Tuning the N-intermediate adsorption of Cu catalysts for efficient electroreduction of nitrate to ammonia

Abstract

The electrocatalytic nitrate reduction reaction (NO₃RR) is a promising technique for both removal of harmful nitrates and sustainable NH₃ production. As yet, developing an electrocatalyst with high activity and stability remains a significant challenge. Herein, a novel electrocatalyst consisting of Cu nanoparticles dispersed on boron (B) and nitrogen (N) co-doped hollow carbon fibers (Cu/BNHCFs) was successfully fabricated. This was achieved through the stereoselective assembly of a Cu-containing zeolitic imidazolate framework onto electrospun fiber films, followed by pyrolysis. The optimized Cu/BNHCFs catalyst achieves a remarkable faradaic efficiency of 94.2% for NH₃ with a yield rate of 32.35 mg h⁻¹ mg_cat⁻¹ at −0.7 V vs. reversible hydrogen electrode. Electrochemical in situ characterization reveals that the reaction pathway on Cu/BNHCFs proceeds from *NO to *NH₂OH. Theoretical calculations further indicate that the B, N co-doped carbon support modulates the D-band center of Cu, effectively optimizing the adsorption/desorption processes of key nitrogen-containing intermediates and thus leading to the excellent catalytic performance. This work provides a design strategy for modifying the electronic structure of transition metal catalysts to achieve efficient nitrate reduction to ammonia.

---

1 Introduction

Ammonia (NH₃) is utilized as a chemical feedstock for producing fertilizers, medicine and nitrogen-containing chemicals, as well as a viable energy carrier owing to its high energy density and zero-carbon emission characteristics.¹ Currently, large-scale synthesis of NH₃ has been dominated by the energy-intensive and environmentally unfriendly Haber-Bosch method, which consumes 1–2% of the global energy and occupies 1% of global carbon dioxide (CO₂) emission.^2,3 Consequently, there is an increasing demand to explore ecofriendly and sustainable technologies for NH₃ production. Electrochemical NH₃ production, relying on renewable electricity for power (e.g., solar energy, wind energy), has shown great advantages due to its environmental safety, remote automated control, and mild operating conditions.^4,5 Considering the relatively low dissociation energy of the N=O bond (204 kJ mol⁻¹) and high-water-solubility of NO₃⁻, the electrocatalytic nitrate (NO₃⁻) reduction reaction (NO₃RR) has been regarded as a potential route for NH₃ synthesis.^6,7 On the other hand, NO₃⁻ is an environmental pollutant widely existing in agricultural and industrial wastewater with a high concentration, posing a serious threat to ecological balance and human health.⁸ Taken together, the NO₃RR offers an attractive route for both generating valuable ammonia and solving NO₃⁻ pollution.

In essence, the NO₃RR is a complex eight-electron coupled nine-proton transfer process, involving multiple intermediates.⁹ Therefore, designing and constructing an efficient catalyst with high Faraday efficiency (FE) and yield rate of NH₃ is crucial to the NO₃RR, yet remains a formidable challenge. Copper (Cu) has been reported as a highly active cathode catalyst for the NO₃RR due to its easy injection of d-electrons into the lowest unoccupied molecular π* orbital (LUMO π*) of NO₃⁻.^10,11 However, naked metal Cu particles tend to agglomerate and sinter during high-temperature treatment due to their large surface energy with active broken bonds, which leads to the incomplete exposure of active sites and significant loss of activity.^12,13 Additionally, another issue associated with Cu is the accumulation of some active intermediates (e.g., *NO₂ and *NO) during the NO₃RR, eventually resulting in sluggish kinetics of the specific steps and unsatisfactory FE of NH₃.^14–16 To overcome such shortcomings of pure Cu, the rational design of supported Cu catalysts is highly desirable. It is found that loading the active metal catalysts on appropriate supports (e.g., zinc oxide nanowire, graphdiyne, boron–carbon–nitrogen) is expected to facilitate their dispersion and alleviate the issue of aggregation during synthesis and electrocatalysis processes.^12,17,18 What's more, the strong metal–support interaction not only increases the material structural stability, but also facilitates electron transfer through the strengthened metal–support bonding.^19,20 Hence, constructing Cu composite electrocatalysts would be a promising resolution, which prevents aggregation of Cu nanoparticles and promotes the exposure of efficient active sites.

Carbon fibers (CFs), especially hollow carbon fibers, have been widely used as support materials for active components in electrocatalysis, thermocatalysis and energy-related fields due to their low price, abundant sources, high chemical stability, favorable mechanical strength, and enlarged contact areas.^21–23 The unique one-dimensional (1D) structures not only favor the orderly and long-range growth of active species along the fiber matrix and ensure the high accessibility of the catalytic sites, but also offer effective ion diffusion pathways and a continuous conductive framework.^24,25 Meanwhile, the hollow structures provide opportunities for the NO₃RR through nano-constrained microenvironments. Importantly, the incorporation of controlled heteroatoms (e.g., N, B, P, S) gives hollow CF-based materials a degree of freedom for compositions, and changes the electronic characteristics and conductivity of carbon, which facilitate the electrochemical process.^26–29 For example, Zhao et al. reported that the B–N bonds in B, N co-doped carbon materials can induce the generation of local electric fields and thus affect their NO₃RR performance.³⁰ This is attributed to the formation of coordination bonds between the electron-deficient B atoms (2s²2p¹) and electron-rich N atoms (2s²2p³). Thus, integrating Cu with heteroatom-doped hollow carbon fiber supports is expected to realize the cooperative effect of electronic structure modulation and structural functionalization, thereby achieving highly active and durable NO₃RR catalysts.

Inspired by the above, a novel and scalable fabrication strategy is presented to construct Cu nanoparticles dispersed onto boron and nitrogen co-doped hollow carbon fibers (Cu/BNHCFs) via in situ generation of a Cu-containing zeolitic imidazolate framework on electrospun fibers, followed by pyrolysis treatment for efficient NO₃RR. This synthetic approach harnesses the inherent benefits of material systems by leveraging the elevated carbon content, excellent fiber-forming properties and nitrogen-rich composition of polyacrylonitrile (PAN), in conjunction with the channel-forming capabilities of polymethyl methacrylate (PMMA) for hollow architecture engineering. The optimized Cu/BNHCFs catalyst exhibits excellent NH₃ conversion performance and achieves a FE of 94.2% at −0.7 V (vs. reversible hydrogen electrode, RHE) and a corresponding NH₃ yield rate of 32.35 mg h⁻¹ mg_cat⁻¹, being on par with many reported copper electrocatalysts. Meanwhile, the catalyst demonstrates a good long-term durability and stability. Finally, evolutionary behavior of NO₃⁻ to NH₃ conversion is proved by in situ Raman spectroscopy and electrochemical in situ mass spectrometry (DEMS). Theoretical calculations confirm that the B, N co-doped carbon support induces a shift in the D-band center of Cu in Cu/BNHCFs, enabling moderate adsorption ability and promoting the desorption of N-containing species, thereby reducing the energy barrier required for the formation of *NO and *HNOH and facilitating a more efficient reaction pathway. This work provides a strategy for optimizing transition metal catalysts in the NO₃RR.

2 Results and discussion

The Cu/BNHCFs catalyst was synthesized via a multistep procedure (Fig. 1a), including electrospinning technology, a post-modification with a Cu-containing zeolitic imidazolate framework (Cu-ZIF-L) precursor and a carbonization process. Initially, the core–shell PAN-PMMA@H₃BO₃ fibers consisted of a polymethyl methacrylate (PMMA)/H₃BO₃ core and a polyacrylonitrile (PAN) shell were prepared through the coaxial electrospinning technique. The scanning electron microscopy (SEM) image shows that PAN-PMMA@H₃BO₃ exhibits a typical fibrous structure with a smooth surface (Fig. 1b). Subsequently, a uniform shell of Cu-ZIF-L was grown in situ on PAN-PMMA@H₃BO₃ fibers via the electrostatic attraction and coordination reactions at room temperature (Cu-ZIF-L/PAN-PMMA@H₃BO₃). As illustrated in Fig. 1c–e, Cu-ZIF-L nanosheets with a leaf-like structure are assembled in an orderly manner along the PAN-PMMA@H₃BO₃ fiber films. After pyrolysis at high temperature, Cu-ZIF-L/PAN-PMMA@H₃BO₃ was converted into Cu nanoparticles dispersed onto B, N co-doped hollow carbon fibers (Cu/BNHCFs) via a carbonization process. The carbon fibers derived from PAN maintain continuous fibrous morphology without structural collapse. Meanwhile, the Cu-ZIF-L is converted into Cu–N–C and coated on the surface of the carbon fibers, retaining the leaf-like structure of Cu-ZIF-L (Fig. 1g–i). During this process, PMMA, as a self-sacrificial template, is the key to forming the hollow structure because of its low pyrolysis temperature. PAN undergoes transformations involving intramolecular cyclization, intermolecular cross-linking, dehydrogenation, and oxidation, which collectively facilitate the progressive formation of thermally stable carbon fiber skeletons.³¹ Meanwhile, some reducing gases generated from the composite films favor the formation of active Cu nanoparticles. As a comparison, BNHCFs were also synthesized through the direct carbonization of PAN-PMMA@H₃BO₃ fibers (Fig. 1f).

Fig. 1 (a) Schematic illustration of the synthesis of the Cu/BNHCFs. SEM images of (b) PAN-PMMA@H₃BO₃, (c) Cu-ZIF-L/PAN-PMMA@H₃BO₃-0.05, (d) Cu-ZIF-L/PAN-PMMA@H₃BO₃-0.1, (e) Cu-ZIF-L/PAN-PMMA@H₃BO₃-0.2, (f) BNHCFs, (g) Cu/BNHCFs-0.05, (h) Cu/BNHCFs-0.1, and (i) Cu/BNHCFs-0.2.

To obtain more detailed structural information about the catalysts, transmission electron microscopy (TEM) measurements were performed. The TEM images of Cu/BNHCFs-0.1 clearly show hollow structures of carbon fibers and the formation of leaf-like Cu carbon nanosheets on BNHCFs (Fig. 2a). This hollow-confined microenvironment could provide opportunities for the NO₃RR. Furthermore, it can be clearly seen that the Cu nanoparticles with an average size of 3.10 nm are uniformly distributed on the hollow carbon fibers (Fig. 2b and S1). In the high-resolution TEM (HRTEM) image (Fig. 2c), the Cu nanoparticles are encapsulated by carbon layers and the lattice spacings of 0.209 nm and 0.345 nm correspond to the (111) and (200) planes of the Cu phase, respectively. Meanwhile, the energy-dispersive X-ray spectroscopy (EDS) elemental mapping of Cu/BNHCFs-0.1 confirms the presence and distribution of Cu, C, N and B elements in the material (Fig. 2d). Moreover, the copper content in the Cu/BNHCFs was determined through inductively coupled plasma optical emission spectrometry (ICP-OES). As shown in Table S1, the Cu mass fraction increased from 2.78 wt% (Cu/BNHCFs-0.05) to 3.95 wt% (Cu/BNHCFs-0.1) and further reached 6.82 wt% (Cu/BNHCFs-0.2).

Fig. 2 (a and b) TEM images, (c) HRTEM image, and (d) the corresponding EDS elemental mapping of Cu/BNHCFs-0.1.

The structures of the as-synthesized materials were characterized by X-ray diffraction (XRD). As displayed in Fig. 4a, a broad diffraction peak is observed at about 25° for all materials, which is assigned to the (002) plane of amorphous carbon.³² Additionally, the characteristic peaks of Cu/BNHCFs-0.1 and Cu/BNHCFs-0.2 located at 43.30°, 50.43°, and 74.13° can be indexed to the (111), (200) and (220) crystal planes of metallic Cu (PDF#04-0836).³³ No diffraction peaks of metallic Cu are detected in Cu/BNHCFs-0.05, which may be due to low percentages and high dispersion of Cu species. The functional groups of Cu/BNHCFs and BNHCFs samples were further investigated by Fourier transform infrared (FTIR) spectroscopy (Fig. 3b). Two bands at 798 and 1390 cm⁻¹ are assigned to the out-of-plane bending vibration and in-plane stretching of B–N, respectively.^34,35 Characteristic peaks observed at 1173 cm⁻¹, 1482 cm⁻¹, and 1600 cm⁻¹ are ascribed to the vibration of B–C, C–N and C–C.^36,37 Besides, a broad band in the 3000–3600 cm⁻¹ region corresponds to the stretching vibration of N–H or O–H. Moreover, the graphitic crystallinity and defect density of the catalysts were further characterized via the ratio of D-band to G-band in Raman spectra (Fig. 3c). The two broad bands centered at 1350 and 1590 cm ⁻¹ are indexed to the D band and G band, reflecting the presence of disordered carbon and crystalline carbon in a graphene layer.³⁸ Higher I_D/I_G of all the as-prepared Cu/BNHCFs than that of BNHCFs means the presence of more defects. The presence of defect sites has been proved to be conducive to modulating the electronic structures of carbon materials, inducing electronic redistribution, thus enhancing the electrocatalytic performance of the catalysts.^39,40 Nitrogen adsorption–desorption measurements were carried out to characterize the pore structure and specific surface area of the catalysts. The N₂ adsorption–desorption isotherms of Cu/BNHCFs and BNHCFs are shown in Fig. 3d and S2. The Brunauer–Emmett–Teller (BET) specific surface areas of BNHCFs, Cu/BNHCFs-0.05, Cu/BNHCFs-0.1, and Cu/BNHCFs-0.2 are calculated to be 22.5, 517.3, 795.3 and 715.8 m² g⁻¹, respectively. Notably, the pore size distribution curve of Cu/BNHCFs-0.1 confirms the coexistence of mesopores and micropores (Fig. 3e). The rich pore structures and high specific surface area are not only beneficial to expose more active sites and facilitate rapid ion transport, but also promote sufficient contact between the electrolyte and the electrocatalyst. Additionally, the interconnected network of carbon fibers enables efficient electrolyte penetration and rapid gas diffusion.

Fig. 3 (a) XRD patterns, (b) FTIR spectra, (c) Raman spectra of BNHCFs and Cu/BNHCFs. (d) N₂ adsorption–desorption isotherms of Cu/BNHCFs. (e) Pore size distribution curve of Cu/BNHCFs-0.1. (f) Cu 2p, (g) C 1s, (h) N 1s, and (i) B 1s XPS spectra of Cu/BNHCFs-0.1.

X-ray photoelectron spectrum (XPS) analysis was conducted to elucidate the elemental compositions and chemical states of Cu/BNHCFs. The high-resolution Cu 2p spectra (Fig. 3f) display two peaks at binding energies of 932.49 and 952.40 eV, corresponding to Cu 2p_3/2 and Cu 2p_1/2 of metallic copper (Cu⁰), and the characteristic peaks with binding energies of 934.78 (Cu 2p_3/2) and 954.93 eV (Cu 2p_1/2) are assigned to Cu²⁺ (unavoidable surface oxidation).³⁸ In Fig. 3g, the high-resolution C 1s spectra shows three peaks: the sp²C at 284.76 eV, the C–N bond at 285.84 eV, and the C–O centered at 288.40 eV.^41,42 Regarding the N 1s spectra (Fig. 3h), four characteristic peaks at 398.50 (N–B bonds/pyridinic N), 399.56 (pyrrolic N), 400.97 (graphitic-N) and 403.18 eV (oxidized N) are identified, respectively.^27,43,44 Additionally, in the B 1s spectra (Fig. 3i), the peaks at 189.90, 190.73, and 191.98 eV are attributed to BN₂C/BC₃, BC₂O/B–N and BCO₂ bonds, respectively.^18,27,45 As confirmed by the above characterization results, Cu/BNHCFs is successfully prepared via a three-step method.

The electrocatalytic NO₃RR activity of the fabricated catalysts was evaluated in a H-type cell. Linear scanning voltammetry (LSV) curves show a significant increase of current density over the four samples after the introduction of nitrate, suggesting that all samples exhibit catalytic activity toward the NO₃RR (Fig. S3). Meanwhile, Cu/BNHCFs-0.1 possesses the highest current density in the presence of nitrate, indicating the highest electrocatalysis activity for the NO₃RR (Fig. 4a). For BNHCFs, the current density is much smaller than that of the other three Cu/BNHCFs catalysts, which implies that the BNHCFs substrate contributes relatively little to the NH₃ production and Cu species play a dominant role in nitrate reduction. According to the LSV results, chronoamperometry tests (electrolysis for 1 hour) were conducted at five potentials from −0.5 to −0.9 V vs. RHE to further evaluate the performance of BNHCFs, Cu/BNHCFs-0.05, Cu/BNHCFs-0.1 and Cu/BNHCFs-0.2 for nitrate electroreduction (Fig. S4). The obtained products (NO₂⁻, NH₃, N₂H₄) were detected and quantified using colorimetric methods, with the corresponding standard curves presented in Fig. S5–S8. With the negative shift of the applied potential, the FE of these four materials exhibits a volcano trend owing to the increasing competition from the hydrogen evolution reaction (HER) at negative potentials, while the yield rate of NH₃ gradually increases (Fig. 4b and S9–S11). As illustrated in Fig. 4c, the NH₃ FE of all samples follows the order of Cu/BNHCFs-0.1 > Cu/BNHCFs-0.05 > Cu/BNHCFs-0.2 > BNHCFs at −0.7 V versus RHE, and Cu/BNHCFs-0.1 shows the maximum NH₃ FE of 94.2%, corresponding to an NH₃ yield rate of 32.35 mg h⁻¹ mg_cat⁻¹ (Fig. 4d). Meanwhile, the partial current density is calculated and follows the order of Cu/BNHCFs-0.1 > Cu/BNHCFs-0.05 > Cu/BNHCFs-0.2 > BNHCFs.

Fig. 4 (a) LSV curves of BNHCFs and Cu/BNHCFs in 1 M KOH with 0.1 M NO₃⁻. (b) NH₃ FE and NH₃ yield rates of Cu/BNHCFs-0.1. (c) NH₃ FE, (d) NH₃ yield, (e) NH₃ partial current densities, and (f) NO₂⁻ FE of BNHCFs and Cu/BNHCFs. (g) NH₃ yield of Cu/BNHCFs-0.1 under different conditions. (h) ¹H NMR measurements using ¹⁴NO₃⁻/¹⁵NO₃⁻ as nitrogen sources. (i) NO₃RR performance of the Cu/BNHCFs-0.1 at −0.7 V vs. RHE in 1 M KOH with different NO₃⁻ concentrations.

Since the electrocatalytic NO₃RR to NH₃ is an intricate eight-electron process, some by-products (NO₂⁻, N₂H₄) may be usually formed. As shown in Fig. 4f, the FE of NO₂⁻ is less than 5% over the whole potential range. Of note, the UV spectrum results demonstrate that no N₂H₄ is produced during the NO₃RR process (Fig. S12). To trace the origin of nitrogen in NH₃ production and exclude possible contamination from the external environment, electrolytes and reagents, some control experiments were carried out. As observed in Fig. 4g, the NH₃ yield rates after electrolysis for 1 h at open circuit potential (OCP) or without nitrate are almost negligible. Furthermore, the isotopic labeling test combined with ¹H nuclear magnetic resonance (NMR) was conducted to further verify that the ammonia originated from NO₃⁻. Only triple peaks of ¹⁴NH₄⁺ are detected by using ¹⁴NO₃⁻ as the N source, whereas double peaks of ¹⁵NH₄⁺ are detected in the electrolyte containing ¹⁵NO₃⁻ (Fig. 4h). All the above results clearly confirm that the production of NH₃ in the electrolyte originates through the electrochemical NO₃RR process rather than from external contamination. Besides, the NO₃RR performance of Cu/BNHCFs-0.1 was also evaluated using different nitrate concentrations. As depicted in Fig. 4i, the NH₃ yield rate increases with the increase of initial nitrate concentration, indicating that catalyst activity is directly influenced by nitrate concentration. When the nitrate concentration reaches 500 mM, the NH₃ yield rate is up to 38.65 mg h⁻¹ mg_cat⁻¹. At a low concentration of 10 mM, the FE of NH₃ is only 25.6%, which may be caused by the limitation of nitrate diffusion and the competition from the HER. Although excessive nitrate (200 mM, 500 mM) leads to a slight decrease in NH₃ FE due to insufficient H⁺ supply,⁴⁶ the catalyst still maintains a high selectivity of NH₃ (FE > 90%). This demonstrates the promising application potential of Cu/BNHCFs-0.1 for treating different concentrations of NO₃⁻.

Based on cyclic voltammetry (CV) curves recorded at various voltage scan rates (Fig. S13), electrochemical double-layer capacitance (C_dl) is calculated to determine the electrochemical active surface area (ECSA), with higher C_dl demonstrating a larger ECSA.^6,12 As shown in Fig. 5a, the C_dl of Cu/BNHCFs-0.1 is calculated to be 22.73 mF cm⁻², larger than those of Cu/BNHCFs-0.05 (16.74 mF cm⁻²), Cu/BNHCFs-0.2 (14.24 mF cm⁻²) and BNHCFs (7.60 mF cm⁻²), which implies that Cu/BNHCFs-0.1 has more electrochemical active sites and higher electrocatalyst activity. Next, the electrochemical impedance spectroscopy (EIS) measurement results also indicate that Cu/BNCFs-0.1 possesses the smallest arc radius (Fig. 5b), indicating that Cu/BNCFs-0.1 has the smallest charge transfer resistance and excellent charge transfer ability in the NO₃RR process. Moreover, the Tafel plots obtained from the LSV curves of the four samples are displayed in Fig. 5c, where the Tafel slope of Cu/BNHCFs-0.1 (157.9 mV dec⁻¹) is significantly smaller than those of Cu/BNHCFs-0.2 (185.5 mV dec⁻¹), Cu/BNHCFs-0.05 (249.8 mV dec⁻¹) and BNHCFs (334.3 mV dec⁻¹), demonstrating the fastest NO₃RR kinetics of Cu/BNHCFs-0.1. The long-term electrochemical stability of Cu/BNHCFs for the NO₃RR was investigated by conducting continuous 10 cycles of electrolysis tests at −0.7 V versus RHE. As depicted in Fig. 5d, both FE and yield rate of NH₃ in the reaction exhibited minor fluctuations but remained essentially stable, demonstrating that Cu/BNHCFs has excellent electrochemical durability and stability. The composition and morphology of Cu/BNHCFs-0.1 after electrolysis were further characterized by XPS and SEM. There are no significant changes in the chemical states of Cu, C, N, and B, and the peak observed at 292.50 eV in the C 1s spectrum corresponds to the C–F bond, originating primarily from the Nafion solution used during the preparation of the working electrode (Fig. S14). Furthermore, the SEM image shows that the morphology of Cu/BNHCFs-0.1 remains essentially unchanged after electrolysis (Fig. S15). Notably, Cu/BNHCFs-0.1 exhibits strong competitiveness in the principal NO₃RR performance parameters when compared with the other reported catalysts (Fig. 5e and Table S1).

Fig. 5 (a) Calculated C_dl, (b) EIS spectra, and (c) Tafel plots of BNHCFs and Cu/BNHCFs. (d) Cycling test of Cu/BNHCFs-0.1 at −0.7 V. (e) Comparison of NH₃ yield rate and FE of Cu/BNHCFs-0.1 with reported catalysts.

To gain a deep insight into the NO₃RR reaction process over the Cu/BNHCFs catalyst, in situ measurements were employed to elucidate the reaction mechanisms and identify the possible intermediates and products. In situ Raman spectra at different potentials are illustrated in Fig. 6a; the characteristic peaks appearing at 1350 and 1590 cm⁻¹ are assigned to the D-band and G-band, respectively. At OCP, the only sharp peak at 1047 cm⁻¹ corresponds to the symmetric NO₃⁻ stretching mode, which confirms that Cu/BNHCFs has robust adsorption for nitrate.⁴⁷ Upon applying the potential shift, the intensity of characteristic NO₃⁻ stretching vibration weakens, concurrent with the emergence of a vibrational mode at 1159 cm⁻¹ assignable to the symmetric stretching of NO₂, suggesting NO₂⁻ production from NO₃⁻ reduction.⁴⁸ At −0.9 V, the peak at 1386 cm⁻¹ is typically assigned to the antisymmetric stretching vibration of the NO₂⁻ group.^2,49 Additionally, the broad peak centered at 1525 cm⁻¹ is likely attributable to the N=O vibration of the HNO* species and the antisymmetric bending vibration of the HNH (NH₃).^2,50 Online differential electrochemical mass spectrometry (DEMS) is employed to record the mass-to-charge (m/z) ratio signals of molecular intermediates during the NO₃RR process. As shown in Fig. 6b and c, the m/z signals at 46, 30, 31, 33, and 17 correspond to NO₂, NO, NOH (or HNO), NH₂OH, and NH₃ species, respectively. In addition, the signal of H₂ (m/z = 2) is also detected, likely originating from the side reactions of the HER. The above in situ measurement results provide comprehensive evidence for the main reaction pathways of the NO₃RR over the Cu/BNHCFs catalysts.

Fig. 6 (a) In situ Raman spectra. (b and c) Electrochemical online DEMS over Cu/BNHCFs. (d) Difference charge density analysis of NO₃⁻ on Cu/BNHCFs (Cyan and yellow isosurfaces represent charge depletion and accumulation). (e) Gibbs free energy diagram for the NO₃RR on Cu/BNHCFs and Cu. (f) Schematic diagram of the reaction pathway on Cu/BNHCFs. (g) Gibbs free energy of *NH₃ → NH₃ on Cu/BNHCFs and Cu. (h) The adsorption energy comparisons of *NO₂ and *NO on Cu/BNHCFs and Cu. (i) PDOS of the Cu d band center for Cu/BNHCFs and Cu.

Density functional theory (DFT) calculations were further performed to illustrate the catalytic mechanism involved for NO₃⁻ on Cu/BNHCFs. The structural models of Cu/BNHCFs and pure Cu are illustrated in Fig. S16 and S17. It is known that adsorption and activation of NO₃⁻ is the initial and key step in the NO₃RR process.⁵¹ Analysis of charge density differences displays electron transfer from Cu/BNHCFs to NO₃⁻ (0.63 e⁻), indicating an interaction between them, which is effective for proceeding the subsequent reduction of NO₃⁻ (Fig. 6d). Meanwhile, the Gibbs free energy (ΔG) of each intermediate for the NO₃RR pathway is computed according to the in situ measurement data (Fig. 6e). The associated adsorption configurations are shown in Fig. S18 and S19. On the pure Cu surface, *NO₂ deoxygenation to *NO undergoes an uphill with an energy barrier of 0.30 eV and the potential determining step (PDS) is *HNO hydrogenation to *HNOH (0.58 eV). In contrast, *NO₂ deoxygenation for Cu/BNHCFs shows a distinct downhill trend and the PDS is also the hydrogenation of *HNO, with an energy barrier of 0.12 eV. Note that *NO follows hydrogenation steps to generate *NOH (or *HNO) and the selection of *HNO or *NHO intermediates is according to the energy barrier. By comparing the ΔG for *NO → *HNO and *NO → *NHO pathways, it can be understood that *HNO is more favorable than *NHO for Cu/BNHCFs and pure Cu (Fig. S18 and S19). Therefore, the following reaction pathway for the NO₃RR process on Cu/BNHCFs is *NO₃ → *NO₂ →*NO → *HNO → *NHOH → *NH₂OH → *NH₂ → *NH₃ (Fig. 6f). *NH₃ desorption is the thermodynamic energy barrier for both Cu/BNHCFs and pure Cu. As shown in Fig. 6g, Cu/BNHCFs promotes this desorption relative to pure Cu, thereby facilitating surface active site reactivation. Additionally, the adsorption strength of *NO₂ and *NO on Cu/BNHCFs weakens compared to pure Cu, a consequence of BNHCFs altering the Cu electronic structure (Fig. 6h). This effect is evidenced by the projected density of states (PDOS) of the Cu atoms, where the D-band center of Cu/BNHCFs shifts away from the Fermi level (Fig. 6i). This shift weakens interactions with *NO₂ and *NO, consequently lowering energy barriers for their desorption, which prevents Cu surface poisoning and facilitates subsequent hydrogenation. Specifically, the BNHCF modification could significantly reduce the overall energy barriers for the reaction, thereby accelerating the conversion of NO₃⁻ to NH₃.

3 Conclusions

In this work, a facile and general method has been developed to fabricate a novel NO₃RR electrocatalyst consisting of uniformly dispersed Cu nanoparticles supported on B, N co-doped hollow carbon fibers. The hollow carbon fiber not only serves as a stable support for the orderly and controllable growth of Cu nanoparticles but also creates a hollow-confined microenvironment beneficial for the NO₃RR. The optimized Cu/BNHCFs exhibited an excellent electrochemical performance toward the NO₃RR. The maximum NH₃ FE of Cu/BNHCFs achieves 94.2%, with a yield rate of 32.35 mg h⁻¹ mg_cat⁻¹ at −0.7 V vs. RHE. The theoretical calculations elucidate that the B, N co-doped carbon support induces a downshifted D-band center of Cu, thereby optimizing adsorption strength for N-containing species while substantially lowering the energy barrier, enabling a more energetically favorable reaction pathway. This work not only provides a useful strategy for electronic modulation of transition metal catalysts in the NO₃RR, but also gives a deep insight into the mechanism.

Author contributions

Ruifang He: Writing – original draft, methodology, formal analysis, data curation, software. Lu Sun: Software. Ke Ren: Software. Xiaona Li: Formal analysis. Peng Tian: Supervision. Junwei Ye: Writing – review & editing, supervision, funding acquisition.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting the findings of this study are available within the article and its supplementary information (SI). Supplementary information: experiment details; particle size distribution; N₂ adsorption–desorption isotherms; LSV curves; and i–t curves; standard curves for nitrate, ammonia, nitrite, and hydrazine; analysis of electrochemical products; ECSA characterizations; XPS spectrum; SEM image; calculated structures; ICP summary. See DOI: https://doi.org/10.1039/d5ta06775h.

Acknowledgements

The authors acknowledge financial support from the National Natural Science Foundation of China (22378053), the Science and Technology Plan Project of Liaoning Province (2023JH2/101700294), and the Fundamental Research Funds for the Central Universities (DUT24ZD115). The authors gratefully acknowledge the assistance of Instrumental Analysis Center, Dalian University of Technology.

References

1. Y. N. Zhang, S. Li, W. Zheng and X. Wang, Nano Res. Energy, 2023, 2, e9120068.
2. Y. W. Wang, S. Wang, Y. F. Fu, J. Q. Sang, P. F. Wei, R. T. Li, D. F. Gao, G. X. Wang and X. H. Bao, Nat. Commun., 2025, 16, 897.
3. M. Yang, M. Chen, Y. Wang, J. Ma, L. Zhang, C. Hu, L. Zhuo, Y. Feng and X. Liu, Commun. Chem., 2025, 8, 267.
4. M. Yang, X. Hou, R. Fu, Z. Wang, C. Hu, G. Hu, L. Zhuo and X. Liu, Nano Res., 2025, 18, 94907778.
5. X. Yang, Y. Zhang, P. Sun and C. Peng, Smart Mol., 2024, 2, e20240019.
6. Y. Li, J. Wei, H. Lin, Y. Guo, X. Lu, S. Liu, H. Liu, M. Tang, J. Zhou and Y. Li, ACS Catal., 2025, 15, 1672–1683.
7. S. Wu, J. Yan, D. Zhao, Z. Cai, J. Yu, R. Li, Q. Li and G. Fan, J. Colloid Interface Sci., 2024, 668, 264–271.
8. Y.-T. Yang, X.-H. Wang, Z.-N. Zhang, X. Ai, X. Xiao, Y. Chen and S.-N. Li, J. Mater. Chem. A, 2024, 12, 29969–29977.
9. Q. L. Hong, B. Q. Miao, T. J. Wang, F. M. Li and Y. Chen, Energy Lab, 2023, 1, 220022.
10. R. Jia, X. Zhang, L. Gan, M. Tahir, Z. Huang, L. Pan, R. Gao, C. Shi, X. Zhang, G. Yang and J. Zou, J. Mater. Chem. A, 2025, 13, 5732–5743.
11. X. H. Wang, Q. L. Hong, L. Y. Shao, Q. G. Zhai, Y. C. Jiang, X. Ai, Y. Chen and S. N. Li, Adv. Funct. Mater., 2024, 34, 2408834.
12. A. Feng, Y. Hu, X. Yang, H. Lin, Q. Wang, J. Xu, A. Liu, G. Wu and Q. Li, ACS Catal., 2024, 14, 5911–5923.
13. R. Belgamwar, R. Verma, T. Das, S. Chakraborty, P. Sarawade and V. Polshettiwar, J. Am. Chem. Soc., 2023, 145, 8634–8646.
14. W. Jang, D. Oh, J. Lee, J. Kim, J. Matthews, H. Kim, S. Lee, S. Lee, Y. Xu, J. Yu, S. Hwang, T. Jaramillo, J. Jang and S. Cho, J. Am. Chem. Soc., 2024, 146, 27417–27428.
15. Y. Liu, J. Wei, Z. W. Yang, L. R. Zheng, J. K. Zhao, Z. M. Song, Y. H. Zhou, J. J. Cheng, J. Y. Meng, Z. G. Geng and J. Zeng, Nat. Commun., 2024, 15, 3619.
16. Y. Fu, S. Wang, Y. Wang, P. Wei, J. Shao, T. Liu, G. Wang and X. Bao, Angew. Chem., Int. Ed., 2023, 62, e202303327.
17. X. Feng, J. Liu, Y. Kong, Z. Zhang, Z. Zhang, S. Li, L. Tong, X. Gao and J. Zhang, Adv. Mater., 2024, 36, e2405660.
18. Z. Wang, S. Xia, X. Deng, G. Baryshnikov, A. Kuklin, H. Agren and H. Zhang, J. Colloid Interface Sci., 2024, 664, 84–95.
19. M. Xu, M. Peng, H. Tang, W. Zhou, B. Qiao and D. Ma, J. Am. Chem. Soc., 2024, 146, 2290–2307.
20. G. D. Wu, Y. X. Liu and J. H. Wang, Acc. Chem. Res., 2023, 56, 911–923.
21. J. Y. Han, N. P. Deng, H. Chi, G. Wang, Y. L. Wang, Q. Zeng, Z. Z. Peng, B. Cheng, B. M. Zhou and W. M. Kang, J. Energy Chem., 2024, 98, 334–363.
22. P. Liu, A. Klyushin, P. Chandramathy Surendran, A. Fedorov, W. Xie, C. Zeng and X. Huang, ACS Nano, 2023, 17, 24395–24403.
23. C. Zhang, S. Gao, X. Chen, D. Yu, L. Wang, X. Fan, Y. Cheng, X. Yu and Z. Zhao, Smart Mol., 2025, 3, e20240022.
24. F. Zhang, J. Luo, J. Chen, H. Luo, M. Jiang, C. Yang, H. Zhang, J. Chen, A. Dong and J. Yang, Angew. Chem., Int. Ed., 2023, 62, e202310383.
25. W. Yang, B. Guo, Z. Guo, X. Wang, Y. Zeng, J. Guo and M. Wu, ACS Sustain. Chem. Eng., 2023, 11, 8884–8892.
26. Z. Tian, Q. Zhang, L. Thomsen, N. Gao, J. Pan, R. Daiyan, J. Yun, J. Brandt, N. Lopez-Salas, F. Lai, Q. Li, T. Liu, R. Amal, X. Lu and M. Antonietti, Angew. Chem., Int. Ed., 2022, 61, e202206915.
27. L. Wang, Y. Wang, M. Wu, Z. Wei, C. Cui, M. Mao, J. Zhang, X. Han, Q. Liu and J. Ma, Small, 2018, 14, e1800737.
28. S. Shanmugapriya, P. Zhu, M. Ganeshbabu, Y. Sung Lee, X. Zhang and R. Kalai Selvan, Mat. Sci. Eng. B, 2022, 284, 115880.
29. M. Chen, J. Ma, Y. Feng, Y. Wu, G. Hu and X. Liu, Coord. Chem. Rev., 2025, 535, 216612.
30. S. Ajmal, A. Kumar, M. Mushtaq, M. Tabish, Y. Zhao, W. Zhang, A. Khan, A. Saad, G. Yasin and W. Zhao, Small, 2024, 20, e2310082.
31. T. Li, Y. Liu, Y. Huang, L. Zhang, Z. Chen, W. Yang, G. Shi, J. Zhou, R. Zou, J. Gan, L. Zhong and X. Peng, Small, 2024, 20, e2311675.
32. R. He, D. Zhang, G. Xu, C. Li and J. Bai, Inorg. Chem. Front., 2020, 7, 2783–2790.
33. X. Zhang, W. Li, J. Zhang, H. Zhou and M. Zhong, Chin. J. Catal., 2025, 68, 404–413.
34. J. Zhu, R. Pitcheri, T. Kang, C. Jiao, Y. Guo, J. Li and Y. Qiu, J. Electroanal. Chem., 2019, 833, 151–159.
35. M. Zhou, X. Jiang, W. Kong, H. Li, F. Lu, X. Zhou and Y. Zhang, Nanomicro Lett., 2023, 15, 166.
36. X. Li, X. Z. Zhao, Y. Zhou, J. Hu, H. Zhang, X. Hu and G. Hu, Appl. Surf. Sci., 2022, 584, 152556.
37. X. Zhao, G. Hu, F. Tan, S. Zhang, X. Hu, A. Kuklin, G. Baryshnikov, H. °Agren, X. Zhou and H. Zhang, J. Mater. Chem. A, 2021, 9, 23675–23686.
38. J. Guo, X. Wang, Z. Guo, B. Guo, Z. Wang, Z. Shang, J. Ma and M. Wu, Carbon, 2024, 228, 119294.
39. C. Zhang, W. Shen, K. Guo, M. Xiong, J. Zhang and X. Lu, J. Am. Chem. Soc., 2023, 145, 11589–11598.
40. H. N. Fernandez-Escamilla, J. Guerrero-Sanchez, E. Contreras, J. M. Ruiz-Marizcal, G. Alonso-Nunez, O. E. Contreras, R. M. Felix-Navarro, J. M. Romo-Herrera and N. Takeuchi, Adv. Energy Mater., 2020, 11, 2002459.
41. S. W. Gao, Y. X. He, G. C. Yue, H. K. Li, S. A. Li, J. C. Liu, B. B. Miao, J. Bai, Z. M. Cui, N. Wang, Q. F. Zhang, L. Jiang and Y. Zhao, Carbon Energy, 2023, 5, e319.
42. J. Cao, X. Zhang, Y. Wang, S. Sha, Q. Wu, X. Sun and H.-G. Wang, Chem. Eng. J., 2024, 502, 157873.
43. X. Liu, Y. Li, H. Wang, J. Liu, J. Fu, J. Liu and S. Li, Carbon Letters, 2024, 34, 1995–2011.
44. R. He, L. Sun, K. Ren, X. Li, P. Tian, G. Ning and J. Ye, ChemSusChem, 2025, 18, e202500059.
45. M. Wu, X. Zhao, J. Gao, J. Guo, J. Xiao and R. Chen, Surf. Interfaces, 2022, 31, 101983.
46. X. Ouyang, W. Qiao, Y. Yang, B. Xi, Y. Yu, Y. Wu, J. Fang, P. Li and S. Xiong, Angew. Chem., Int. Ed., 2025, 64, e202422585.
47. C. Wang, J. Fan, M. Sun, L. Xu, P. Xia, Q. He, Z. Ye and X. Duan, Appl. Catal., B, 2025, 367, 125104.
48. J. Zhang, Y. Lou, Z. Wu, X. Huang and S. Sun, J. Am. Chem. Soc., 2024, 146, 24966–24977.
49. W. Gao, K. Xie, J. Xie, X. Wang, H. Zhang, S. Chen, H. Wang, Z. Li and C. Li, Adv. Mater., 2023, 35, e2202952.
50. Z. Niu, S. Fan, X. Li, J. Duan and A. Chen, Appl. Catal., B, 2023, 322, 122090.
51. Y. Wang, F. Hao, M. Sun, M. Liu, J. Zhou, Y. Xiong, C. Ye, X. Wang, F. Liu, J. Wang, P. Lu, Y. Ma, J. Yin, H. Chen, Q. Zhang, L. Gu, H. Chen, B. Huang and Z. Fan, Adv. Mater., 2024, 36, e2313548.

---