
==== Front
Natl Sci Rev
Natl Sci Rev
nsr
National Science Review
2095-5138
2053-714X
Oxford University Press

10.1093/nsr/nwae193
nwae193
RESEARCH ARTICLE
MATERIALS SCIENCE
Special Topic: Functional and Smart Fibers
Nsr/4
AcademicSubjects/MED00010
AcademicSubjects/SCI00010
Fe-N co-doped carbon nanofibers with Fe3C decoration for water activation induced oxygen reduction reaction
https://orcid.org/0009-0009-6034-1345
Li Shaoxiong College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

Xing Gengyu College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

Zhao Sheng College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

Peng Jian Institute for Superconducting and Electronic Materials Australian Institute for Innovative Materials, University of Wollongong Innovation Campus, North Wollongong, NSW 2522, Australia

Zhao Lingfei Institute for Superconducting and Electronic Materials Australian Institute for Innovative Materials, University of Wollongong Innovation Campus, North Wollongong, NSW 2522, Australia

Hu Feng College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

Li Linlin College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

Wang Jiazhao Institute for Superconducting and Electronic Materials Australian Institute for Innovative Materials, University of Wollongong Innovation Campus, North Wollongong, NSW 2522, Australia

Ramakrishna Seeram Department of Mechanical Engineering, National University of Singapore, Singapore  117583, Singapore

https://orcid.org/0000-0003-1591-1301
Peng Shengjie College of Materials Science and Technology, Nanjing University of Aeronautics and Astronautics, Nanjing 210016, China

Corresponding author. E-mail: pengshengjie@nuaa.edu.cn
10 2024
04 6 2024
04 6 2024
11 10 nwae19307 5 2024
25 5 2024
31 5 2024
01 7 2024
© The Author(s) 2024. Published by Oxford University Press on behalf of China Science Publishing & Media Ltd.
2024
https://creativecommons.org/licenses/by/4.0/ This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.

ABSTRACT

Proton activity at the electrified interface is central to the kinetics of proton-coupled electron transfer (PCET) reactions in electrocatalytic oxygen reduction reaction (ORR). Here, we construct an efficient Fe3C water activation site in Fe-N co-doped carbon nanofibers (Fe3C-Fe1/CNT) using an electrospinning-pyrolysis-etching strategy to improve interfacial hydrogen bonding interactions with oxygen intermediates during ORR. In situ Fourier transform infrared spectroscopy and density functional theory studies identified delocalized electrons as key to water activation kinetics. Specifically, the strong electronic perturbation of the Fe–N4 sites by Fe3C disrupts the symmetric electron density distribution, allowing more free electrons to activate the dissociation of interfacial water, thereby promoting hydrogen bond formation. This process ultimately controls the PCET kinetics for enhanced ORR. The Fe3C-Fe1/CNT catalyst demonstrates a half-wave potential of 0.83 V in acidic media and 0.91 V in alkaline media, along with strong performance in H2-O2 fuel cells and Al-air batteries.

Delocalized electrons driving dissociation of interfacial water promotes hydrogen bonding, thus improving the oxygen reduction kinetics and enabling Fe3C-Fe1/CNT nanofibers to exhibit excellent performance in H2-O2 fuel cells and Al-air batteries.

electrospinning
nanofibers
Fe–N–C
oxygen reduction reaction
Al-air batteries
National Natural Science Foundation of China 10.13039/501100001809 22075141 22101132 Scientific and Technological Innovation Special Fund for Carbon Peak Carbon Neutrality of Jiangsu Province BK20220039
==== Body
pmcINTRODUCTION

Exploring new generation energy storage and conversion technology is crucial for the sustainable development of human society [1]. Proton exchange membrane fuel cells (PEMFCs) and metal-air batteries, known for their high energy efficiency and environmentally friendly features, hold great potential for sustainable energy applications [2]. Oxygen reduction reaction (ORR) electrocatalysts are essential for the development of these technologies, such as metal-air batteries and PEMFCs [3–6]. However, the ORR exhibits slow kinetics and often relies on costly and rare platinum group metals (PGMs) as catalysts, which poses a significant bottleneck due to their high cost and limited availability, hindering large-scale application [7–9]. Consequently, there has been considerable interest in developing precious-metal–free catalysts as alternatives to expensive PGM-based ORR electrocatalysts [10–14]. Transition metal/nitrogen co-doped carbon-based materials (M–N–C) have garnered particular attention as emerging non–precious-metal catalysts due to their high intrinsic activity, well-defined active sites, and good stability [15–18]. Specifically, Fe–N–C catalysts, where iron is coordinated with nitrogen in an Fe–N4 configuration, demonstrate Pt-like behavior in O2 adsorption and the subsequent O=O bond breaking during the ORR catalytic process [19–21]. Nevertheless, the unsatisfactory adsorption energy of Fe–N4 sites for oxygen intermediates has led to limited intrinsic activity and severe degradation, especially in acidic media [2,22].

Various strategies to enhance the intrinsic activity of Fe-based catalysts include doping heteroatoms (such as S, P, etc.) into the carbon matrix, introducing neighboring M–N4 sites (where M = Co, Ni, Mn, etc.), and constructing axial coordination groups at the metal active center (such as O, OH, Cl, etc.) [23–27]. Particularly, transition metal compounds modified with atomically dispersed Fe–N4 catalysts are considered promising candidates to improve ORR performance and stability in acidic media [20]. However, the modulation of the intrinsic activity of Fe–N4 sites in acidic media by transition metal compounds remains poorly understood. For instance, the critical impact of water activation on the mechanism, thermodynamics, and kinetics of electrocatalytic reactions has been largely overlooked. Additionally, the strong electric field polarization effect under acidic conditions aligns interfacial water molecules around the metal center into an O-down configuration [28]. This alignment disrupts hydrogen bonding between oxygenated intermediates and interfacial water molecules, impeding proton transfer in the proton-coupled electron transfer (PCET) step and ultimately slowing ORR reaction kinetics. Promoting hydrogen bond formation between oxygen-containing intermediates and interfacial water by constructing efficient water activation catalytic sites is an effective strategy to accelerate the PCET process [29]. The role of platinum species in activating water for various electrocatalytic processes in aqueous systems is well-documented [30,31]. Thus, it is hypothesized that Fe3C, with a Pt-like electronic configuration, could serve as an efficient catalytic site for water activation in proton generation from water molecules, complementing the traditional Fe–N4 site to enhance ORR activity. Moreover, electrospinning has emerged as a fascinating and scalable technique for fabricating energy-functional composite nanofibers, enabling the incorporation of both single atomic and metallic compounds [32].

In this work, we successfully fabricated the Fe3C-modified Fe–N4 enriched with a carbon nanotube system (Fe3C-Fe1/CNT) on carbon nanofibers to achieve efficient and stable ORR using a unique electrospinning technology. The obtained catalyst features inter-crosslinked CNT and a three-dimensional (3D) porous network structure, which enhances active site exposure and facilitates the movement of reactants and products into and out of the active sites, as well as electron transport throughout the structural network. Additionally, the strong electronic perturbation of the Fe–N4 active center by Fe3C leads to d-orbital electron rearrangement, optimizing the binding energy of intermediates in the rate-determining step. Notably, the Fe3C-Fe1/CNT exhibits a half-wave potential of up to 0.83 V in acidic media and 0.91 V in alkaline media. Moreover, delocalized electrons can drive the activation of interfacial water, accelerating the entire PCET process. As a result, the Fe3C-Fe1/CNT demonstrates a high power density of 724 mW cm–2 in H2-O2 fuel cells and a discharge performance of 1.65 V at 1 mA cm–2 in Al-air batteries. This study provides guidance for the rational design of highly active electrospun M–N–C electrocatalysts.

RESULTS AND DISCUSSION

Design, synthesis and characterization

The fabrication approach of Fe3C-Fe1/CNT is described in Fig. 1a. Typically, the precursor solution was electrospun into a three-dimensional (3D) membrane composed of Fe-embedded 1D carbon nanofibers (Fe-CNF, Fig. S1 in the online Supplementary data). After that, pre-oxidation was performed to stabilize the microstructure and avoid the fusion of the carbon fiber in the subsequent carbonization process. Finally, thermal treatments and acid etching were implemented to form Fe3C nanoparticles (NPs) and atomically dispersed Fe-anchored N-doped CNT. Incorporating CNT roughened the carbon nanofiber surface, enhancing active site exposure and facilitating ORR electron/ion transport, while melamine-free heat treatment yielded a smooth surface devoid of CNT structures (Fig. 1b and c, Fig. S2) [33]. The transmission electron microscopy (TEM) images in Fig. 1d and e further confirm the graphite-coated Fe3C NPs loading on carbon fiber intimate contact with the bamboo-like CNT structures in Fe3C-Fe1/CNT. No NPs are observed for Fe1/CNT revealing that Fe3C NPs are removed after the HNO3 acid leaching and the Fe species present in an atomically dispersed state (Fig. S3). Aberration-corrected high-angle annular dark-field STEM (AC-HAADF-STEM) is suitable for further detecting the status of metal species. As shown in Fig. 1f, the carbon substrate contains both Fe3C NPs and Fe atomic sites. In addition, the corresponding energy-dispersive X-ray spectroscopy (EDS) shows that the Fe, N, and C elements are distributed uniformly in Fe3C-Fe1/CNT nanofibers (Fig. 1g). The above results demonstrate that the electrospinning process, thermal treatment, and acid etching route successfully produce the CNT structure, in which Fe3C NPs and atomically dispersed Fe sites co-exist in the carbon matrix.

Figure 1. Synthesis and morphological characterization of catalysts. (a) Schematic of the synthesis process of Fe3C-Fe1/CNT and Fe1/CNT. (b–c) SEM images of Fe3C-Fe1/CNT. (d) TEM and (e) HRTEM images of Fe3C-Fe1/CNT. (f) AC-HAADF-STEM image of Fe3C-Fe1/CNT and (g) the corresponding EDS maps of Fe, N, C.

The catalysts are analyzed using X-ray diffraction (XRD) patterns to determine their chemical makeup and crystal structure. As illustrated in Fig. 2a, the typical Fe3C-Fe1/CNT diffraction peaks at around 26.6° are attributed to the graphitic carbon structures’ (002) plane. Insignificant Fe3C-related diffraction peaks are seen from 42–46° (JCPDS No. 75–910), indicating that a significant portion of the Fe-containing species was removed after H2SO4 treatment. After further treatment with HNO3, the Fe3C peaks vanish, leaving just the typical graphitic carbon peak in Fe1/CNT [34], which suggests the graphitic carbon is well retained, whereas the NPs are being leached off. For the other control group, distinct graphitized carbon peaks are observed in Fe3C-Fe1/CNT compared to Fe3C-Fe1, indicating that the formation of CNT enhances electron transport and corrosion resistance. (Fig. S4). The D-band peak for lattice defects at 1345 cm−1 and the G-band peak for sp2-hybridized carbon at 1590 cm−1 are present for all samples observed by Raman spectroscopy and are analyzed in Fig. 2b. The D-band and G-band peak intensity ratios of Fe3C-Fe1/CNT (ID/IG ≈1.145) are larger than those of Fe1/CNT (ID/IG ≈1.016). This demonstrates that the carbon matrix of Fe3C-Fe1/CNT contains many more carbon nanostructures with defects, which can cooperate with Fe–N–C active centers and enhances electronic conductivity, modulates local electron density, facilitates oxygen intermediate adsorption, and improves electron transfer kinetics, thereby significantly boosting overall catalytic efficiency and performance. The Brunauer–Emmett–Teller (BET) specific surface area of Fe3C-Fe1/CNT is measured by isothermal N2 adsorption-desorption analysis (Fig. 2c). The specific surface area of Fe3C-Fe1/CNT is 399.7 m2 g–1, higher than that of Fe1/CNT (363.4 m2 g–1). Fe3C-Fe1/CNT shows many mesopores during pyrolysis due to the presence of Zn atoms, facilitating the penetration and transport of reaction products during electrocatalysis. X-ray photoelectron spectroscopy (XPS) is per performed to investigate the electron and coordination structure of catalysts (Fig. 2d–f, Fig. S5). The C–N in Fe3C-Fe1/CNT experiences a decrease in binding energy, suggesting that the electronic structure of the Fe–N4 symmetry is disrupted and transferable to the carbon substrate. The peak at 706.8 eV is ascribed to Fe with mixed valence states suggesting the existence of Fe3C, which may activate the carbon layer outer surface to promote its involvement in electrocatalysis [35]. In comparison with the Fe1/CNT, the Fe–N species in Fe3C-Fe1/CNT show a shift toward lower binding energy. These results evidence that strong electronic perturbation of Fe3C NPs lead to an asymmetric distribution of Fe–N4 electron density. Significantly, Fe3C-Fe1/CNT exhibits larger contents of pyridinic-N, Fe–N, and graphitic-N species (Table S1). Pyridinic-N may lower the energy barrier for O2 to adsorb on nearby carbon atoms, which can behave as an efficient active site to speed up the generation of oxygen-containing intermediates in the ORR process [36,37].

Figure 2. Fine structure characterization of the active sites. (a) XRD patterns. (b) Raman spectra. (c) BET spectra. XPS spectra of (d) C 1s, (e) Fe 2p, (f) N 1s. (g) Normalized Fe K-edge XANES spectra. (h) Fourier transform k3-weighted Fe K-edge EXAFS spectra. (i) Wavelet transform of Fe3C-Fe1/CNT and Fe1/CNT.

Local electronic changes can be further observed in the fine structure of X-ray absorption (XAFS). The Fe3C-Fe1/CNT absorption edge falls between the absorption edges of FePc and Fe2O3, indicating that the oxidation state of Fe lies in the range of +2 to +3, that is, Fe2+ and Fe3+ coexist in the Fe3C-Fe1/CNT catalyst (Fig. 2g). In Fe–N–C catalysts, only the Fe3+–N4 structure is generally obtained [38]. One of the main reasons here is that during the material synthesis process, Fe2+ in the iron source is easily oxidized to Fe3+ to form the Fe3+–N4 structure [39]. In fact, the Fe2+–N4 structure is much more catalytically active than the Fe3+–N4 structure for ORR [40]. Furthermore, Fe3C-Fe1/CNT displays a negative absorption edge compared to Fe1/CNT, suggesting that the Fe center has a reduced valence and a larger electron density. The presence of more electronegative groups like Fe3C has an impact on the active center of Fe1/CNT, according to the positive absorption edge compared to FePc. Thus, the introduction of Fe3C leads to the conversion of Fe3+ to Fe2+ and the asymmetric distribution of Fe–N4 electron density, which is consistent with the XPS analysis. The Fe K-edge Fourier-transformed extended XAFS (FT-EXAFS) spectrum provides significant structural information (Fig. S6, Table S2). As shown in Fig. 2h, the peak position contributed to Fe–N of Fe3C-Fe1/CNT in R-space is significantly negatively shifted, which implies a shortening of the Fe–N bond length and a change in the local structure of the Fe–N4 active sites [41,42]. The Fe–Fe bonds are represented by peaks in Fe3C-Fe1/CNT at 2.2 Å, which is consistent with the presence of Fe3C species [43]. Wavelet-transform EXAFS (WT-EXAFS) of the corresponding catalysts are performed (Fig. 2i, Fig. S7). The WT-EXAFS analyses of Fe1/CNT show the greatest intensity at ∼1.51 Å and 3.57 Å−1, confirming the coordination structure of Fe–N. In contrast, Fe3C-Fe1/CNT show maximum peak intensity at shorter 1.41 Å and longer 4.88 Å–1, suggesting the superposition of the Fe–N bond and Fe–C bond. Combining XPS and XANES/EXAFS analyses, we successfully obtained a composite structure of Fe3C/Fe–N4 with an asymmetric distribution of the electron density of Fe–N4 around the Fe active center.

ORR performance evaluation of Fe3C-Fe1/CNT

The electrocatalytic activity of each resulting catalyst toward the ORR was assessed utilizing a conventional three-electrode rotating ring-disk electrode (RRDE) system in the O2-saturated 0.5 M H2SO4. The cyclic voltammetry (CV) curves unmistakably display a reduction peak, while the N2-saturated solution lacks a matching peak (Fig. S8), proving the electrochemical response of Fe3C-Fe1/CNT toward the ORR. The linear sweep voltammetry (LSV) tests are performed to evaluate the catalysts ORR activity (Fig. 3a). The ultimate Fe3C-Fe1/CNT exhibits a Tafel slope of 75.9 mV dec−1, which is similar to that of Pt/C and smaller than Fe1/CNT (Fig. 3b). Fe3C-Fe1/CNT exhibits an onset potential (Eonset) of 0.93 V and a half-wave potential (E1/2) of 0.83 V, which is preferable to Fe1/CNT, and somewhat lesser activity than Pt/C catalyst (Fig. 3c). In addition, the resulting Fe3C-Fe1/CNT ORR activity is substantially better than most of the previously reported M–N–C catalysts (Table S3). The Koutecky-Levich (K–L) equations might be used to determine the number of electrons transported (n) (Fig. 3d, Fig. S9). The n value for Fe3C-Fe1/CNT is determined to be 3.93 according to the linearity of the K–L plots in the potential of 0.3 V, 0.4 V, 0.5 V. Compared to Fe1/CNT, Fe3C-Fe1/CNT exhibits a higher double-layer capacitance, which is correlated with its electrochemical surface area (Fig. 3e, Fig. S10), leading to improved electrochemical performance. Additionally, no current degradation is seen after adding methanol to the electrolyte, which displays remarkable methanol resistance (Fig. 3f). As shown in Fig. 3g, the H2O2 yields stay below 5%, indicating a 4-electron ORR route over the Fe3C-Fe1/CNT. After 8000 CV cycles, the E1/2 of Fe3C-Fe1/CNT only negatively shifts 39 mV, whereas the 85 mV negative shifts for Pt/C (Fig. 3h). As a result, Fe3C NPs adorned Fe–N4 with higher electron cloud density are more successful than the Fe–N4 sites in lowering the ORR reaction energy barrier, which results in a lower overpotential (Fig. 3i).

Figure 3. Electrocatalytic activity evaluation. (a) LSV curves of Fe3C-Fe1/CNT, Fe1/CNT, and 20% Pt/C in 0.5 M H2SO4. (b) Tafel slopes for the corresponding catalysts. (c) Eonset and E1/2. (d) ORR polarization curves of Fe3C-Fe1/CNT at different rotating sweeps. Inset: the fitted K-L plots and electron transfer numbers. (e) Plots of current densities (at 0.04 V vs. RHE) as functions of scan rates. (f) Methanol tolerance of the catalysts in 0.5 M H2SO4. (g) H2O2 yield and electron transfer number of the catalysts. (h) Stability tests. (i) Schematic representation of ORR properties versus oxidation state.

Fe3C-Fe1/CNT not only performs well in acid solution but also exhibits remarkable ORR performance in 0.1 M KOH solution. Fe3C-Fe1/CNT with an E1/2 of 0.91 V outperforms other catalysts in terms of alkaline ORR activity (Figs S11 and S12), and the majority of M–N–C electrocatalysts described thus far (Table S4). In addition, as shown in Figs S13–S15, the Fe3C-Fe1/CNT exhibits high selectivity for 4-electron ORR, the excellent stability and resistance to methanol. Fe3C-Fe1/CNT was also used as the Al-air battery cathode catalyst given the optimized electrocatalytic activity for alkaline ORR. The solid electrolyte with a polyacrylic acid (PAA) base facilitates flexibility while concurrently offering sufficient mechanical support (Fig. S16a). The Fe3C-Fe1/CNT based Al-air battery can sustain a discharge voltage of 1.65 V for 28 hours at a discharge current density of 1 mA cm–2, which is higher than that of Pt/C, Fe1/CNT and the majority of solid-state Al-air batteries that have been previously documented (Fig. S16b, Table S5). Fe3C-Fe1/CNT produces a greater voltage than Fe1/CNT, particularly during elevated discharge current densities, as demonstrated by the rate performance measurement (Fig. S16c). Due to the electron transfer process occurring more quickly, discharge polarization shows that Fe3C-Fe1/CNT may accomplish a peak power density of 44 mW cm–2, which is higher than Fe1/CNT (26 mW cm–2), Pt/C (28 mW cm–2), and outperform most of their cathode catalyst counterparts’ all-solid-state Al-air batteries (Fig. S16d–f).

PEMFC performance evaluation of Fe3C-Fe1/CNT

To further explore the application potential of Fe3C-Fe1/CNT, we assemble a PEMFC to assess efficacy and stability (Fig. 4a, Fig. S17). With Fe3C-Fe1/CNT catalyst loading of 4 mg cm−2 shows an open-circuit voltage of 0.91 V, close to the 0.94 V of the Pt/C-based cell at the usual 1.0 bar H2/O2 at 80°C. Additionally, the PEMFC could produce current densities of 1.39 A cm−2 at 0.5 V and 2.40 A cm−2 at 0.2 V (Fig. 4b). The Fe3C-Fe1/CNT-assembled PEMFC exhibits high performance with a peak power density of 716 mW cm−2, significantly higher than that of Fe1/CNT (517 mW cm−2). Despite showing comparable ORR-catalytic activity on RRDE, Fe3C-Fe1/CNT and Pt/C catalysts exhibit contrasting behaviors in PEMFC. Fe3C particle sintering may disrupt Nafion-ionomer distribution and diminish the effectiveness of the cathode's three-phase reaction interface. Variations in catalyst structure, including Fe loading and pore size distribution, contribute to noticeable current density differences between Fe3C-Fe1/CNT and Fe1/CNT. It is evident that when metal loading rises, peak power density increases initially and then declines, reaching its maximum value at 4 mg cm−2 (Fig. 4c). A suitable loading could ensure that electrocatalytic activity develops sufficiently since high loading results in an increasingly thick catalyst layer that may obstruct mass transportation and charge transmission. Polarization plots before and after an aggressive square-wave accelerated durability test (ADT) further documented cell voltage loss (Fig. 4d). Once more, the Fe1/CNT assembled PEMFC exhibits the greatest performance deterioration, with a cell voltage loss of 296 mV, which is significantly higher than that of Fe3C-Fe1/CNT (74 mV loss). Significantly, Fe3C-Fe1/CNT based PEMFC could discharge steadily for 100 hours at 0.6 V, demonstrating the well-durability (Fig. 4e). As shown in Fig. 4f and Table S6, the Fe3C-Fe1/CNT exhibited excellent stability and competitive activity levels compared with previously reported Fe–N–C catalysts. No significant changes in the valence states of the corresponding elements were observed in Fe3C-Fe1/CNT after ADT, as determined by XPS analysis (Fig. S18). Furthermore, the morphology was well maintained and no particle aggregation was detected, and the uniform distribution of Fe3C nanoparticles and Fe single atoms was well-preserved, as evidenced by the AC-HAADF-STEM image obtained post-ADT (Fig. S19), confirming the retention of the highly stable ordered structure. The foregoing superior performance and durability qualities point to the Fe3C-Fe1/CNT potential for application in fuel cells.

Figure 4. H2-O2 PEMFC performance. (a) Structure illustration of the H2-O2 fuel cell. (b) Polarization and power density curves by using the catalysts as cathodes. (c) Comparison of different loads of Fe3C-Fe1/CNT. (d–e) Stability tests. (f) Comparison of PEMFC performance with reported advanced catalysts.

The origin of enhanced activity of Fe3C-Fe1/CNT

To verify our conjecture, in situ Fourier transform infrared (FTIR) spectroscopy measurements were used to probe the interfacial water structure and further explore the water activation mechanism. As shown in Fig. 5a and b, O−H stretching modes (∼2800–3600 cm−1) and H−O−H bending modes (∼1600–1700 cm−1) of interfacial water can be observed, which provides clear evidence for the involvement of surface H2O in the ORR. The H−O−H and O−H stretching strength at each potential in Fe3C-Fe1/CNT is much larger than Fe1/CNT, indicating that the Fe3C sites have a significant affinity for H2O activation and form interfacial hydrogen bonds (Fig. 5d) [44]. More importantly, the O−H stretching peaks at each potential in Fe3C-Fe1/CNT are mainly located near 3330 cm−1 redshifted by 50 cm−1 compared with Fe1/CNT (∼3380 cm−1), suggesting a strong hydrogen-bonding interaction between the interfacial water molecules and the ORR oxygen intermediates [29]. The H2O bending vibration has a higher wave number on Fe3C-Fe1/CNT than on Fe1/CNT, showing that the presence of Fe3C strengthens interfacial hydrogen bonding (Fig. 5e) [45]. The in-situ experimental results confirmed the mechanism of water activation (Fig. 5c), which was further investigated in-depth through density functional theory (DFT) calculations. The transition state (TS) formation on Fe3C/Fe–N4 and Fe–N4 in the hydrolysis and dissociation process is endothermic, confirming it as the rate-determining step (RDS). In this RDS, the ΔG on Fe3C/Fe–N4 is 3.29 eV, significantly smaller than that on Fe–N4 at 4.88 eV (Fig. 5f). The remarkably low energy barriers in RDS explain the outstanding activity of Fe3C/Fe–N4 in water activation and the facile dissociation of water into protons for the ORR. Further insights into the water interface mechanism were gained through theoretical calculations, including evaluations of the electron localization function (ELF) to measure excess kinetic energy density due to Pauli repulsion [46]. As a result, compared to Fe–N4, Fe3C/Fe–N4 shows enhanced polarization in the O−H bonding of adsorbed H2O, illustrating the regulatory influence of Fe3C on water molecule bonding states, and suggesting that strong polarity is advantageous for hydrogen dissociation (Fig. 5g). Specifically, O2 preferentially binds to the Fe−N4 sites to form Fe−N4-O2. The spontaneous dissociation of O2 molecules is aided by interfacial water hydrogen bonding rather than just the presence of Fe−N4 sites. H2O preferentially binds to the Fe3C site to form Fe3C-OH2 and generates OH− by 1e– reduction. Surface protons can be transferred from Fe3C-OH2 to neighboring Fe−N4-O2, leading to the formation of Fe3C-OH and Fe−N4-OOH, which subsequently undergoes 1e– reduction to produce Fe−N4=O and H2O. Then Fe−N4=O undergoes the same PCET process and returns to the initial state (Fig. 5h). Meanwhile, we observed similar intermediate products in alkaline media, which highlights the pH-universal water activation induced ORR mechanism for our catalyst (Fig. S20). The extraordinary feature of this mechanism includes the interfacial activated water by the turnover of Fe3C-OH/Fe3C-OH2 and the surface proton transfer between the proximate Fe3C and Fe−N4 sites.

Figure 5. In-situ characterization and reaction mechanism. (a–b) In situ FTIR spectra of the Fe3C-Fe1/CNT and Fe1/CNT catalysts for ORR. (c) Scheme of the structure-activity-potential relationship. (d) The detailed variations of H−O−H and O−H with potential. (e) The variations of wavenumber with potential. (f) The energy configuration of water activation on the catalysts. (g) ELF evaluations for H2O adsorption on the Fe site of the catalysts. (h) ORR mechanism of water activation for Fe3C-Fe1/CNT.

DFT calculations

In the Fe3C/Fe–N4 model, O2 can be readily adsorbed as well as activated, and the O–O bond has the largest stretching length (Fig. 6a), which implies that the O–O bond is relatively the easiest to break resulting in higher ORR activity [47]. Supported by the above analysis, the optimized models of ORR on Fe3C/Fe–N4, Fe–N4, and Fe3C are established with the free energy ladder images (Fig. 6b, Fig. S21). For the Fe–N4 and Fe3C models, since the production of *OOH and *OH intermediate is endothermic in contrast to the other stages, the development and breakdown of them constitute the RDS of ORR. For the Fe3C/Fe–N4 model, the free energy values for the *OOH step are spontaneous, implying that from O2 to *OOH is more advantageous in terms of thermodynamics. Although the significant adsorption of *OH intermediates on Fe3C/Fe–N4 model, the energy barrier of *OH on Fe3C/Fe–N4 (0.81 eV) is lower than Fe–N4 (1.41 eV) and Fe3C (1.18 eV), which is essential for speeding up the ORR activity. The analysis of the partial density of states (DOS) for O2 molecules adsorbed in various models may be employed to assess the degree of binding between reactants and metal centers. The Fe3C/Fe–N4 model exhibits a Fe d-band center energy of −2.92 eV, compared to the Fe–N4 (−1.37 eV), indicating weaker interaction and easier dissociation of intermediates (Fig. 6c, Fig. S22). This strong electronic interaction between Fe3C species and Fe–N4 sites leads to down-shifted d-band centers in catalysts, moderating bonding strength of oxygen-related intermediates and enhancing ORR activities.

Figure 6. Theoretical calculation of the catalytic activity. (a) Adsorption energies for H2O and O2. (b) Free energy diagram of ORR steps on Fe3C/Fe–N4 and Fe–N4. (c) DOS of Fe3C/Fe–N4 and Fe–N4. (d) COHP analysis of Fe−O bond after *OH adsorption for Fe3C/Fe–N4. (e) Schematic representation of the regulatory effect of Fe3C toward ΔG*OH on Fe−N4 sites. The red and green zones represent the charge accumulation and dispersion, respectively.

To elucidate active site mechanisms, the projected density of states (PDOS) of Fe 3d orbitals (dxz, dxy, dyz, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} ${{d}_{{{{\mathrm{x}}}^{\mathrm{2}}}{\mathrm{ - }}{{{\mathrm{y}}}^{\mathrm{2}}}}}$\end{document}, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} ${{d}_{{{{\mathrm{z}}}^{\mathrm{2}}}}}$\end{document}) is shown in Fig. S23. Most of the electron orbitals involved in non-homogeneous catalysis concentrate near the Fermi level. In Fe3C/Fe–N4, \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} ${{d}_{{{{\mathrm{x}}}^{\mathrm{2}}}{\mathrm{ - }}{{{\mathrm{y}}}^{\mathrm{2}}}}}$\end{document} is minimally influential, while non-localized dxz, dyz, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} ${{d}_{{{{\mathrm{z}}}^{\mathrm{2}}}}}$\end{document} orbitals show homogeneous Fermi level distribution, underscoring their role in catalytic efficiency. To identify the reason for the high intrinsic catalytic activity center, the dxz-p, dyz-p, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} ${{d}_{{{{\mathrm{z}}}^{\mathrm{2}}}}}$\end{document}-p were further computed using projected crystal orbital Hamilton population (pCOHP). For Fe–N4, significant localized electronic effects are observed in the dyz-p and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} ${{d}_{{{{\mathrm{z}}}^{\mathrm{2}}}}}$\end{document}-p bonding orbitals and spin-down antibonding orbitals, respectively, and exhibit unusually high intensity (Fig. S24). As a result, the intermediations are more affected by the very localized enhancing interaction, which makes it challenging to dissociate. On the contrary, Fe3C/Fe–N4 could observe significant electron delocalization. As shown in Fig. 6d, the dxz-p, dyz-p, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} ${{d}_{{{{\mathrm{z}}}^{\mathrm{2}}}}}$\end{document}-p have non-localized enhancement interactions in spin up or spin down, meaning that the *O is easily formed into *OH by the moderately intense localized enhancement interaction acting on it. Consequently, a greater limiting potential (0.93 V) is observed for Fe3C-Fe1/CNT in experimental results. The Fe3C NPs altered the electronic arrangement of the Fe center orbitals on the Fe−N4 site, thereby disrupting the symmetric electronic structure of the Fe−N4 site and significantly altering the degree of electronic delocalization of the spin up and spin down in the dxz-p, dyz-p, and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{upgreek} \usepackage{mathrsfs} \setlength{\oddsidemargin}{-69pt} \begin{document} ${{d}_{{{{\mathrm{z}}}^{\mathrm{2}}}}}$\end{document}-p orbitals enhancing the Fe−O bonding strength (Fig. 6e). Thus, based on in situ FTIR as well as DFT, results indicate that the increased activity at ORR-related potentials is mainly dependent on the dissociation of interfacial water induced by delocalized electrons, thereby strengthening the hydrogen bonding interactions of interfacial water molecules with ORR oxygen intermediates.

CONCLUSION

In summary, we successfully constructed Fe3C-modified, atomically dispersed Fe-N co-doped CNT anchored on nanofibers using an electrospinning-pyrolysis-etching route. The strong electronic perturbation of the Fe−N4 active center by Fe3C leads to electronic rearrangement in the Fe center orbitals, enhancing electron delocalization in the d orbitals and achieving an asymmetric electron density distribution at the Fe−N4 sites. Delocalized electrons can activate the dissociation of water molecules and accelerate PCET processes by forming hydrogen bonds with surface oxygen intermediates. The Fe3C-Fe1/CNT catalyst demonstrated a high power density of 716 mW cm–2 in H2-O2 fuel cells and a high discharge performance of 1.65 V at 1 mA cm–2 in Al-air batteries. This strategy may inspire the development of more efficient single-atom electrocatalysts, providing a new and insightful perspective on improving the inherent catalytic performance of electrospun M−N−C catalysts.

Supplementary Material

nwae193_Supplemental_File

ACKNOWLEDGEMENT

The authors thank the Synchrotron Radiation Research Center, Hsinchu, for conducting XAFS measurements.

FUNDING

This work was supported by the National Natural Science Foundation of China (22075141 and 22101132), and the Scientific and Technological Innovation Special Fund for Carbon Peak and Carbon Neutrality of Jiangsu Province (BK20220039).

AUTHOR CONTRIBUTIONS

S.P. conceived the idea, wrote the paper, supervised the entire project and is responsible for the infrastructure and project direction. S.L., G.X. and S.Z. performed the experiments, and collected and analysed the data. J.P., L.Z. and J.W. performed the AC-TEM measurements. L.L., F.H. and S.R. supervised the whole experimental procedure and co-wrote the paper. All authors discussed the results and commented on and revised the manuscript.

Conflict of interest statement. None declared.
==== Refs
REFERENCES

1. Xie  H, Xie  X, Hu  G  et al.  Ta–TiOx nanoparticles as radical scavengers to improve the durability of Fe–N–C oxygen reduction catalysts. Nat Energy  2022; 7 : 281–9.10.1038/s41560-022-00988-w
2. Yuan  K, Lützenkirchen-Hecht  D, Li  L  et al.  Boosting oxygen reduction of single iron active sites via geometric and electronic engineering: nitrogen and phosphorus dual coordination. J Am Chem Soc  2020; 142 : 2404–12.10.1021/jacs.9b11852 31902210
3. Yang  Y, Li  P, Zheng  X  et al.  Anion-exchange membrane water electrolyzers and fuel cells. Chem Soc Rev  2022; 51 : 9620–93.10.1039/D2CS00038E 36345857
4. Gao  F-Y, Gao  M-R.  Nickel-based anode catalysts for efficient and affordable anion-exchange membrane fuel cells. Acc Chem Res  2023; 56 : 1445–57.10.1021/acs.accounts.3c00071 37170082
5. Han  X, Ling  X, Wang  Y  et al.  Generation of nanoparticle, atomic-cluster, and single-atom cobalt catalysts from zeolitic imidazole frameworks by spatial isolation and their use in zinc–air batteries. Angew Chem Int Ed  2019; 58 : 5359–64.10.1002/anie.201901109
6. Dou  Y, Xie  Z, Wei  Y  et al.  Redox mediators for high-performance lithium–oxygen batteries. Natl Sci Rev  2022; 9 : nwac040.10.1093/nsr/nwac040 35548381
7. Zhu  S, Sun  M, Mei  B  et al.  Intrinsic spin shielding effect in platinum–rare-earth alloy boosts oxygen reduction activity. Natl Sci Rev  2023; 10 : nwad162.10.1093/nsr/nwad162 37900058
8. Zhou  Y, Lu  R, Tao  X  et al.  Boosting oxygen electrocatalytic activity of Fe–N–C catalysts by phosphorus incorporation. J Am Chem Soc  2023; 145 : 3647–55.10.1021/jacs.2c12933 36744313
9. Pan  Y, Qian  Y, Zheng  X  et al.  Precise fabrication of single-atom alloy co-catalyst with optimal charge state for enhanced photocatalysis. Natl Sci Rev  2021; 8 : nwaa224.10.1093/nsr/nwaa224 34691561
10. Tian  H, Song  A, Zhang  P  et al.  High durability of Fe–N–C single-atom catalysts with carbon vacancies toward the oxygen reduction reaction in alkaline media. Adv Mater  2023; 35 : 2210714.10.1002/adma.202210714
11. Cai  J, Zhang  H, Zhang  L  et al.  Hetero-anionic structure activated Co–S bonds promote oxygen electrocatalytic activity for high-efficiency zinc–air batteries. Adv Mater  2023; 35 : 2303488.10.1002/adma.202303488
12. Xiong  W, Li  H, You  H  et al.  Encapsulating metal organic framework into hollow mesoporous carbon sphere as efficient oxygen bifunctional electrocatalyst. Natl Sci Rev  2020; 7 : 609–19.10.1093/nsr/nwz166 34692080
13. Zhuang  Z, Li  Y, Li  Y  et al.  Atomically dispersed nonmagnetic electron traps improve oxygen reduction activity of perovskite oxides. Energy Environ Sci  2021; 14 : 1016–28.10.1039/D0EE03701J
14. Zhang  W, Cai  G, Wu  R  et al.  Templating synthesis of metal–organic framework nanofiber aerogels and their derived hollow porous carbon nanofibers for energy storage and conversion. Small  2021; 17 : 2004140.10.1002/smll.202004140
15. Xiong  Y, Li  H, Liu  C  et al.  Single-atom Fe catalysts for Fenton-like reactions: roles of different N species. Adv Mater  2022; 34 : 2110653.10.1002/adma.202110653
16. Huang  H, Yu  D, Hu  F  et al.  Clusters induced electron redistribution to tune oxygen reduction activity of transition metal single-atom for metal–air batteries. Angew Chem Int Ed  2022; 61 : e202116068.10.1002/anie.202116068
17. Miao  Z, Wang  X, Zhao  Z  et al.  Improving the stability of non-noble-metal M–N–C catalysts for proton-exchange-membrane fuel cells through M–N bond length and coordination regulation. Adv Mater  2021; 33 : 2006613.10.1002/adma.202006613
18. Tong  M, Sun  F, Xie  Y  et al.  Operando cooperated catalytic mechanism of atomically dispersed Cu−N4 and Zn−N4 for promoting oxygen reduction reaction. Angew Chem Int Ed  2021; 60 : 14005–12.10.1002/anie.202102053
19. Wan  X, Liu  Q, Liu  J  et al.  Iron atom–cluster interactions increase activity and improve durability in Fe–N–C fuel cells. Nat Commun  2022; 13 : 2963.10.1038/s41467-022-30702-z 35618792
20. Wei  X, Song  S, Cai  W  et al.  Tuning the spin state of Fe single atoms by Pd nanoclusters enables robust oxygen reduction with dissociative pathway. Chem  2023; 9 : 181–97.10.1016/j.chempr.2022.10.001
21. Cui  T, Wang  Y-P, Ye  T  et al.  Engineering dual single-atom sites on 2D ultrathin N-doped carbon nanosheets attaining ultra-low-temperature zinc-air battery. Angew Chem Int Ed  2022; 61 : e202115219.10.1002/anie.202115219
22. Liu  Q, Liu  X, Zheng  L  et al.  The solid-phase synthesis of an Fe-N-C electrocatalyst for high-power proton-exchange membrane fuel cells. Angew Chem Int Ed  2018; 57 : 1204–8.10.1002/anie.201709597
23. Huang  H, Huang  A, Liu  D  et al.  Tailoring oxygen reduction reaction kinetics on perovskite oxides via oxygen vacancies for low-temperature and knittable zinc–air batteries. Adv Mater  2023; 35 : 2303109.10.1002/adma.202303109
24. Zhu  Z, Yin  H, Wang  Y  et al.  Coexisting single-atomic Fe and Ni sites on hierarchically ordered porous carbon as a highly efficient ORR electrocatalyst. Adv Mater  2020; 32 : 2004670.10.1002/adma.202004670
25. Peng  L, Yang  J, Yang  Y  et al.  Mesopore-rich Fe–N–C catalyst with FeN4–O–NC single-atom sites delivers remarkable oxygen reduction reaction performance in alkaline media. Adv Mater  2022; 34 : 2202544.10.1002/adma.202202544
26. Meng  H, Wu  B, Zhang  D  et al.  Optimizing electronic synergy of atomically dispersed dual-metal Ni–N4 and Fe–N4 sites with adjacent Fe nanoclusters for high-efficiency oxygen electrocatalysis. Energy Environ Sci  2024; 17 : 704–16.10.1039/D3EE03383J
27. Wang  X, Yu  M, Feng  X.  Electronic structure regulation of noble metal-free materials toward alkaline oxygen electrocatalysis. eScience  2023; 3 : 100141.10.1016/j.esci.2023.100141
28. Li  P, Jiao  Y, Ruan  Y  et al.  Revealing the role of double-layer microenvironments in pH-dependent oxygen reduction activity over metal-nitrogen-carbon catalysts. Nat Commun  2023; 14 : 6936.10.1038/s41467-023-42749-7 37907596
29. Wang  T, Zhang  Y, Huang  B  et al.  Enhancing oxygen reduction electrocatalysis by tuning interfacial hydrogen bonds. Nat Catal  2021; 4 : 753–62.10.1038/s41929-021-00668-0
30. Zhang  J, Mück-Lichtenfeld  C, Studer  A. Photocatalytic phosphine-mediated water activation for radical hydrogenation. Nature  2023; 619 : 506–13.10.1038/s41586-023-06141-1 37380779
31. Chung  M, Maalouf  JH, Adams  JS  et al.  Direct propylene epoxidation via water activation over Pd-Pt electrocatalysts. Science  2024; 383 : 49–55.10.1126/science.adh4355 38175873
32. Ji  D, Lin  Y, Guo  X  et al.  Electrospinning of nanofibres. Nat Rev Methods Primers  2024; 4 : 1.10.1038/s43586-023-00278-z
33. Li  X, Huang  X, Xi  S  et al.  Single cobalt atoms anchored on porous N-doped graphene with dual reaction sites for efficient Fenton-like catalysis. J Am Chem Soc  2018; 140 : 12469–75.10.1021/jacs.8b05992 30165734
34. Fan  L, Liu  PF, Yan  X  et al.  Atomically isolated nickel species anchored on graphitized carbon for efficient hydrogen evolution electrocatalysis. Nat Commun  2016; 7 : 10667.10.1038/ncomms10667 26861684
35. Yang  H, Gong  L, Wang  H  et al.  Preparation of nickel-iron hydroxides by microorganism corrosion for efficient oxygen evolution. Nat Commun  2020; 11 : 5075.10.1038/s41467-020-18891-x 33033245
36. Shen  H, Gracia-Espino  E, Ma  J  et al.  Atomically FeN2 moieties dispersed on mesoporous carbon: A new atomic catalyst for efficient oxygen reduction catalysis. Nano Energy  2017; 35 : 9–16.10.1016/j.nanoen.2017.03.027
37. Rao  P, Wu  D, Wang  T-J  et al.  Single atomic cobalt electrocatalyst for efficient oxygen reduction reaction. eScience  2022; 2 : 399–404.10.1016/j.esci.2022.05.004
38. Liu  S, Li  C, Zachman  MJ  et al.  Atomically dispersed iron sites with a nitrogen–carbon coating as highly active and durable oxygen reduction catalysts for fuel cells. Nat Energy  2022; 7 : 652–63.10.1038/s41560-022-01062-1
39. Chen  Z, Niu  H, Ding  J  et al.  Unraveling the origin of sulfur-doped Fe-N-C single-atom catalyst for enhanced oxygen reduction activity: effect of iron spin-state tuning. Angew Chem Int Ed  2021; 60 : 25404–10.10.1002/anie.202110243
40. Asset  T, Atanassov  P.  Iron-nitrogen-carbon catalysts for proton exchange membrane fuel cells. Joule  2020; 4 : 33–44.10.1016/j.joule.2019.12.002
41. Wu  Y-J, Wu  X-H, Tu  T-X et al.  Controlled synthesis of FeNx-CoNx dual active sites interfaced with metallic Co nanoparticles as bifunctional oxygen electrocatalysts for rechargeable Zn-air batteries. Appl Catal, B  2020; 278 : 119259.10.1016/j.apcatb.2020.119259
42. Shang  H, Zhou  X, Dong  J  et al.  Engineering unsymmetrically coordinated Cu-S1N3 single atom sites with enhanced oxygen reduction activity. Nat Commun  2020; 11 : 3049.10.1038/s41467-020-16848-8 32546781
43. Zhao  L, Zhang  Y, Huang  L-B  et al.  Cascade anchoring strategy for general mass production of high-loading single-atomic metal-nitrogen catalysts. Nat Commun  2019; 10 : 1278.10.1038/s41467-019-09290-y 30894539
44. Cheng  Z, Huang  B, Pi  Y  et al.  Partially hydroxylated ultrathin iridium nanosheets as efficient electrocatalysts for water splitting. Natl Sci Rev  2020; 7 : 1340–8.10.1093/nsr/nwaa058 34692162
45. Wang  Y, Yang  Y, Jia  S  et al.  Synergistic Mn-Co catalyst outperforms Pt on high-rate oxygen reduction for alkaline polymer electrolyte fuel cells. Nat Commun  2019; 10 : 1506.10.1038/s41467-019-09503-4 30944328
46. Farias  SAS, Martins  JBL.  Bonding and electronic structure of sillenites. Chem Phys Lett  2012; 533 : 78–81.10.1016/j.cplett.2012.03.030
47. Li  S, Zhao  S, Hu  F  et al.  Exploring the potential Ru-based catalysts for commercial-scale polymer electrolyte membrane water electrolysis: a systematic review. Prog Mater Sci  2024; 145 : 101294 10.1016/j.pmatsci.2024.101294
