Ammonia (NH₃), being a crucial industrial chemical, finds extensive applications in the textile, plastics, medicine, and other industries. Moreover, its remarkable energy density (3000 kW/kg) and high hydrogen storage capacity (17.7 wt%) render it a promising next-generation energy carrier [1-3]. However, the current large-scale production of NH₃ heavily relies on the conventional Haber-Bosch process which leads to energy consumption and pollution issues. Besides, nitrate (NO₃⁻) is a prevalent nitrogen source, particularly in surface and groundwater contaminants, and its excessive accumulation poses a grave threat to public health. The removal of NO₃⁻ contaminants has significant implications for maintaining the balance of the global N-cycle [4-6]. Electrochemical nitrate reduction reaction (eNO₃RR) has emerged as an ingenious method that not only enables effective treatment of nitrate pollutants but also facilitates NH₃ production with high added value [7-10]. Nevertheless, eNO₃RR involves intricate reaction pathways encompassing multiple electron and proton transfers, resulting in various undesired by-products during the multiple reaction pathways [11]. Consequently, developing electrocatalysts with superior activity and selectivity for eNO₃RR-to-NH₃ conversion, which is a desirable but challenging task.

In recent years, supported metal catalysts composed of rigid support (e.g., metal oxides, carbon-based materials, zeolites) and anchored active metals have attracted wide attention due to their unique geometry and electronic structure, which exhibit excellent in the realm of catalysis [12-14]. Supported Cu-based catalysts have been preferentially employed as primary catalysts for eNO₃RR owing to their variable valence state, favorable NO₃⁻ adsorption, effective inhibition of hydrogen evolution reaction (HER), and cost-effectiveness [15-20]. However, the scaling up of NH₃ production for practical applications is still not desirable with a few urgent issues, such as the cumulative effect of nitrite (NO₂⁻) on Cu, as well as high overpotential and poor stability [18].

Notably, the electronic metal-support interaction (EMSI) between the active metal and functiornal support significantly influences adsorption property and catalytic activity [21]. In this regard, the immobilizing Cu onto a reducible oxide with abundant surface oxygen vacancies, investigating their structure-activity relationship, and modulating the EMSI would constitute a feasible strategy for enhancing catalytic performance of eNO₃RR. The obvious EMSI is caused by abundant oxygen vacancy, which can optimize the electronic structure and make the charge density of active metal highly delocalized, resulting in positive charge, as well as the enhancement of the formation of highly active metal species and the exposed active sites amount [22-24]. The phenomenon can facilitate charge transfer and modulate the surface adsorption/desorption behavior of intermediate molecules during the reaction process, while effectively suppress the HER and reduce the cumulation of NO₂⁻, thereby achieving the industrial goal for NH₃ synthesis with high efficiency [25-32].

In this work, we designed and synthesized a Cu-based supported catalyst (Cu-ZrO₂) for eNO₃RR-to-NH₃. As demonstrated by electrochemical studies, the Cu-t-ZrO₂ had the excellent performance with a 97.54% FE of NH₃ and the corresponding NH₃ yield of 33.64 mg h⁻¹ mg_cat⁻¹ and partial current density of 45.79 mA/cm² at −0.4 V vs. RHE, and the electrocatalyst exhibited stability over up to seven cycles/20 h of testing. The detailed characterization clearly demonstrated the presence of an EMSI effect between the supported metal Cu and the t-ZrO₂ support. The t-ZrO₂ exhibited a higher concentration of oxygen vacancies compared to m-ZrO₂ and electrons were more likely to be transferred from Cu to t-ZrO₂, facilitating charge transfer during the reaction process. Additionally, the lower work function of Cu results in electron deficiency and formation of highly active Cu species, which enhanced NO₃⁻ adsorption and selective NH₃ conversion. This study provides new inspiration for constructing eNO₃RR electrocatalysts with EMSI effect.

The synthetic procedure of the Cu-ZrO₂ catalyst is schematically shown in Fig. 1a. Two types of ZrO₂ supports (monoclinic-ZrO₂ and tetragonal-ZrO₂) were initially synthesized using a facile co-precipitation method, and then we loaded Cu on m-, and t-ZrO₂ supports via the impregnation method (details in the Preparation of Catalysts section of Supporting information), which was subsequently annealed at 400 ℃ to yield Cu-m-ZrO₂ or Cu-t-ZrO₂. From the scanning electron microscopy (SEM) images, it can be observed that both Cu-m-ZrO₂ and Cu-t-ZrO₂ catalysts exhibited a morphology of irregularly shaped lumps aggregates, characterized by rough and porous surface (Fig. S1 in Supporting information). The transmission electron microscopy (TEM) images illustrated that the catalyst of Cu-t-ZrO₂ exhibited a uniform dispersion of CuO (Figs. 1b and c). Moreover, the high-resolution TEM images revealed well-defined lattice fringes of CuO (002) with a spacing of 0.243 nm, indicating that the predominantly exposed facet of CuO corresponds to the (002) plane. Similarly, the lattice fringes of ZrO₂ (011), were observed at a spacing of 0.313 nm (Figs. 1d and e). Energy-dispersive spectroscopy (EDS) elemental mapping also showed that Zr, Cu and O were both uniformly distributed in Cu-t-ZrO₂ (Figs. 1f and i-k), which was initially proved that Cu was successfully doped to t-ZrO₂ supports and existed in the form of CuO. The X-ray photoelectron spectroscopy (XPS) spectra of catalysts also proved the existence of Cu (Fig. S2 in Supporting information).

X-ray diffraction (XRD) was performed to investigate the crystalline phase structure of t-ZrO₂, m-ZrO₂, Cu-t-ZrO₂ and Cu-m-ZrO₂ (Fig. 2a). Characteristic diffraction peaks of ZrO₂ were observed in the resulting Cu-ZrO₂. The absence of diffraction peaks for Cu can be attributed to its amorphous structure and/or low loading content. The actual Cu content in Cu-t-ZrO₂ and Cu-m-ZrO₂ was quantified as 0.183 wt% and 0.145 wt%, respectively, using Inductively Coupled Plasma (ICP) analysis (Table S1 in Supporting information). Then the high-resolution Zr 3d XPS spectra were analyzed for all catalysts, with the spectra displayed in Fig. 2b. The XPS spectra of Zr 3d_5/2 in ZrO₂ can be deconvoluted into two peaks, assigned to partially reduced Zr^(4-x)+ (related to oxygen vacancy) and stoichiometric ZrO₂ (Zr⁴⁺). Compared with Cu-m-ZrO₂ (26.45%), Cu-t-ZrO₂ contains more Zr^(4-x)+ species (34.39%), there was a possibility of an increased presence of oxygen vacancies. At the same time, combined with the O 1s XPS spectra shown in Fig. 2c. The two prominent peaks centered at 532.0 eV and 533.3 eV can be attributed to oxygen vacancies and the hydroxyl species of surface-adsorbed water molecules, respectively. The peak area ratios of oxygen vacancies are 49.24%, 52.34%, 53.76% and 61.68% for m-ZrO₂, Cu-m-ZrO₂, t-ZrO₂ and Cu-t-ZrO₂, respectively. Therefore, Cu-t-ZrO₂ had more oxygen vacancies than Cu-m-ZrO₂. Cu-m-ZrO₂ and Cu-t-ZrO₂ were further investigated using the electron paramagnetic resonance (EPR). The EPR spectra exhibited axial signals with the g value of 2.003 (Fig. S3 in Supporting information), further corroborating the presence of oxygen vacancies, and Cu-t-ZrO₂ demonstrated a more pronounced axial signal compared to Cu-m-ZrO₂. This implied that Cu-t-ZrO₂ exhibited a higher concentration of oxygen vacancies, which was consistent with the XPS findings.

Furthermore, to investigate the interaction between Cu and ZrO₂ supports, the XPS of Zr 3d and Cu 2p were analysed. As shown in Fig. 2b, for m-ZrO₂, after loading Cu (Cu-m-ZrO₂), the binding energy of Zr 3d shifted from 181.74 eV to 181.62 eV, and a negative shift of 0.12 eV occurred. At the same time, the Zr 3d binding energy of Cu-t-ZrO₂ exhibited a more negative shift by 0.15 eV Fig. 2d showed that the Cu 2p XPS spectra exhibited significantly distinct Cu²⁺ features, and Cu⁺ or Cu⁰ species existed in Cu-t-ZrO₂ and Cu-m-ZrO₂, but it was hard to distinguish from XPS spectra due to the similarity in the binding energies. Therefore, the Cu LMM Auger spectra were also performed to provide more insight into the electronic structures and surface state (Fig. S4 in Supporting information), indicating the coexistence of Cu⁺ and Cu⁰ species on Cu-t-ZrO₂ and Cu-m-ZrO₂. Besides, the Cu 2p spectra revealed a slight positive shift in the binding energies of Cu-t-ZrO₂ for both Cu 2p_3/2 and Cu 2p_1/2, compared to those of Cu-m-ZrO₂. The results suggested the presence of electronic interaction between Cu and ZrO₂ support. The shift in binding energy also indicated that the electrons of Cu move towards Zr, resulting in an enlargement of the electron cloud on Zr. Additionally, in Cu-t-ZrO₂, the electron cloud surrounding Cu demonstrated a more pronounced shift towards Zr compared to Cu-m-ZrO₂. The semi-in situ XPS was further carried out to understand the relationship between EMSI and catalytic behavior. The Cu 2p semi-XPS spectra and Zr 3d semi-XPS spectra were measured after 10 min and 30 min operation at −0.4 V vs. RHE (Fig. S5 in Supporting information). During the progression of the reaction, the binding energies of Zr^(4-x)+ and Cu²⁺ in Cu-t-ZrO₂ shifted in opposite directions. This significant displacement signified a heightened electronic metal-support interaction in Cu-t-ZrO₂, which correlated with a superior eNO₃RR activity.

The interaction between Cu and ZrO₂ support was further investigated through ultraviolet photoelectron spectroscopy (UPS), wherein the evaluation of the work function (WF) of the supported Cu. The WF of Cu-m-ZrO₂ and Cu-t-ZrO₂, as shown in Figs. 2e and f respectively, were determined to be 3.22 eV and 2.97 eV using the E_cut-off method (indicated by the dotted lines intercepting at the x axis). Compared with Cu-m-ZrO₂, the smaller WF of Cu-t-ZrO₂ means that electrons escape more easily on the catalyst surface, indicating that more oxygen vacancies could enhance the EMSI between the Cu and t-ZrO₂ support. The resulting EMSI effect order of Cu-t-ZrO₂ > Cu-m-ZrO₂ is in good agreement with the XPS results. Besides, the lower work function of Cu results in electron deficiency and formation of highly active Cu species, which may be considered as the main active sites for eNO₃RR.

The electrocatalytic reduction of NO₃⁻ measurements was evaluated under ambient conditions in 1 mol/L KOH electrolyte containing 0.1 mol/L KNO₃ using a typical electrochemical H-cell with a standard three-electrode system. The reactants and products were quantified by colorimetric methods based on the linear relationship between concentrations and absorbance, including NH₃, NO₂⁻, and NO₃⁻ (Figs. S6-S8 in Supporting information). All potentials were expressed against the RHE unless special explanation. The electrocatalytic performance was initially assessed through linear sweep voltammetry (LSV). Fig. 3a illustrates the LSV curves measured in a 1 mol/L KOH electrolyte, both with and without the addition of NO₃⁻. It was evident that the current density significantly increased when NO₃⁻ was added, as compared to its absence, indicating the remarkable electrocatalytic response of Cu-ZrO₂ towards NO₃⁻ reduction. Then, the electrocatalytic kinetics of all catalysts were further studied. As shown in Fig. 3b, the tafel slope of Cu-t-ZrO₂ was calculated to be 413.9 mV/dec, much smaller than those of Cu-m-ZrO₂ (439.1 mV/dec) and t-ZrO₂ (700.1 mV/dec), respectively, demonstrating a faster eNO₃RR rate with lower overpotential for Cu-t-ZrO₂ over those of other catalysts.

To further probe the origin of high performance of Cu-t-ZrO₂, the measured double-layer capacitance was used to infer the electrochemically active surface area (ECSA) of different catalysts (Fig. S9 in Supporting information). As shown in Fig. 3c, Cu-t-ZrO₂ had the largest ECSA, which was about 1.5 and 2.3 times that of Cu-m-ZrO₂ and t-ZrO₂, respectively, suggesting that Cu-t-ZrO₂ may provide more exposed catalytical active sites and conducive to the adsorption and transformation of NO₃⁻ in the process of eNO₃RR, thereby confirming that Cu-t-ZrO₂ had high intrinsic catalytic performance. In addition, electrochemical impedance spectroscopy (EIS) measurements confirmed the enhanced charge transfer of Cu-t-ZrO₂ in comparison to Cu-m-ZrO₂ and t-ZrO₂, which promoted the kinetics of eNO₃RR reaction (Fig. S10 in Supporting information). The findings suggested that Cu-t-ZrO₂ with EMSI facilitated enhanced eNO₃RR activity by promoting the generation of more active sites and accelerating charge transfer in reaction [33, 34]. Subsequently, chronoamperometry (i-t) testing was conducted for 2 h at various potentials to evaluate the eNO₃RR performance over the Cu-t-ZrO₂ catalyst. As shown in Fig. S11 (Supporting information), the current density remained stable during the reaction and enhanced with the increase of potential. Figs. 3d and e showed FEs and NH₃ yields at the corresponding potentials. The NH₃ yields of three catalysts gradually increased from −0.3 V to −1.1 V, while the FE exhibited a magnificent volcanic-shaped trend. At −0.4 V, the maximum FE of NH₃ on Cu-t-ZrO₂ reached an astonishing 97.54%, and the NH₃ yield was 9.143 mg h⁻¹ mg_cat⁻¹. The current density of t-ZrO₂, Cu-m-ZrO₂ and Cu-t-ZrO₂ catalysts increased with the increase of potential (Fig. S12 and Table S2 in Supporting information). The electroreduction measurements were also conducted for Cu-t-ZrO₂ and Cu-m-ZrO₂ under identical conditions, revealing their inferior catalytic performance compared to Cu-t-ZrO₂. Clearly, Cu-t-ZrO₂ had the highest NH₃ yield at a potential of −1.1 V, reaching 33.64 mg h⁻¹ mg_cat⁻¹, which was higher than Cu-m-ZrO₂ and t-ZrO₂ (25.58 mg h⁻¹ mg_cat⁻¹ and 10.32 mg h⁻¹ mg_cat⁻¹, respectively). Subsequently, we tested the Cu-SiO₂ and Cu-C (inert supports) under the same conditions, the results about FEs and NH₃ yield were shown in Fig. S13 (Supporting information). By comparing the FEs and NH₃ yields, it was further proved that Cu-t-ZrO₂ had significantly high eNO₃RR activity. This was mainly due to the strong EMSI effect between the supported metal Cu and t-ZrO₂ support, which enhanced NO₃⁻ adsorption and selective NH₃ conversion. In addition, during the time-varying electrolysis of eNO₃RR (Fig. 3f), the concentration of NO₃⁻-N decreased sharply, the concentration of NH₃—N increased significantly, and the concentration of NO₂⁻-N was almost negligible. The results showed that Cu-t-ZrO₂ had high eNO₃RR selectivity and it did not cause NO₂⁻ accumulation during the whole reaction process. We further investigated the selectivity of other by-products in the eNO₃RR process (Fig. S14 in Supporting information). Cu-t-ZrO₂ exhibited low FE for NO₂⁻ and H₂, with no detection of N₂, indicating the exceptional selectivity of Cu-t-ZrO₂ towards the conversion of NO₃⁻ to NH₃.

To eliminate potential interference from the electrocatalyst itself or external factors, a comparative test was conducted. Electrolyte without NO₃⁻ addition and without applying a potential were also tested. As shown in Fig. 4a, the electrochemical measurements conducted in a blank electrolyte containing 1 mol/L KOH yielded insignificant quantities of NH₃ and the same result was shown without applying potential, proving that NH₃ came from electrocatalytic NO₃⁻ reduction and not from the system itself or other sources of nitrogen pollution [35, 36].

Based on the above results, we evaluated the eNO₃RR on the Cu-t-ZrO₂ catalyst in 1 mol/L KOH electrolytes containing various concentrations of NO₃⁻ (Fig. 4b) [37, 38]. When the NO₃⁻ concentration was gradually reduced from 2000 ppm to 100 ppm, the FE of NH₃ showed a slight decline, however, the FE of NH₃ can still exceed 80% even when the concentration of NO₃⁻ was further reduced (Fig. S15 in Supporting information). The slight decrease in catalytic performance at low concentrations may be due to the limitation of NO₃⁻ diffusion and the competition of HER. Therefore, the results showed that the Cu-t-ZrO₂ catalyst had a wide range of applications. Moreover, stability was another critical parameter for evaluating the practical application of a catalyst [39, 40]. The cycling electrolysis experiments showed that the FE and NH₃ yield were no obvious fluctuations after seven consecutive recycling tests using the same piece of electrode (Fig. 4c). It shows that Cu-t-ZrO₂ had good cycle stability. After continuous electrolysis for 20 h (Fig. 4d), the FEs and NH₃ yield of Cu-t-ZrO₂ catalyst remained at 97.5% and ~9.2 mg h⁻¹ cm⁻¹, respectively. At the same time, Fig. S16 (Supporting information) showed that the FE and NH₃ yields of different batches of Cu-t-ZrO₂ catalysts remained stable without significant differences. The results indicated that Cu-t-ZrO₂ catalyst exhibited good long-term stability and reproducibility in eNO₃RR. As shown in Fig. 4e, this work was compared with the performance of some previously reported electrocatalysts for NO₃⁻ reduction. Notably, the FEs and NH₃ yield of Cu-t-ZrO₂ were superior to other values reported in the literature (see detailed contents in Table S3 in Supporting information).

The valence band spectra were conducted to elucidate the charge-transfer mechanism involved in NO₃⁻ adsorption and activation on Cu-t-ZrO₂ during electrolysis. Fig. S17 (Supporting information) illustrated the downward shift of the valence band edge of Cu-t-ZrO₂ following NO₃⁻ reduction, primarily attributed to the strong adsorption of NO₃⁻, as well as the occurrence of charge transfer on Cu-t-ZrO₂ during eNO₃RR. The conversion of eNO₃RR-to-NH₃ involves a complex eight-electron transfer process (NO₃⁻ + 9H⁺ + 8e⁻ → NH₃ + 3H₂O). The reaction commences with the initial adsorption of NO₃⁻ onto the Cu sites of electrode (Cu-t-ZrO₂) surface. Subsequently, the adsorbed NO₃⁻ is reduced to NO₂⁻, followed by a series of protonation processes leading to the ultimate formation of NH₃. As shown in Fig. 4f, the rich oxygen vacancy t-ZrO₂ support provides abundant interfaces for the activation of NO₃⁻. The EMSI between Cu and t-ZrO₂ support enables the efficient electron transfer from Cu to t-ZrO₂ support and triggers interfacial charge polarization, which makes Cu atom sites electron deficient, facilitating the enrichment of NO₃⁻ and the formation of the key intermediates (*NO₂ and *NO). The cathodic discharge of Zr atom sites could catalyze the dissociation of H₂O molecules to form adsorbed H species, which are then transferred from the Zr site to the *NO adsorption intermediate located at the Cu site, thus facilitating the hydrogenation step of NH₃ synthesis [41-44].

In summary, Cu-based supported catalyst (Cu-t-ZrO₂) with EMSI effect has been successfully synthesized for efficient eNO₃RR to NH₃. The highest FE and NH₃ yield could reach 97.54% and 33.64 mg h⁻¹ mg_cat⁻¹, respectively. Moreover, the catalyst also showed good stability. The study shows that the presence of abundant oxygen vacancy enhances the EMSI effect, which can accelerate charge transfer during the reaction and improve the adsorption and activation of NO₃⁻ and key intermediates, thus promoting the reduction of NO₃⁻. This work presents a promising approach for achieving high-efficiency eNO₃RR, which has the potential to inspire new explorations and principles in the design of electrocatalysts for industrial-level NH₃ synthesis.

Declaration of competing interests

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Weiwei Guo reports financial support was provided by Natural Scientific Foundation of China (Nos. 22127803, 22174110, 22203050). Weiwei Guo reports financial support was provided by Natural Scientific Foundation of Shandong (No. ZR2022QB002). Weiwei Guo reports financial support was provided by China Postdoctoral Science Foundation (No. 2020T130331). If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT authorship contribution statement

Doudou Liu: Validation. Weiwei Guo: Writing – review & editing, Methodology, Funding acquisition. Guoliang Mei: Writing – review & editing. Youpeng Dan: Writing – review & editing. Rong Yang: Writing – review & editing. Chao Huang: Writing – review & editing. Yanling Zhai: Supervision, Funding acquisition. Xiaoquan Lu: Supervision, Funding acquisition.

Acknowledgements

This work was supported by the Natural Scientific Foundation of China (Nos. 22127803, 22174110, 22203050), Natural Scientific Foundation of Shandong (No. ZR2022QB002), and China Postdoctoral Science Foundation (No. 2020T130331).

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110578.