Carbon‐Confined Fe3C/Fe3N Janus Interfaces for Selective Nitric Oxide‐to‐Ammonia Electroreduction Authors: Jialing Song, Ziqi Wei, Haotian Huang, Lupeng Han, Biyi Huang, Evangelina Pensa, Sam Fong Yau Li, Eslam Hamed, Fun Man Fung, Emiliano Cortés, Dengsong Zhang Published: 2026 ABSTRACT ABSTRACTElectrocatalytic nitric oxide reduction to ammonia couples pollutant valorization with sustainable nitrogen conversion, but high activity and selectivity require concurrent control of NO transport, NO activation, and hydrogenation at the gas‐liquid‐solid interface. Here, we report a self‐supported, noble‐metal‐free Fe3C/Fe3N@C catalyst, composed of earth‐abundant Fe, C, and N, featuring defect‐rich Fe3C/Fe3N Janus nanostructures confined within graphitic carbon. The catalyst achieves an NH3 yield rate of 468.3 µmol h−1 cm−2 with a Faradaic efficiency of 94.2% at −0.6 V versus RHE, placing it among the most efficient reported NORR electrocatalysts. Mechanistic studies reveal that the graphitic carbon shell facilitates NO diffusion by alleviating the steric and dynamic constraints imposed by the hydrogen‐bonded water network. At the Janus interface, Fe3C sites preferentially adsorb and activate NO, whereas nitrogen‐vacancy‐rich Fe3N sites promote H2O dissociation to supply reactive *H for subsequent hydrogenation. This spatial coupling of mass‐transfer promotion, NO activation, and interfacial *H generation enables efficient and selective NO‐to‐NH3 electroreduction. These findings establish carbon‐confined, earth‐abundant carbide/nitride Janus interfaces as a promising design principle for high‐performance NORR catalysts. Introduction Ammonia (NH3) is one of the most important commodity chemicals, used in the manufacture of fertilizers, refrigerants, and carbon‐free energy carriers [1, 2]. The primary route of industrial NH3 synthesis relies on the Haber‐Bosch process, which operates under harsh conditions and emits substantial amounts of CO2 [3, 4]. As a promising alternative, electrocatalytic reduction of nitrogen‐containing species, such as nitric oxide (NO), nitrate (NO3 −), and nitrogen (N2), has gained increasing attention due to its green and sustainable advantages [5, 6]. The strong N≡N triple bond renders N2 chemically inert, posing major challenges for the N2 reduction reaction (NRR) [7, 8, 9]. The electroreduction of NO3 − to NH3 (NO3 −RR) involves complex pathways and readily yields various nitrogenous by‐products [10]. In contrast, the NO reduction reaction (NORR) is more favorable than NRR due to its weaker N═O bond and higher solubility in electrolytes [11]. Moreover, NORR features a simpler reaction route and fewer side reactions than NO3 −RR. Recently, NORR has garnered increasing interest, as it not only offers a promising alternative for NH3 synthesis but also provides an efficient means of removing NO pollutants from industrial exhaust. However, the electrocatalytic NORR to NH3 is a complex reaction occurring at the gas‐liquid‐solid interface, involving NO diffusion to the catalyst surface, adsorption/activation, proton supply via H2O dissociation, stepwise hydrogenation, and NH3 product desorption [12, 13]. Due to the low solubility of NO and the steric and kinetic hindrance of the hydrogen bonding network of water, NO mass transfer limitation represents a critical and frequently neglected challenge for NORR [14, 15, 16]. Moreover, insufficient adsorption and activation of NO as well as inadequate proton supply at catalyst active sites further complicate the problem [17]. Therefore, synergistically enhancing NO diffusion, adsorption, and proton supply is key to achieving a high NH3 yield rate and NH3 Faradaic efficiency (FE) in NORR. Recently, porous carbon materials have shown potential as supports for optimizing NO transport pathways and improving the NO diffusion kinetics in water [14, 18]. In addition, most transition metals exhibit weak binding affinity toward NO and tend to form metal‐H bonds, which intensifies the competing hydrogen evolution reaction (HER) [19, 20, 21]. To overcome this limitation, materials with metal‐like electronic properties, such as metal carbides (e.g., Mo2C and TiC), have attracted significant attention owing to their excellent conductivity and hydrogenation activity [19, 22, 23]. Furthermore, metal nitrides with nitrogen vacancies (e.g., Cu3N and LaN) can effectively modulate the electronic structure of catalysts to enhance reactant adsorption and activation [24, 25, 26]. Taken together, porous carbon materials, metal carbides, and metal nitrides, provide viable platforms to improve NO diffusion, adsorption, and H2O dissociation. However, achieving an effective integration of their distinct catalytic functions remains a critical challenge that demands further exploration. Beyond activity and selectivity, the elemental composition of NORR catalysts is also important for practical deployment. Catalysts based on earth‐abundant elements are preferable to noble‐metal systems from the perspectives of cost, resource availability, and scalability. In this regard, iron carbides and iron nitrides are attractive platforms because they combine the abundance of Fe with metallic conductivity, tunable electronic structures, and catalytic functions relevant to NO activation and water dissociation. In this study, a novel self‐supported defect‐rich Fe3C/Fe3N@C catalyst with a graphitic carbon‐encapsulated Fe3C/Fe3N Janus nanostructure was designed and synthesized via a top‐down electrospinning approach. Such a catalyst displayed a high NH3 yield rate of 468.26 µmol h−1 cm−2 with an FE of 94.18% at −0.6 V versus RHE, outperforming most reported NORR electrocatalysts in the literature. The superior NORR performance of Fe3C/Fe3N@C is attributed to the synergistic coupling of enhanced NO diffusion, adsorption and proton supply within the graphitic carbon‐encapsulated Fe3C/Fe3N Janus nanostructure. The graphitic carbon framework facilitates NO mass transfer to the catalyst surface, while the Fe3C/Fe3N Janus structure simultaneously promotes NO adsorption and proton supply via accelerated H2O dissociation. This study provides a viable strategy for the rational design of highly efficient NORR electrocatalysts by integrating the functions of porous carbon shells, metal carbides, and metal nitrides. Results and Discussion Figure 1a illustrates the schematic diagram of the Fe3C/Fe3N@C synthesis via a top‐down electrospinning approach. First, a self‐supported membrane was prepared by electrospinning a precursor solution containing iron acetylacetonate and polyacrylonitrile (PAN) (Figure S1). The use of PAN facilitates the introduction of nitrogen defects, as charge transfer readily occurs from PAN functional groups to metal‐ion complexes, leading to the formation of surface defects [27, 28, 29]. The self‐supported membrane was then subjected to thermal treatment at 280°C in air, undergoing cyclization and dehydrogenation to form a ladder‐like polymer structure (Fe‐PAN) [30]. Subsequently, Fe3C/Fe3N@C was obtained after the Fe‐PAN membrane was carbonized at 900°C under an N2 atmosphere, which promoted the formation of Fe3N phases through the reaction between iron and nitrogen atoms. Concurrently, the Fe3N phases reacted with carbon to generate Fe3C while simultaneously producing nitrogen vacancies. Furthermore, during high‐temperature carbonization, carbon atoms continued to reorganize around the iron‐based particles, gradually forming a graphitic carbon shell that encapsulates the surface. This shell not only stabilizes the Fe3C and Fe3N phases but also enhances the electrical conductivity and structural integrity of the composite. Scanning electron microscopy (SEM) images of Fe3C/Fe3N@C reveal that continuous and crosslinked nanofibers form a hierarchical three‐dimensional architecture with abundant nanopores (Figure 1b). The average diameter of nanofibers is approximately ∼220 nm. Interestingly, a large quantity of tiny nanotube arrays is observed on the rough carbon fiber surface with nanoparticles situated at the upper ends of the nanotubes (Figure S2). Transmission electron microscopy (TEM) revealed the morphology of graphitic carbon‐encapsulated nanoparticles (Figure 1c), and the corresponding element mappings confirmed the even distribution of Fe, C, and N within these structures (Figure S3). Rietveld‐refined x‐ray diffraction analysis (Rietveld‐refined XRD) further confirms the co‐existence of Fe3C and Fe3N, in which Fe3N accounts for approximately 9% of the iron sites (Figure 1d and Table S1, S2) [31]. The high‐resolution TEM image of Fe3C/Fe3N@C exhibits a distinct Janus structure, displaying lattice spacings of 0.225 nm corresponding to the (031) plane of Fe3C and 0.215 nm corresponding to the (111) plane of Fe3N (Figure 1e,f). Furthermore, the aberration‐corrected high‐angle annular dark‐field scanning transmission electron microscopy (AC‐HAADF‐STEM) of Fe3C/Fe3N@C indicates the presence of nitrogen vacancies in Fe3N (111) and interlayer spacing of 0.325 nm corresponding to the C (002) crystal plane (Figure 1g). As a comparison, Fe‐PAN membrane was also carbonized at 500°C (denoted as Fe3N@C), and the material only contained Fe3N phases as confirmed by XRD results (JCPDS No. 49–1662, Figure S4). In addition, due to the lower carbonization temperature, the Fe3N nanoparticles in Fe3N@C are encapsulated by amorphous carbon rather than graphitic carbon (Figure S5). Besides, Fe3C@C was also prepared with a similar process as Fe3C/Fe3N@C using a precursor solution containing iron acetylacetonate and polyvinylidene difluoride (without N elements). Fe3C@C exhibits the XRD peaks of Fe3C (JCPDS No. 77–0255, Figure S6) and the morphology of graphitic carbon‐encapsulated Fe3C nanoparticles (Figure S7). Combining the XRD and TEM results, we speculate that iron species within the nanofibers are initially converted to Fe3N at 500°C, which further reacts partially with carbon at 900°C to form Fe3C while releasing N2, resulting in nitrogen vacancies. Morphology and structure of catalysts. (a) Schematic illustration of the synthesis of Fe3C/Fe3N@C. (b) SEM and (c) HR‐TEM images of Fe3C/Fe3N@C. (d) The Rietveld XRD pattern of Fe3C/Fe3N@C. (e, f) HAADF‐STEM images of Fe3C/Fe3N@C. (g) AC‐HAADF‐STEM image of Fe3C/Fe3N@C. Fourier transform infrared spectroscopy (FTIR) was used to analyze the structural evolution from Fe3N@C to Fe3C/Fe3N@C. As shown in Figure 2a, the peaks corresponding to the bending vibrations of N‐N and the stretching vibrations of N–H/C═N become less intense in Fe3C/Fe3N@C compared to Fe3N@C, indicating the loss of lattice N and the formation of N vacancies in Fe3C/Fe3N@C [32]. According to the x‐ray photoelectron spectroscopy (XPS) results (Figure 2b), the N content in Fe3C/Fe3N@C is significantly lower compared to that in Fe3N@C (Table S3 and Figure 2c), which is also consistent with the inductively coupled plasma optical emission spectrometry (ICP‐OES) results (Table S4). The N 1s spectrum can be deconvoluted into oxidized N (405.2 eV), graphitic N (402.3 eV), pyrrolic N (400.6 eV), Fe‐N (399.7 eV), and pyridinic N (398.6 eV) [33]. Generally, pyridinic N atoms located at defects or edges of the carbon matrix are more reactive due to their unsaturated coordination state [34, 35]. The pyridinic N species are preferentially eliminated in the form of N2 during high‐temperature carbonization, leading to the formation of nitrogen vacancies. Consequently, the pyridinic N species in Fe3C/Fe3N@C decrease notably compared to those in Fe3N@C. In addition, the content of Fe‐N species in Fe3C/Fe3N@C is also significantly lower than that in Fe3N@C. This suggests that some of the Fe‐N species in Fe3C/Fe3N@C react with carbon to form Fe3C and N2, thereby creating nitrogen vacancies. Brunauer‐Emmett‐Teller (BET) analysis shows that Fe3C/Fe3N@C has a much larger specific surface area (312.95 m2 g−1) than Fe3N@C (109.61 m2 g−1) (Figure S8 and Table S5), which is attributed to the existence of more defect sites. Besides, the C 1s XPS spectra confirm more graphitic carbon in Fe3C/Fe3N@C than Fe3N@C (Figure S9). Raman spectroscopy was also employed to probe the structural evolution of the carbon matrix [36, 37]. The characteristic D band (∼1350 cm−1) and G band (∼1580 cm−1) corresponding to lattice defects/disorder and graphitic sp2 carbon were observed, respectively. The ratio of the D band area to the G band area (AD/AG) serves as a key indicator of defect density within the carbon framework [38]. As shown in Figure 2d, the AD/AG values followed a distinct order: Fe3C/Fe3N@C (2.24) > Fe3C@C (1.66) > Fe3N@C (1.44). This trend aligns with the established mechanism that intense nitrogen volatilization in nitrogen‐containing precursors at ∼900°C generates maximal lattice disorder [27]. Specifically, Fe3C/Fe3N@C undergoes the most significant nitrogen removal. This process not only etches the carbon framework to produce a high defect density but also drives the essential phase separation to form the Janus heterostructure and nitrogen vacancies. In contrast, Fe3N@C prepared at 500°C exhibits milder nitrogen loss and thus a lower defect density, while Fe3C@C derived from a nitrogen‐free polymer at 900°C lacks this disruptive volatilization pathway entirely, resulting in a more graphitized carbon support. Therefore, the elevated AD/AG ratio in the Janus catalyst serves as a direct spectroscopic signature of the vacancy‐generating chemistry, which simultaneously constructs the active interface and optimizes the conductive carbon host for enhanced electrocatalysis. Electron paramagnetic resonance (EPR) spectroscopy further confirms that Fe3C/Fe3N@C possesses more nitrogen vacancies than Fe3N@C (Figure 2e), as evidenced by the stronger signal intensity at g = 2.003 that is widely reported for nitrogen vacancies in carbon‐based metal nitrides and N‐doped carbons [24, 39, 40]. Based on the above results, the proposed atomic structures during the thermal tuning process at 500°C and 900°C are illustrated in Figure 2f and Figure S10, respectively. Structure analysis of catalysts. (a) FTIR spectra, (b) high‐resolution N 1s XPS spectra, (c) comparison of N species content of Fe3C/Fe3N@C and Fe3N@C. (d) Raman spectra of Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C. (e) EPR spectra of Fe3C/Fe3N@C and Fe3N@C. (f) Proposed atomic structure during the pyrolysis process at 900 °C. (g) Normalized Fe K‐edge XANES spectra and (h) Fourier transform k 3‐weighted Fe K‐edge EXAFS spectra of Fe3C/Fe3N@C, Fe3C@C, Fe3N@C, Fe foil, FeO, and Fe2O3. The inset is the enlarged pre‐edge region. (i) Wavelet transforms EXAFS contour plots of Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C. The chemical states and local environments of Fe atoms in Fe3C/Fe3N@C were evaluated using x‐ray absorption near‐edge structure (XANES) and extended x‐ray absorption fine structure (EXAFS) spectroscopy. The Fe K‐edge adsorption position of Fe3C/Fe3N@C lies between those of FeO and Fe2O3, indicating an average valence state of Fe between +2 and +3 (Figure 2g). The edge energy of Fe3C/Fe3N@C is higher than that of the Fe3C reference, suggesting a higher Fe valence state in Fe3C/Fe3N@C, in line with Fe 2p XPS results (Figure S11). This shift is ascribed to the incorporation of Fe3N, where the more electronegative N (relative to C) withdraws more electron density from Fe. To elucidate the coordination environment of Fe atoms in these samples, the EXAFS spectra were Fourier‐transformed into R‐space (Figure 2h). Fe3N@C displays Fe‐N (1.36 Å) and Fe‐Fe (2.05 Å) bonds, while Fe3C@C shows Fe‐C (1.60 Å) and Fe‐Fe (2.10 Å) bonds, respectively. In comparison, Fe3C/Fe3N@C exhibits a significantly intensified peak at ∼1.48 Å, which is attributed to the superimposed scattering from Fe‐N and Fe‐C bonds. The FT‐EXAFS spectra of Fe3C/Fe3N@C, Fe3N@C, and Fe3C@C can be well‐fitted using the theoretical structure models (Figures S12–S15, Table S6). Quantitative fitting reveals a total first‐shell Fe coordination number of ∼3.8 (from Fe‐N and Fe‐C), substantially higher than the values for the single‐phase references of Fe‐N (2.0) in Fe3N@C and Fe‐C (1.4) in Fe3C@C, thus confirming the coexistence of Fe‐N and Fe‐C bonds in Fe3C/Fe3N@C [41, 42]. In addition, Fe3C/Fe3N@C shows a lower Fe‐Fe coordination number (4.3) than Fe3N@C (10.2) and Fe3C@C (7.3), which indicates a higher dispersion of Fe atoms and, consequently, an increased number of accessible active sites. The wavelet transform (WT) contour plot of Fe3C/Fe3N@C further displays overlapping signals of Fe‐N and Fe‐C relative to those of Fe3N@C and Fe3C@C (Figure 2i, Figures S16–S18). Collectively, the above results demonstrate that Fe3C/Fe3N@C features a Fe3C/Fe3N Janus structure encapsulated by graphitic carbon. The electrocatalytic NORR measurements of all the catalysts were conducted in NO‐saturated 0.5 M Na2SO4 electrolyte using a gastight H‐type cell. Prior to introducing NO, the electrolyte was purged with high‐purity argon for 30 min to eliminate oxygen from the system. The linear sweep voltammetry (LSV) curves (Figure 3a) show that Fe3C/Fe3N@C exhibits notably higher current in the NO‐saturated electrolyte than in the Ar‐saturated electrolyte, and its current is also larger than those of Fe3N@C and Fe3C@C, indicating the superior NORR activity of Fe3C/Fe3N@C. Tafel slope analysis (Figure 3b) further revealed an earlier current response and smaller slopes for Fe3C/Fe3N@C, suggesting that the introduction of nitrogen vacancies and the Janus structure enhanced its NORR kinetics. The NH3 yield was determined by the indophenol blue spectrophotometric (IBS) method with the corresponding calibration curve (Figure S19). As shown in Figure 3c, the self‐supported Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C catalysts outperform their powdered counterparts at −0.6 V versus RHE. This enhanced performance can be attributed to the integrated conductive network, where the in‐situ formed carbon shells promote efficient inter‐component electron transfer, while the binder‐free 3D architecture maximizes the exposure of active sites and facilitates NO diffusion. As shown in Figure 3d, the NH3 yield and FE of Fe3N@C, Fe3C@C, and Fe3C/Fe3N@C display a volcano‐shaped curve from −0.3 to −0.9 V. When the potential increased from −0.3 to −0.6 V, the NH3 yield and FE gradually increased due to the accelerated reaction kinetics of NORR and a higher *H utilization rate during the hydrogenation steps. However, when the applied potential exceeds −0.6 V versus RHE, the NH3 yield rate and FE decline due to competing HER. Among the three catalysts, Fe3C/Fe3N@C achieves the highest NH3 yield of 468.26 µmol h−1 cm−2 and FE of 94.18% at −0.6 V versus RHE. The possible liquid byproducts hydroxylamine (NH2OH) and hydrazine hydrate (N2H4) were not detected during the NORR process (Figures S20–S23). A trace amount of H2 was detected, while N2 and N2O were not observed by differential electrochemical mass spectrometry (DEMS), establishing H2 as the primary NORR by‐product (Figure S24). Notably, the relative intensity of the H2 signal follows the order Fe3C/Fe3N@C < Fe3C@C < Fe3N@C (Figure S25), consistent with the progressively decreasing NH3 selectivity. This trend is further supported by gas chromatography (GC) results, which show corresponding Faradaic efficiencies for H2 of 2.55% (Fe3C/Fe3N@C), 12.83% (Fe3C@C), and 28.75% (Fe3N@C), respectively (Figure S26). The NH3 yield and FE determined by 1H nuclear magnetic resonance (NMR) show similar product concentrations to those obtained from UV‐vis measurements, demonstrating the accuracy of the quantitative analysis in this work (Figures S27–S28). Subsequently, a series of control experiments were conducted to determine the nitrogen source of the produced NH3. Trace amounts of NH3 products were detected in the NO‐saturated electrolyte without an applied potential and in the Ar‐saturated electrolyte with an applied potential, thus eliminating possible interference from nitrogen‐containing contaminants (Figure S29). To further confirm the origin of the generated NH3, isotopic labeling experiment using 15NO and 14NO were performed and analyzed by 1H NMR. As shown in Figure 3e, the NMR spectrum exhibits a characteristic 14NH4 + triplet signal with a coupling constant of 52 Hz when 14NO is used as the N precursor, whereas a doublet signal with a coupling constant of 72 Hz associated with 15NH4 + appears when 15NO is used. These results demonstrate that the generated NH3 originates from the electroreduction of NO rather than from any nitrogen‐containing impurities in the electrolytic system. NORR performance. (a) LSV curves of Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C with a scanning rate of 5 mV s−1 recorded in 0.5 M Ar‐saturated and NO‐saturated Na2SO4 (with 90% iR‐correction). (b) Tafel slopes of Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C. (c) NH3 yield rate and FE on Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C. (d) The comparison of NH3 yield rate and FE for Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C at different applied potentials. (e) NMR spectra of isotopic labeling experiments on Fe3C/Fe3N@C. (f) The Cdl comparison of Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C. (g) Cycle stability tests of Fe3C/Fe3N@C at −0.6 V versus RHE. Each cycle requires a replacement of the 0.5 M Na2SO4 electrolyte. (h) Comparison of NH3 electrosynthesis performance of Fe3C/Fe3N@C with previously reported electrocatalysts. To evaluate the catalytic activity of the catalysts, we calculated the double‐layer capacitances (C dl) and accessed the electrochemical active surface area (ECSA) based on cyclic voltammetry (CV) curves at scan rates ranging from 10 to 100 mV s−1 (Figure 3f and S30). Fe3C/Fe3N@C exhibits a higher Cdl (6.4 mF cm−2) than those of Fe3C@C (3.9 mF cm−2) and Fe3N@C (2.2 mF cm−2), indicating a larger exposed catalytic surface during NORR. Besides, the turnover frequency (TOF) value of Fe3C/Fe3N@C was calculated to be 0.59 s−1 at −0.6 V versus RHE, which is higher than those of Fe3N@C (0.46 s−1) and Fe3C@C (0.51 s−1) (Figure S31). These results suggest that Fe3C/Fe3N@C exhibits higher intrinsic activity than Fe3N@C and Fe3C@C. Electrochemical impedance spectroscopy measurement (EIS) (Figure S32) illustrates that Fe3C/Fe3N@C exhibits the smallest charge transfer resistance (2.62 Ω) compared to those of Fe3N@C (4.68 Ω) and Fe3C@C (5.08 Ω), indicating faster electron transfer kinetics during catalysis. Moreover, the NH3 yield and FE of Fe3C/Fe3N@C do not show significant fluctuation after 15 cycles (Figure 3g), and the system also retained its original current density during a 100‐h test, demonstrating the exceptional electrochemical stability of Fe3C/Fe3N@C (Figure S33). TEM, XRD, and XPS analysis after the stability tests reveal that the crystal structure and chemical composition of Fe3C/Fe3N@C remain unchanged, demonstrating its structural stability (Figures S34–S38). It is worth noting that Fe3C/Fe3N@C is superior to most reported electrochemical NORR catalysts in terms of NH3 yield, FE, and applied potential (Figure 3h and Table S7). In addition, we assembled a proof‐of‐concept Zn‐NO battery, which can simultaneously eliminate nitrogen oxides, generate NH3, and output electrical energy [43]. Surprisingly, the calculated maximum power density of the Zn‐NO battery is 16.28 mW cm−2, and the NH3 yield rate is 519.8 µg h−1 cm−2 at 8 mA cm−2 (Figure S39 and Table S8), showing great potential for practical applications. To elucidate the reasons behind the excellent NORR performance of Fe3C/Fe3N@C, a series of computational studies were conducted. Efficient NORR is often limited by the diffusion of NO molecules to the catalyst surface, a process often hindered by the hydrogen‐bonded network of surrounding water molecules [23, 44]. Using density functional theory (DFT), the migration pathway of NO molecules in solution was simulated for Fe3C/Fe3N Janus nanoparticles with and without carbon encapsulation (Figure 4a,b). For Fe3C/Fe3N nanoparticles without carbon encapsulation, the hydrogen‐bonded network of water imposes steric and dynamic constraints that force NO to navigate through narrow interstitial spaces, resulting in a convoluted migration pathway and significant mass transport limitation (Figure 4b). In contrast, when the Fe3C/Fe3N nanoparticles are encapsulated by a graphitic carbon shell, the carbon surface attracts and accumulates NO molecules, forming a more direct and unobstructed diffusion route toward the active sites at the solid‐liquid interface (Figure 4a). DFT‐calculated diffusion energies further confirm this mechanism: the maximum migration energy barrier decreases from 1.26 eV without carbon capsulation to only 0.38 eV with it (Figure 4c). This “carbon‐bridge” effect not only enhances NO mass transport and increases its surface coverage, which facilitates the overall NORR process, but it also suppresses the competing HER by limiting H‐H coupling, thereby improving the FE for NH3 production. NORR mechanism. Simulated NO diffusion path in the electrolyte with (a) Fe3C/Fe3N@C and (b) Fe3C/Fe3N as the catalysts. (c) Corresponding NO migration energy barriers of Fe3C/Fe3N@C and Fe3C/Fe3N. (d) In situ NO adsorption breakthrough curves of Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C. (e) The calculated NO adsorption energy of N‐end mode on Fe sites within Fe3C, Fe3N, and Fe3N‐VN. (f) The COHP analysis of Fe‐N bond during NO adsorption on Fe3C, Fe3N, and Fe3N‐VN sites within Fe3C/Fe3N@C. (g) EPR spectra recorded on Fe3C/Fe3N@C, Fe3C@C, and Fe3N@C upon electrolysis in the absence of NO. (h) Gibbs free energy for H2O dissociation on Fe3C/Fe3N@C with/without N vacancies. (i) Potential‐dependent population of interfacial water on Fe3C/Fe3N@C. Once NO molecules overcome the diffusion barrier and reach the catalyst surface, the next crucial step for efficient NORR is the adsorption and activation of NO at the active sites. The NO adsorption capacities of the three catalysts were first explored using in‐situ gas adsorption breakthrough tests (Figure 4d). The NO breakthrough curve of Fe3C/Fe3N@C displays a longer saturation time compared to those of Fe3N@C and Fe3C@C, and quantitative analysis shows that the NO adsorption amount of Fe3N@C, Fe3C@C, and Fe3C/Fe3N@C is 69, 211, and 428 µmol g−1, respectively. These results suggest that Fe3C sites inherently favor NO capture more effectively than Fe3N sites. To elucidate the thermodynamic origin of this behavior, DFT calculations were performed on Fe3C, Fe3N, and N‐vacancy‐containing Fe3N (Fe3N‐VN) sites within the heterostructure (Figure 4e and S40). The N‐end adsorption mode is more favorable than the O‐end mode due to its more negative adsorption energy on Fe3C (−5.01 vs. −3.91 eV, Figure S41). After structure optimization, the N‐end adsorption energies of NO on Fe3C, Fe3N, and Fe3N‐VN are −5.01, −3.7, and −4.14 eV, respectively. The stronger adsorption on Fe3C arises from the higher electron‐donating ability of C atoms, which increases the electron density at Fe, strengthens π back‐donation to NO, and facilitates its activation. Notably, while Fe3C offers the strongest individual adsorption affinity, the NO adsorption capacity of the Fe3C/Fe3N@C heterostructure (428 µmol g−1) is further governed by geometric factors. As confirmed by BET analysis (Table S5), Fe3C/Fe3N@C possesses a much larger specific surface area (312.95 m2 g−1) and pore volume (0.28 cm3 g−1) compared to Fe3C@C (233.08 m2 g−1 and 0.18 cm3 g−1, respectively). This increased surface area and pore volume ensure a higher density of accessible active sites, thereby maximizing the total NO adsorption capacity of Fe3C/Fe3N@C. Further insights from crystal orbital Hamilton population (COHP) analysis reveal that Fe3C exhibits a more negative integrated COHP (ICOHP) value for the Fe‐N bond (−5.52 eV) than Fe3N (−3.01 eV) and Fe3N‐VN (−2.79 eV), confirming the stronger Fe‐N bond on Fe3C (Figure 4f and Figure S42). Overall, these findings demonstrate that Fe3C sites in Fe3C/Fe3N@C serve as the primary NO adsorption centers, while the presence of Fe3N‐VN sites further improves NO adsorption capability. Following efficient NO adsorption and activation at the catalyst surface, the availability and reactivity of surface‐bound hydrogen (*H) play a crucial role in driving the hydrogenation and stepwise conversion of NO intermediates toward NH3 formation [45, 46]. As a result, the catalysts’ ability to dissociate H2O, which governs the *H supply, was investigated. EPR spectroscopy was first used to probe *H formation in Na2SO4 electrolyte (Figure 4g). The observed DMPO‐H signal with an intensity ratio close to 1:1:2:1:2:1:2:1:1 matches the calculated hyperfine coupling constants (a N = 16.5 G and a H = 21.3 G). The overall EPR signal intensities reveal that Fe3C@C exhibits limited *H generation, Fe3N@C demonstrates significantly improved *H generation, and Fe3C/Fe3N@C displays the strongest *H generation ability among the three. These results suggest that the Fe3N sites within Fe3C/Fe3N@C serve as the primary sites for H2O dissociation to produce *H. To explore the role of nitrogen vacancies in boosting *H supply, DFT calculations were employed to model Fe3C/Fe3N@C with (Fe3C/Fe3N@C) or without (Pristine Fe3C/Fe3N@C) N vacancies for comparison (Figures S43, S44). The calculated transition state (TS) activation energy barriers for H2O dissociation on Fe3C/Fe3N@C and Fe3C/Fe3N‐NV@C are 0.38 eV and 0.97 eV, respectively (Figure 4h and Figure S45), indicating that the presence of N vacancies makes H2O dissociation more kinetically favorable and thus enhances *H generation for the NORR process. To further investigate the H2O dissociation ability of Fe3C/Fe3N@C, in‐situ Raman spectroscopy measurements were conducted under various applied potentials. The interfacial water species consist of 4‐coordinated hydrogen‐bonded water (4‐HB·H2O), 2‐coordinated hydrogen‐bonded water (2‐HB·H2O), and Na+ ion‐hydrated water (Na·H2O). Among these, Na·H2O represents weakly hydrogen‐bonded and more easily dissociable water molecules that act as key *H sources during NORR [47, 48, 49]. Therefore, we performed Gaussian fitting of the broad O‐H stretching band from 2800 to 3800 cm−1, and three distinct peaks corresponding to 4‐HB·H2O, 2‐HB·H2O, and Na·H2O were resolved from low to high wavenumbers [50]. As shown in Figure 4i, increasing the negative potential results in a substantial growth of Na·H2O content on Fe3C/Fe3N@C (from 8.53% to 18.55%), which is much higher than that of Fe3C@C (from 6.44% to 10.49%) and Fe3N@C (from 4.94% to 12.59%) (Figures S46–S50). This trend demonstrates the superior ability of Fe3C/Fe3N@C to activate water and generate *H. Overall, the synergistic design of Fe3C/Fe3N@C enables efficient H2O dissociation and abundant H* production, which, combined with its enhanced NO diffusion, adsorption, and activation capabilities, ensures fast and continuous NO hydrogenation while simultaneously suppresses competing HER [49]. To clarify the possible reaction intermediates and pathway of NORR, in‐situ attenuated total reflection surface‐enhanced infrared absorption spectroscopy (ATR‐SEIRAS) measurements were conducted in the potential range from OCP to −0.9 V (vs. RHE). With increasing negative potential, characteristic peaks corresponding to NH4 + (1438 cm−1), NH3 (1633 and 1240 cm−1), N–H bending modes (1510 and 1066 cm−1), and *HNO (1125 cm−1) species become more intense on Fe3C/Fe3N@C (Figure 5a) [51, 52]. Meanwhile, a broad band ranging from 3100 to 3700 cm−1, attributed to O–H species, increases in intensity, demonstrating the consumption of protons during NO hydrogenation [53, 54]. The time‐resolved in situ ATR‐SEIRAS spectra further confirm the generation of these reaction intermediates on Fe3C/Fe3N@C (Figure S51). For both Fe3N@C (Figure 5b) and Fe3C@C (Figure 5c), similar peaks corresponding to NH3, N‐H bending modes, *HNO, and O‐H species are also observed, but their intensities remain significantly weaker, especially at less negative potentials, thus implying their lower capacities for NO adsorption, activation, and subsequent hydrogenation compared to Fe3C/Fe3N@C. Reaction pathway and mechanism. In situ ATR‐SEIRAS spectra recorded on (a) Fe3C/Fe3N@C, (b) Fe3N@C, and (c) Fe3C@C in 0.5 M Na2SO4 at applied potentials ranging from OCP to −0.9 V versus RHE. (d) Online DEMS measurements on Fe3C/Fe3N@C at −0.6 V versus RHE. (e) The calculated NORR Gibbs free energy diagrams on Fe3C/Fe3N@C, Fe3N@C, and Fe3C@C. To further confirm the key reaction intermediates, differential electrochemical mass spectrometry (DEMS) measurements were performed at −0.6 V versus RHE. As shown in Figure 5d, signals corresponding to *NOH, *N, *NH, *NH2 and *NH3 are detected on Fe3C/Fe3N@C. The NO signal gradually decreases as the reaction proceeds, demonstrating continuous consumption of NO during NORR (Figure S52). The intermediates observed in DEMS agree with those identified by in‐situ ATR‐SEIRAS measurements. Based on these results, the NORR pathway on Fe3C/Fe3N@C is proposed to follow a step‐by‐step hydrogenation process: NO → *NO → *NOH → *N → *NH → *NH2 → *NH3 → NH3. The reaction energy diagrams of NORR for Fe3C/Fe3N@C, Fe3N@C, and Fe3C@C are displayed in Figure 5e (the structural models of intermediates are shown in Figures S53–S55). The initial hydrogenation of NO preferentially forms *NOH rather than *HNO, while the second hydrogenation of NOH* to N* is more favorable than its hydrogenation to HNOH* (Figure S56). The hydrogenation of *N to *NH is the rate‐determining step (RDS) for all catalysts, with an uphill free energy of 0.89, 1.42, 1.50 eV on Fe3C/Fe3N@C, Fe3N@C, and Fe3C@C, respectively. Besides, for Pristine Fe3C/Fe3N@C (Figures S57–S58), the *NO → *NOH and *NH → *NH2 steps exhibit much higher energy barriers of 2.67 and 2.70 eV, respectively, indicating that nitrogen vacancies play a vital role in facilitating these transformations. Collectively, these findings demonstrate that the conversion of NO to NH3 is thermodynamically and kinetically more favorable on Fe3C/Fe3N@C with nitrogen vacancies than on Fe3C@C, Fe3N@C, and Fe3C/Fe3N@C without nitrogen vacancies. Based on the above analysis, the NORR reaction mechanism of Fe3C/Fe3N@C is illustrated in Figure 5e. The overall NORR process on Fe3C/Fe3N@C involves several synergistic steps: efficient NO diffusion to the catalyst surface, robust NO adsorption and activation, rapid *H generation via H2O dissociation, and stepwise hydrogenation. The graphitic carbon‐encapsulated Fe3C/Fe3N Janus nanostructure promotes all these key processes simultaneously, leading to significantly enhanced electrochemical NORR performance. Conclusion In summary, we developed a self‐supported Fe3C/Fe3N@C catalyst composed of defect‐rich, graphitic‐carbon‐confined Fe3C/Fe3N Janus nanostructures for efficient electrochemical NO‐to‐NH3 conversion. The catalyst achieves a high NH3 yield rate of 468.3 µmol h−1 cm−2 with a Faradaic efficiency of 94.2% at −0.6 V versus RHE, demonstrating the effectiveness of this multifunctional architecture. Beyond the performance improvement, this study shows that the key kinetic requirements of NORR can be addressed within a single catalyst design. The graphitic carbon shell promotes NO transport across the interfacial water environment, while the Fe3C/Fe3N Janus interface provides complementary active sites for NO activation and H2O dissociation. In this architecture, Fe3C favors NO adsorption and activation, whereas nitrogen‐vacancy‐rich Fe3N facilitates *H generation for stepwise hydrogenation toward NH3. These results establish carbon‐confined, earth‐abundant carbide/nitride Janus interfaces as a useful platform for coupling mass‐transfer regulation with interfacial reactivity control, offering a practical design principle for advanced NORR catalysts and other gas‐involving electrocatalytic transformations. Author Contributions Jialing Song: investigation, writing – original draft, methodology, visualization, formal analysis. Ziqi Wei: investigation, writing – review and editing, methodology. Haotian Huang: investigation, visualization, writing – review and editing. Lupeng Han: methodology, formal analysis, writing – review and editing. Biyi Huang: methodology, writing – review and editing. Evangelina Pensa: writing – review and editing, data curation, formal analysis. Sam Fong Yau Li: funding acquisition, writing – review and editing, formal analysis, data curation, resources, supervision. Eslam Hamed: formal analysis, writing – review and editing, investigation, validation, data curation. Fun Man Fung: writing – review and editing, validation, data curation. Emiliano Cortés: writing – review and editing, funding acquisition, conceptualization, supervision, resources, validation. Dengsong Zhang: conceptualization, investigation, writing– review and editing, funding acquisition, supervision, resources, project administration. Conflicts of Interest The authors declare no conflicts of interest. Supporting information The detailed experimental section and additional figures and tables are listed in the Supporting Information file. The authors have cited additional references within the Supporting Information.Supporting File: anie73787‐sup‐0001‐SuppMat.docx.