English

• Nitrogen reduction into ammonia (NH₃) is a pivotal chemical transformation for modern economy sustainable development because NH₃ is not only an important chemical product but also an attractive carbon-neutral energy carrier to cope with the impending fossil fuel crisis [1,2]. So far, the conventional industrial NH₃ production is primarily governed by the Haber-Bosch technology, which relies on the harsh reaction conditions, such as high pressure (150−200 atm) and high temperature (300−500 ℃). This process is responsible for ca. 1.6%–3% of global fossil fuel-derived CO₂ emissions and accounts for over 1% of total global energy consumption [3,4]. Therefore, exploring a green and energy-conserving alternative route to achieve ammonia synthesis is particularly urgent.

  Electrocatalytic nitrogen reduction reaction (NRR) at ambient conditions provides an attractive approach for sustainable ammonia production, but it requires highly active electrocatalysts to reduce the thermodynamic resistance [5]. Until now, highly efficient NRR catalysts mostly rely on precious metals, such as Ru, Au, Pd, Pt, and Rh, but their high price and limited accessibility significantly hinder their large-scale practical applications [6–8]. In comparison, low-cost transition metals and their derivatives leveraging natural abundance have been developed as promising NRR electrocatalysts owing to their rich bonding orbitals to N₂ [9,10]. Inspired by the Fe active centers in the structure of natural nitrogenases and traditional catalyst for Haber-Bosch process, Fe-based nanomaterials, such as iron oxide (Fe₂O₃), have attracted particular attention in electrochemical NRR since the partially filled d orbitals and widely variable oxidation state of Fe can promote strong N₂ adsorption by forming the Fe-N bond and generating oxygen vacancies [11–13]. Although various Fe₂O₃-based materials have been fabricated [14–16], their electrocatalytic NRR performance in terms of ammonia generation rate and reaction efficiency remains far from satisfaction because of the difficulty of N₂ adsorption and activation, as well as thermodynamically favored hydrogen evolution reaction (HER) competitive process [17–25]. Therefore, it is imperative to construct more effective Fe-based active sites that can accelerate the reaction kinetics and balance the competition between N₂ and H, thereby optimizing NRR electrocatalytic activity and selectivity.

  Cobalt (Co), with unique 3d electronic structure, plays a crucial role in the NRR electrocatalytic system. Its abundant oxidation states and high electronegativity are beneficial for generating oxygen vacancies and optimizing the electronic structure of neighboring atoms [26,27]. Additionally, the hybridization of 3d orbitals during the interaction between Co and Fe can enhance the chemical adsorption of N₂ and facilitate electron transfer to N-related intermediate species absorbed on active sites, thereby accelerating the hydrogenation process [10,28]. Inspired by the above considerations, in this paper, we successfully engineered and modulated the oxygen vacancies in rhombohedral Fe₂O₃ by incorporating Co dopants for electrocatalytic NRR. The resulting Co-doped oxygen vacancy-rich Fe₂O₃ (Vo:Co-Fe₂O₃–5) exhibits excellent electrocatalytic N₂-to-NH₃ conversion efficiency under ambient conditions, achieving a high NH₃ yield rate of 30.4 μg h^-1 mg^-1_cat. and a large FE of 16.4%. These values outperform those of pristine Fe₂O₃ (13.8 μg h^-1 mg^-1_cat. and 7.5%) and most reported similar NRR electrocatalysts. Comprehensive characterization analysis discloses that Co dopants not only create abundant oxygen vacancies but also induce the attenuation of electron cloud in Fe sites to well strengthen the N₂ adsorption and activation. Furthermore, Co dopants regulate the electronic structure and surface characteristics of Fe₂O₃, generating more active sites, facilitating electron transfer, and further promoting the NRR electrocatalytic activity by suppressing the competitive HER process.

  The synthesis process of Vo:Co-Fe₂O₃ is depicted in Fig. 1a, where the amount of the Co dopant can be easily regulated by adjusting the moles of cobalt in the raw materials. The phase composition and purity of the as-fabricated catalysts were characterized by powder X-ray diffraction (XRD) analysis. As depicted in Fig. 1b, all characteristic diffraction peaks of Vo:Co-Fe₂O₃–5 match the rhombohedral α-Fe₂O₃ phase (JCPDS No. 33–0664). No additional peaks related to cobalt oxides or other impurity phases were observed in the XRD patterns (Fig. S1 in Supporting information), demonstrating that Co was successfully incorporated into the Fe₂O₃ lattice without generating other crystalline phase. Whereas, the reflection peaks of Vo:Co-Fe₂O₃–5 shift slightly to higher degree compared to pristine Fe₂O₃ (inset in Fig. 1b), likely due to lattice distortion or stress induced by Co doping [29]. The refined lattice parameters (Figs. S2 and S3 and Table S1 in Supporting information) exhibit that the crystal cell volume of Vo:Co-Fe₂O₃–5 is slightly larger than that of Fe₂O₃, demonstrating that the Fe-O bond lengths have been prolonged due to distortion of the FeO₆ octahedron [30]. This observation is further supported by the Raman spectra (Fig. S4 in Supporting information).

  Figure 1

  

  Figure 1.  (a) Synthesis schematic illustration of Vo:Co-Fe₂O₃. (b) XRD patterns of Fe₂O₃ and Vo:Co-Fe₂O₃–5. (c-e) SEM images, (f) TEM, (g) HRTEM (the inset shows SAED pattern) and (h-l) the corresponding elemental mapping images of Vo:Co-Fe₂O₃–5.

  DownLoad: Full-Size Img PowerPoint

  The morphology and microstructure of the synthesized samples were examined by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). As shown in Fig. S5 (Supporting information), pure Fe₂O₃ exhibits a typical spindle-shaped structure. The SEM images of Vo:Co-Fe₂O₃ (Figs. 1c-e and Fig. S6 in Supporting information) confirm that this spindle-shaped structure is retained after Co doping. Energy dispersive X-ray (EDX) elemental analysis (Figs. S7 and S8 in Supporting information) reveals that Fe, O, and Co are uniformly distributed in Vo:Co-Fe₂O₃–5, with the Co atomic content increasing with doping concentration. Similarly, the same spindle-shaped structure is observed in the TEM image of Vo:Co-Fe₂O₃–5 and Fe₂O₃ (Fig. 1f and Fig. S9 in Supporting information). High-resolution TEM (HRTEM) image (Fig. 1g) shows clear lattice fringes with a spacing of 0.233 nm, corresponding to the (104) crystal planes of Fe₂O₃. The inset in Fig. 1g depicts the selected area electron diffraction (SAED) pattern of Vo:Co-Fe₂O₃–5, where the diffraction spots align well with rhombohedral α-Fe₂O₃, proving its high-quality single-crystalline nature. The corresponding 3D lattice fringe mapping (Fig. S10 in Supporting information) reveals an uneven surface of Fe₂O₃ after Co doping, indicating the presence of abundant defect sites. TEM elemental mappings (Figs. 1h-l) further demonstrate the uniform distribution of Fe, O, and Co in Vo:Co-Fe₂O₃–5, confirming the successful introduction of Co.

  X-ray photoelectron spectroscopy (XPS) was performed to investigate the elemental states and electronic structure of Fe₂O₃ and Vo:Co-Fe₂O₃–5. As shown in the Fe 2p spectrum of Fe₂O₃ (Fig. 2a), the peaks at 710.71 eV (Fe 2p_3/2) and 724.25 eV (Fe 2p_1/2) correspond to Fe²⁺, while the doublet at 712.48 eV (Fe 2p_3/2) and 726.47 eV (Fe 2p_1/2) is attributed to Fe³⁺ [31,32]. Compared to Fe₂O₃, the binding energies of Fe²⁺ and Fe³⁺ in Vo:Co-Fe₂O₃–5 exhibit positive shifts, which suggests the electronic interaction between Co and Fe ions and the attenuation of electron cloud, inducing more positively charged Fe³⁺ ions in Vo:Co-Fe₂O₃–5. Consequently, the Fe ions in Vo:Co-Fe₂O₃–5 own lesser electrons than in pristine Fe₂O₃, which increases the electron affinity of Fe ions and thus facilitating N₂ reduction via "acceptance-donation" mechanism [33]. Quantitative analysis of the Fe 2p spectra (Table S2 in Supporting information) reveals that the contents of Fe³⁺ and Fe²⁺ in Vo:Co-Fe₂O₃–5 are 63.7% and 37.3%. The increased Fe²⁺ content relative to pristine Fe₂O₃ is attributed to the generation of oxygen vacancy. The optimized Fe³⁺/Fe²⁺ ratio establishes a faster electron transport channel through Fe²⁺-O^2--Fe³⁺ super exchange, promoting electron transport in the NRR process [34]. The O 1s spectra (Fig. 2b and Fig. S11 in Supporting information) are deconvoluted into three components at 530.2, 532.2, and 533.4 eV, which are assigned to the lattice oxygen (O_L), vacancy oxygen (O_V), and chemisorbed oxygen species (O_C), respectively [35]. According to the integral area of O 1s spectra, the vacancy oxygen density increases with the amount of Co doping. The O_V/(O_L + O_V + O_C) ratio increases from 56.1% in Fe₂O₃ to 68.6% in Vo:Co-Fe₂O₃–5, confirming enhanced oxygen vacancy concentration, which is further supported by EPR spectroscopy (Fig. S12 in Supporting information). As such, the combination of more positively charged Fe³⁺ and increased oxygen vacancies in Vo:Co-Fe₂O₃–5 enhances its Lewis acid characteristic, benefiting N₂ adsorption and reduction [36]. The Co 2p spectrum (Fig. 2c) confirms the presence of Co²⁺ and Co³⁺ in Vo:Co-Fe₂O₃–5 [37].

  Figure 2

  

  Figure 2.  (a) Fe 2p, (b) O 1s and (c) Co 2p XPS spectra of Fe₂O₃ and Vo:Co-Fe₂O₃–5. (d) XANES spectra, (e, f) Fourier transform spectroscopy and (g, h) Cauchy wavelet transform contour plots of Fe₂O₃ and Vo:Co-Fe₂O₃–5.

  DownLoad: Full-Size Img PowerPoint

  The oxidation state of Fe and its local coordination environment in Fe₂O₃ and Vo:Co-Fe₂O₃–5 were further probed using X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) spectroscopy. As shown in Fig. 2d, the Fe K-edge XANES spectra show that the absorption intensity at the absorption edge onset position of Vo:Co-Fe₂O₃–5 is consistent with Fe₂O₃, revealing the coexistence of Fe²⁺ and Fe³⁺ in the samples, which aligns with the XPS analysis [38,39]. Meanwhile, the white-line peak of Vo:Co-Fe₂O₃–5 is observed to be elevated, which may be attributed to the Fe (1s) → Fe (4p) transition. This result suggests that the (1s) → (4p) transition of Fe³⁺ in Vo:Co-Fe₂O₃–5 typically requires higher energy compared to pristine Fe₂O₃, because the energy level of the 4p orbital is upshifted in higher oxidation states, which further confirms that the incorporation of Co induces more positively charged Fe³⁺ in Vo:Co-Fe₂O₃–5 [40]. The radial distribution function curve of Vo:Co-Fe₂O₃–5 (Fig. 2e) shows a significant decrease in the intensity of the Fe-O and Fe-O-Fe shell peaks, further confirming the presence of defects. These defects likely result from local structural perturbations due to the introduction of Co dopants, which in turn affect the coordination environment of Fe (Fig. 2f and Table S3 in Supporting information). Further observations reveal that the Fe-O bond length in Vo:Co-Fe₂O₃–5 is longer than that of pristine Fe₂O₃, demonstrating that the interaction between Fe and oxygen atoms is weakened, which suggests the Fe-O bonds in Vo:Co-Fe₂O₃–5 are easier to be broken and consequently enhancing its reducibility [39]. Figs. 2g and h present the Cauchy wavelet transform (WT) EXAFS spectra of Fe₂O₃ and Vo:Co-Fe₂O₃–5. As depicted in WT contour plots, Vo:Co-Fe₂O₃–5 exhibits higher signal intensity at K values of ca. 2.5 Å and 3 Å than pristine Fe₂O₃, corresponding to the Fe-O and Fe-O-Fe bonds. This further indicates that the introduction of Co dopants regulates the local coordination environment of Fe, consistent with the EXAFS results [41–43].

  The NRR electrochemical experiments were carried out under ambient conditions in a H-type electrolytic cell configuration using N₂-saturated 0.1 mol/L Na₂SO₄ as electrolyte. The indophenol blue method was used to quantitatively analyze the NH₃ concentration (Fig. S13 in Supporting information). Linear scanning voltammetry (LSV) curves were acquired in Ar-/N₂-saturated electrolyte. As shown in Fig. 3a, Vo:Co-Fe₂O₃–5 exhibits a higher current density in N₂ atmosphere than in Ar atmosphere, suggesting the possible occurrence of NRR [44]. Moreover, the observed current density of Vo:Co-Fe₂O₃–5 is higher than that of Fe₂O₃ under the same potential in the N₂-bubbled electrolyte, revealing a better NRR activity. The current difference in two electrolytes is significantly larger for Vo:Co-Fe₂O₃–5 than for pristine Fe₂O₃, with a proportion (△1/△2) of 1.403, suggesting that NRR becomes dominant while HER is suppressed due to the Co doping in the Fe₂O₃ catalyst. To quantitatively determine the NRR performance of Vo:Co-Fe₂O₃–5, chronoamperometry (Fig. 3b) and UV–vis tests (Fig. 3c) were conducted within the potential range of −0.30 V to −0.70 V vs. RHE. As shown in Fig. 3d, Vo:Co-Fe₂O₃–5 achieves the optimal NH₃ yield rate (30.4 μg h^-1 mg_cat^-1) and FE (16.4%) at −0.50 V vs. RHE. When the potential is more negative than −0.50 V vs. RHE, the NH₃ yield rates and FEs of Vo:Co-Fe₂O₃–5 decreased significantly, likely due to the competitive HER dominating and blocking the surface-active sites, thereby inhibiting NRR [45]. For getting clearer comparison, the NH₃ yields and FEs at an applied potential of −0.50 V vs. RHE were evaluated for pristine Fe₂O₃ and other Vo:Co-Fe₂O₃ samples. As shown in Fig. 3e, pristine Fe₂O₃ shows the low NH₃ formation rate and FE. In contrast, the electrocatalytic activity of Vo:Co-Fe₂O₃ is significantly enhanced. The NH₃ yields and FEs initially increase and then decrease with increasing Co doping content. Among all samples, Vo:Co-Fe₂O₃–5 achieves the highest NH₃ yield and maximum FE, leading to a 2.2-fold increase compared to Fe₂O₃, and also outperforming most recently reported similar electrocatalysts (Table S4 in Supporting information). This improvement is ascribed to the optimal balance between Co doping ratio and surface oxygen vacancies density [46]. N₂ temperature-programmed desorption (N₂-TPD) measurements were conducted to access the N₂ adsorption capabilities of Vo:Co-Fe₂O₃–5 and pristine Fe₂O₃. As illustrated in Fig. 3f, both samples display two peaks in the range of 50–500 ℃, corresponding to the desorption of physically and chemically adsorbed N₂, respectively. Notably, Vo:Co-Fe₂O₃–5 shows higher desorption temperatures and larger peak areas for both physical and chemical adsorption of N₂, highlighting the critical role of Co dopants in strengthening N₂ adsorption. The wettability of the catalysts was evaluated via contact angle tests (Fig. S14 in Supporting information). The contact angle of Fe₂O₃ increases from ~37° to ~47° after Co doping, demonstrating improved hydrophobicity of Vo:Co-Fe₂O₃–5. The variation in surface wettability promotes N₂ molecules and reduces direct proton adsorption during the electroreduction process, thereby improving NRR efficiency by suppressing the competitive HER.

  Figure 3

  

  Figure 3.  (a) LSV curves in 0.1 mol/L Na₂SO₄ electrolyte saturated with N₂/Ar. (b) Chronoamperometry curves, (c) UV–vis spectra of electrolytes stained with the indophenol indicator. NH₃ yields and FEs of (d) Vo:Co-Fe₂O₃–5 at different potentials and (e) various samples at the potential of −0.5 V. (f) N₂-TPD profiles of Fe₂O₃ and Vo:Co-Fe₂O₃–5. (g) Amount of NH₃ generated after charging at −0.5 V for 2 h under various conditions. (h) NH₃ yields and FEs alternately cycled between the N₂- and Ar saturated electrolytes for 2 h of Vo:Co-Fe₂O₃–5 at −0.5 V. (i) NMR spectra of the standard ¹⁵NH₄⁺ and ¹⁴NH₄⁺ samples and the electrolyte fed by ¹⁵N₂ and ¹⁴N₂ after electrolysis on Vo:Co-Fe₂O₃–5 at −0.5 V.

  DownLoad: Full-Size Img PowerPoint

  To verify the origin of the produced NH₃ and eliminate the influence of possible NH₃ contaminations, comparative experiments and ¹⁵N isotopic labeling experiments were carried out. As described in Fig. 3g, Vo:Co-Fe₂O₃–5 exhibits good NRR performance in 0.1 mol/L Na₂SO₄ electrolyte saturated with N₂ at −0.50 V vs. RHE. However, only a trace amount of NH₃ is detected under other controlled conditions. The alternately cycled electrocatalytic experiments (Fig. 3h) further demonstrate that NH₃ is generated exclusively in N₂-saturated electrolyte, confirming that the nitrogen in generated NH₃ originates from the electrochemical reduction of N₂ and Vo:Co-Fe₂O₃–5 plays a crucial role in this process [47]. ¹⁵N isotopic labeling experiment was performed to confirm the nitrogen source of the produced NH₃. As shown in Fig. 3i, when using ¹⁴N₂ as the feeding gas, the ¹H nuclear magnetic resonance (NMR) spectra of the electrolyte display typical triplet peaks of ¹⁴NH₄⁺ with a coupling constant of 52 Hz, matching the ¹H NMR spectrum of standard ¹⁴NH₄Cl. By feeding ¹⁵N₂ gas during the electrocatalytic NRR, characteristic doublet peaks with a coupling constant of 73 Hz are observed in the ¹H NMR spectra, consistent with the ¹H NMR spectrum of standard ¹⁵NH₄Cl. The ¹⁵N isotopic labeling experiment results firmly prove that the produced NH₃ originates from the electrocatalytic NRR rather than NH₃ contamination.

  To further investigate the electrochemical reaction kinetics, Tafel plots were obtained. The lowest Tafel slope for Vo:Co-Fe₂O₃–5 implies its fastest interfacial reaction kinetics (Fig. S15 in Supporting information). Electrochemical impedance spectroscopy (EIS) reveals that the charge transfer impedance of Vo:Co-Fe₂O₃ is smaller than that of Fe₂O₃, suggesting that Co doping facilitates charge transfer kinetics (Fig. S16 in Supporting information) [48]. The electrochemical active surface area (ECSA) was further assessed via double-layer capacitance (C_dl) [49]. The C_dl of Vo:Co-Fe₂O₃–5 is larger than that of Fe₂O₃ (Figs. S17 and S18 in Supporting information), indicating more exposed active sites, which aligns with the Brunauer-Emmett-Teller (BET) surface area (Fig. S19 in Supporting information). The ECSA-normalized polarization curves (Fig. S18b) exhibit that Vo:Co-Fe₂O₃–5 shows a higher normalized current density than that of Fe₂O₃, suggesting its more favorable intrinsic activity. The competitive HER was monitored using in situ differential electrochemical mass spectrometry (DEMS) [50]. As illustrated in Fig. S20 (Supporting information), the H₂ signal intensity for Vo:Co-Fe₂O₃–5 is lower than that for Fe₂O₃, confirming that Co doping effectively suppresses HER. The stability and reproducibility of Vo:Co-Fe₂O₃–5 were also investigated. As presented in Fig. S21 (Supporting information), the current density remains stable during the 50 h of continuous electrolysis, with negligible decline in NH₃ yield and FE (inset in Fig. S21). Inductively coupled plasma mass spectrometry (ICP-MS) analysis (Fig. S22 in Supporting information) confirms no significant leaching of Co and Fe after the electrochemical process. After the stability tests, the structure, composition, and valence states of Vo:Co-Fe₂O₃–5 are well maintained (Figs. S23-S25 in Supporting information), indicating its good stability. Six consecutive cyclic tests show no significant decrease in NH₃ yield or FE (Fig. S26 in Supporting information), demonstrating its excellent reproducibility.

  Density functional theory (DFT) calculations were performed to corroborate the experimental findings. The results suggest that N₂ molecule preferentially adsorb on the oxygen vacancy of Vo:Co-Fe₂O₃ with a higher adsorption energy of −2.584 eV (Fig. 4a and Fig. S27 in Supporting information) compared to −2.323 eV for pure Fe₂O₃. This demonstrates that the Vo:Co-Fe₂O₃ surface has a stronger adsorption capacity than the Fe₂O₃ surface, in agreement with the experimental results. Compared to a free N₂ molecule, the N—N bond of the adsorbed N₂ is lengthened to 1.15172 Å, whereas on the Fe₂O₃ surface, the N—N bond length is 1.15107 Å. These results suggest that the N₂ adsorption and activation are more favorable on the Vo:Co-Fe₂O₃ surface. This is further evidenced by the charge density difference plots in Fig. 4b, which show a weakened N≡N bond and electron enrichment between the nitrogen atoms. The Gibbs free-energy values of hydrogen adsorption (∆G_H*) on the surface of Vo:Co-Fe₂O₃ and Fe₂O₃ are shown in Fig. 4c. The calculated ΔG_H* values are −2.611 and −2.928 eV for Vo:Co-Fe₂O₃ and Fe₂O₃, demonstrating that the H* binding energy was regulated by Co doping. The above computational findings collectively confirm that the synergistic effect of Co doping and oxygen vacancies can significantly optimize the NRR performance of Vo:Co-Fe₂O₃ by regulating the HER and NRR.

  Figure 4

  

  Figure 4.  (a) Adsorption energy and structures of N₂ adsorbed on Vo:Co-Fe₂O_3–Ov and Fe₂O_3–Ov surface. (b) Charge density difference plots of N₂ adsorption on the Vo:Co-Fe₂O_3–Ov surface. (c) Free energy profile of HER on the Vo:Co-Fe₂O₃ and Fe₂O₃ surfaces. (d) Mass spectra obtained during NRR tests with Vo:Co-Fe₂O₃–5 at −0.5 V vs. RHE in N₂-saturated 0.1 mol/L Na₂SO₄. (e) Possible reaction pathways of NRR on the surface of Vo:Co-Fe₂O₃–5.

  DownLoad: Full-Size Img PowerPoint

  Based on the above discussion, the improved NRR performance of Vo:Co-Fe₂O₃–5 is rationalized as follows: (1) The more positively charged Fe³⁺ ions and oxygen vacancies serve as Lewis acid sites to strength the chemisorption of N₂ molecules [11,36]; (2) under electrocatalytic conditions, Fe³⁺ ions in the Vo:Co-Fe₂O₃–5 lattice acquire electrons to form Fe²⁺ ions, which is conducive to injecting electrons into the π* antibonding orbitals of N₂, weakening the N≡N bond and promoting nitrogen activation [51]; (3) under the influence of Co doping and oxygen vacancies, the molar ratio of Fe³⁺/Fe²⁺ was optimized. The Fe³⁺/Fe²⁺ redox pathway establishes fast electron channels via super-exchange between Fe²⁺-O^2--Fe³⁺ (Fig. S28 in Supporting information), enhancing hydrogenation efficiency at NRR active sites [34]. Finally, volatile intermediates and products during the electrochemical NRR were monitored using online DEMS. As shown in Fig. S29 (Supporting information), the absence of m/z signals at 27, 31, and 33 under −0.5 V vs. RHE in N₂-saturated 0.1 mol/L Na₂SO₄ rules out N₂H₄ formation. Therefore, the m/z signal at 15 confirms the generation of NH₃ (Fig. 4d) [52]. Importantly, no ammonia-related fragments were detected in Ar-saturated electrolyte, and the detected signal at m/z = 18 is assigned to H₂O (Fig. S30 in Supporting information), indicating that the produced NH₃ is derived from N₂. Moreover, the signal corresponded to the N₂H₂ intermediate is detected at m/z = 30. Combining these results with the research by Shao et al., the possible reaction pathway for NRR on Vo:Co-Fe₂O₃–5 is proposed and illustrated in Fig. 4e. The absorbed N₂ molecule on defective sites first initiates a two-electron reduction to form N₂H₂* intermediates, which subsequently undergo desorption and cleavage to form NH₃ (3N₂H₂ → 2NH₃ + 2N₂) [53]. Nevertheless, the competing distal associative route cannot be completely excluded, which involves the sequential protonation at the terminal N atom, followed by NH₃ desorption and subsequent hydrogenation for another NH₃ molecule [54–56].

  In summary, defect-engineered Fe₂O₃ with tunable oxygen vacancy concentrations was successfully synthesized through a facile one-step hydrothermal method via cobalt incorporation, which was proven to serve as an efficient noble metal-free catalyst to achieve the electrochemical N₂-to-NH₃ conversion under ambient conditions. The Vo:Co-Fe₂O₃–5 catalyst with the optimal Co dopant amount exhibited a high NH₃ yield rate of 30.4 μg h^-1 mg^-1_cat. and a large FE of 16.4% at −0.5 V vs. RHE in 0.1 mol/L Na₂SO₄, which is significantly better than bare Fe₂O₃. The superior NRR electrocatalytic performance of Vo:Co-Fe₂O₃–5 results from the rich active sites induced by Co dopants and the balance of Co doping ratio and surface oxygen vacancies density. Specifically, the doping of Co atoms generated abundant oxygen vacancy and regulated the electronic structures of Fe₂O₃, which provided rich active sites for N₂ chemisorption and subsequent activation by promoting the interfacial electron transfer from catalysts to the adsorbed N₂ molecules. Moreover, the competitive HER process could be effectively restrained, responsible for the remarkable electrocatalytic activity of Vo:Co-Fe₂O₃–5.

  CRediT authorship contribution statement

  Yating Yang: Writing – original draft, Software, Methodology, Investigation. Ming Xu: Software, Investigation. Ze Gao: Software, Methodology. Ting Wang: Writing – review & editing, Software, Resources, Funding acquisition. Zhaoxing Li: Software, Investigation. Yuqi Wang: Software, Investigation. Daming Yang: Software. Ming Feng: Writing – review & editing, Supervision, Funding acquisition.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  This work is financially supported by the Program for the Development of Science and Technology of Jilin Province (No. 20240101004JJ), the National Natural Science Foundation of China (No. 52471229), and the Science and Technology Project of Jilin Provincial Education Department (No. JJKH20250931KJ).

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2026.112580.

  
  
• 1. [1]

     W.C. Xu, G.L. Fan, J.L. Chen, et al., Angew. Chem. Int. Ed. 59 (2020) 3511–3516. doi: 10.1002/anie.201914335

  2. [2]

     X.H. Li, T.S. Li, Y.J. Ma, et al., Adv. Energy Mater. 8 (2018) 1801357. doi: 10.1002/aenm.201801357

  3. [3]

     S.H. Wu, M. Zhang, S.M. Huang, et al., Chin. Chem. Lett. 34 (2023) 107282. doi: 10.1016/j.cclet.2022.03.005

  4. [4]

     S.B. Zhang, Y.K. Zha, Y.X. Ye, et al., Nano-Micro Lett. 16 (2024) 9. doi: 10.1007/s40820-023-01217-z

  5. [5]

     M. Zhang, J. Dai, S.M. Huang et al., Chin. Chem. Lett. 33 (2022) 1001–1005. doi: 10.1016/j.cclet.2021.08.069

  6. [6]

     H. Zhao, D. Zhang, Z.C. Wang, et al., Appl. Catal. B. 265 (2020) 118481. doi: 10.1016/j.apcatb.2019.118481

  7. [7]

     M.L. Yang, Z.X. Jin, X.X. Cao, et al., Tungsten 6 (2024) 428–437. doi: 10.1007/s42864-023-00206-4

  8. [8]

     X.T. Feng, J.Y. Liu, L. Chen, et al., J. Am. Chem. Soc. 145 (2023) 10259–10267. doi: 10.1021/jacs.3c01319

  9. [9]

     F.H. Jin, H.Q. Yin, R. Feng, et al., J. Colloid Interface Sci. 647 (2023) 354–363. doi: 10.1016/j.jcis.2023.05.108

  10. [10]

      G.X. Song, R. Gao, Z. Zhao, et al., Appl. Catal. B 301 (2022) 120809. doi: 10.1016/j.apcatb.2021.120809

  11. [11]

      T. Takashima, H. Fukasawa, T. Mochida, et al., ACS Appl. Nano Mater. 6 (2023) 23381–23389. doi: 10.1021/acsanm.3c04712

  12. [12]

      Y.Y. Ji, Q. Hu, W.D. Cheng, et al., ACS Sustainable Chem. Eng. 12 (2024) 176806–6810.

  13. [13]

      Y.Y. Ji, L. Li, W.D. Cheng, et al., Inorg. Chem. Front. 8 (2021) 3156. doi: 10.1039/d1qi00419k

  14. [14]

      X.J. Xiang, Z. Wang, X.F. Shi, et al., ChemCatChem 10 (2018) 4530–4535. doi: 10.1002/cctc.201801208

  15. [15]

      J. Li, X.J. Zhu, T. Wang, et al., Inorg. Chem. Front. 6 (2019) 2682–2685. doi: 10.1039/c9qi00968j

  16. [16]

      Y. Li, Y. Kong, Y. Hou, et al., ACS Sustain. Chem. Eng. 7 (2019) 8853–8859. doi: 10.1021/acssuschemeng.9b00852

  17. [17]

      L. Chen, C.Z. He, R. Wang, et al., Chin. Chem. Lett. 32 (2021) 53–56. doi: 10.1016/j.cclet.2020.11.013

  18. [18]

      M.S. Iqbal, Z.B. Yao, Y.K. Ruan, et al., Rare Met. 42 (2023) 1075–1097. doi: 10.1007/s12598-022-02215-7

  19. [19]

      J.M. Liu, L.H. He, S.R. Zhao, et al., Adv. Sci 10 (2023) 2205786. doi: 10.1002/advs.202205786

  20. [20]

      H.M. He, H.M. Wen, P. Li, et al., Chem. Eng. J. 476 (2023) 146760. doi: 10.1016/j.cej.2023.146760

  21. [21]

      H.M. He, H.M. Wen, H.K. Li, et al., Adv. Sci. 10 (2023) 2206933. doi: 10.1002/advs.202206933

  22. [22]

      J.X. Wang, Q.Q. Zhu, J.X. Wang, et al., Angew. Chem. Int. Ed. 64 (2025) e202415208. doi: 10.1002/anie.202415208

  23. [23]

      H.M. He, Q.Q. Zhu, Y. Yan, et al., Appl. Catal. B 302 (2022) 120840. doi: 10.1016/j.apcatb.2021.120840

  24. [24]

      H.M. He, H.K. Li, Q.Q. Zhu, et al., Appl. Catal. B: Environ. 316 (2022) 121673. doi: 10.1016/j.apcatb.2022.121673

  25. [25]

      X.Y. Chi, J.X. Li, X.X. Wang, et al., Chem. Eng. J. 498 (2024) 155424. doi: 10.1016/j.cej.2024.155424

  26. [26]

      S.H. Jiang, R.Y. Zhang, H.X. Liu, et al., J. Am. Chem. Soc. 142 (2020) 6461–6466. doi: 10.1021/jacs.9b13915

  27. [27]

      J. Zhang, X.Y. Tian, M.J. Liu, et al., J. Am. Chem. Soc. 141 (2019) 19269–19275. doi: 10.1021/jacs.9b02501

  28. [28]

      J. Yu, X.H. Ren, J.W. Lu, et al., J. Alloys Compd. 902 (2022) 163862. doi: 10.1016/j.jallcom.2022.163862

  29. [29]

      J.X. Wang, X.Y. Leng, S. Kan, et al., J. Alloys Compd. 976 (2024) 172990. doi: 10.1016/j.jallcom.2023.172990

  30. [30]

      N. Kirik, A. Krylov, A. Boronin, et al., Materials 16 (2023) 4466. doi: 10.3390/ma16124466

  31. [31]

      Z. Chen, W. Xu, W. Wang, et al., Small 21 (2024) 2409325.

  32. [32]

      T. Wang, B.W. Li, P. Wang, et al., J. Colloid Interface Sci. 672 (2024) 715–723. doi: 10.1016/j.jcis.2024.06.079

  33. [33]

      K. Wang, C. Ni, X. Qian, et al., Chem. Sci. 16 (2025) 7467–7476. doi: 10.1039/d4sc08789e

  34. [34]

      H.X. Zhang, W.C. Zhou, L.Q. Hu, et al., J. Colloid Interface Sci. 677 (2025) 264–272. doi: 10.1016/j.jcis.2024.07.226

  35. [35]

      M. Yu, X.Q. Mu, W.T. Meng, et al., Renewables 1 (2023) 465–473. doi: 10.31635/renewables.023.202300025

  36. [36]

      D.Y. Wang, L.Z. Zhang, J.Y. Xu, et al., Chem. Eng. J. 512 (2025) 162687. doi: 10.1016/j.cej.2025.162687

  37. [37]

      P. Zeng, Y. Meng, Z. Liu, et al., Small 19 (2023) 2301279. doi: 10.1002/smll.202301279

  38. [38]

      W.Y. Du, Z.Y. Sun, K. Chen, et al., Adv. Energy Mater. 14 (2024) 2401765. doi: 10.1002/aenm.202401765

  39. [39]

      K. Chen, D.Y. Ma, Y. Zhang, et al., Adv. Mater. 36 (2024) 2402160. doi: 10.1002/adma.202402160

  40. [40]

      Y.L. Zhao, H.C. Chen, X.L. Ma, et al., Adv. Mater. 36 (2024) 2308243. doi: 10.1002/adma.202308243

  41. [41]

      Z.Y. Sun, R. Niu, S.Y. Shang, et al. Nat. Commun. 16 (2025) 8837. doi: 10.1038/s41467-025-63923-z

  42. [42]

      Q.B. Tian, J.L. Chang, X.M. Peng, et al., Angew. Chem. Int. Ed. 64 (2025) e202503995. doi: 10.1002/anie.202503995

  43. [43]

      Q.P. Huang, G.J. Xia, B. Huang, et al., Energy Environ. Sci. 7 (2024) 5260–5272. doi: 10.1039/d4ee01588f

  44. [44]

      Z.H. Ni, N. Liu, C.H. Zhao, et al., Polyoxometalates 3 (2024) 9140044. doi: 10.26599/pom.2023.9140044

  45. [45]

      G. Ren, J. Zhao, Z. Zhao, et al., Angew. Chem. Int. Ed. 63 (2024) e202314408. doi: 10.1002/anie.202314408

  46. [46]

      T. Li, J. Xia, Q. Chen, et al., ACS Appl. Mater. Interfaces 13 (2021) 40724–40730. doi: 10.1021/acsami.1c12279

  47. [47]

      X. Xu, X. Liu, J. Xu, et al., Chem. Eng. J. 437 (2022) 135417. doi: 10.1016/j.cej.2022.135417

  48. [48]

      X. Zhu, L.Y. Chen, Y.G. Liu, et al., Polyoxometalates 2 (2023) 9140031. doi: 10.26599/pom.2023.9140031

  49. [49]

      Y. Yang, R. Su, D.R. Wang, et al., Renewables 2 (2024) 250–263. doi: 10.31635/renewables.024.202400063

  50. [50]

      Z. Gao, R. Liu, D. Yang, et al., J. Alloys Compd. 1010 (2025) 177106. doi: 10.1016/j.jallcom.2024.177106

  51. [51]

      R.F. Li, W. Teng, Y.X. Shi, et al., Appl. Catal. B 373 (2025) 125335. doi: 10.1016/j.apcatb.2025.125335

  52. [52]

      D.Z. Yao, C. Tang, L.Q. Li, et al., Adv. Energy Mater. 10 (2020) 2001289. doi: 10.1002/aenm.202001289

  53. [53]

      Y. Yao, S. Zhu, H. Wang, et al., Angew. Chem. Int. Ed. 59 (2020) 10479. doi: 10.1002/anie.202003071

  54. [54]

      T.Y. Dai, T.H. Wang, Z. Wen, et al., Adv. Funct. Mater. 34 (2024) 2400773. doi: 10.1002/adfm.202400773

  55. [55]

      H.B. He, X.L. Ding, Y.Y. Wang, et al., ChemPhysChem 25 (2024) e202400143. doi: 10.1002/cphc.202400143

  56. [56]

      Y.H. Zhao, J.Y. Qu, H.B. Li, et al., Nano Lett. 22 (2022) 4475–4481. doi: 10.1021/acs.nanolett.2c01185

• 

  Figure 1  (a) Synthesis schematic illustration of Vo:Co-Fe₂O₃. (b) XRD patterns of Fe₂O₃ and Vo:Co-Fe₂O₃–5. (c-e) SEM images, (f) TEM, (g) HRTEM (the inset shows SAED pattern) and (h-l) the corresponding elemental mapping images of Vo:Co-Fe₂O₃–5.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) Fe 2p, (b) O 1s and (c) Co 2p XPS spectra of Fe₂O₃ and Vo:Co-Fe₂O₃–5. (d) XANES spectra, (e, f) Fourier transform spectroscopy and (g, h) Cauchy wavelet transform contour plots of Fe₂O₃ and Vo:Co-Fe₂O₃–5.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  (a) LSV curves in 0.1 mol/L Na₂SO₄ electrolyte saturated with N₂/Ar. (b) Chronoamperometry curves, (c) UV–vis spectra of electrolytes stained with the indophenol indicator. NH₃ yields and FEs of (d) Vo:Co-Fe₂O₃–5 at different potentials and (e) various samples at the potential of −0.5 V. (f) N₂-TPD profiles of Fe₂O₃ and Vo:Co-Fe₂O₃–5. (g) Amount of NH₃ generated after charging at −0.5 V for 2 h under various conditions. (h) NH₃ yields and FEs alternately cycled between the N₂- and Ar saturated electrolytes for 2 h of Vo:Co-Fe₂O₃–5 at −0.5 V. (i) NMR spectra of the standard ¹⁵NH₄⁺ and ¹⁴NH₄⁺ samples and the electrolyte fed by ¹⁵N₂ and ¹⁴N₂ after electrolysis on Vo:Co-Fe₂O₃–5 at −0.5 V.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) Adsorption energy and structures of N₂ adsorbed on Vo:Co-Fe₂O_3–Ov and Fe₂O_3–Ov surface. (b) Charge density difference plots of N₂ adsorption on the Vo:Co-Fe₂O_3–Ov surface. (c) Free energy profile of HER on the Vo:Co-Fe₂O₃ and Fe₂O₃ surfaces. (d) Mass spectra obtained during NRR tests with Vo:Co-Fe₂O₃–5 at −0.5 V vs. RHE in N₂-saturated 0.1 mol/L Na₂SO₄. (e) Possible reaction pathways of NRR on the surface of Vo:Co-Fe₂O₃–5.

  下载: 全尺寸图片 幻灯片
•