Nitrogen, essential in biogeochemistry, manifests in forms such as ammonia (NH₃), hydrazine (N₂H₄), and nitrite/nitrate (NO₂^–/NO₃^–), vital for Earth's sustainability. NH₃ stands out, widely used in agriculture and as a high-energy, zero-emission green energy source [1-4]. Traditionally, NH₃ production has relied on the energy-intensive Haber-Bosch process, resulting in significant energy consumption and carbon emissions [5]. The electrochemical N₂ reduction reaction (NRR) has gained considerable attention as a sustainable alternative for NH₃ synthesis, operating at room temperature and driven by renewable energy sources. Nevertheless, NRR faces challenges related to the robustness of the N≡N triple bond, N₂ solubility in water, and the simultaneous hydrogen evolution reaction (HER), limiting its broad applicability [6-15]. In contrast, nitrate (NO₃^–) features a lower bond energy, high solubility in aqueous solutions, and a superior theoretical reaction potential, rendering the electrochemical NO₃^– reduction reaction (NO₃^–RR) a more appealing option than NRR [16-20]. Moreover, excessive NO₃^– in wastewater from agricultural and factory wastewater sources poses health risks. Remediation techniques, like ion exchange and reverse osmosis, help, but often leave high NO₃^– level [18,21,22]. The NO₃^–RR process is thus promising, transforming NO₃^– into valuable NH₃ and fostering global nitrogen equilibrium, embodying sustainability. However, NH₃ production via NO₃^–RR involves an eight-electron transfer process, potentially leading to by-product formation. Given these considerations, the development of efficient and highly selective electrocatalysts for NO₃^–RR is an urgent research imperative.

Noble metal-based catalysts (e.g., Rh [23], Ru [24], Pt [25]) proficiently facilitate the conversion of NO₃^– to NH₃, yet their prohibitive costs and scarce availability constrain their practical utilization. In recent years, many studies have focused on developing transition metal-based catalysts with abundant resources, low cost, and high electrocatalytic activity as efficient NO₃^–RR [26,27]. In this context, cobalt (Co) has emerged as a notable contender, with recent theoretical and empirical studies showing its efficacy in facilitating NH₃ synthesis via NO₃^–RR among transition metals [28-30]. Co-based catalysts, specifically oxide [31], sulfide [32], and phosphide [33], have showcased exceptional selectivity and catalytic efficiency in this conversion process. Distinguished among its counterparts, Co-based selenide boasts near-metallic electrical conductivity, drawing considerable interest for its operational durability, and remarkable hydrogen adsorption capacity, which are pivotal for catalytic hydrogenation reactions [34-37]. Moreover, selenides have emerged as attractive and multi-functional electrocatalysts due to their high catalytic activity, tunable band gap, and good durability [38,39]. Shen et al. reported selenium vacancy-rich WSe₂ (WSe_2-x) nanosheets as electrocatalysts capable of electrochemically converting NO₃^– to NH₃ with a high FE of 92.7% [40]. From a morphological perspective, the one-dimensional nanowire structure provides specific anti-aggregation ability, anisotropy, and an uninterrupted electron transfer pathway, reducing charge transfer resistance and enhancing catalytic performance [41,42]. Therefore, it is anticipated that Co-based selenide nanowire could facilitate a more efficient generation of NH₃ through the NO₃^–RR process.

Herein, we present a CoSe₂ nanowire array on carbon cloth (CoSe₂/CC) as a highly efficient electrocatalyst for NH₃ production via NO₃^–RR. Operational in a 0.1 mol/L NaOH electrolyte containing 0.1 mol/L NO₃^–, this CoSe₂/CC catalyst delivered an impressive NH₃ yield of 517.7 µmol h^–1 cm^–2 and achieved a Faradaic efficiency (FE) of 97.6% at –0.6 V versus the reversible hydrogen electrode (RHE). Notably, the CoSe₂/CC maintained exceptional stability during continuous electrochemical operations. Additionally, insights into the reaction mechanism of NO₃^–RR on CoSe₂ were gleaned through comprehensive density functional theory (DFT) calculations, further elucidating the catalytic process.

The CoSe₂ nanowire array on CC was fabricated via straightforward hydrothermal and selenization processes (Fig. 1a), with details provided in Supporting information. The X-ray diffraction (XRD) analysis, shown in Fig. 1b, validates the characteristic peaks of CoSe₂ (JCPDS No. 88–1712) [43], with the broad diffraction peak at ~26° corresponding to the CC substrate (Fig. S1 in Supporting information) [44]. As illustrated in Fig. S2 (Supporting information), CoSe₂/CC can still be obtained by changing the annealing temperature or the heating rate during the selenization process, while the product becomes CoSe₂@Co₃Se₄@CoO/CC when less Se powder was added. Scanning electron microscopy (SEM) images depicted in Fig. S3 (Supporting information) demonstrate the integral encapsulation of the CC by Co(OH)F nanowire. After selenization, the CoSe₂/CC maintains its nanowire array with an enhanced thickness and roughness of the nanowires (Figs. 1c and d). Meanwhile, CoSe₂/CC-500 and CoSe₂/CC-10 acquired by annealing at different temperatures and heating rates, respectively, still maintained the nanowire array structure (Fig. S4 in Supporting information). The energy-dispersive X-ray (EDX) mapping, illustrated in Fig. 1e, confirms the uniform distribution of Co and Se elements within the nanowire arrays. Moreover, the EDX spectrum shows the absence of F element in the prepared CoSe₂/CC (Fig. S5 in Supporting information). Transmission electron microscopy (TEM) image in Fig. 1f further show a nanowire characteristic. High-resolution TEM (HRTEM) image shows a lattice spacing of 0.26 nm, matching the (210) facet of CoSe₂ (Fig. 1g). The surface elemental composition and valence states of CoSe₂ were analyzed by X-ray photoelectron spectroscopy (XPS). As illustrated in Fig. S6 (Supporting information), the Co 2p spectrum can be roughly separated into Co 2p_3/2 and Co 2p_1/2 peaks. The peaks at 778.4 eV and 779.3 eV in Co 2p_3/2 correspond to Co³⁺ and Co²⁺, respectively, while the Co 2p_1/2 peaks are at 793.4 eV for Co³⁺ and 796.0 eV for Co²⁺ [45,46]. The presence of Co²⁺ indicates that the material is a CoSe₂ crystal, while the formation of Co³⁺ corresponds to the partial surface oxidation of CoSe₂ in air. In aspects of Se 3d (Fig. S7 in Supporting information), two peaks for Se 3d_3/2 (55.6 eV) and Se 3d_5/2 (54.6 eV) are attributed to Se^2–. Moreover, the additional peaks at the interval 57–65 eV all belong to Co 3p_1/2, Co 3p_3/2, and SeO₂ [47]. Using XPS analysis for element F, there was no characteristic peak of F detected in the F 1s spectrum, further verifying that no residual F element exists in CoSe₂/CC (Fig. S8 in Supporting information). The above findings demonstrate that the preparation of CoSe₂ loaded on carbon cloth was successful.

The electrocatalytic NO₃^–RR performance of CoSe₂/CC was evaluated in both NO₃^–-free and NO₃^–-containing 0.1 mol/L alkaline electrolyte. The possible NO₃^–RR products, like NH₃, N₂H₄, and NO₂^–, were quantified using distinct colorimetric methods, with their respective UV–vis spectra and calibration curves detailed in Figs. S9-S11 (Supporting information). Initial assessment of CoSe₂/CC's NO₃^–RR electrocatalytic efficiency employed linear sweep voltammetry (LSV) curves. As depicted in Fig. 2a, the current density (j) of CoSe₂/CC significantly increased upon the introduction of 0.1 mol/L NO₃^– into the 0.1 mol/L NaOH electrolyte. Comparative LSV tests of pure CC, the precursor Co(OH)F/CC, and CoSe₂ powder/CC under identical conditions (Figs. S12 and S13 in Supporting information) revealed lower j for both pure CC (–71.2 mA/cm²), Co(OH)F/CC (–97.6 mA/cm²), and CoSe₂ powder/CC (–128.0 mA/cm²), as compared to CoSe₂/CC (–207.6 mA/cm²) at –1.0 V. We further compared the three catalysts (CoSe₂/CC, CoSe₂/CC-500, and CoSe₂/CC-10) generated under different preparation conditions for NO₃^–RR. Preliminarily, it was found that the j of these three catalysts did not differ much within the alkaline electrolyte with and without NO₃^– (Fig. S14 in Supporting information). Chronoamperometry (CA) tests, conducted for 1 h at various potentials ranging from –0.3 V to –0.9 V (Fig. 2b), further probed CoSe₂/CC's NO₃^–RR capabilities. The UV–vis spectra post-CA tests (Fig. S15 in Supporting information) exhibited progressively intensifying absorption peak intensities with increased negative potentials, indicating a preference for NH₃ synthesis at higher negative potentials. Fig. 2c illustrates that the NH₃ yield of CoSe₂/CC increased consistently with more negative operating potentials, reaching a peak yield of 818.0 µmol h^–1 cm^–2 at –0.9 V. Additionally, the FE curve displayed a volcano-shaped profile, achieving a maximum FE of 97.6% at –0.6 V, correlating with an NH₃ yield of 517.7 µmol h^–1 cm^–2, thereby confirming the superior NO₃^–RR performance for NH₃ synthesis. The electrochemical NO₃^– reduction to NH₃ capabilities of Co(OH)F/CC, CC, and CoSe₂ powder/CC were compared under identical conditions at –0.6 V (Fig. 2d and Fig. S16 in Supporting information). The results showed that the NH₃ yields and FEs for Co(OH)F/CC (292.6 µmol h^–1 cm^–2, 65.3%), CC (19.7 µmol h^–1 cm^–2, 23.6%), and CoSe₂ powder/CC (150.1 µmol h^–1 cm^–2, 71.9%) were significantly lower than those of CoSe₂/CC (517.7 µmol h^–1 cm^–2, 97.6%). Also, the NH₃ yields and FEs of CoSe₂–500 (498.0 µmol h^–1 cm^–2, 90.0%) and CoSe₂/CC-10 (438.2 µmol h^–1 cm^–2, 83.4%), prepared under different conditions, were only slightly lower than that of CoSe₂/CC (Fig. S17 in Supporting information). Remarkably, the highest NH₃ yield and FE of CoSe₂/CC surpass those of some recently reported NO₃^–RR electrocatalysts (Fig. 2e and Table S1 in Supporting information). This exceptional NO₃^–RR performance of CoSe₂/CC substantiates the role of CoSe₂ as the active phase for efficient and highly selective electrocatalytic conversion of NO₃^– to NH₃.

Since the electrochemical synthesis of NH₃ from NO₃^– involves an eight-electron transfer process, the formation of by-products (e.g., NO₂^–, N₂H₄, and H₂) is possible. Remarkably, our study observed no production of N₂H₄ across all tested potentials (Fig. S18 in Supporting information). Furthermore, the yields and FEs of NO₂^– and H₂ generated during the electrocatalytic NO₃^–RR were minimal, almost negligible compared to NH₃ (Fig. 3a). This low yield of H₂ and its FEs indicate that the CoSe₂/CC catalyst effectively suppresses the competing HER. The partial current densities (j_partial) of different products in the electrolyte are shown in Fig. 3b. At –0.9 V, the j_partial for NH₃ (–172.1 mA/cm²) is significantly higher than that for NO₂^– (–1.3 mA/cm²) and H₂ (–5.2 mA/cm²), further demonstrating CoSe₂/CC's exceptional selectivity for the electroreduction of NO₃^– to NH₃. To confirm the source of the detected NH₃, control experiments were conducted on CoSe₂/CC. The NH₃ yield after 1-h electrolysis at open-circuit potential (OCP) was comparable to that in fresh electrolyte (Fig. S19 in Supporting information), ruling out contamination from experimental equipment or reagents. Alternating tests in NO₃^–-containing and NO₃^–-free alkaline electrolytes at –0.6 V (Fig. 3c and Fig. S20 in Supporting information) revealed high NH₃ yields and FEs only in NO₃^–-containing solutions. Additionally, ¹⁵N isotope labeling experiments, analyzed via ¹H nuclear magnetic resonance (¹H NMR) spectra, validated the nitrogen origin in NH₃. The ¹H NMR spectra (Fig. 3d) showed distinct peaks for ¹⁵NH₄⁺ and ¹⁴NH₄⁺, confirming that the NH₃ originated from the electrocatalytic reduction of NO₃^–, not external sources. Stability is crucial for assessing an electrocatalyst's performance. We conducted a 12-h continuous NO₃^–RR test on CoSe₂/CC at –0.6 V. Fig. S21 (Supporting information) shows consistent j around –100 mA/cm² throughout the test, indicating stable long-term operation. Additionally, 12 repeated cycles at the optimal NH₃-producing potential (–0.6 V) demonstrated CoSe₂/CC's excellent reproducibility. The NH₃ yields and FEs, based on overlapping CA curves, accumulated charges, and corresponding UV–vis spectra, exhibited only minor variations (Fig. 3e and Fig. S22 in Supporting information). Significantly, XRD and SEM of CoSe₂/CC after 12-h electrolysis (Figs. S23 and S24 in Supporting information) show that its crystalline phase and nanowire array structure have remained almost unchanged. Those conclusions demonstrate the durable electrocatalytic activity of CoSe₂/CC in the NO₃^–RR.

DFT computations were performed to elucidate the exceptional NO₃^–RR properties of CoSe₂. The CoSe₂(210) slab model in Fig. S25 (Supporting information) was constructed based on the HRTEM results. As indicated by the calculated electronic density of states (DOS) in Fig. 4a, CoSe₂ exhibits a local density of states near the Fermi level, confirming its metallic behavior as established in prior studies [48]. Fig. 4b illustrates that the electronic state of CoSe₂(210) with adsorbed NO₃^– is more pronounced around the Fermi level compared to the bare CoSe₂(210), implying that NO₃^– adsorption boosts the electrical conductivity of CoSe₂, thereby facilitating electron transfer during the electrocatalytic process. This is further corroborated by charge density differences and Bader charge analysis (Fig. 4c), which reveal that CoSe₂(210) donates a charge of 0.66 e^– to *NO₃, enhancing the activation of N—O bonds and catalyzing the NO₃^–RR process. The free energy landscape of NO₃^–RR on CoSe₂(210) surface was mapped out to gain deeper insights into the mechanism (Fig. 4d), with the optimized configurations of NO₃^–RR intermediates (Fig. S26 in Supporting information). The potential-determining step (PDS) of the NO₃^–RR involves the protonation of *NO to *ONH, exhibiting a relatively low energy barrier of 0.65 eV. This underscores the facilitation of NO₃^–RR on the CoSe₂(210) surface. Additionally, the competitive HER was also investigated, given its prominence as a competing reaction in NO₃^–RR [49,50]. The free energy of *NO₃ (–0.37 eV) on the CoSe₂(210) surface, depicted in Fig. 4e, is significantly more negative than that of *H (0.68 eV), suggesting that the CoSe₂(210) surface effectively suppresses the HER side reaction, thereby enhancing the overall efficiency and selectivity of NO₃^–RR to NH₃ conversion.

In summary, we have successfully synthesized CoSe₂/CC and demonstrated its an effective electrocatalyst for NH₃ synthesis via the NO₃^–RR process. The catalyst of CoSe₂/CC achieves a substantial NH₃ yield of 517.7 µmol h^–1 cm^–2 and a FE of 97.6% at –0.6 V. DFT calculations indicate that CoSe₂ possesses a higher energy barrier for H₂ evolution and a lower PDS energy barrier in NO₃^–RR. This research not only introduces a cost-effective and efficient catalyst for NH₃ synthesis but showcases the potential for advancing metal selenides for NO₃^–RR applications.

Declaration of competing interest

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

CRediT authorship contribution statement

Ting Xie: Writing – original draft, Data curation. Xun He: Data curation. Lang He: Data curation. Kai Dong: Data curation. Yongchao Yao: Data curation. Zhengwei Cai: Investigation. Xuwei Liu: Investigation. Xiaoya Fan: Investigation. Tengyue Li: Investigation. Dongdong Zheng: Investigation. Shengjun Sun: Validation. Luming Li: Validation. Wei Chu: Validation. Asmaa Farouk: Validation. Mohamed S. Hamdy: Validation. Chenggang Xu: Writing – review & editing, Conceptualization. Qingquan Kong: Writing – review & editing, Software. Xuping Sun: Writing – review & editing, Conceptualization.

Acknowledgment

The authors extend their appreciation to the Deanship of Scientific Research at King Khalid University for funding support through large group research project (No. RGP2/119/45).

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110005.