Since Jasinski suggested the possibility of metal phthalocyanine as an ORR catalyst, there has been discussion about approaches based on nitrogen-containing macrocyclic metal complexes (Fig. 1) [36]. At first, metal phthalocyanine-promoted carbon black was used as an electrode, however, in an acidic environment, the macrocycles decomposed and the center metals were leached out, showing low stability [37,38]. To achieve stable MNCs, heat treatments of macrocyclic metal complexes with carbon black were proposed [39]. In addition, Gupta et al. separated the N precursor and metal source by pyrolysis of polyacrylonitrile with metal salts, enabling the development of a more flexible synthesis method [40]. Based on these efforts, researches on methodologies of forming active sites using metals and N-containing carbon sources such as pyrrole and phenanthroline were conducted [41–43]. However, unlike the development of the synthesis method, the understanding of the evolution mechanism of the active site of pyrolyzed MNC was still insufficient.

Jaouen et al. suggested that M-N_x active sites are hosted in micropores surrounded by nitrogen by observing changes in the physical and electrochemical properties of the catalyst depending on heat treatment conditions [44]. Then, approaches have mainly been attempted for pyrolysis of metal-chelated composites such as polyaniline-metal complexes or forming active sites with a microporous carbon support [45]. In the latter, the metal-organic framework (MOF), which is one of the most frequently used methods for the synthesis of MNCs, was included [46]. MNC using MOF was first reported in 2010 and 2011, the PEMFC performance containing a ZIF-8-based Fe-N-C cathode was reported to have a peak power density of 0.91 W cm⁻² in H₂-O₂ condition, which accelerated research on MOF-based MNCs [47–49]. Pyrolysis of ZIF-8 and Fe or Co precursors composite fixed Fe or Co ions to the N-hosted micropore in ZIF which was developed by gasification of Zn (b.p. of Zn: 907°C) to form an M-N₄ site [31,50].

However, since the agglomeration of metal particles was inevitable during the heat treatment to form and stabilize the metal-N_x sites, it was attempted to suppress the growth of metal particles by adding inhibitors such as silica and chelators to prevent metal particle agglomeration [51–53]. Nevertheless, the amount of catalytically active metals in MNC was limited to several wt.%, and electrochemical behaviors in a single cell were not yet fully understood [5,33]. Therefore, the performance of MNC in a single cell was inferior to that of a Pt catalyst even though the higher activity of MNCs in half-cell than that of Pt/C is achieved in alkaline conditions. To overcome performance limitations due to low metal content, studies on morphology and active sites of MNCs have been conducted to improve ORR kinetics and mass transport [5]. In the next chapter, first, approaches to control the morphology of the catalyst, which attributes mass transport with hierarchical pore structure for facilitated oxygen and charge transfer, were introduced. Then, it presented recent advances in the design of active sites of MNC to improve reaction kinetics by modulation of the electronic state of the active site with heteroatom introduction and defect engineering, which shifts the adsorption energy between the reaction intermediate and the active site.

First of all, in the pore structure of MNC, micropores, mesopores, and macropores have different roles. Micropores are involved in the formation of an active site because Fe is fixed to a defect surrounded by multiple Ns and forms the M-N_x moiety [44]. Mesopore increases the electrolyte-wettable area of the catalyst. Macropores contribute to mass transport such as O₂ transport [54]. For highly active FeNC, a hierarchical pore structure where all three pores are developed is ideal. The AEMFC performance using MNC catalysts with such characteristics is summarized in Table 1.

Adabi et al. reported the improved performance of AEMFC with commercial FeNC cathode, peak power density of 2.05 W cm⁻² and current density of 7.4 A cm⁻² @0.1 V (H₂/O₂ flow) of AEMFC (Cathode: FeNC 1 mg cm⁻²; Anode PtRu/C 0.6 mg_PtRu cm⁻²) using commercial FeNC with hierarchical pore structure from Varipore^TM method which uses fumed silica as a hard template [55]. It was achieved by optimizing the operating conditions, backpressure, relative humidity, and temperature of the cell to control the water balance of the MEA. The performance of 0.66 V at 1 A cm⁻², exceeding the 2022 U.S. DOE milestone (0.65 V at 1 A cm⁻²), was observed in MEA using a low precious metal anode of 0.125 mg_PtRu cm⁻² meeting U.S. DOE target set in 2020 [4]. It was found that when the temperature of the cell and humidified inlet gas was assigned to a relatively low temperature of 60°C, there was a problem of water accumulation in the anode, resulting in swelling of the anode ionomer. It was suppressed by controlling appropriate temperature and humidity conditions, and backpressure was applied to the cathode to remove water inside the cell.

Designing catalyst morphology to have a hollow structure was considered one of the approaches to increase active site utilization. For this purpose, open-mouthed particles of hollow cobalt/nitrogen-doped carbon with mesoporous shells (OMH-Co/NC) were proposed [69]. Zn and Co sulfate were dissolved in methanol to form a metal sulfate sphere, and then 2-methylimidazole was added to coordinate with the metal sulfate on the surface (Fig. 2a). OMH-Co/NC was obtained by heat-treating the formed raspberry-shaped R-Zn/Co-ZIF (Fig. 2b,c). OMH-Co/NC was confirmed to have a mesopore developed in the form of a hollow porous sphere with an opening at one end (Fig. 2b–e). When ORR activity was evaluated in a half-cell (0.1 M KOH), a half-wave potential of 0.865 V and a limiting current density of 5.7 mA cm⁻² at 1600 rpm were obtained (Fig. 2f). The hollow structure and open hole of OMH-Co/NC contributed to low mass transport resistance and charge transfer resistance (Fig. 2g,h).

Persky et al. reported a bioinspired Fe-Cu porphyrrole aerogel catalyst [70]. The authors explained that the catalyst has a high specific surface area and a hierarchical pore structure, giving it advantages in density and mass transport at the active site. FeCl₂, 5, 10, 15, 20-(tetra-4-aminophenyl) porphyrin, and Cu(II) 5, 10, 15-tris(4-formylphenyl) corrole were chelated to create a gel and then dried using a critical point dryer. Prepared aerogel was carbonized in an inert gas atmosphere at 800°C to prepare HT-FeCu porphyrrole aerogel. HT-FeCu porphyrrole aerogel showed a half-wave potential of 0.80 V and an onset potential of 0.94 V vs RHE in 0.1 M KOH. An onset potential of 0.97 V and a peak power density after iR-correction performance of 0.51 W cm⁻² was reported in AEMFC. Understanding the origin of ORR catalytic activity of HT-FeCu in detail using DFT calculations, adsorption free energy of OH^* can be applied as a descriptor of ORR kinetics. It was reported to have the highest ORR activity when the free energy of OH^* is in the range of 0.7–1.0 eV [42]. Based on this, as a result of calculating the active site of HT-FeCu porphyrrole aerogel, FeCu-porphyrrole would have an OH^* adsorption free energy of 0.796 eV, which was 0.173 eV higher than 0.623 eV of Fe-porphyrin structure.

The particle size of MNC would affect the ORR activities in the aspect of the electrochemically active area. Lee et al. reported that mesoporous FeNC catalysts with different particle sizes would affect the electrochemically active area of catalysts [71]. Mesoporous FeNC catalysts were prepared by a hard-template method using mesoporous silica spheres varying in diameters as 1173, 723, 337, and 151 nm (Fig. 3a,b,d). Mesoporous silica spheres were prepared by modifying Stöber’s method using n-hexa-decylamine as a soft template. The size of silica spheres was controlled by the added amount of ammonium hydroxide solution. Iron chloride and 1, 10-phenanthroline were infiltrated into the pores of silica templates and pyrolyzed at 900^oC. The silica template prevented the agglomeration of iron nanoparticles during pyrolysis and led to the development of the mesopore structure of FeNC. It was concluded that the smaller particle size of the catalyst enabled enhanced electrolyte accessibility and led to improved ORR activities (Fig. 3c,e). It was maximized at the diameter of 337 nm with 0.918 V of half-wave potential and 5.881 mA cm⁻² of limiting current density. Agglomeration of metal particles at the diameter of 151 nm resulted in the hindrance of M-N_x site formation. It indicated that FeNC with smaller particle sizes would have a high electrolyte accessibility, nevertheless, it could be vulnerable to aggregation.

High active site density and porosity of the catalyst are essential for high ORR activity. Chen et al. reported high accessible active site density and hierarchically porous SA-Fe-NHPC by heat treating 2,6- diaminopyridine/ZnFe/SiO₂ complex [72]. The high density of the active site was secured through two pyrolysis processes involving Fe precursor. Silica colloid contributed to mesopore formation, and Zn suppressed the formation of iron carbide. SA-Fe-NHPC created through this has a hierarchical pore structure and a high density of Fe-N₄ sites (Fig. 4a–c). Comparing SA-Fe-NHPC to NP-Fe-NHPC without Zn, it had a higher porosity, and the effect of Zn on pore formation was clear (Fig. 4d,e). This showed high durability, maintaining performance at a high half-wave potential of 0.93 V and an accelerated stress test of 10,000 cycles (Fig. 4f,g).

What determines the kinetics of a catalyst is the adsorption energy between the active site and the reaction intermediate [73]. If the binding is too weak, the intermediate will be desorbed before the reaction proceeds. If the binding is too strong, the intermediate will not be desorbed and the active site will be inactive. Binding energy can be adjusted by adding electron donating or withdrawing chemical species to FeNC. High ORR activity can be achieved when the binding energy is at the proper range.

Tian et al. reported a Fe_H-N-C catalyst with carbon vacancy engineering with superior durability [74]. In synthesizing FeNC using ZIF-8, the zeolitic imidazole framework (ZIF) was etched with tannic acid to remove unstable chemical species, and it induced the formation of carbon defects during pyrolysis (Fig. 5a). The hollow structure of Fe_H-N-C, which was completed by heat treatment in an inert gas atmosphere (Fig. 5b,c). It showed high activity and durability, with the half wave potential decreasing by about 30 mV from 0.91 V to 0.88 V even after a durability test of 100,000 cycles in O₂ purged 0.1 M KOH electrolyte solution, which was unprecedented durability for MNC catalysts (Fig. 5d,e). As a result of computations, high stability of the active site was expected because the Fe dissolution energy between Fe and N in Fe-N₄ located at the defective carbon of Fe_H-N-C is higher than those of pristine FeNC (Fig. 5f).

Jiang et al. formed binary single-atomic Fe/Co-N_x sites in a porous shell through interfacial assembly (FeCo-NCH) [60]. FeCo-NCH was prepared by pyrolyzing Fe and Co salts loaded ZIF-8 with poly-dopamine coating (Fig. 6a). As a result, FeCo-NCH was synthesized in the form of porous hollow particles containing C, N, Fe, and Co (Fig. 6b). Both Fe and Co have a coordination environment similar to metal phthalocyanine in a metal-N coordinated form rather than nanoparticles, and it was confirmed that they existed as the Fe-N₅ and Co-N₄ sites through EXAFS analysis (Fig. 6c). These active sites were highly accessible from oxygen and electrolyte and have a high half-wave potential of 0.89 V at 0.1 M KOH because the Fe/Co-N_x site has favorable electronic modulation for ORR (Fig. 6d). When applied as an AEMFC cathode, it showed a high peak power density of 750.7 mW cm⁻², showing the synergistic effect of a highly accessible active site with high ORR kinetics (Fig. 6e). Liu et al. reported that in the case of a highly coordinated structure in which the axial sulfur atom was located in the first shell and the phosphorus was located in the second shell, the adsorption energy of oxygen intermediates was weakened by down-shift of the d-band center [75]. The addition of the 4, 4’-sulfonyldiphenol, and phos-phonitrilic chloride trimer to the ZIF-8 synthesis procedure led to the introduction of S and P into the ZIF-8. The Fe-N₄ site located in the hollow carbon nanocage containing N, P, and S showed high half-wave potentials of 0.912 and 0.814 V in alkaline and acidic media, respectively.

The electronic state of the active site can be adjusted by introducing other metals or heteroatoms. FeCuNC was synthesized using grain-shaped SBA-15, an ordered mesoporous silica template by carbonization of FeCl₃, CuSO₄, 1,10-phenanthroline, citric acid, and phosphoric acid impregnated SBA-15 (Fig. 7a) [61]. The use of the SBA-15 template was advantageous for mesopore structure and active site formation by limitation of the Fe agglomeration (Fig. 7b). Cu, S, and P shown in EDS mapping contributed to the positive shift of oxidation state of the active site despite being located above the second shell of the Fe-N₃C₁ site (Fig. 7c,d). The coordination environment of Cu was similar to that of Fe as Cu-N₃C₁, however, the shift of the oxidation state of Cu due to heteroatoms was negligible. Additionally, XPS and XAS revealed that S and P did not have a direct interaction with metals. Thus, it was concluded that heteroatom doping in the carbon matrix could shift the electronic state of Fe even located at the second or higher shell of the active metal. ORR activity was 0.92 V at −3 mA cm⁻² in half-cell (0.1 M KOH) and current density was 0.490 A cm⁻² at 0.6 V in AEMFC single cell (Fig. 7e). DFT calculations on catalyst model based on the results of XAS suggested that introduction of Cu, S, and P into FeNC improved ORR kinetics controlling the binding energy of OH^* and OOH^*, reaction intermediates (Fig. 7f).