Multi-Functional Amorphous Nickel Phosphide Electrocatalytic Reduction of Nitrate for Ammonia Production: Unraveling the Anode-Driven Enhancement Mechanism

Abstract

The electrocatalytic reduction of nitrate (ERN) to ammonia offers a promising route to address energy shortages and environmental pollution, but its practical application is hindered by low selectivity due to complex eight-electron transfer pathways and high energy consumption (EC) from the kinetically sluggish oxygen evolution reaction (OER). This study proposes a dual strategy: (1) designing a multi-functional self-supported ANP electrode via vapor deposition to enhance ERN activity and (2) replacing the OER with thermodynamically favorable anodic reactions (urea oxidation reaction (UOR), sodium metabisulfite oxidation reaction (S(IV)OR), sulfite and urea oxidation reaction (S(IV)/UOR)) to reduce EC. The ANP cathode achieved a nitrate removal rate (R%) of 97.7%, ammonia selectivity (SE%) of 91.8%, and Faradaic efficiency (FE) of 97.3% at −1.2 V, with an ammonia yield of 0.0616 mmol h⁻¹ mg⁻¹ and an EC of 8.239 kWh/kg, while in situ-generated atomic hydrogen (*H) was identified as key to improving nitrate removal and selectivity. Replacing the OER with alternative anodic reactions significantly improved system efficiency: the UOR reduced EC by 17.5%, S(IV)OR saved 27.6% energy with 7.1% higher ammonia yield, and a hybrid S(IV)/UOR system achieved a 32.1% lower EC and a 12.6% greater ammonia yield than the OER. These differences stemmed from variations in cell voltage and ammonia production rates. This work provides a viable approach for selective nitrate-to-ammonia conversion and guides the design of energy-efficient electrocatalytic systems for sustainable nitrogen recovery.

1. Introduction

Ammonia (NH₃), as a low-cost chemical raw material, has the advantages of high energy density, no CO₂ generated by combustion, and easy compression, storage, and transportation. At present, ammonia is predominantly generated by the Haber–Bosch process, which consumes fossil energy and releases CO₂. Among the methods of preparing ammonia, electrochemical technology can be driven by electricity generated by environmentally benign energy (e.g., photovoltaic energy), and utilization of the interconversion of H in H₂O to store and transport energy, basically without carbon emissions [1,2]. At the same time, elevated ${\mathrm{N}\mathrm{O}}_3^{-}$ concentrations in water bodies endanger human well-being [3]. Therefore, the ERN for the ammonia production process under mild conditions is of great significance [4,5,6,7]. However, current ERN systems face three intertwined challenges: (i) the competing hydrogen evolution reaction (HER) limits Faradaic efficiency for NH₃ production, (ii) the eight-electron transfer pathway leads to kinetic bottlenecks, and (iii) the high overpotential of the anodic oxygen evolution reaction (OER) which may dominate energy consumption (EC) [8,9,10]. Critically, no existing electrocatalyst simultaneously resolves these issues while operating in neutral media.

Single- or multi-component electrocatalysts have been reported based on noble metals (Pd, Pt, Au, Ru, Ir, etc.), non-noble metals (Cu, Fe, Ti, Co, etc.), and carbon-based materials [10,11,12,13]. Among them, Pd, Ru, and Cu exhibit superior electrocatalytic performance [14,15,16]. The reduction mechanism of the ERN for ammonia production proceeds via direct electron reduction at the cathode surface or indirect reduction pathways (reactive reducing species). In the indirect reduction process, ${\mathrm{N}\mathrm{O}}_3^{-}$ is reduced by atomic hydrogen (*H) (E⁰ = 2.10 V vs. RHE). *H is generated through the Volmer step via the reduction of water or protons and is generally considered key to improving purification performance and reducing energy consumption in electrocatalytic reduction systems [17,18].

Compared to metal oxides (e.g., Cu₂O, Co₃O₄) widely used in catalysis, transition metal phosphides (TMPs) have emerged as a novel class of electrocatalysts in recent years [12,18]. Given their high electrical conductivity, mixed orbital states, and economic viability, TMPs have become a research focus in the field of electrocatalysis [19,20,21]. In TMPs, the high electronegativity of phosphorus (P) atoms enables them to attract electrons from adjacent metals, while P (^δ−) also captures protons (^δ+) as a substrate [22,23]. This optimizes the binding state between the catalyst surface and reaction intermediates/products, resulting in superior catalytic activity [24]. Nickel-based phosphides are particularly attractive in photocatalysis and electrocatalysis due to their dual functionality: P sites act as proton acceptors to suppress the HER; Ni sites facilitate *H generation for nitrate hydrogenation [24,25]. Yet, crystalline TMPs suffer from limited active sites and poor stability under operational conditions [26,27]. Here, we propose that amorphous nickel phosphide (ANP) overcomes these limitations via the ERN for ammonia production.

The overall energy consumption cost of ERN for ammonia production is a critical yet often overlooked issue. The anodic half-reaction of ERN is primarily the OER, which is kinetically sluggish and requires overcoming a high energy barrier [9,28]. Replacing the OER with thermodynamically more favorable oxidizable substances (e.g., ethanol, methanol, glycerol, urea, etc.) can reduce the EC of hydrogen production via water electrolysis [29,30]. Further, this study proposes a strategy of replacing the OER with more easily oxidized substances to lower the required anodic potential and improve the energy efficiency of the entire ERN for the ammonia production system. Research has shown that the outstanding performance of Ni-based catalysts (nickel oxide and nickel hydroxide) largely depends on their ability to transform into NiOOH during the reaction (NiOOH serves as the active site for THE UOR catalysis) [31,32].

In view of the needs of improving the efficiency of NH₃ synthesis and reducing EC via the ERN, starting from the improvement in the most critical performance factors, this paper rationally designs the multi-functional electrodes for NH₃ synthesis via the ERN, and the strategy of substituting the OER with energetically favorable anodic processes (e.g., UOR and S(IV)OR) to reduce the EC of NH₃ synthesis. To simplify fabrication and ensure cost-effectiveness, a multi-functional ANP electrode served as both cathode and anode in the ERN for NH₃ synthesis. This study paves the way for designing high-performance electrocatalysts toward selective nitrate-to-ammonia valorization, providing a theoretical basis for clean carbon-free energy and nitrogen recovery.

2. Materials and Methods

2.1. Materials

NaH₂PO₂ H₂O, NaNO₃, Na₂SO₄, NaOH, NaNO₂, HCl, sulfanilamide, and N-(1-naphtyl)-ethylenediamine dihydrochloride (98%, and t-butyl alcohol (t-BuOH) were purchased from Aladdin (Shanghai, China). Ni foam (NF) was sourced from Shanghai Macklin Biochemical Co., Ltd. (Shanghai, China). Deionized water was obtained by purifying Millipore ultrapure water (resistivity = 18.2 MΩ·cm⁻¹, 298 K).

2.2. Preparation of Electrocatalyst

The preparation of amorphous nickel phosphide (ANP) followed our previous work [33]. Briefly, ANP was synthesized via a one-step vapor-phase phosphidation process. Pretreated NF (1 × 1 × 0.15 cm) was positioned between two NaH₂PO₂·H₂O (molar ratio ~ 5:1:5) in a tube furnace with 10 cm spacing. The system was heated under N₂ using a two-stage program: (1) 250 °C for 1 h, then (2) ramped at 2 °C/min to 300 °C for 90 min. Controlled N₂ flow rates (5 mL/min) below 300 °C promoted amorphous Ni₂P formation. The resulting black product (ANP) was rinsed with water/ethanol and then vacuum-dried at 60 °C for 24 h. To prepare the Pd/C catalyst solution, 40 μL of a 5% Nafion solution, 240 μL of DW, and 720 μL of anhydrous ethanol were combined. Next, 15 mg of Pd/C was sonicated in 1 mL of this mixture to create a uniform black slurry. A pretreated nickel foam (NF) substrate (1 × 1 cm²) was coated evenly with 200 μL of the slurry and left to air-dry. The catalyst loading was calculated by weighing the weight difference before and after the NF load. The calculated loading was 2.21 mg cm⁻² for Pd/C.

ANP was tested as both the cathode and anode in systems combining the ERN with OER, or UOR, or S(VI)OR, or S(IV)/UOR. The cathode chamber contained a 1 M NaOH solution with nitrate ions, while the anode chamber included 0.5 M urea, 0.05 M Na₂SO₃, or a mixture of the two kinds of solution. After the reactions, the degradation of urea in the anode chamber was analyzed.

2.3. Electrochemical Testing

Electrocatalytic activity was evaluated via linear sweep voltammetry (LSV), cyclic voltammetry (CV), and chronoamperometry. These tests were carried out with a CHI 660D electrochemical workstation (Shanghai Chenhua Instruments Co., Ltd., Shanghai, China) in a conventional three-electrode setup.

The working and counter electrodes each exposed an area of 1 × 1 cm², and an undivided electrochemical cell was used. Unless otherwise stated, all potentials were measured against the Hg/HgO reference electrode. For consistency, these were normalized to the reversible hydrogen electrode (RHE) potential using the equation (Equation (1)):

E_RHE = E_Hg/HgO + 0.098 V + 0.059 pH

(1)

The Electrochemical Surface Area (ECSA) was calculated using the double-layer capacitance (C_dl) method. Cyclic voltammograms were recorded at a series of scan rates within a non-Faradaic potential range. The capacitive current at the midpoint potential was regressed as a function of scan rate, and the linear fit yielded the C_dl value. The ECSA was determined using the formula: ECSA = C_dl(catalyst)/(C_dl(NF) × per ECSA)_(NF). The Tafel slope, reflecting reaction kinetics, was determined using the Tafel equation (Equation (2)):

η = blog j + a

(2)

where η is the overpotential (V), b is the Tafel slope (mV dec⁻¹), and j represents the current density (mA cm⁻²).

2.4. Electrolysis Experiments and Analysis

The ERN process was conducted at room temperature using an H-type electrolyzer (Gaoshi Rui Lian, Wuhan Gaossunion Technology Co., Ltd., Wuhan, China) with a dual-compartment cell divided by a Nafion 117 membrane. The Nafion 117 membrane underwent standard pretreatment before use (boiling 5% H₂O₂ for 20 min, cleaning with ultrapure water, soaking in 1 M H₂SO₄) to ensure no degradation products interfere. The potentials were controlled by an electrochemical station (CHI 660D, CH Instruments, Inc., Shanghai, China) with a working electrode (ANP, 1 × 1 cm²), a reference electrode (Hg/HgO electrode), and a counter electrode (Pt foil, 1 × 1 cm²). The cathode chamber contained nitrate ions, while both chambers were filled with electrolyte solutions. A controlled potential was applied to drive the reaction, with continuous stirring to ensure uniform mixing. Throughout the process, the cell voltage and current were recorded. Samples were periodically collected from both chambers and the ammonia recovery tank (H₂SO₄ solution) for analysis of total nitrogen, NO₃⁻, NO₂⁻, NH₄⁺/NH₃, and N₂H₄ concentrations. The ${\mathrm{N}\mathrm{O}}_3^{-}$ concentration was determined using Nessler’s method by UV-vis spectroscopy (Mapada UV-6100S.) by measuring the absorbance at 220 nm and 275 nm (A = A_{220 nm}−2A_{275 nm}). The NH₃ concentration was evaluated by Nessler’s reagent by absorbance at 420 nm. Gas samples were taken during the electrolysis and analyzed by gas chromatography (GC). The content of ${\mathrm{N}\mathrm{O}}_2^{-}$ was assayed spectrophotometrically at 540 nm, due to the azo dye generated via the Griess reaction.

The results of the standard curves are shown in Figure 1. A standard curve of NO₃⁻ was generated using the equation: y = 0.24492x + 0.00688, (R² = 0.9998). The standard curve of NO₂⁻ using the equation: y = 3.47756x + 0.00123, (R² = 0.9999). The standard curve of ${\mathrm{N}\mathrm{H}}_4^{+}$ was plotted with the equation: y = 0.17504x + 0.0003, (R² = 0.9999). The standard curve for Total nitrogen (TN) was expressed as: y = 0.24328x − 0.0013, (R² = 0.9995). The urea concentration was determined using high-performance liquid chromatography (HPLC).

HPLC measurements were conducted on a Waters 2695 system (Waters Corp., Milford, MA, USA) equipped with a Symmetry C18 column (250 mm × 4.6 mm, 5 μm). Ultrapure water was the mobile phase at a flow rate of 1 mL/min at 20 °C. Detection was carried out at 190 nm. Under these conditions, urea’s retention time was 2.731 min.

The performance of the ERN was evaluated using the following metrics: nitrate removal rate (R%), ammonia selectivity (SE%), ammonia yield (yield of NH₃), Faradaic efficiency (FE), and energy consumption (EC).

R% = (C₀ − C_t/C₀) × 100%

(3)

SE% = (X_t/(C₀ − C_t)) × 100%

(4)

yield of NH₃ = (C_NH3 × V)/(M_NH3 × t × m)

(5)

FE = (8F × C_NH3 × V)/(M_NH3 × Q) × 100%

(6)

EC = U_cellIt ×10⁻³/([C₀ − C_t] × V × M)

(7)

U_cell = φ_e,a + η_a − φ_e,c + η_c + IR

(8)

• where: C₀: initial concentration of ${\mathrm{N}\mathrm{O}}_3^{-}$, (mg/L, in N);
• C_t: concentration of ${\mathrm{N}\mathrm{O}}_3^{-}$, at time t (mg/L, in N);
• X_t: concentration of product X at time t (mg/L);
• C_NH3: concentration of NH₃ (mg/L);
• m: mass of catalyst (mg);
• V: volume of dissolved liquid (L); t: electrolysis time;
• F: Faraday’s constant (96,485 C mol⁻¹);
• Q: total charge through the electrode;
• U_cell: tank voltage (V); I: average current (mA);
• φ_e,a: the anode equilibrium potential, η_a: the anode overpotential, φ_e,c: the cathode equilibrium potential, η_c: the cathode overpotential; R: the total resistance of the entire system (including electrode, electrolyte, and membrane resistance)
• M: relative molecular mass.

2.5. Free Radical Quenching and Detection

To study the role of atomic hydrogen, free radical quenching experiments were performed using t-BuOH as a hydrogen quencher. A 100 mM t-BuOH solution was added and compared with a control group. Electron Spin Resonance (ESR) spectroscopy, with DMPO (5,5-dimethyl-1-pyrroline-N-oxide) as a radical scavenger, was used to detect active free radicals in the reaction system. Samples from the electrocatalytic nitrate degradation experiment were mixed with DMPO, shaken, left to stand for five minutes, and analyzed using ESR. The parameters included a central field strength of 3380.00 G, 100 kHz frequency, 0.2 mT modulation, 10.0 mW power, and a 60-s scan time.

2.6. Nitrogen Isotope Labeling

For nitrogen isotope labeling, sodium nitrate (Na¹⁵NO₃) was used under conditions similar to other electrochemical tests. To detect ¹⁵NH₃, ¹H NMR spectroscopy (600 MHz) was employed, with maleic acid (C₄H₄O₄) as the external standard. The samples were adjusted to slightly acidic conditions with 4 M H₂SO₄, and the external standard was added to a final concentration of 300 ppm. A calibration curve was created by comparing the ¹H NMR peak area ratios of ¹⁵NH₄⁺ and the standard, which was then used to determine concentrations in the samples.

2.7. In Situ Differential Electrochemical Mass Spectrometry (DEMS)

In situ DEMS was used to monitor volatile products during the electrocatalytic reaction in real time. Ar gas was bubbled through the electrolyte throughout the experiment. The working electrode varied between the prepared cathode material, a platinum (Pt) sheet, and Hg/HgO. The electrolyte used was 10 mM NaNO₃ in 1 M NaOH solution. LSV or CV tests were conducted from −0.5 V to −1.3 V at a scan rate of 5 mV/s, and mass spectrometry signals were recorded during the tests.

2.8. Characterization Techniques

The morphological analysis was performed using scanning electron microscopy (SEM, Hitachi S-4800, Hitachi High-Tech, Tokyo, Japan) and transmission electron microscopy (TEM, JEOL JEM-2100FkV, FEI, JEOL Ltd., Tokyo, Japan). The chemical states of samples were obtained from powder X-ray diffraction (XRD, Bruke D8 Advance, Bruker Corporation, Billerica, MA, USA) patterns, Raman spectrometer (In Via Reflex, Renishaw plc. Wotton-under-Edge, UK), and X-ray photoelectron spectroscopy (XPS, Thermal Fisher Scientific, Waltham, MA, USA). An inductive coupled plasma atomic mission spectrometer (ICP-MS, Agilent 7700, Agilent Technologies, Santa Clara, CA, USA) was quantified to metal leaching, and an electron spin resonance spectrometer (ESR, JES-FA200, JEOL Ltd., Tokyo, Japan) was employed to identify reactive intermediates radical species. The products were analyzed by Gas Chromatography (GC-14B, Shimadzu Corporation, Kyoto, Japan), Ion Chromatography (ICS-1100, Thermal Fisher Scientific, Waltham, MA, USA), and Ultraviolet-visible diffuse reflection spectrum (UV-6100S, Meipuda, Shanghai, China).

3. Results and Discussion

The ECSA of ANP was evaluated using CV in the non-Faradaic potential region. CV curves for both nickel foam (NF) and ANP were obtained at varying scan rates (Figure 2a,b), and the C_dl was calculated from the linear relationship between scan rate and current density. The C_dl values in Figure 1c for ANP and NF were 6.04 mF/cm² and 0.52 mF/cm², respectively, with the higher value for ANP attributed to its amorphous nickel phosphide structure, which enhanced its ECSA. The ECSA of ANP was calculated to be 11.62 cm² (Figure 2d).

To assess the feasibility of the ERN on the ANP cathode, LSV measurements were conducted for NF and ANP electrodes in an electrolyte containing 100 mM ${\mathrm{N}\mathrm{O}}_3^{-}$ and 1 M NaOH. The LSV curves (Figure 3a) showed that NF exhibited weak hydrogen evolution activity at potentials below −1.17 V, indicating poor activity for the ERN. In contrast, the ANP electrode demonstrated a significant current response, which increased with more negative potentials, suggesting effective ${\mathrm{N}\mathrm{O}}_3^{-}$ reduction. As the initial ${\mathrm{N}\mathrm{O}}_3^{-}$ concentration in the electrolyte increased, both current density and the area enclosed by the LSV curve grew (Figure 3b), indicating that higher reactant concentrations improved mass transfer and accelerated ERN for ammonia production.

The distribution of reactants and products during the ERN process at an applied potential of −1.2 V was investigated over time (Figure 3c). Over 180 min, the ${\mathrm{N}\mathrm{O}}_3^{-}$ concentration decreased sharply from 60.00 mg/L to 1.40 mg/L, achieving an R% of 97.7% and an SE% of 91.8%. During the reaction, the intermediate ${\mathrm{N}\mathrm{O}}_2^{-}$ initially increased, peaking at 3.19 mg/L after 60 min, before gradually decreasing. The ammonia concentration increased steadily, reaching a yield rate of 0.0616 mmol h⁻¹·mg⁻¹ by the end of the reaction. No hydrazine (N₂H₄) was detected, and the only gaseous nitrogen product identified was nitric oxide (NO), at a concentration of 3.21 mg/L (5.4% of nitrogen products). This suggested that NO was an intermediate in ammonia formation, supporting a reaction mechanism similar to the results of the reported article [25].

The catalytic performance of the ANP electrode was compared to NF, nickel oxide on NF (NiO/NF), and a commercial Pd/C powder electrode (Figure 3d). The R% was 13.2% for NF, 17.9% for NiO/NF, 97.7% for ANP, and 80.2% for Pd/C, while the SE% was 8.9% for NF, 11.8% for NiO/NF, 91.8% for ANP, and 87.4% for Pd/C. ANP exhibited the highest performance, surpassing the Pd/C electrode. Unlike Pd/C, which suffered from high resistance, limited active sites, and catalyst loss due to its powdery form, the self-supported ANP electrode offers a seamless connection between the catalyst and conductive substrate, enhancing conductivity and maximizing exposure of active sites. Its three-dimensional structure further improved electrolyte penetration and mass transfer, resulting in superior catalytic activity. These results confirmed that the ANP electrode was a highly effective alternative to precious metal catalysts for ERN ammonia production. It combined high activity, practicality, and stability, making it a promising material for scalable applications.

To further investigate the effect of the applied potential on the ERN performance of the ANP cathode, we examined the R%, SE%, and FE at different applied potentials: −0.9 V, −1.1 V, −1.2 V, −1.3 V, and −1.5 V, as shown in Figure 4a. The corresponding R% values were 47%, 89.3%, 97.7%, 97.1%, and 73%, respectively; SE% values were 60.3%, 82.3%, 91.8%, 90.1%, and 89.3%; and FE values were 57.1%, 88.9%, 97.1%, 95%, and 79.5%, respectively.

It could be observed that as the applied potential increased, the R% significantly increased from 47% to 97.7%. This was because, with higher applied potentials, the supply of electrons per unit time increased, allowing more electrons to transfer to ${\mathrm{N}\mathrm{O}}_3^{-}$ and water molecules, thereby facilitating both direct and indirect electro-reduction of ${\mathrm{N}\mathrm{O}}_3^{-}$. However, after −1.2 V, there was no further increase in the R% when higher potentials were applied. This was due to the applied potential exceeding the critical potential for water splitting, leading to side reactions such as the HER. Moreover, the excessive production of hydrogen gas bubbles interfered with the diffusion of ${\mathrm{N}\mathrm{O}}_3^{-}$ to the ANP surface, and the accumulation of gas bubbles hindered the mass transfer process, resulting in a decrease in R%.

In summary, the R% exhibited a volcano-shaped trend with changes in the applied potential, with the optimal R% occurring at −1.2 V. Therefore, in subsequent studies, unless otherwise specified, an applied potential of −1.2 V was used. Additionally, as shown in Figure 4a, the SE% of the ANP cathode maintained superior performance across a wide potential range (−1.0 V to −1.5 V), with the maximum SE% of 91.8% occurring at −1.2 V.

Considering the practical application of the ERN in water treatment, the EC of the process was evaluated. The results are illustrated in Figure 4b. The EC for ammonia production at various applied potentials of ANP cathodes was as follows: 10.6846, 8.975, 8.239, 10.794, and 33.558 kWh/kg at potentials of −0.9, −1.1, −1.2, −1.3, and −1.5 V, respectively. This trend was attributed to the fact that as the applied potential increased, the electron transfer between the electrodes increased, leading to more H* generation, which was more favorable for the ERN to produce ammonia.

At the optimal applied potential of −1.2 V, the R% was highest, and the EC was the lowest. However, when the applied potential exceeded the HER potential, side reactions such as the HER were intensified, reducing ammonia production efficiency and increasing EC. Therefore, the EC at −1.2 V (8.239 kWh/kg) was the lowest. At an applied potential of −1.5 V, the EC was 4.1 times higher than at −1.2 V, indicating that a higher applied potential corresponds to higher energy consumption.

As shown in Figure 4c, the initial concentrations of ${\mathrm{N}\mathrm{O}}_3^{-}$ in the electrolyte were 60, 120, 300, and 480 mg/L, with a corresponding R% of 97.7%, 91.3%, 84.3%, and 73.6%, and an SE% of 91.8%, 90.3%, 91.4%, and 90.6%, respectively. As the initial concentration of ${\mathrm{N}\mathrm{O}}_3^{-}$ increased, R% gradually decreased, while SE% remained stable.

With increasing ${\mathrm{N}\mathrm{O}}_3^{-}$ concentration, the number of anionic by-products, such as ${\mathrm{N}\mathrm{O}}_2^{-}$, also increased. At this point, by-products and ${\mathrm{N}\mathrm{O}}_3^{-}$ formed a competition for adsorption sites, which inhibited the ${\mathrm{N}\mathrm{O}}_3^{-}$ electro-reduction reaction, and led to a decrease in the removal efficiency. When the initial ${\mathrm{N}\mathrm{O}}_3^{-}$ concentration reached 480 mg/L, the selectivity for ammonia remained almost unchanged, indicating that the ANP cathode’s ability to reduce nitrate to ammonia can be applied over a wide range of ${\mathrm{N}\mathrm{O}}_3^{-}$ concentrations. In summary, for the reaction system at this scale, the optimal initial ${\mathrm{N}\mathrm{O}}_3^{-}$ concentration for effective treatment was 60 to 120 mg/L. Therefore, in practical applications, to treat high concentrations of ${\mathrm{N}\mathrm{O}}_3^{-}$, the reactor can be scaled up by increasing the working area of the electrodes or the volume of the electrolytic cell, among other factors.

In the ANP cathode ERN process system, *H was qualitatively detected. The ESR spectrum results (Figure 5a) showed a signal from DMPO-H with nine peaks, with an intensity ratio of 1:1:2:1:2:1:1 (ɑ_N = 144 G and ɑ_H^β = 22.40 G), confirming that *H was generated in the system. Furthermore, the R% of the ERN process was analyzed at −1.2 V. As shown in Figure 5b, the addition of t-BuOH significantly inhibited the ERN process. At a tert-butanol concentration of 10 mM, the R% of the ANP cathode in the ERN was 49.3%, which was a 42.5% decrease compared to the R% without t-BuOH. This was because t-BuOH quenched some of the *H species, further indicating that *H generated in the system without a quenching agent was the main reducing species responsible for the ERN process.

Thus, the ERN process generally includes both direct and indirect reduction pathways. *H was generated through the electrolysis of H₂O or H⁺ on the ANP cathode via the Volmer reaction, with the reaction: ANP + H₂O + e⁻→ANP + *H + OH⁻ (Volmer).

Since both ${\mathrm{N}\mathrm{O}}_3^{-}$ in the solution and N₂ in the atmosphere can serve as potential nitrogen sources for ammonia production in electrochemical reactions, further confirmation of the nitrogen source in the product (ammonia) was necessary. As shown in Figure 6a, using isotopically labeled ¹⁴${\mathrm{N}\mathrm{O}}_3^{-}$ as the reactant, the ¹⁴${\mathrm{N}\mathrm{H}}_4^{+}$ product was qualitatively and quantitatively analyzed using ¹H NMR. The ¹H NMR spectrum showed the typical three peaks for ¹⁴${\mathrm{N}\mathrm{H}}_4^{+}$ (δ values at 94, 7.03, and 7.12 ppm). The quantitative result, shown in Figure 6b, revealed that the generated ¹⁴${\mathrm{N}\mathrm{H}}_4^{+}$ concentration was 54.50 ppm. Similarly, using Na¹⁵NO₃ as the reactant, the ¹H NMR spectrum for the electrolyte showed double peaks for ¹⁵${\mathrm{N}\mathrm{H}}_4^{+}$ at δ = 97 ppm and δ = 7.09 ppm (Figure 6c), with a concentration of 57.18 ppm (Figure 6d). The ammonia yield, based on the colorimetric method and ¹H NMR spectral quantification, is summarized in

Table 1. Comparison of the yield of NH₃ by different quantitative methods with the ANP cathode.

Test Conditions	Quantitative Method	Yield of NH3 (mmol h−1 mg−1)
14${\mathrm{N}\mathrm{O}}_3^{-}$, 25 °C; −1.2V (3 h)	UV-Vis spectrometry	0.0621
14${\mathrm{N}\mathrm{O}}_3^{-}$, 25 °C; −1.2V (3 h)	1H NMR	0.0614
15${\mathrm{N}\mathrm{O}}_3^{-}$, 25 °C; −1.2V (3 h)	1H NMR	0.0627

. The ¹⁵${\mathrm{N}\mathrm{H}}_4^{+}$ and ¹⁴${\mathrm{N}\mathrm{H}}_4^{+}$ concentrations roughly matched the NH₃ content previously detected by UV-vis, indicating that the nitrogen in the ammonia came entirely from ${\mathrm{N}\mathrm{O}}_3^{-}$.

To further explore the reaction pathway and gaseous products, DEMS tests were also conducted. The results in Figure 7 show signals corresponding to a mass-to-charge ratio (m/z) of 30, 17, and 2, which correspond to NO, NH₃, and H₂, respectively. This indicated that NO was an intermediate species in the ERN. The DEMS results were consistent with the gas-phase product GC detection, confirming that NO was an intermediate in the reaction pathway. Moreover, since only Ar was passed through during the test, with no external nitrogen source, the detected NO and NH₃ must originate from the nitrogen in ${\mathrm{N}\mathrm{O}}_3^{-}$.

The stability of the ANP cathode was a critical factor for practical applications, necessitating an investigation into its reusability and durability during the ERN process. As shown in Figure 8, both SE% and yield of NH₃ increased slightly during the initial cycles, likely due to a reduction in nickel oxide content during the ERN process. Subsequently, the performance stabilized, indicating that the ANP cathode demonstrated excellent reusability and stability throughout the ERN process.

To further assess the long-term stability of both the ANP cathode and anode, their morphology and elemental valence states were analyzed after extended ERN and OER tests using SEM, TEM, XRD, Raman spectroscopy, and XPS. Figure 9a,c show that the ANP cathode retained its uniform nanorod microemulsion structure, with only minor separation of nanorods after prolonged reaction in solution. In contrast, the ANP anode in Figure 9b,d exhibited slight nanorod corrosion and a transformation from a microemulsion structure to a paste-like, protruding morphology after extended durability testing.

The XRD patterns (Figure 9e) for both the cathode and anode showed no significant changes, confirming their structural stability during the ERN and OER processes. Raman spectra (Figure 9f) revealed that the cathode’s features remained stable post-testing, whereas the anode exhibited new peaks at 479.7 cm⁻¹ and 560.8 cm⁻¹, corresponding to NiOOH. This suggested partial oxidation of the anode into NiOOH during the oxidation reaction, which was beneficial for the OER [34].

Further insights were obtained through XPS analysis. For the ANP cathode (Figure 10), ANP active sites remained largely intact. Surface oxidation components decreased by approximately 11%, likely due to the reduction of nickel oxide back to metallic nickel during cathodic polarization. This reduction initially enhanced electrochemical activity before stabilizing over subsequent cycles. ICP-MS analysis confirmed no phosphorus leaching (0 mg/L) of the ANP cathode, which was consistent with the XPS results of the ANP cathode, suggesting the non-pollution of the ANP cathode. For the ANP anode, high-resolution Ni 2p XPS spectra (Figure 10c) revealed additional peaks at binding energies of 853.3 eV and 871.2 eV (Niᵟ⁺) and 857.1 eV and 875.6 eV (NiOOH) [35]. This indicated partial oxidation of the anode to NiOOH or NiO during the OER process. The transition from Ni²⁺ to Ni³⁺ during oxidation caused the Ni-P bond to rupture, releasing a small amount of phosphorus into the electrolyte. ICP-MS analysis confirmed phosphorus leaching of the ANP anode, with a measured concentration of 0.023 mg/L.

The P 2p XPS spectrum (Figure 10d) after the OER test showed peaks at 130.1 eV and 129.1 eV, corresponding to P 2p_1/2 and P 2p_3/2, associated with transition metal phosphides (Pᵟ⁻). Additionally, a peak at 133.2 eV was observed, indicating oxidized phosphorus species formed through surface oxidation. The proportion of oxidized phosphorus increased compared to the pre-reaction ANP, suggesting that part of the phosphorus was converted to an oxidized state during the OER. These findings implied that the primary active sites in the ANP anode during the OER process were high-valent nickel species, such as NiOOH, which are well-documented as active components in both the OER and UOR processes [31,34].

The stability of the ANP cathode was largely attributed to its preparation method, which utilizes self-sacrificial NF as both the substrate and nickel source. This approach created a superhydrophilic structure that facilitated efficient electrolyte diffusion, enhanced electron transfer, and maximized the exposure to active sites [33].

To evaluate the UOR catalytic performance of the ANP anode, experiments were conducted using ANP as the working electrode in a three-electrode setup with different electrolytes: 0.5 M urea, 1 M NaOH, and a mixture of 0.5 M urea and 1 M NaOH. In Figure 11a, no anodic current density was observed in the 0.5 M urea solution alone, indicating that the UOR cannot proceed without NaOH as an electrolyte. In 1 M NaOH, the ANP anode displayed OER activity at 1.515 V vs. RHE to achieve a current density of 10 mA cm⁻². However, in the combined presence of 1 M NaOH and 0.5 M urea, a significant anodic current response was observed at 1.335 V vs. RHE to reach 10 mA cm⁻². The reduction of 180 mV in overpotential compared to the OER demonstrated that the UOR was more energy-efficient, requiring less voltage. These findings confirmed the high effectiveness of ANP as an electrocatalyst for the UOR.

The UOR performance of the ANP anode with commercial Pt and NF anodes was compared in Figure 11b. At a current density of 10 mA cm⁻², the ANP anode exhibited a lower overpotential of 1.335 V vs. RHE compared to Pt (1.562 V vs. RHE) and NF (1.423 V vs. RHE), highlighting its superior UOR activity. To further understand reaction kinetics, Tafel slopes were analyzed for ANP, Pt, and NF (Figure 11c). The ANP anode demonstrated a Tafel slope of 44.2 mV dec⁻¹, outperforming most reported electrocatalysts [29,35,36,37,38,39] (

Table 2. Comparison of the UOR activity of ANP with other anodes at 10 mA cm⁻²

Anode	Electrolytes	ηa (V vs. RHE)	Reference
Ni-MIL-77	1 M KOH + 0.33 M urea	1.48	[36]
V-doped Ni3N/NF	1 M KOH + 0.5 M urea	1.36	[35]
NiFeMo hybrid	1 M KOH + 0.33 M urea	1.38	[37]
Ni-bza-900	1 M KOH + 0.33 M urea	1.38	[38]
NiF3/Ni2P@CC	1 M KOH + 0.33 M urea	1.36	[39]
NiOx/CNx	1 M KOH + 0.33 M urea	1.38	[29]
ANP	1.33	2.08	This work

). This value was significantly lower than NF (95.4 mV dec⁻¹) and Pt (235.5 mV dec⁻¹), indicating better catalytic activity and faster reaction kinetics for the UOR.

The long-term stability of the ANP anode was also assessed for practical applications. Figure 11d showed that the ANP anode at 1.37 V vs. RHE retained 84.1% of its initial current density (20 mA cm⁻²) after 32 h of chronoamperometry, demonstrating excellent stability. Additional polarization curve tests (Figure 12a) conducted before and after five cycles of the ERN + UOR reactions revealed no significant performance degradation, further confirming the durability and reliability of the ANP anode for real-world applications.

If the OER was used as the anodic half-reaction, optimizing and improving the OER catalytic performance of the anode material did not reduce the anode potential below the theoretical value required for the OER (1.23 V vs. RHE). Therefore, further optimization of the anodic oxidation reaction kinetics was necessary to lower the anodic oxidation potential (φ_e,a + η_a) and enhance the NH₃ production efficiency in the ERN. In this regard, the oxidation performance of the ANP anode in sodium metabisulfite (Na₂SO₃) solution was investigated. Sodium metabisulfite can be obtained by alkaline absorption of the atmospheric pollutant SO₂, thus promoting the concept of “waste recycling”.

As illustrated in Figure 12b, LSV curves demonstrated a substantial difference between the NaOH electrolyte and sodium metabisulfite-containing solution. The onset oxidation potential for the OER in pure NaOH solution was observed at 1.52 V vs. RHE, whereas the presence of sodium metabisulfite shifted the oxidation onset potential for the S(IV)OR to 0.61 V vs. RHE, representing a 0.91 V reduction in activation potential. Notably, the anodic current density in the sodium metabisulfite electrolyte exceeded that of the OER system by a considerable margin within a potential window of 0.61~1.52 V vs. RHE, suggesting significantly enhanced oxidation kinetics on the electrode surface. This kinetic advantage was further confirmed by the lower anode potential (by approximately 0.3 V) required to achieve 20 mA cm⁻² current density in the sulfite-containing electrolyte compared to conventional OER conditions. The improved reaction kinetics can be attributed to two primary factors: The thermodynamically favorable sulfite oxidation pathway (S(IV)OR: SO₃²⁻ + H₂O → SO₄²⁻ + 2H⁺ + 2e⁻) requires lower activation energy than the OER pathway (2H₂O → O₂ + 4H⁺ + 4e⁻). The generated protons (H⁺) migrated through the proton exchange membrane to the cathode compartment under the applied electric field, thereby simultaneously facilitating the cathodic reduction reactions.

Based on the excellent catalytic activity of the ANP anode for the OER, UOR, and S(IV)OR, it holds potential for enhancing the ERN performance at the ANP cathode. Therefore, ANP was used as both the cathode and anode material to assemble three systems: ERN + UOR, ERN + S(IV)OR, and ERN + S(IV)/UOR (Figure 13a–c), with an initial ${\mathrm{N}\mathrm{O}}_3^{-}$ concentration of 60 mg/L. The efficiency of enhancing the ANP cathode for ammonia production in the ERN process was studied at −1.2 V. In Figure 13d–f, the corresponding R% for the ERN + UOR, ERN + S(IV)OR, and ERN + S(IV)/UOR systems were 98.5%, 98.8%, and 99.1%, respectively, with an SE% of 92.1%, 92.8%, and 93.4%, respectively, and reaction times of 3, 2.6, and 2.5 h, respectively. Intermediate products ${\mathrm{N}\mathrm{O}}_2^{-}$ and the final product, ammonia, were generated in all systems. The generation of ${\mathrm{N}\mathrm{O}}_2^{-}$ initially increased and then gradually decreased, indicating that ${\mathrm{N}\mathrm{O}}_2^{-}$ could be further reduced to ammonia. Notably, the time at which the maximum ${\mathrm{N}\mathrm{O}}_2^{-}$ was generated in the ERN + S(IV)/UOR system occurred 30 min earlier, demonstrating that the ERN + S(IV)/UOR system can accelerate the ammonia production process on the ANP cathode. Compared to the ERN + OER system discussed earlier, all three systems can enhance the ammonia production on the ANP cathode to some extent, with the ERN + S(IV)/UOR system showing the best performance.

In reference to the ERN + OER system, where the ANP cathode performed best at −1.2 V with the lowest EC, the ammonia production performance of the ERN + UOR, ERN + S(IV)OR, and ERN + S(IV)/UOR systems was further explored at this potential. As shown in Figure 14a, the FE for these three systems was similar, all around 97.3%. The ammonia yield for the three systems was 0.0630, 0.0659, and 0.0693 mmol h⁻¹ mg⁻¹, which were 2.2%, 7.1%, and 12.6% higher, respectively, than the ammonia yield in the ERN + OER system (Figure 14b). These results indicated that both the UOR and S(IV)OR performances at the anode contribute to enhancing the ammonia yield on the ANP cathode, and the combination of the UOR and S(IV)OR has an even more pronounced promotive effect.

The average current during the reactions in the ERN + UOR, ERN + S(IV)OR, and ERN + S(IV)/UOR systems was 19, 15, and 18 mA, respectively. The EC for ammonia production in these three systems was calculated using the energy consumption formula, resulting in values of 7.957, 5.964, and 5.590 kWh/kg, respectively (Figure 14c). The EC of the ERN + S(IV)/UOR system was only 67.86% of that in the ERN + OER system, achieving a 32.14% energy saving. The difference in the EC for ammonia production in the three systems can be attributed to the different anodic half-reaction potentials, which ultimately resulted in different energy consumption levels. The cell voltages (Figure 14d) for the ERN + UOR, ERN + S(IV)OR, and ERN + S(IV)/UOR systems showed a gradual decrease, corresponding to 94.41%, 87.43%, and 80.3% of the ERN + OER system, respectively.

These results demonstrated that ANP, as a multi-functional catalyst, offers an effective strategy for enhancing the ammonia production on the ANP cathode in the ERN process. By replacing the OER with the UOR, S(IV)OR, and S(IV)/UOR on the ANP anode, this approach significantly enhances the ammonia production efficiency.

In addition, the degradation of urea in the anode chamber under the ERN + UOR and ERN + S(IV)/UOR systems was investigated. As shown in Figure 15, the urea removal rates for the ERN + UOR and ERN + S(IV)/UOR systems were 11.82% and 38.71%, respectively. After the addition of sodium metabisulfite, the urea degradation rate accelerated, and the ionic products gradually changed from ${\mathrm{N}\mathrm{H}}_4^{+}$ to ${\mathrm{N}\mathrm{O}}_3^{-}$. The gaseous N₂ also increased from 12.31% to 40.16%. This was because, during the oxidation process of sodium metabisulfite, sulfate and hydroxyl radicals were generated, which can accelerate the degradation of urea through indirect oxidation pathways [40]. This suggests that adding sodium metabisulfite can enhance the electrochemical oxidation performance in the anode chamber, providing a potential approach for the electrochemical degradation of more organic pollutants.

4. Conclusions

ANP, serving as a multi-functional self-supported electrocatalyst, exhibited excellent catalytic activity in the ERN, OER, UOR, and S(IV)OR reactions. ANP achieved an R% of 97.7%, an SE% of 91.8%, and an ammonia yield of 0.0616 mmol h⁻¹ mg⁻¹ at −1.2 V. In situ generated *H on the ANP cathode significantly improved nitrate removal efficiency and ammonia selectivity. ANP showed excellent UOR performance with a starting potential of 1.335 V vs. RHE at 10 mA cm⁻² and a Tafel slope of 44.2 mV dec⁻¹. Compared to the OER, UOR reduced the overpotential by 180 mV. For S(IV)OR, the starting oxidation potential was significantly lower at 0.61 V vs. RHE, compared to 1.52 V vs. RHE for the OER. Furthermore, replacing the OER with thermodynamically favorable anodic reactions (UOR, S(IV)OR, and S(IV)/UOR) further enhanced ammonia production efficiency. Results revealed that compared to the OER, the UOR reduced energy consumption by 17.5%; the S(IV)OR achieved a 27.6% reduction with a 7.1% increase in ammonia production; and the S(IV)/UOR delivered a 32.1% energy saving with a 12.6% increase in ammonia production. Differences in energy consumption were primarily attributed to variations in cell voltage and ammonia yield.

Notably, several challenges remain to be addressed for practical implementation: (1) Applicability to real wastewater systems: The current experiments were conducted using simulated wastewater (NaNO₃ solution). Future work should evaluate the catalyst’s performance in real wastewater containing complex pollutants, such as organic compounds, heavy metals, and competing anions (e.g., Cl⁻, SO₄²⁻), which may influence reaction kinetics and selectivity. (2) Long-term stability and durability: The current H-cell tests were limited to 9 cycles, which was insufficient to assess industrial viability. Extended stability tests (e.g., >100 cycles) in flow-cell or membrane electrode assembly configurations are necessary to evaluate degradation mechanisms and optimize catalyst longevity in the future.

In short, this study provides a novel strategy for developing an efficient multi-functional electrocatalyst for the selective conversion of ${\mathrm{N}\mathrm{O}}_3^{-}$ into high-value ammonia, offering theoretical support for carbon-free energy systems and nitrogen resource recovery.