Fe-N co-doped carbon nanofibers with Fe3C decoration for water activation induced oxygen reduction reaction

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.


INTRODUCTION
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 atten-tion 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-N 4 configuration, demonstrate Ptlike behavior in O 2 adsorption and the subsequent O = O bond breaking during the ORR catalytic process [19 -21 ].Nevertheless, the unsatisfactory adsorption energy of Fe-N 4 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-N 4 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-N 4 catalysts are considered promising candidates to improve ORR performance and stability in acidic media [20 ].However, the modulation of the intrinsic activity of Fe-N 4 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 Fe 3 C, 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-N 4 site to enhance ORR activity.Moreover, electrospinning has emerged as a fascinating and scalable technique for fabricating energyfunctional composite nanofibers, enabling the incorporation of both single atomic and metallic compounds [32 ].
In this work, we successfully fabricated the Fe 3 C-modified Fe-N 4 enriched with a carbon nanotube system (Fe 3 C-Fe 1 /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 Fe 3 C leads to d-orbital electron rearrangement, optimizing the binding energy of intermediates in the rate-determining step.Notably, the Fe 3 C-Fe 1 /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 Fe 3 C-Fe 1 /CNT demonstrates a high power density of 724 mW cm -2 in H 2 -O 2 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.

Design, synthesis and characterization
The fabrication approach of Fe 3 C-Fe 1 /CNT is described in Fig. 1 a.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 Fe 3 C 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. 1 b and c, Fig. S2) [33 ].The transmission electron microscopy (TEM) images in Fig. 1 d and e further confirm the graphite-coated Fe 3 C NPs loading on carbon fiber intimate contact with the bamboo-like CNT structures in Fe 3 C-Fe 1 /CNT.No NPs are observed for Fe 1 /CNT revealing that Fe 3 C NPs are removed after the HNO 3 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. 1 f, the carbon substrate contains both Fe 3 C 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 Fe 3 C-Fe 1 /CNT nanofibers (Fig. 1 g).The above results demonstrate that the electrospinning process, thermal treatment, and acid etching route successfully produce the CNT structure, in which Fe 3 C NPs and atomically dispersed Fe sites co-exist in the carbon matrix.
The catalysts are analyzed using X-ray diffraction (XRD) patterns to determine their chemical makeup and crystal structure.As i l lustrated in Fig. 2 a, the typical Fe 3 C-Fe 1 /CNT diffraction peaks at around 26.6°are attributed to the graphitic carbon structures' (002) plane.Insignificant Fe 3 C-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 H 2 SO 4 treatment.After further treatment with HNO 3 , the Fe 3 C peaks vanish, leaving just the typical graphitic carbon peak in Fe 1 /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 Fe 3 C-Fe 1 /CNT compared to Fe 3 C-Fe 1 , indicating (BET) specific surface area of Fe 3 C-Fe 1 /CNT is measured by isothermal N 2 adsorption-desorption analysis (Fig. 2 c).The specific surface area of Fe 3 C-Fe 1 /CNT is 399.7 m 2 g -1 , higher than that of Fe 1 /CNT (363.4 m 2 g -1 ).Fe 3 C-Fe 1 /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. 2 d-f, Fig. S5).The C-N in Fe 3 C-Fe 1 /CNT experiences a decrease in binding energy, suggesting that the electronic structure of the Fe-N 4 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 Fe 3 C, which may activate the carbon layer outer surface to promote its involvement in electrocatalysis [35 ].In comparison with the Fe 1 /CNT, the Fe-N species in Fe 3 C-Fe 1 /CNT show a shift toward lower binding energy.These results evidence that strong electronic perturbation of Fe 3 C NPs lead to an asymmetric distribution of Fe-N 4 electron density.Significantly, Fe 3 C-Fe 1 /CNT exhibits larger contents of pyridinic-N, Fe-N, and graphitic-N species ( Table S1).Pyridinic-N may lower the energy barrier for O 2 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 ].
Local electronic changes can be further observed in the fine structure of X-ray absorption (XAFS).The Fe 3 C-Fe 1 /CNT absorption edge falls between the absorption edges of FePc and Fe 2 O 3 , indicating that the oxidation state of Fe lies in the range of + 2 to + 3, that is, Fe 2 + and Fe 3 + coexist in the Fe 3 C-Fe 1 /CNT catalyst (Fig. 2 g).In Fe-N-C catalysts, only the Fe 3 + -N 4 structure is generally obtained [38 ].One of the main reasons here is that during the material synthesis process, Fe 2 + in the iron source is easily oxidized to Fe 3 + to form the Fe 3 + -N 4 structure [39 ].In fact, the Fe 2 + -N 4 structure is much more catalytically active than the Fe 3 + -N 4 structure for ORR [40 ].Furthermore, Fe 3 C-Fe 1 /CNT displays a negative absorption edge compared to Fe 1 /CNT, suggesting that the Fe center has a reduced valence and a larger electron density.The presence of more electronegative groups like Fe 3 C has an impact on the active center of Fe 1 /CNT, according to the positive absorption edge compared to FePc.Thus, the introduction of Fe 3 C leads to the conversion of Fe 3 + to Fe 2 + and the asymmetric distribution of Fe-N 4 electron density, which is consistent with the XPS analysis.The Fe K-edge Fourier-transformed extended X AFS (FT-EX AFS) spectrum provides significant structural information ( Fig. S6, Table S2).As shown in Fig. 2 h, the peak position contributed to Fe-N of Fe 3 C-Fe 1 /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-N 4 active sites [41 ,42 ].The Fe-Fe bonds are represented by peaks in Fe 3 C-Fe 1 /CNT at 2.2 Å, which is consistent with the presence of Fe 3 C species [43 ].Wavelettransform EXAFS (WT-EXAFS) of the corresponding catalysts are performed (Fig. 2 i, Fig. S7).The WT-EXAFS analyses of Fe 1 /CNT show the greatest intensity at ∼1.51 Å and 3.57 Å −1 , confirming the coordination structure of Fe-N.In contrast, Fe 3 C-Fe 1 /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 X ANES/EX AFS analyses, we successfully obtained a composite structure of Fe 3 C/Fe-N 4 with an asymmetric distribution of the electron density of Fe-N 4 around the Fe active center.

ORR performance evaluation of Fe 3 C-Fe 1 /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 O 2 -saturated 0.5 M H 2 SO 4 .The cyclic voltammetry (CV) curves unmistakably display a reduction peak, while the N 2 -saturated solution lacks a matching peak ( Fig. S8), proving the electrochemical response of Fe 3 C-Fe 1 /CNT toward the ORR.The linear sweep voltammetry (LSV) tests are performed to evaluate the catalysts ORR activity (Fig. 3 a).The ultimate Fe 3 C-Fe 1 /CNT exhibits a Tafel slope of 75.9 mV dec −1 , which is similar to that of Pt/C and smaller than Fe 1 /CNT (Fig. 3 b).Fe 3 C-Fe 1 /CNT exhibits an onset potential ( E onset ) of 0.93 V and a half-wave potential ( E 1/2 ) of 0.83 V, which is preferable to Fe 1 /CNT, and somewhat lesser activity than Pt/C catalyst (Fig. 3 c).In addition, the resulting Fe 3 C-Fe 1 /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. 3 d, Fig. S9).The n value for Fe 3 C-Fe 1 /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 Fe 1 /CNT, Fe 3 C-Fe 1 /CNT exhibits a higher double-layer capacitance, which is correlated with its electrochemical surface area (Fig. 3 e, 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. 3 f).As shown in Fig. 3 g, the H 2 O 2 yields stay below 5%, indicating a 4-electron ORR route over the Fe 3 C-Fe 1 /CNT.After 80 0 0 CV cycles, the E 1/2 of Fe 3 C-Fe 1 /CNT only negatively shifts 39 mV, whereas the 85 mV negative shifts for Pt/C (Fig. 3 h).As a result, Fe 3 C NPs adorned Fe-N 4 with higher electron cloud density are more successful than the Fe-N 4 sites in lowering the ORR reaction energy barrier, which results in a lower overpotential (Fig. 3 i).
Fe 3 C-Fe 1 /CNT not only performs well in acid solution but also exhibits remarkable ORR performance in 0.1 M KOH solution.Fe 3 C-Fe 1 /CNT with an E 1/2 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 Fe 3 C-Fe 1 /CNT exhibits high selectivity for 4electron ORR, the excellent stability and resistance to methanol.Fe 3 C-Fe 1 /CNT was also used as the Alair battery cathode catalyst given the optimized electrocatalytic activity for alkaline ORR.The solid electrolyte with a polyacrylic acid (PAA) base facilitates flexibi lity whi le concurrently offering sufficient mechanical support ( Fig. S16a).The Fe 3 C-Fe 1 /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, Fe 1 /CNT and the majority of solid-state Al-air batteries that have been previously documented ( Fig. S16b, Table S5).Fe 3 C-Fe 1 /CNT produces a greater voltage than Fe 1 /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 Fe 3 C-Fe 1 /CNT may accomplish a peak power density of 44 mW cm -2 , which is higher than Fe 1 /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 Fe 3 C-Fe 1 /CNT
To further explore the application potential of Fe 3 C-Fe 1 /CNT, we assemble a PEMFC to assess efficacy and stability (Fig. 4 a, Fig. S17).With Fe 3 C-Fe 1 /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 H 2 /O 2 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. 4 b).The Fe 3 C-Fe 1 / CNT-assembled PEMFC exhibits high performance with a peak power density of 716 mW cm −2 , significantly higher than that of Fe 1 /CNT (517 mW cm −2 ).Despite showing comparable ORR-catalytic activity on RRDE, Fe 3 C-Fe 1 /CNT and Pt/C catalysts exhibit contrasting behaviors in PEMFC.Fe 3 C 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 Fe 3 C-Fe 1 /CNT and Fe 1 /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. 4 c).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 squarewave accelerated durability test (ADT) further documented cell voltage loss (Fig. 4 d).Once more, the Fe 1 /CNT assembled PEMFC exhibits the greatest performance deterioration, with a cell voltage loss of 296 mV, which is significantly higher than that of Fe 3 C-Fe 1 /CNT (74 mV loss).Significantly, Fe 3 C-Fe 1 /CNT based PEMFC could discharge steadily for 100 hours at 0.6 V, demonstrating the welldurability (Fig. 4 e).As shown in Fig. 4 f and Table S6, the Fe 3 C-Fe 1 /CNT exhibited excellent stability and competitive activity levels compared w ith prev iously reported Fe-N-C catalysts.No significant changes in the valence states of the corresponding elements were observed in Fe 3 C-Fe 1 /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 Fe 3 C 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 Fe 3 C-Fe 1 /CNT potential for application in fuel cells.

The origin of enhanced activity of Fe 3 C-Fe 1 /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. 5 a and b, O −H stretching modes ( ∼280 0-360 0 cm −1 ) and H −O −H bending modes ( ∼160 0-170 0 cm −1 ) of interfacial water can be observed, which provides clear evidence for the involvement of surface H 2 O in the ORR.The H −O −H and O −H stretching strength at each potential in Fe 3 C-Fe 1 /CNT is much larger than Fe 1 /CNT, indicating that the Fe 3 C sites have a significant affinity for H 2 O activation and form interfacial hydrogen bonds (Fig. 5 d) [44 ].More importantly, the O −H stretching peaks at each potential in Fe 3 C-Fe 1 /CNT are mainly located near 3330 cm −1 redshifted by 50 cm −1 compared with Fe 1 /CNT ( ∼3380 cm −1 ), suggesting a strong hydrogen-bonding interaction between the interfacial water molecules and the ORR oxygen intermediates [29 ].The H 2 O bending vibration has a higher wave number on Fe 3 C-Fe 1 /CNT than on Fe 1 /CNT, showing that the presence of Fe 3 C strengthens interfacial hydrogen bonding (Fig. 5 e) [45 ].The in-situ experimental results confirmed the mechanism of water activation (Fig. 5 c), which was further investigated in-depth through density functional theory (DFT) calculations.The transition state (TS) formation on Fe 3 C/Fe-N 4 and Fe-N 4 in the hydrolysis and dissociation process is endothermic, confirming it as the rate-determining step (RDS).In this RDS, the G on Fe 3 C/Fe-N 4 is 3.29 eV, significantly smaller than that on Fe-N 4 at 4.88 eV (Fig. 5

DFT calculations
In the Fe 3 C/Fe-N 4 model, O 2 can be readily adsorbed as well as activated, and the O-O bond has the largest stretching length (Fig. 6 a), 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 Fe 3 C/Fe-N 4 , Fe-N 4 , and Fe 3 C are established with the free energy ladder images (Fig. 6 b, Fig. S21).

CONCLUSION
In summary, we successfully constructed Fe 3 Cmodified, atomically dispersed Fe-N co-doped CNT anchored on nanofibers using an electrospinningpyrolysis-etching route.The strong electronic perturbation of the Fe −N 4 active center by Fe 3 C 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 −N 4 sites.Delocalized electrons can activate the dissociation of water molecules and accelerate PCET processes by forming hydrogen bonds with surface oxygen intermediates.The Fe 3 C-Fe 1 /CNT catalyst demonstrated a high power density of 716 mW cm -2 in H 2 -O 2 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.

Figure 1 .
Figure 1.Synthesis and morphological characterization of catalysts.(a) Schematic of the synthesis process of Fe 3 C-Fe 1 /CNT and Fe 1 /CNT.(b-c) SEM images of Fe 3 C-Fe 1 /CNT.(d) TEM and (e) HRTEM images of Fe 3 C-Fe 1 /CNT.(f) AC-HAADF-STEM image of Fe 3 C-Fe 1 /CNT and (g) the corresponding EDS maps of Fe, N, C.

Figure 3 .
Figure 3. Electrocatalytic activity evaluation.(a) LSV curves of Fe 3 C-Fe 1 /CNT, Fe 1 /CNT, and 20% Pt/C in 0.5 M H 2 SO 4 .(b) Tafel slopes for the corresponding catalysts.(c) E onset and E 1/2 .(d) ORR polarization curves of Fe 3 C-Fe 1 /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 H 2 SO 4 .(g) H 2 O 2 yield and electron transfer number of the catalysts.(h) Stability tests.(i) Schematic representation of ORR properties versus oxidation state.

Figure 4 .
Figure 4. H 2 -O 2 PEMFC performance.(a) Structure illustration of the H 2 -O 2 fuel cell.(b) Polarization and power density curves by using the catalysts as cathodes.(c) Comparison of different loads of Fe 3 C-Fe 1 /CNT.(d-e) Stability tests.(f) Comparison of PEMFC performance with reported advanced catalysts.

Fe 3 C
/Fe-N 4 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-N 4 , Fe 3 C/Fe-N 4 shows enhanced polarization in the O −H bonding of adsorbed H 2 O, i l lustrating

Figure 5 .
Figure 5. In-situ characterization and reaction mechanism.(a-b) In situ FTIR spectra of the Fe 3 C-Fe 1 /CNT and Fe 1 /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 H 2 O adsorption on the Fe site of the catalysts.(h) ORR mechanism of water activation for Fe 3 C-Fe 1 /CNT.
the regulatory influence of Fe 3 C on water molecule bonding states, and suggesting that strong polarity is advantageous for hydrogen dissociation (Fig.5g).Specifically, O 2 preferentially binds to the Fe −N 4 sites to form Fe −N 4 -O 2 .The spontaneous dissociation of O 2 molecules is aided by interfacial water hydrogen bonding rather than just the presence of Fe −N 4 sites.H 2 O preferentially binds to the Fe 3 C site to form Fe 3 C-OH 2 and generates OH − by 1e -reduction.Surface protons can be transferred from Fe 3 C-OH 2 to neighboring Fe −N 4 -O 2 , leading to the formation of Fe 3 C-OH and Fe −N 4 -OOH, which subsequently undergoes 1e -reduction to produce Fe −N 4 = O and H 2 O. Then Fe −N 4 = O undergoes the same PCET process and returns to the initial state (Fig. 5 h).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 Fe 3 C-OH/Fe 3 C-OH 2 and the surface proton transfer between the proximate Fe 3 C and Fe −N 4 sites.

Figure 6 .
Figure 6.Theoretical calculation of the catalytic activity.(a) Adsorption energies for H 2 O and O 2 .(b) Free energy diagram of ORR steps on Fe 3 C/Fe-N 4 and Fe-N 4 .(c) DOS of Fe 3 C/Fe-N 4 and Fe-N 4 .(d) COHP analysis of Fe −O bond after *OH adsorption for Fe 3 C/Fe-N 4 .(e) Schematic representation of the regulatory of Fe 3 C toward G *OH on Fe −N 4 sites.The red and green represent the charge accumulation and dispersion, respectively.