.5 V versus RHE in 0.1 M HClO 4 .
Nevertheless, stable M-N 4 -C structures do not always show the best activity and selectivity for the electrochemical synthesis of H 2 O 2 [75]. Selecting suitable coordinating atoms to optimize the electronic structure of the metal center is crucial for breaking the deadlock. Li and colleagues designed and prepared carbon-based SACs with a biomimetic FeN 3 S 1 asymmetric catalytic center (Figure 3b-d) [76]. The newly designed Fe-NS/C catalyst displayed an excellent selectivity of~90% for synthesizing H 2 O 2 . At a high current density of 100 mA cm −2 , it achieved an accumulation concentration of 5.8 wt.% H 2 O 2 , enabling medical-grade H 2 O 2 disinfectant electro-synthesis. Experiments along with DFT calculations amply confirmed that the remarkable activity was attributed to the introduction of S species in the FeN 3 S 1 site, which broke the symmetry of the FeN 4 structure, accelerated proton transfer, promoted the formation of *OOH intermediates, and facilitated product desorption.
Zhao and colleagues investigated an array of carbon-based Pt-SACs featuring different adjacent coordinating atoms [77]. The results showed that, compared to Pt-C 4 and Pt-N 4 sites, the Pt-S 4 site with the S coordination exhibited an excellent performance in reducing O 2 to H 2 O 2 , owing to the optimized electronic structure that provided proper binding strength to the intermediates. atoms to optimize the electronic structure of the metal center is crucial for breaking the deadlock. Li and colleagues designed and prepared carbon-based SACs with a biomimetic FeN3S1 asymmetric catalytic center (Figure 3b-d) [76]. The newly designed Fe-NS/C catalyst displayed an excellent selectivity of ~90% for synthesizing H2O2. At a high current density of 100 mA cm −2 , it achieved an accumulation concentration of 5.8 wt.% H2O2, enabling medical-grade H2O2 disinfectant electro-synthesis. Experiments along with DFT calculations amply confirmed that the remarkable activity was a ributed to the introduction of S species in the FeN3S1 site, which broke the symmetry of the FeN4 structure, accelerated proton transfer, promoted the formation of *OOH intermediates, and facilitated product desorption. Zhao and colleagues investigated an array of carbon-based Pt-SACs featuring different adjacent coordinating atoms [77]. The results showed that, compared to Pt-C4 and Pt-N4 sites, the Pt-S4 site with the S coordination exhibited an excellent performance in reducing O2 to H2O2, owing to the optimized electronic structure that provided proper binding strength to the intermediates.  Apart from N atoms, O atoms, which have higher electronegativity than N, can also serve as excellent coordinating atoms [78,79]. Jia et al. tuned the coordinating environment of Zn single atoms through the alteration of functional groups in the MOF precursor, leading to the synthesis of a unique ZnO 3 C catalyst with O and C coordinations (Figure 3e-g) [36]. Compared to the traditional Zn-N 4 structure, this catalyst achieved near-zero overpotential and high selectivity for 2e − ORR in 0.1 M KOH. DFT calculations revealed that the outstanding 2e − ORR efficacy of ZnO 3 C was owing to the greater electronegativity of the O atoms, which diluted the electron distribution around the central Zn atom, lowered its d-band center, and thus weakened its adsorption of *OOH intermediates.

The Number of Coordination Atoms
Compared to the investigation of changing coordination atom types, the impact of altering the number of coordination atoms on 2e − ORR for H 2 O 2 synthesis has been less studied. This can be attributed to the instability of low or high coordination structures compared to the classical M-L 4 configuration, resulting in limited synthetic pathways [80]. From the perspective of coordination chemistry, changes in the number of coordinating atoms significantly affect the oxidation state of the metal center, altering the geometric structure and electronic configuration of the d-orbitals, finally influencing the coupling with the 2p-orbitals of the intermediate O atoms [63]. Therefore, adjusting the number of coordinating atoms around the metal center has become a viable approach for optimizing the 2e − ORR performance.
Zhang and co-workers constructed Co-O-C active sites on the surface of graphene oxide (GO) through a two-step acid etching method (Figure 4a,b) [81]. This catalyst displayed an onset potential of 0.91 V versus. RHE and an exceptional H 2 O 2 selectivity of 81.4% in 0.1 M KOH. Through DFT calculations, they confirmed that the low-coordination Co-O 3 -C was the actual catalytic center. Compared to the Co-O 4 -C site, where the Co atom had a large bandgap at the Fermi level, resulting in electron transfer inhibition, the Co-O 3 -C structure had a smaller bandgap, promoting electron transfer between the catalytic center and intermediates, leading to its outstanding 2e − ORR efficiency. Gong et al. synthesized a Co-N-C catalyst with low-coordination CoN 2 sites and abundant epoxy groups by a one-step microwave thermal shock technique, achieving the modulation of the coordination number of atomically dispersed Co sites and neighboring oxygen functional groups (Figure 4c,d) [82]. This catalyst exhibited high selectivity (91.3%) for H 2 O 2 synthesis, along with a high mass activity (44.4 A g −1 ) and high kinetic current density (11.3 mA cm −2 ) at 0.65 V versus RHE. The stability of this low-coordinated catalyst, which required particular attention, was verified through a 10 h chronoamperometric test with no significant decline in the current and selectivity. Subsequent DFT calculations explained that the excellent activity and selectivity of Co-N 2 -C/HO originated from its distinctive electronic structure, which was synergistically formed by the low-coordination configuration and the existence of oxygen functional groups.

Axial Coordination
As is well-known, the metal centers within M-N4 sites are in an unsaturated coordination state. With the d-orbitals not being fully occupied, there is an opportunity for an interaction with external ligands along the axial direction [35,83,84]. Therefore, in addition to adjusting the in-plane coordination atoms and numbers in the first coordination sphere, controlling axial coordination may become a novel avenue for advancing the future of carbon-based SACs. Therefore, it is necessary to discuss this aspect separately.
Zhao and colleagues reported Co-SAC with an axial N coordination (Figure 5a-c) [85]. SACs with CoN5 sites achieved a substantial H2O2 yield of 3.4 mol gcatalyst −1 h −1 in the flow electrolysis cell with an electrolyte of a 0.5 M NaCl solution at 50 mA cm −2 for a longterm stability period of 24 h. Experiments and calculations showed that the Co-N5 structure was a more favorable coordination type compared to Co-N4, as it not only optimized the projected density of Co 3d-orbitals but also exhibited faster reaction kinetics in the 2e − ORR pathway.
Similarly, Fan et al. designed a hybrid SAC(Co-N5) by connecting cobalt phthalocyanine molecules with pyridine-functionalized carbon nanotubes through pyridine ligands [86]. The introduction of pyridine ligands disrupted the typical planar symmetry and symmetric charge distribution of Co-N4 sites. The lone-pair electrons in pyridinic N caused a repulsion concerning the electrons in the Co-3dz 2 orbital when forming coordination bonds with CoPc. This induced electron accumulation above the Co center, enhanced the interaction between Co-3dz 2 and *OOH's p orbitals, and thus reduced ∆G*OOH, resulting in an onset potential of 0.85 V versus RHE and a high selectivity of 95%. Furthermore, the team extended the study to demonstrate tunable activity and

Axial Coordination
As is well-known, the metal centers within M-N 4 sites are in an unsaturated coordination state. With the d-orbitals not being fully occupied, there is an opportunity for an interaction with external ligands along the axial direction [35,83,84]. Therefore, in addition to adjusting the in-plane coordination atoms and numbers in the first coordination sphere, controlling axial coordination may become a novel avenue for advancing the future of carbon-based SACs. Therefore, it is necessary to discuss this aspect separately.
Zhao and colleagues reported Co-SAC with an axial N coordination (Figure 5a-c) [85]. SACs with CoN 5 sites achieved a substantial H 2 O 2 yield of 3.4 mol g catalyst −1 h −1 in the flow electrolysis cell with an electrolyte of a 0.5 M NaCl solution at 50 mA cm −2 for a long-term stability period of 24 h. Experiments and calculations showed that the Co-N 5 structure was a more favorable coordination type compared to Co-N 4 , as it not only optimized the projected density of Co 3d-orbitals but also exhibited faster reaction kinetics in the 2e − ORR pathway. Similarly, Fan et al. designed a hybrid SAC(Co-N 5 ) by connecting cobalt phthalocyanine molecules with pyridine-functionalized carbon nanotubes through pyridine ligands [86]. The introduction of pyridine ligands disrupted the typical planar symmetry and symmetric charge distribution of Co-N 4 sites. The lone-pair electrons in pyridinic N caused a repulsion concerning the electrons in the Co-3dz 2 orbital when forming coordination bonds with CoPc. This induced electron accumulation above the Co center, enhanced the interaction between Co-3dz 2 and *OOH's p orbitals, and thus reduced ∆G *OOH , resulting in an onset potential of 0.85 V versus RHE and a high selectivity of 95%. Furthermore, the team extended the study to demonstrate tunable activity and selectivity through the introduction of electron-donating or electron-withdrawing groups on the pyridine lig-ands, showcasing the versatility of controlling the activity and selectivity through the axial ligand coordination.
Moreover, small-molecule ligands and halide ions exhibit robust coordination capabilities in the M-N 4 sites and show a critical impact on enhancing the activity [87]. The preparation of SACs with additional coordination can be achieved through low-temperature thermal treatment. Kim and co-workers reported the reversible control of SACs' coordination structures through ligand exchange reactions to modify the reaction activity and selectivity of ORR (Figure 5d) [88]. By thermally treating carbon carriers impregnated with RhCl 3 in CO or an NH 3 gas flow at 200 • C, they obtained atomically dispersed Rh catalysts with CO or NH 2 coordinations. The ORR activity of CO-coordinated Rh-SACs is roughly 30-times greater than that of NH 2 -coordinated Rh-SACs, while the H 2 O 2 selectivity of NH 2 -coordinated Rh-SACs is about three-times higher than the former. The key innovation of this study lies in the reversible exchange of CO and NH 2 ligands and the change in Rh's oxidation state through different thermal treatment atmospheres. This achieved the reversible adjustment of ORR activity and selectivity, demonstrating the essential function of small-molecule coordination in regulating the ORR performance of SACs.
Cao et al. used DFT calculations with *OOH binding energy as a descriptor to screen transition metal SACs for 2e − ORR and guide the catalyst synthesis. They synthesized a Cophthalocyanine catalyst (CoPc-OCNT) with axial oxygen coordination in oxidized carbon nanotube carriers [89]. The axial oxygen coordination optimized the *OOH binding energy of CoPc, allowing it to be positioned at the top of the volcano curve. Finally, it exhibited a negligible energy barrier, thus demonstrating an excellent 2e − ORR performance. Analogously, Xiao and colleagues synthesized SACs with a super-coordinated Ni 1 N 4 O 2 structure on carboxylated multi-walled carbon nanotubes (OCNTs) through thermal decomposition (Figure 5e) [90]. Comprehensive structural analysis indicated that two oxygen atoms in OCNTs anchored the NiN 4 sites, which facilitated the detachment of *OOH and accelerated the reaction kinetics of 2e − ORR. This ultimately resulted in a high H 2 O 2 electro-synthesis rate of 5.7 mmol cm −2 h −1 and a high Faradaic efficiency of 96% at an industrial-relevant current density of 200 mA cm −2 .
Energies 2023, 16, x FOR PEER REVIEW 8 of 15 selectivity through the introduction of electron-donating or electron-withdrawing groups on the pyridine ligands, showcasing the versatility of controlling the activity and selectivity through the axial ligand coordination. Moreover, small-molecule ligands and halide ions exhibit robust coordination capabilities in the M-N4 sites and show a critical impact on enhancing the activity [87]. The preparation of SACs with additional coordination can be achieved through lowtemperature thermal treatment. Kim and co-workers reported the reversible control of SACs' coordination structures through ligand exchange reactions to modify the reaction activity and selectivity of ORR (Figure 5d) [88]. By thermally treating carbon carriers impregnated with RhCl3 in CO or an NH3 gas flow at 200 °C, they obtained atomically dispersed Rh catalysts with CO or NH2 coordinations. The ORR activity of COcoordinated Rh-SACs is roughly 30-times greater than that of NH2-coordinated Rh-SACs, while the H2O2 selectivity of NH2-coordinated Rh-SACs is about three-times higher than the former. The key innovation of this study lies in the reversible exchange of CO and NH2 ligands and the change in Rh's oxidation state through different thermal treatment atmospheres. This achieved the reversible adjustment of ORR activity and selectivity, demonstrating the essential function of small-molecule coordination in regulating the ORR performance of SACs.
Cao et al. used DFT calculations with *OOH binding energy as a descriptor to screen transition metal SACs for 2e − ORR and guide the catalyst synthesis. They synthesized a Co-phthalocyanine catalyst (CoPc-OCNT) with axial oxygen coordination in oxidized carbon nanotube carriers [89]. The axial oxygen coordination optimized the *OOH binding energy of CoPc, allowing it to be positioned at the top of the volcano curve. Finally, it exhibited a negligible energy barrier, thus demonstrating an excellent 2e − ORR performance. Analogously, Xiao and colleagues synthesized SACs with a supercoordinated Ni1N4O2 structure on carboxylated multi-walled carbon nanotubes (OCNTs) through thermal decomposition (Figure 5e) [90]. Comprehensive structural analysis indicated that two oxygen atoms in OCNTs anchored the NiN4 sites, which facilitated the detachment of *OOH and accelerated the reaction kinetics of 2e − ORR. This ultimately resulted in a high H2O2 electro-synthesis rate of 5.7 mmol cm −2 h −1 and a high Faradaic efficiency of 96% at an industrial-relevant current density of 200 mA cm −2 .

Adjustment of the Second and Higher Coordination Layers
In contrast to the direct adjustment of atomic type or coordination number in the first coordination layer, doping non-metallic heteroatoms in the second and higher coordination layers may also facilitate electrocatalytic H 2 O 2 synthesis by modulating the charge distribution around the metal sites through long-range delocalization effects [91]. Liu et al. studied the effects of introducing P heteroatoms into the second coordination layer of CoN 4 sites in carbon-based SACs using both experimental and DFT calculations (Figure 6a,b) [92]. They found that the introduction of P atoms to the second coordination layer of the Co center in the form of C-P bonds caused tensile strain on the Co-N bonds in the CoN 4 site. This induced a reduction in the electron density of the central Co atom, optimizing the adsorption strength of *OOH on the CoN 4 -PC active site, ultimately enhancing the efficiency of H 2 O 2 synthesis. The exceptional electrochemical performance, including a high onset potential of 0.81 V, over 97% selectivity, and an H 2 O 2 production rate of 11.2 mol g −1 catalyst h −1 during 110 h of durability testing, makes it highly promising for large-scale H 2 O 2 production.
Hyeon and colleagues also highlighted that the catalytic performance of CoN 4 sites can be modulated by precisely adjusting the atomic arrangement in the second coordination layer [93]. They calculated the effects of adsorbing electron-poor species (4H*/2H*) and electron-rich species (O*/2O*) neighboring the CoN 4 sites on the electronic structure and ∆G *OOH (Figure 6c). The final reports revealed that the adsorption of electron-rich species slightly increased ∆G *OOH due to the electron state correction of the Co atom, thereby favoring the 2e − ORR pathway. Likewise, Zhang et al. investigated the electrochemical/chemical treatment to adjust the nearby oxygen functional groups surrounding atomically dispersed Co sites for high-activity and selective H 2 O 2 synthesis under acidic conditions (Figure 6d,e) [94]. The incorporation of neighboring C-O-C functional groups adjusted the electronic structure of the CoN 4 portion, significantly weakening the *OOH interaction with the Co atom. Finally, it exhibited the exceptional activity and selectivity of 2e − ORR. The overpotential was found to be lower than 0.02 V, comparable to or even lower than the state-of-the-art PtHg 4 catalysts. This study highlighted the significance of the synergistic effect between the CoN 4 portion and adjacent epoxy functional groups.
Furthermore, Liu et al. utilized β-substituted cobalt porphyrin to adsorb on carbon nanotube substrates to construct multiphase Co-SACs (Figure 6f-h) [95]. Through DFT calculations of different substituted SACs' d-orbital electronic structures and adsorption energies with intermediates, they found that the electron-withdrawing β-substituted group F significantly weakened the Co-O interaction, providing the best adsorption energy for intermediates and higher activity in generating H 2 O 2 . As a reward, the CoPorF/CNT catalyst showed over 94% H 2 O 2 selectivity and a turnover frequency (TOF) of 3.51 s −1 . In a dual-electrode electrolytic cell, its rate of H 2 O 2 yield reached 10.76 mol g catalyst −1 h −1 . Additionally, Yan et al. fabricated SACs of antimony (Sb-NSCF) supported on nitrogen and sulfur co-doped nanofibers using nanostructured Sb 2 S 3 as a template [96]. By doping S atoms in the second coordination layer, they adjusted the oxidation state of Sb atoms, impacting their interactions with *OOH intermediates and promoting the enhanced activity (114.9 A g catalyst −1 at 0.65 V) and selectivity (97.2%) of Sb-N 4 portions for 2e − ORR.

Conclusions and Outlook
Carbon-based SACs with their unique structures and excellent catalytic performance hold great promise as commercial catalysts for electrochemical H2O2 synthesis. The Figure 6. (a) Optimized structure of CoN 4 -PC with two PC 4 sites near the CoN 4 site (CoN 4 -P 2 ); (b) ORR reaction pathways and free energy diagrams for 2e − and 4e − ORRs at the CoN 4 -P 2 sites, the red and green spheres represent O and H atoms, respectively, while the rest remains the same with figure (a). Reproduced with permission from [92], Elsevier, 2023. (c) Differential charge densities of Co-N 4 /graphene after 4H*, 2H*, O*, or 2O* was adsorbed near the cobalt atom. Yellow and cyan isosurfaces (±0.003 Bohr −3 ) show the electron gain and loss, respectively. Reproduced with permission from [93], Springer Nature, 2020. (d) The optimized structures of bare CoN 4 and CoN 4 sites with different epoxy oxygen coverages, the gray, blue, red and orange spheres represent C, N, O and Co atoms, respectively; (e) calculated Sabatier volcano plot for 2e − ORR to H 2 O 2 for bare CoN 4 and CoN 4 sites featuring various epoxy oxygen coverages obtained from DFT simulations. Reproduced with permission from [94], Springer Nature, 2020. (f) Division of two distinct substitution sites in the cobalt porphyrin structure; (g) schematic illustration showing the preparation of HMC for H 2 O 2 electrosynthesis from ORR; (h) structure of the investigated cobalt porphyrin molecules. Reproduced with permission from [95], RSC, 2023.

Conclusions and Outlook
Carbon-based SACs with their unique structures and excellent catalytic performance hold great promise as commercial catalysts for electrochemical H 2 O 2 synthesis. The coordination environment of the metal center is a key factor determining its intrinsic catalytic activity and selectivity. By summarizing the recent efforts in regulating the coordination environment for optimizing 2e − ORR's performance, we described the influence of coordination environments on electronic structures, leading to changes in the reaction tendencies. Various relevant descriptors were proposed to qualitatively and quantitatively correlate the differences in electronic structures with catalytic performance, including adsorption energies, metal center oxidation states, spin states, and others. Selecting a universal descriptor from numerous descriptors that can qualitatively determine various performances will save a significant amount of effort and costs in catalyst developments [97]. This review elucidated the core concept that the coordination engineering of SACs alters the d-orbital structure of the metal center through the suitable manipulation of the coordination environment, which in turn affects the binding strength with *OOH, thereby modulating the activity and selectivity of the electrocatalytic ORR. The exploration and discovery of the correlation between structure and performance demonstrated the immense potential of coordination engineering in SACs.
For future development, the research on the coordination engineering of SACs should be pursued from both experimental and theoretical aspects. On the experimental side, special attention should be directed towards hydrogen peroxide synthesis under acidic conditions, given its higher stability and stronger oxidative nature in such environments, which holds significant importance for market applications. In addition to carbon-based supports, exploring other single-atom catalyst substrates, such as metal-based compounds, may present new avenues for the development of SACs. Moreover, developing universal synthesis strategies for SACs with uniformly distributed active sites and the precise control of the coordination environment remains challenging but essential. For example, leveraging the different affinities of metal atoms to various heteroatoms can help achieve specific coordination structures [98,99].
Characterization is also a crucial cornerstone for studying the coordination engineering of SACs. At present, SACs' structural identification relies mainly on X-ray absorption fine structure (XAFS) combined with high-resolution electron microscopy. However, this method has its limitations and may not provide precise microscopic structural information. Therefore, efforts should be made to develop new structural characterization techniques. Additionally, active catalytic sites in ORR reactions are not static but dynamically evolving during the reaction. Hence, advanced in situ characterization techniques, such as in situ XAFS and in situ Fourier-transform infrared spectroscopy [100], should be actively studied to track the dynamic catalytic behavior of active sites during the catalytic reaction, revealing reaction mechanisms to provide more insightful guidance for the development of improved catalysts.
On the theoretical calculation side, from interpreting the experimental results from the past to producing increasingly instructive experiments at present, the advantages of theoretical calculations in terms of their simplicity and low cost are becoming more evident. Using DFT calculations combined with high-throughput screening to predict catalytic performance can save a significant amount of trial-and-error experimental costs. Thus, computational tools should continue to be fully utilized to rationally design the coordination environment of SACs for excellent 2e − ORR performances in upcoming studies. Nevertheless, it is vital to note that theoretical calculations are conducted under simplified and idealized conditions, and they always present differences compared to the actual complex catalytic reactions. Consequently, efforts should be made to develop new theoretical models, considering solvent effects and electrode potential, thus narrowing the gap between theory and experiment to make more accurate predictions or interpretations of the experimental results.
In conclusion, although we achieved certain accomplishments in the coordination engineering of SACs for electrocatalytic H 2 O 2 synthesis, challenges and opportunities coexist. Innovations in both the experimental and computational approaches hold great potential for the further development of SAC coordination engineering and offer promising prospects for the commercial application of electrocatalytic H 2 O 2 production.