Electrosynthesis of NH₃ from low-concentration NO on cascade dual-site catalysts in neutral media

Abstract

Electrosynthesis of NH₃ from low-concentration NO (NORR) in neutral media offers a sustainable nitrogen fixation strategy but is hindered by weak NO adsorption, slow water dissociation, and sluggish hydrogenation kinetics. Herein, we propose an intriguing strategy that successfully overcomes these limitations through using an electron-donating motif to modulate NO-affinitive catalysts, thereby creating dual active site with synergistic functionality. Specifically, we integrate electron-donating nanoparticles into a Fe single-atom catalyst (Fe_SAC), where Fe sites ensure strong NO adsorption, while electron-donating motifs promote water dissociation and NO hydrogenation. In situ X-ray absorption spectroscopy (XAS), in situ attenuated total reflection-infrared spectroscopy (ATR-IR), and theoretical calculations reveal that electron-donating motifs increase Fe site electron density, strengthening NO adsorption. Additionally, these motifs also promote water dissociation, supplying protons to lower the NO hydrogenation barrier. This synergistic interplay enables a cascade reaction mechanism, delivering a remarkable Faradaic efficiency (FE) of 90.3% and a NH₃ yield rate of 709.7 µg h⁻¹ mg_cat.⁻¹ under 1.0 vol% NO in neutral media, outperforming pure Fe_SAC (NH₃ yield rate: 444.2 µg h⁻¹ mg_cat.⁻¹, FE: 56.6%) and prior to systems operating under high NO concentrations. Notably, the high NH₃ yield of 3207.7 μg h⁻¹ mg_cat.⁻¹ is achieved in a membrane electrode assembly (MEA) electrolyzer under a 1.0 vol% NO. This work establishes an inspirational paradigm in NORR by simultaneously enhancing NO adsorption, water dissociation, and hydrogenation kinetics, providing a scalable route for efficient NH₃ electrosynthesis from dilute NO sources.

Weak NO adsorption, slow water dissociation, and sluggish hydrogenation hinder neutral-media NH₃ electrosynthesis from low-concentration NO. Here, the authors report an electron-donating motif modulating a NO-affinitive catalyst to create dual active sites, thus facilitating NO-to-NH₃ conversion.

Introduction

Nitric oxide (NO) is a harmful atmospheric pollutant primarily emitted from industrial processes and vehicle exhaust, posing severe environmental and health risks^1–6. Selective catalytic reduction (SCR) is the predominant method for mitigating NO emissions^6–10. Nevertheless, it requires temperatures of 300-400 °C and uses ammonia as a reducing agent, thus restricting its efficiency under near-ambient conditions and failing to generate any value-added products. Electrochemical NO reduction (NORR) under mild conditions, utilizing renewable electricity and water, offers a sustainable route to both NO removal and ammonia (NH₃) production—an essential fertilizer and potential hydrogen carrier—via a carbon-free process^11–13. Large-scale NORR requires neutral water conditions to enable earth-abundant transition metal catalysts, which are unstable in acidic media, and to facilitate direct seawater use without desalination^14–18. However, its efficiency in neutral media is severely restricted by low NO solubility, sluggish water dissociation, and slow NO hydrogenation kinetics.

Most current research focuses on engineering electron-deficient sites to strengthen NO adsorption for NORR under neutral media, even at low NO concentrations^19–26. Among these catalysts, Fe-based catalysts exhibit remarkable efficiency in enhancing the performance of low-concentration NORR^22,25,26. This becomes particularly effective when Fe sites are in a low oxidation state with sharp d-state defect features, enabling energy-aligned orbital configurations that facilitate the adsorption and activation of NO molecules²³. Moreover, as one of the most cost-effective and abundant elements in the earth²⁷, Fe also plays a pivotal role in biological nitrogenases, enabling efficient natural nitrogen fixation²⁸. However, the current approach of enhancing NO affinity on catalysts overlooks the critical challenge of inherently slow water dissociation and NO hydrogenation kinetics in neutral media.

Electron-donating materials have demonstrated significant catalytic efficiency in water dissociation reactions^29,30, which are essential for supplying protons to promote NO hydrogenation^21–24. Inspired by aforementioned guidance, we propose an innovative and universal strategy to improve neutral NORR performance by integrating electron-donating motifs into NO-affinitive catalysts, forming a dual-active-site architecture with synergistic functionality. To demonstrate this strategy and its underlying mechanism, we selected Pt nanoparticles as the electron donor, incorporated into an Fe single-atom catalyst (Fe_SAC) with NO-affinitive ability, as a representative case for detailed analysis. In situ X-ray absorption spectroscopy (XAS), attenuated total reflection-infrared spectroscopy (ATR-IR), and density functional theory (DFT) calculations confirmed that Pt_NPs serve as electron donors, increasing the electron density at the Fe sites for stronger NO adsorption, while also facilitating water dissociation to supply protons and significantly promoting the subsequent hydrogenation of NO intermediate. The synergistic interplay between these dual sites enables a cascade reaction mechanism. As a result, the Pt_NPs/Fe_SAC catalyst achieved a NH₃ yield rate of 709.7 μg h⁻¹ mg_cat.⁻¹ and a Faraday efficiency (FE) of 90.3% at −0.6 V under 1.0 vol% NO/Ar in neutral media, outperforming Fe_SAC alone (NH₃ yield rate: 444.2 μg h⁻¹ mg_cat.⁻¹, FE: 56.6%) and surpassing previously reported NORR systems under high NO concentrations (above 10%) in various medias. Notably, in a membrane electrode assembly (MEA) electrolyzer, the system achieved the NH₃ yield of 3207.7 μg h⁻¹ mg_cat.⁻¹. This approach was successfully extended to other electron-donating nanoparticles, such as Au, highlighting its broad applicability for efficient NORR in neutral media.

Results and discussion

Theoretical calculations

DFT calculations were performed to understand the mechanism underlying the enhanced NORR activity. Fe single atoms coordinated with pyridine-4N in graphene (FeN₄) were chosen as the catalyst model due to their high stability (Fig. 1a)³¹. To simulate the nanoparticle environment, nine Pt atoms were introduced into the FeN₄ structure (FeN₄-Pt, Fig. 1b). Before Pt loading, Fe single atom in FeN₄ exhibited partially unoccupied orbitals above the Fermi level (Fig. 1c). After Pt loading, these unoccupied orbitals shifted towards the Fermi level, indicating electron injection into Fe (Fig. 1d). Bader charge analysis confirmed that Fe received ~0.16e^- (Fig. 1e), increasing its electron density and reactivity³². In Supplementary Fig. 2, the charge density difference analysis of Fe site in FeN₄-Pt demonstrated markedly increased electron density accumulation relative to those in FeN₄, suggesting that Pt species exhibit electron-donating characteristics that enhance the electronic environment at Fe active centers. Therefore, the incorporation of Pt nanoparticle into FeN₄ creates a dual active site (Fig.1f).

Fig. 1. Theoretical calculations.

Atom configurations of (a) FeN₄ and (b) FeN₄-Pt. Atom color-coding: brown, iron; blue, nitrogen; grey, carbon; cyan, platinum. Corresponding DOS plots of Fe atom in (c) FeN₄ and (d) FeN₄-Pt. e Bader charge analysis of Fe atom in FeN₄ and FeN₄-Pt, respectively. f Schematic diagram of dual site catalyst. Source data are provided as a Source Data file.

We firstly investigated water dissociation on both FeN₄ and FeN₄-Pt sites (Figs. 2a, b). On FeN₄ site, the dissociation of H₂O into *H and *OH required a free energy of 1.45 eV. However, in FeN₄-Pt, H₂O adsorbed at Pt sites, reducing the dissociation energy to 0.99 eV, demonstrating the role of Pt in facilitating proton supply for NORR.

Fig. 2. Reaction mechanism.

The free energy of water dissociation on (a) FeN₄ and (b) FeN₄-Pt. Atom color-coding: brown, iron; blue, nitrogen; grey, carbon; cyan, platinum; white, hydrogen; red, oxygen. c The reaction free energies of NORR pathways on FeN₄ and FeN₄-Pt. d Comparation between traditional and cascade mechanisms. Source data are provided as a Source Data file.

Next, the free energy pathway for NO to NH₃ conversion was calculated and a significant enhancement in the NO adsorption capability on FeN₄-Pt (−1.36 eV) compared to FeN₄ (0.05 eV) was revealed (Fig. 2c). The first proton step (*NO → *HNO) was energetically favorable on both catalysts, but the activation barrier was reduced from 1.50 eV (FeN₄) to 0.54 eV (FeN₄-Pt). The second protonation led to the formation of *NH₂O (FeN₄-Pt, ΔG = 0.55 eV) instead of *HNOH (FeN₄, ΔG = 0.40 eV), thus shifting the reaction pathway. The third protonation on FeN₄-Pt led to *NH₂OH formation, but its high desorption energy (3.18 eV) made *NH₂OH an unlikely byproduct. Instead, *NH₂OH underwent hydrogenation to *NH₂ + H₂O, followed by an exothermic step to NH₃ (ΔG = 0.23 eV).

Overall, Pt_NPs enhance NORR by donating electrons to Fe, strengthening NO adsorption, and accelerating water dissociation to supply protons for NO hydrogenation. This synergistic cascade mechanism significantly boosts NH₃ synthesis efficiency in neutral low-concentration NO conditions (Fig. 2d).

Catalyst characterization

The Fe_SAC and Pt_NPs/Fe_SAC catalysts were synthesized based on computational guidance. Fe_SAC was synthesized through pyrolysis followed by acid etching, while Pt_NPs were subsequently loaded onto Fe_SAC by a thermal reduction method. High-aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of Fe_SAC (Fig. 3a) revealed that isolated atomic sites are randomly dispersed on nitrogen doped carbon (NC) without observable metal clusters or particles, consistent with the X-ray diffraction (XRD, Supplementary Fig. 3) result. Inductively coupled plasma optical emission spectrometer (ICP-OES) confirmed a Fe content of 0.66 wt% in Fe_SAC (Supplementary Table 1).

Fig. 3. Structural characterizations of catalysts.

HAADF-STEM images of (a) Fe_SAC and (b, c) Pt_NPs/Fe_SAC. d EDX mapping images of Pt_NPs/Fe_SAC. e Fe K-edge XANES spectra of Fe foil, FePc, Fe₂O₃, Fe_SAC and Pt_NPs/Fe_SAC. f FT-EXAFS spectra at Fe K-edge. g XPS spectrum of Pt_NPs/Fe_SAC in Pt 4 f region. Source data are provided as a Source Data file.

In contrast, HAADF-STEM images of Pt_NPs/Fe_SAC showed Pt_NPs with a size of ~3 nm and a lattice distance of 0.214 nm, corresponding to the (111) crystal plane (Fig. 3b). XRD pattern further confirmed the presence of Pt_NPs (Supplementary Fig. 3). As illustrated in Fig. 3c, Fe_SAC are densely distributed around Pt nanoparticles, and energy dispersion X-ray spectroscopy (EDX, Fig. 3d) mapping confirmed the uniform dispersion of Fe, C, and N elements, with Pt as the primary nanoparticle component. ICP-OES results showed Fe and Pt contents of 0.62 wt% and 0.84 wt%, respectively (Supplementary Table 1).

To investigate the chemical state and atomic structure, X-ray absorption spectroscopy (XAS) and X-ray photoelectron spectroscopy (XPS) were employed. The Fe K-edge X-ray absorption near-edge structure (XANES) spectra revealed that Fe in Fe_SAC lies between +2 and +3 oxidation states (Fig. 3e). Notably, Pt incorporation induces a negative shift in the Fe K-edge, indicating a reduced Fe oxidation state, consistent with the theoretical predictions shown in Fig. 1.

Fourier transformed (FT) k²-weighted extended X-ray absorption fine structure (EXAFS) analysis (Fig. 3f) confirmed the absence of Fe clusters, as FePc, Fe_SAC and Pt_NPs/Fe_SAC exhibited a prominent Fe-N scattering peak at 1.51 Å^33,34. EXAFS fitting curves revealed that the average coordination numbers of Fe_SAC and Pt_NPs/Fe_SAC are 4.4 and 4.2, respectively, indicating that Fe is coordinated with four nitrogen atoms (Fe-N₄) in both systems. The bond lengths for Fe-N are 1.96 and 1.97 Å, respectively (Supplementary Fig. 4 and Supplementary Table 2). Additionally, wavelet transform (WT) analysis of Fe K-edge EXAFS oscillations (Supplementary Fig. 5) displays a strong signal with a maximum intensity at 3.8 Å⁻¹ for Pt_NPs/Fe_SAC, corresponding to the Fe-N first coordination shell, which is similar to Fe_SAC³³. Impressively, no Fe-Fe signal was detected in WT contour plots for either Fe_SAC or Pt_NPs/Fe_SAC, further confirming the atomically dispersed Fe in both catalysts.

XPS analysis of the Pt 4 f region (Fig. 3g) showed two Pt⁰ peaks at 71.8 (4f_7/2) and 75.1 eV (4f_5/2) along with two Pt²⁺ peaks at 72.7 (4f_7/2) and 76.0 eV (4f_5/2) in Pt_NPs/Fe_SAC (Fig. 2g), likely due to partial surface oxidation or electron donation from adjacent Fe atoms^35,36. The successful synthesis and characterization of Pt_NPs/Fe_SAC confirmed its distinct structural features, including reduced Fe oxidation states upon Pt incorporation and stable Fe-N₄ coordination.

In situ spectroscopy

To demonstrate the cascade reaction mechanism during the dynamic catalytic process of NORR in the Pt_NPs/Fe_SAC dual-site catalyst, in situ electrochemical XAS and attenuated total reflection-infrared (ATR-IR) spectroscopies were conducted. As shown in Fe K-edge XANES spectra of Pt_NPs/Fe_SAC (Fig. 4a), the white-line peak intensity increased as the applied potential decreases from open circuit potential (OCP) to −0.3 V, indicating that NO adsorption on Fe sites leads to an increase in coordination number. A shift of the adsorption edge to higher energy suggested an increased oxidation state of Fe. As the potential further decreased to −0.6 V, the white-line peak intensity diminished, implying NO consumption and reduction of Fe valence state, confirming Fe_SAC as the real active sites.

Fig. 4. In situ XAS and ATR-IR spectra of NORR.

a In situ Fe K-edge XANES spectra of Pt_NPs/Fe_SAC at different potentials (without iR-correction). (Inset) Magnified white-line peak and pre-edge XANES region. b Corresponding FT-EXAFS spectra at Fe K-edge (without iR-correction). Potential-dependent in situ ATR-IR spectra of (c) Fe_SAC and (e) Pt_NPs/Fe_SAC (without iR-correction). Corresponding 2D ATR-IR contour map of (d) Fe_SAC and (f) Pt_NPs/Fe_SAC (without iR-correction). g Reaction mechanisms. Source data are provided as a Source Data file.

The EXAFS spectra (Fig. 4b) further revealed the dynamic evolution of Fe’s local coordination. The Fe-N peak was enhanced with decreasing the applied potentials, suggesting that the intermediates adsorb on Fe sites. A notable negative shift from OCP to −0.3 V, suggested the formation of a shorter Fe-N bond due to NO adsorption. When the potential was further reduced to −0.6 V, the peak moved from 1.53 to 1.58 Å, signaling the occurrence of NORR and aligning well with the XANES results.

Next, in situ ATR-IR spectroscopy was employed to identify the adsorbed intermediates. As shown in Figs. 4c and 4e, both Fe_SAC and Pt_NPs/Fe_SAC display peaks corresponding to NO adsorption, including vertical mode (NO_v) and bent mode NO_b)^22,37. Specifically, for Fe_SAC, the NO_v and NO_b peaks were observed at 1743.4 and 1694.5 cm⁻¹, respectively. Similarly, in the spectra of Pt_NPs/Fe_SAC, the NO_v and NO_b peaks appeared at 1743.3 and 1693.2 cm⁻¹, respectively. No significant peak shift was observed, suggesting NO molecule was adsorbed on the similar active site in both catalysts. As shown in Supplementary Fig. 9, the Pt nanoparticles supported on NC (Pt_NPs) exhibits only a strong H₂O adsorption peak under a 1.0 vol% NO/Ar atmosphere, which indicates that the Fe single atom site remains the primary adsorption site for NO in Pt_NPs/Fe_SAC. Besides, the adsorption energies of NO on Fe sites and Pt sites were systematically evaluated. For Pt sites, two distinct adsorption configurations (Supplementary Fig. 10b,c) were analyzed, revealing positive adsorption free energies of 0.36 eV and 0.53 eV, respectively. In contrast, the adsorption energy of NO on Fe site was determined to be −1.36 eV (Supplementary Fig. 10a). These results strongly suggest that NO has a significantly higher affinity for Fe site than for Pt site.

Impressively, the higher intensity of these peaks in Pt_NPs/Fe_SAC indicated stronger NO adsorption. With the potentials shifted from OCP to −1.0 V, the intensity of NO_b peak increased, suggesting effective NO activation at Fe sites. As illustrated in Figs. 4d and 4f, the NO_b peak intensity in Pt_NPs/Fe_SAC spectra is much stronger than that of Fe_SAC, implying that the lower valence state of Fe site induced by introduction of Pt nanoparticles is more conducive to activate NO, in agreement with the DFT results shown in Figs. 1 and 2. The NO temperature-programmed desorption (NO-TPD) measurement was conducted to evaluate the NO adsorption capability. In Supplementary Fig. 11, the NO chemisorption peak for Fe_SAC appeared at 247.2 °C, while that of Pt_NPs/Fe_SAC was observed at a higher temperature of 266.8 °C. The elevated desorption temperature indicates stronger NO adsorption on Pt_NPs/Fe_SAC, suggesting that the introduction of Pt nanoparticles enhances the NO binding affinity.

Compared to Fe_SAC, the peak of -OH bending vibration at 1648 cm⁻¹ was strengthened in Pt_NPs/Fe_SAC spectra, indicating that the Pt nanoparticles facilitate H₂O electrolysis to produce *H, which is beneficial for subsequent NO hydrogenation (Figs. 4d, f)^38,39. This was further verified by the electron paramagnetic resonance (EPR) and H₂-TPD techniques^40–43. As shown in Supplementary Fig. 12, the Pt_NPs/Fe_SAC and Fe_SAC were electrolyzed in Ar-saturated 0.5 M K₂SO₄ electrolyte for 10 min at −0.6 V and −0.7 V respectively, and then the EPR tests were conducted immediately. The distinct hydrogen radical signals were detected for Pt_NPs/Fe_SAC but not in Fe_SAC, which confirmed that the introduction of Pt nanoparticles facilitated the electrochemical generation of hydrogen radicals. However, when the Pt_NPs/Fe_SAC was tested in NO-saturated 0.5 M K₂SO₄ electrolyte for 10 min at −0.6 V, the signal of hydrogen radicals disappeared, proving that the hydrogen radicals were involved in the NORR reaction (Supplementary Fig. 13). Besides, H₂-TPD experiments were also carried out to investigate the hydrogen spillover effect. The H₂ desorption peak of Pt_NPs/Fe_SAC (303.4 °C) exhibited a positive shift compared to that of Fe_SAC (274.3 °C), indicating a stronger binding strength between hydrogen and Pt_NPs/Fe_SAC catalyst (Supplementary Fig. 14). This finding implies that the presence of Pt nanoparticles increases the hydrogen concentration around the Pt sites, thereby triggering the hydrogen spillover effect that facilitates the NO hydrogenation⁴³. Hydrogenated intermediates NH_x (1412 and 1530 cm⁻¹) were progressively formed at lower potentials, ultimately converting to NH₄⁺ (1465 cm⁻¹)^44–46. Moreover, when Pt_NPs/Fe_SAC was tested under Ar atmosphere, only a strong peak corresponding to H₂O was observed, confirming the signals of hydrogenated intermediates derive from NO (Supplementary Fig. 15). Overall, these results clearly demonstrated Pt_NPs served as electron donors, increasing Fe site electron density to enhance NO adsorption while simultaneously promoting water dissociation to supply protons for efficient NORR. These results clearly demonstrated that a cascade reaction mechanism occurs on the dual-site catalyst, Pt_NPs/Fe_SAC, thereby promoting NORR (Fig. 4g).

NORR performance evaluation

The NORR performance was evaluated in 0.5 M K₂SO₄ electrolyte, using an air-tight H cell. Prior to electrolysis, high-purity Ar was purged into the electrolyte to remove residual oxygen. As shown in Fig. 5a, the linear sweep voltammetry (LSV) curves revealed a significant increase in current density for Pt_NPs/Fe_SAC in 1 vol% NO atmosphere compared to Ar-saturated conditions, confirming effective NO reduction on the catalyst surface. Compare to Fe_SAC, Pt_NPs and NC, the Pt_NPs/Fe_SAC showed the largest current density gap, indicating superior NORR activity (Supplementary Fig. 16).

Fig. 5. Electrochemical NORR performance.

a LSV curves of Fe_SAC and Pt_NPs/Fe_SAC in Ar- and 1 vol% NO-saturated 0.5 M K₂SO₄ (pH =7.1, without iR-correction) at scan rate of 10 mV s⁻¹. b NH₃ yield and FE with 1 vol% NO over Pt_NPs/Fe_SAC at each given potential (without iR-correction). c The NH₃ yield of Pt_NPs, NC, Fe_SAC and Pt_NPs/Fe_SAC. d NH₃ yield of Pt_NPs/Fe_SAC under different conditions. e¹H NMR spectra of the electrolyte fed by ¹⁵NO and ¹⁴NO for NORR over the Pt_NPs/Fe_SAC at −0.6 V (without iR-correction). f NORR cycling stability test over the Pt_NPs/Fe_SAC at −0.6 V (without iR-correction). g LSV curves of Pt_NPs/Fe_SAC test in MEA electrolyzer and H cell at scan rate of 10 mV s⁻¹ (without iR-correction). h Corresponding NH₃ yield and FE of Pt_NPs/Fe_SAC (without iR-correction). i Comparison of NORR performance with other electrocatalysts. Error bars indicate the relative standard deviations of the mean (n = 3). Source data are provided as a Source Data file.

As depicted in Fig. 5b, the NH₃ yield rate and FE of Pt_NPs/Fe_SAC increase steadily with applied potentials from -0.4 to -0.6 V, reaching the maximum values of 709.7 μg h⁻¹ mg_cat.⁻¹ and 90.3% at −0.6 V, outperforming previously reported NORR catalysts (Supplementary Table 4). NH₃ quantification via colorimetric method (Supplementary Fig. 17) and nuclear magnetic resonance (NMR) spectroscopy (Supplementary Fig. 18) yielded consistent results, ensuring measurement accuracy (Supplementary Fig. 19). However, at the applied potentials beyond −0.6 V, NORR performance of Pt_NPs/Fe_SAC declined due to the competing HER (Supplementary Fig. 20). Other liquid byproducts such as N₂H₄ and NH₂OH were not detected (Supplementary Figs. 21 and 22). In comparison, Fe_SAC obtained its highest NORR performance at -0.7 V with a NH₃ yield of 444.2 μg h⁻¹ mg_cat.⁻¹ and FE of 56.6% (Fig. 5c and Supplementary Fig. 23), while Pt_NPs/NC and bare NC exhibited negligible NORR activity (Fig. 5c).

As shown in Supplementary Fig. 26, Pt_NPs/Fe_SAC exhibits a significantly higher NH₃ partial current densities than Fe_SAC, which confirms the superior intrinsic activity of Pt_NPs/Fe_SAC towards NORR. We also conducted electrochemical NORR tests under 10 vol% NO/Ar to explore the mechanistic origin of the activity enhancement. In Supplementary Fig. 27, the Pt_NPs/Fe_SAC achieved an enhanced NH₃ yield rate of 1562.9 µg h⁻¹ mg_cat.⁻¹ and FE of 92.3% under 10 vol% NO/Ar compared to 1.0 vol% NO/Ar condition. The remarkable improvement in performance primarily originates from an acceleration of the NO adsorption step rather than solely the intrinsic catalytic activity, indicating the mass transport and surface NO availability are important limiting factors under lower concentrations. In addition, the turnover frequency (TOF) values of Fe_SAC and Pt_NPs/Fe_SAC are 194.2 h⁻¹ and 344.2 h⁻¹, respectively, under 1.0 vol% NO/Ar atmosphere. The significantly enhanced TOF value suggests that the Pt nanoparticle introduction facilitates a quick water dissociation and thereby accelerates the turnover frequency of the NO reduction sites.

DFT and in-situ electrochemical experiments clearly verified that Pt_NPs served as electron donors, increasing the electron density at the Fe single-atom site to promote NO adsorption, while facilitating water dissociation for proton supply and promoting NO hydrogenation. This dual role enabled efficient NO-to-NH₃ conversion with high activity and selectivity by activating the cascade reaction mechanism.

To validate that NH₃ originated from NO reduction, control experiments were conducted under three conditions: (1) electrolysis at −0.6 V in Ar-saturated electrolyte, (2) testing at OCP in NO-saturated electrolyte, and (3) testing in NO-saturated electrolyte with bare carbon paper (CP) as work electrode (Fig. 5d). Only negligible NH₃ was detected in these cases, confirming that NH₃ formation on Pt_NPs/Fe_SAC exclusively results from NO reduction. Additionally, an isotope labeling experiment was performed to exclude the possibility of nitrogen contamination. The laboratory-produced ¹⁵NO was used as the feeding gas for NORR at −0.6 V for 1 h. As shown in Fig. 5e, the NMR analysis conclusively verified that the generated NH₃ is originated from NO.

Electrocatalyst stability is crucial for long-term energy conversion and storage. Pt_NPs/Fe_SAC demonstrated excellent durability, maintaining stable NH₃ yield and FE over ten consecutive cycles (Fig. 5f). A long-term stability test (>70 h, Supplementary Fig. 30) showed no significant decline in current density, further confirming its robust electrochemical stability. Post-reaction characterization revealed minor catalyst changes: XRD pattern showed weakened Pt diffraction peaks, suggesting partial Pt dissolution (Supplementary Fig. 31). ICP analysis detected trace amounts of Pt in the electrolyte post-NORR (Supplementary Table 3). XANES analysis exhibited a slight increase in Fe valence state after NORR test (Supplementary Fig. 32a). EXAFS spectra revealed no Fe-Fe bond formation, confirming Fe remains atomically dispersed (Supplementary Fig. 32b). XPS spectrum of Pt exhibited two dominant peaks at 72.1 and 74.8 eV, which were attributed to Pt⁰ 4f_7/2 and Pt⁰ 4f_5/2, respectively, indicating that a small amount of Pt oxide species on the catalyst surface were reduced (Supplementary Fig. 33).

To mitigate NO mass transport limitations due to its low solubility (~1.92 mmol L⁻¹ atm⁻¹ in water at 25 °C), NORR was conducted in a membrane electrode assembly (MEA) electrolyzer (Supplementary Fig. 34)^47,48. Compared to H-cell, the MEA significantly enhanced current density, suggesting improved NO utilization (Fig. 5g). As shown in Fig. 5h, increasing the applied cell voltage from 1.2 to 1.6 V vs. cell voltage leads to higher NH₃ yield and FE. Notably, Pt_NPs/Fe_SAC achieved the high NH₃ yield of 3207.7 μg h⁻¹ mg_cat.⁻¹ and FE of 95.6% at 1.6 V vs. cell voltage (Fig. 5h), surpassing all previously reported NORR catalyst even tests under high NO concentration (Fig. 5i).

To explore alternative metal nanoparticles, Au_NPs/Fe_SAC was synthesized for NO-to-NH₃ conversion. TEM images (Supplementary Fig. 35) and XRD patterns (Supplementary Fig. 36) clearly verified that successful Au nanoparticle deposition. The obtained Au_NPs/Fe_SAC achieved optimal NH₃ yield of 595.4 μg h⁻¹ mg_cat.⁻¹ and FE of 78.9% at -0.5 V, further demonstrating the versatility of metal-modified Fe_SAC systems (Supplementary Fig. 37).

In summary, we present an electron-donating particle-mediated strategy that integrates Pt_NPs as the electron donor into the Fe_SAC electrocatalyst to construct dual-active-site architectures, thereby enhancing NO adsorption while promoting water dissociation and NO hydrogenation for significantly improving neutral NORR at low NO concentration (1 vol%). In situ spectro-electrochemical experiments, coupled with theoretical calculations, confirmed that the Pt nanoparticles serve as electron donors, increasing the electron density at the Fe single-atom site to promote NO adsorption, while facilitating water dissociation to provide protons and thus reducing the activation energy of NO hydrogenation. The synergistic interplay between these dual sites enables a cascade reaction mechanism achieving a superior NORR performance with a NH₃ FE up to 90.3% and a high NH₃ yield rate of 709.7 µg h⁻¹ mg_cat.⁻¹ under 1 vol% NO concentration at −0.6 V, outperforming Fe_SAC (NH₃ yield rate: 444.2 µg h⁻¹ mg_cat.⁻¹, FE: 56.6%) and prior to systems operating under high NO concentrations. Notably, in a MEA electrolyzer, the system achieved the NH₃ yield of 3207.7 μg h⁻¹ mg_cat.⁻¹. This work not only offers an attractive earth-abundant nanocatalyst for NH₃ electrosynthesis at low NO concentrations, but also provides an innovative methodology for designing superior electrocatalytic NORR systems through a dual-active-site strategy, paving the way for large scale NH₃ electrosynthesis.

Methods

Chemicals

Ketjen black ecp600JD (KJ), Nafion and carbon paper (CP) were purchased from Suzhou Sinero Thechology Co., Ltd (Suzhou, China). Sodium hydroxide (NaOH, A. R. grade), sodium nitrite (NaNO₂, A. R. grade), p-aminobenzenesulfonic acid (pAA, C₆H₇NO₃S, A. R. grade) m-phenylenediamine (mPDA, C₆H₈N₂, 99.5%) ammonium persulfate ((NH₄)₂S₂O₈, A. R. grade), iron chloride hexahydrate (FeCl₃·6H₂O, 99%), polyvinylpyrrolidone (PVP, Mw24000), sodium hypochlorite (NaClO, 6–14% active chlorine basis) and ethylene glycol (EG, A. R. grade) were obtained from Aldrich Chemical Reagent Co., Ltd. (Shanghai, China). Chloroplatinic acid hexahydrate (H₂PtCl₆·6H₂O, A. R. grade), gold chloride trihydrate (HAuCl₄·3H₂O, A. R. grade), salicylic acid (C₇H₆O₃, A. R. grade), trisodium citrate dihydrate (Na₃C₆H₅O₇·2H₂O, A. R. grade), sodium nitroferricyanide (III) dihydrate (Na₂Fe(CN)₅NO·2H₂O, A. R. grade), dimethyl sulfoxide (DMSO-d6, 99.9%), ammonium chloride (¹⁴NH₄Cl, A. R. grade; ¹⁵NH₄Cl, 99atom %), sodium borohydride (NaNH₄, A. R. grade) and sodium nitrite (Na¹⁵NO₂, 99atom %) were bought from Macklin Chemical Reagent Co., Ltd. (Shanghai, China). Fe powder (99.9%), potassium thiocyanate (KSCN, A. R. grade) and hydrochloric acid (HCl, A. R. grade) were supplied by Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). All the reagents were used as received. The water used throughout all experiments was purchased from Wahaha Group Co., Ltd. (Hangzhou, China).

Preparation of NC

1.0 g of KJ was dispersed in 30 mL H₂O and sonicated in an ice bath for 2 h to labeled A1. Separately, 0.74 g pAA was dissolved in 30 mL H₂O. To this solution, 9 mL of 1 M NaOH, 4 mL of 1 M NaNO₂ and 21 mL of 1 M HCl were added to prepare the diazo salt (labeled A2), maintain the reaction at 0 °C. Subsequently, the A2 was mixed with A1. To initiate surface grafting of the diazonium salt, 0.7 g of reduced Fe powder was introduced. Afterward, 25 mL concentrated HCl, 7.6 g of mPDA, 23 mL of 1 M FeCl₃ and 70 mL of 2 M (NH₄)₂S₂O₈ solution were added, and the reaction was allowed to react overnight. After filtration, the product was washed with water for 3 times and vacuum drying overnight to yield the NC precursor.

Preparation of Fe_SAC

0.6 g of the NC precursor was ultrasonically dispersed in 30 mL H₂O, followed by the addition of 1.8 mL of 1 M FeCl₃ and 6 mL of 1 M KSCN. The solvent was removed by rotary evaporation, and the residue was heated to 950 °C for 1 h under an Ar atmosphere. The obtained black powder was dispersed in 50 mL of 1 M HCl solution overnight at 80 °C. Then, the powder was washed with water 3 times, and drying overnight. Finally, the dried powder was heated at 950 °C for 3 h in an Ar atmosphere to yield Fe_SAC.

Preparation of Pt_NPs/Fe_SAC

100 mg Fe_SAC was dispersed in 20 mL of EG. Next, 5 mg H₂PtCl₆·6H₂O was dispersed in 20 mL of EG and then dripped into Fe_SAC. The obtained suspension was stirred for 2 h at room temperature and then treated at 160 °C for 90 min. Afterward, the final product was washed 3 times and dried overnight. The Pt_NPs was synthesized by replacing the Fe_SAC to NC.

Preparation of Au_NPs/Fe_SAC

100 mg of Fe_SAC and 10 mg of PVP were dispersed in 30 mL H₂O. Then, 400 μL of a 10 mg/mL HAuCl₄ solution was added dropwise, followed by 100 mg of NaNH₄ dissolved in 20 mL H₂O to reduce HAuCl₄. The mixture was stirred for 10 min at room temperature.

Preparation of work electrodes

10 mg catalyst powder was dispersed in 950 μL ethanol and 50 Nafion mixture by sonication for 2 h. Next, 100 homogenous ink was dropped on CP (1*1 cm²) and dried at room temperature. The catalyst loading is 1 mg cm^-2.

Characterizations

XRD patterns were collected by a LabX XRD-6100 X-ray diffractometer with Cu Kα radiation (40 kV, 30 mA) of wavelength 0.154 nm (SHIMADZU, Japan). TEM images were acquired on a Hitachi H-8100 electron microscopy (HITACHI, Japan), equipped with EDS mapping. XPS measurements were conducted on an ESCALABMK II X-ray photoelectron spectrometer using Mg as the exciting source. In situ electrochemical ATR-FTIR spectra were obtained by using a Thermo iS50. The XAS spectra were performed at the BL11B beamline of Shanghai Synchrotron Radiation Facility (SSRF). The incident photons were monochromatized by a Si (111) double-crystal monochromator. The absorbance data of spectrophotometer were measured on SHIMADZU UV-1800 ultraviolet-visible (UV-Vis) spectrophotometer.

Electrochemical measurements

Firstly, the Nafion 211 membrane was protonated in 5 wt% H₂O₂ at 80 °C for 1 h, then treated in 0.5 M H₂SO₄ for 3 h, and finally soaked in water for 6 h. All steps were performed at 80 °C. Electrochemical measurements were performed with a CHI660E electrochemical station (CH Instruments, Inc., Shanghai) in a H-cell under ambient condition, which was separated by Nafion 211 membrane. All the NORR tests were conducted in 0.5 M K₂SO₄ electrolyte (30 mL, pH = 7.1), which was stored at room temperature. The Pt_NPs/Fe_SAC, Ag/AgCl and graphite rod acted as the working electrode (WE), reference electrode (RE) and counter electrode (CE), respectively. The reference electrode was calibrated to the reversible hydrogen electrode (RHE) scale in all measurements using the following equation:

$$E\left(\mathrm{RHE}\right)=E\left(\mathrm{Ag}/\mathrm{AgCl}\right)+\left(0.059\times \mathrm{pH}+0.197\right)\mathrm{V}$$ 1

All data are presented without iR-correction. High purity Ar gas (99.999%) was bubbled into cathode chamber with the flow rate of 30 sccm for 30 min to removal oxygen before NORR. Then, the low-concentration NO (1% v/v) was firstly washed by 4 M KOH and then fed at 30 sccm for 30 min to saturate the electrolyte, maintain a constant flow during the electrochemical tests. The electrocatalytic NORR measurements were conducted at different potentials for 1 h, and then the electrolyte was collected and analysed. For NORR tested in MEA, a piece of anion exchange membrane (AEM, Sustainion, thickness: 50 μm) was used to separate cathode and anode chambers. Before electrochemical test, the AEM was immersed in 1 M KOH for 12 h, then rinsed with deionized water, and finally stored in pure water. The RuO₂ supported on nickel foam (RuO₂/NF) as anode tested in 1 M KOH (30 mL, pH = 14), while Pt_NPs/Fe_SAC was used as the cathode for NORR and the tail gas was absorbed by 1 M HCl. The error bars were the mean values standard deviation according to the obtained data. All experiments were carried out at room temperature (≈25 °C).

Determination of NH₃

The produced NH₃ was quantified by the indophenol blue method⁴⁹ and NMR. For indophenol blue method, standard NH₃ solution with a series of concentrations were used to calibrated the concentration-absorbance curves. The fitting curve (y = 0.3435x + 0.05843, R² = 0.9999) showed good linear relation of absorbance value with NH₃ concentration. All electrolytes were diluted 10 times before testing unless otherwise specified. For ¹H NMR measurements, the electrolyte after electrolysis was diluted 2 times with 1 M HCl to adjust the pH to acidic. Then, 100 mL DMSO-d6 was added in 500 μL acidified electrolyte.

Determination of FE and Yield of NH₃

The FE for NH₃ synthesis was calculated as the amount of electric charge used for NH₃ production divided by the total charge passed through the electrodes during electrolysis. The total amount of NH₃ produced was measured using colorimetric methods. The FE could be calculated as follows:

$$\mathrm{FE}\left({\mathrm{NH}}_3\right)=5\times \mathrm{F}\times \left[{\mathrm{NH}}_3\right]\times V/\left(17\times Q\right)\times 100\%$$ 2

NH₃ yield was calculated using the following equation:

$${\mathrm{NH}}_3\mathrm{yield}=\left[{\mathrm{NH}}_3\right]\times V/\left({\mathrm{m}}_{\mathrm{cat}.}\times \mathrm{t}\right)$$ 3

where F is the Faraday constant (96485 C/mol), [NH₃] is the measured NH₃ concentration, V is the volume of the electrolyte in the cathodic chamber, Q is the total quantity of applied electricity, t is the reduction time, m_cat. is the loaded mass of catalyst on carbon paper.

Determination of FE and yield of gas byproducts

The gas products (N₂O, N₂ and H₂) were quantified on a Gas chromatography (GC 2014 SHIMADZU). The yield and FE of gas products were calculated according to the following equations:

$$r_{\mathrm{gas}}=P_{\mathrm{gas}}\times F_{\mathrm{gas}}\times 3.6\times {10}^6/\left({\mathrm{m}}_{\mathrm{cat}.}\times {\mathrm{V}}_{\mathrm{m}}\right)$$ 4

$${\mathrm{FE}}_{\mathrm{gas}}=n_{\mathrm{gas}}\times r_{\mathrm{gas}}\times \mathrm{t}\times {\mathrm{m}}_{\mathrm{cat}.}\times \mathrm{F}\times {10}^{-6}/\mathrm{Q}\times 100\%$$ 5

Where r_gas is the gas product formation rate, P_gas is the percentage of gas in the total gas flow detected by GC, F_gas is the flow rate of 1 vol% NO/Ar, m_cat. is the loaded mass of catalyst on carbon paper, V_m is the molar volume in the standard condition (V_m = 22.4 L/mol), n_gas is the electron transfer numbers of gas product, t is the reduction time, F is the Faraday constant (96485 C/mol), Q is the total quantity of applied electricity.

TOF calculations

We calculate the TOF according to the following equation:

$$\mathrm{TOF}\left({\mathrm{h}}^{-1}\right)=\frac{{\text{I}}_{\text{product}}/n\text{F}}{{\mathrm{m}}_{\mathrm{cat}.}\times \alpha /{\mathrm{M}}_{\mathrm{metal}}}\times 3600$$ 6

Where I_product is partial current for NH₃, n is number of electrons transferred for NH₃, F is Faradaic constant, m_cat. is catalyst mass in the electrode, α is mass ratio of active atoms in catalysts, M_metal is atomic mass of metal.

Electrochemical in situ ATR-IR measurements

The in situ electrochemical ATR-IR measurements were conducted on a Nicolet iS50 FT-IR spectrometer with a liquid nitrogen-cooled MCT-A detector. The Si prism loaded with catalyst, Pt plate and Ag/AgCl were used as the working electrode, counter electrode and reference electrode, respectively, with 0.5 M K₂SO₄ as electrolyte. During the process of tests, 1 vol% NO was bubbled into the electrolyte with the flow rate of 10 sccm. Prior to testing, the Si prism was coated with Au film. The Si prism was first polished by 100 nm Al₂O₃. Next, the Si prism was soaked in a piranha solution for 30 min to removal organic contaminants. Then, the reflecting surface was immersed in a mixture of the Au plating solution (5.75 mM NaAuCl₄·2H₂O + 0.025 M NH₄Cl + 0.075 M Na₂SO₃ + 0.025 M Na₂S₂O₃ + 0.026 M NaOH) and a 2 wt % HF solution at 60 °C for 5 min. Afterward, the Au film was rinsed with deionized water and dried with N₂. In situ ATR-IR spectra were collected at OCP and different applied potentials.

Electrochemical in situ XAS measurements

The in situ XAS measurements were conducted in the fluorescence mode using a home-made electrochemical cell. The Pt_NPs/Fe_SAC, Ag/AgCl and graphite rod were used as working electrode, reference electrode and counter electrode, respectively. Prior to NORR test, 30 mL 0.5 M K₂SO₄ electrolyte was added into electrochemical cell and purged with Ar gas to removal dissolved oxygen. Subsequently, 1.0 vol% NO was bubbled through the deoxygenated electrolyte for 30 min to ensure saturation. Finally, the XAS spectra were recorded at OCP and different applied potentials.

Computation and model details

All spin-polarized simulations were carried out using density functional theory as implemented in the GPAW software^50,51 version 19.8.1. The exchange-correlation effects were accounted for using the BEEF-vdW-functional, which combines the generalized gradient approximation with the Langreth-Lundqvist van der Waals-functional to achieve accurate adsorption energies. A 2 × 2 × 1 k-point mesh was used because it ensures energy convergence in the calculation (Supplementary Fig. 38). The vacuum layer thickness of 25 Å was employed in the modeling framework (Supplementary Fig. 39). Then, to model the solvent at the electrochemical interface, a hybrid implicit/explicit approach was employed in the vacuum layer, where 40 explicit water molecules surrounded the electrode surface, and the remaining water was modeled using the SCMVD⁵² dielectric continuum model. The positions and orientations of the explicit water molecules were optimized using the minima hopping global optimization method⁵³ as implemented in ASE⁵⁴. To simulate the nanoparticle environment, nine Pt atoms were introduced into the FeN₄ structure.

To investigate NORR electrocatalytic processes involving proton-coupled electron transfer, free energy calculations were performed using the computational hydrogen electrode (CHE) approach. The chemical potential (the free energy per H) for H*  + e was correlated with that of 1/2H₂⁵⁵. The free energies of the reaction intermediates were defined as by ΔG = ΔE + ΔZPE – TΔS, where ΔE, ΔZPE, T, and ΔS represent the reaction energy, zero-point energy, temperature (298.15 K), and the entropy, respectively (Supplementary Table 5 and 6). The optimized computational models are provided in Supplementary Data 1.

Enhanced sampling

The slow growth sampling approach^56,57 in the constrained molecular dynamics simulation method can be used to describe the kinetic energy barrier in the reaction process by setting a suitable collective variable (CV, ξ), which changes from state 1 to state 2 at a certain transformation rate ${\xi }^{\cdot }$. The work performed throughout the entire process from state 1 to state 2 can be calculated using the following formula:

$$w_{state1\to state2}={\int }_{\xi \left(state1\right)}^{\xi \left(state2\right)}\left(\frac{\partial V\left(q\right)}{\partial \xi }\right)\cdot {\xi }^{\cdot }dt$$ 7

Where V(q) represents the free energy, and $\frac{\partial V\left(q\right)}{\partial \xi }$ is calculated using the SHAKE algorithm. When approaching the infinitesimal limit ${\xi }^{\cdot }$, the work $w_{state1\to state2}$ required from state 1 to state 2 corresponds to the difference in free energy. In the SG sampling method, $\partial \xi$ is selected to 0.0005 Å, and the final reaction’s free energy barrier can be obtained by aggregating the free energy distribution diagram.

Author contributions

T.W. and X.G. conceived the idea, wrote the original draft, and collected and analyzed the data. X.G. and T.W. performed the DFT calculations and experiments. C.M. provided the HAADF-STEM characterization. M.L. supervised this project. All authors contributed and reviewed the manuscript.

Peer review

Peer review information

Nature Communications thanks Xianbiao Fu, Rui-Ting Gao, Xin Wang and the other anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Data availability

Full data supporting the findings of this study are available within the article and its Supplementary Information, as well as from the corresponding author upon reasonable request. Source data are provided with this paper. The CIF files of computationally optimized structures are freely accessible on the website at https://github.com/sunyao-coder/Structures. Source data are provided with this paper.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41467-025-63365-7.