Interfacial Confinement‐Programmed Hydrogen Spillover on Ag/CoNiS Boosts Nitrate‐to‐Ammonia Electrosynthesis

ABSTRACT

Electrochemical nitrate reduction (NO₃RR) under ambient conditions offers a sustainable route for ammonia (NH₃) synthesis; however, its efficiency is restricted by the kinetic mismatch between water dissociation and nitrate hydrogenation. Here, we design Ag/CoNiS heterostructures in which Ag loading density programs interfacial confinement to regulate hydrogen spillover from CoNiS water‐activation domains to Ag‐associated nitrate/nitrogen oxide (NO_x) intermediates, thereby coupling ^*H generation, relay, and deep nitrate hydrogenation. The optimized Ag_M/CoNiS achieves an NH₃ yield of 22.31 mg h⁻¹ cm⁻² with 99.13% Faradaic efficiency. In situ Raman, distribution of relaxation times (DRT) analysis, hydrogen/deuterium (H/D) isotope experiments, and tert‐butanol (TBA) perturbation tests reveal that the confined Ag–CoNiS interface regulates interfacial water and establishes a balanced ^*H supply–consumption regime, thereby suppressing competing hydrogen evolution. Density functional theory (DFT) calculations further show that Ag facilitates nitrate deoxygenation, whereas excessive Ag coverage weakens Co/Ni‐centered water activation, explaining the volcano‐type activity trend. Coupling NO₃RR with the sulfide oxidation reaction (SOR) further enables a low‐voltage NO₃RR||SOR electrolyzer, requiring only 0.70 V at 50 mA cm⁻² for energy‐saving co‐production of ammonia and sulfur.

Interfacial confinement‐programmed Ag/CoNiS interfaces regulate active‐hydrogen generation, spillover, and utilization, establishing a balanced ^*H supply–consumption regime that boosts nitrate‐to‐ammonia electrosynthesis while suppressing the hydrogen evolution reaction.

1. Introduction

The escalating imbalance of the global nitrogen cycle, driven by massive fossil fuel consumption and the overuse of fertilizers, has rendered nitrate (NO₃ ⁻) contamination a severe threat to aquatic ecosystems [1]. The electrocatalytic nitrate reduction (NO₃RR) provides a compelling “waste–to–value” strategy that simultaneously removes nitrate and synthesizes ammonia (NH₃) under ambient conditions, offering a decentralized and potentially lower–carbon complement to the energy– and carbon–intensive Haber–Bosch process [2, 3, 4]. However, despite its thermodynamic appeal, NO₃RR is hindered by complex multi‐electron/proton transfer pathways, in which mismatched nitrate activation, water‐derived ^*H supply, and deep hydrogenation often lead to sluggish kinetics and poor selectivity [5, 6].

Mechanistically, NO₃RR in alkaline electrolytes proceeds through consecutive deoxygenation/hydrogenation steps (e.g., ^*NO₃ → ^*NO₂ → ^*NO → ^*N → ^*NH_x → ^*NH₃), in which the availability of surface active hydrogen (^*H) is decisive for N–O bond scission and the hydrogenation of oxygenated intermediates. Yet, under alkaline conditions the proton donor switches from H₃O⁺ to H₂O, making active hydrogen (^*H) supply tightly coupled to sluggish water dissociation (Volmer step) [7, 8, 9]. This creates a central kinetic mismatch: insufficient ^*H leads to stalled hydrogenation and nitrite accumulation, whereas excessive ^*H promotes parasitic hydrogen evolution reaction (HER) and erodes NH₃ selectivity [10, 11]. For example, Luo et al., showed that modulating active hydrogen adsorption at an Fe─N interface strengthens in situ ^*H participation in nitrate hydrogenation while mitigating HER–related inefficiencies [12]. Hydrogen spillover in tandem/interface catalysts has therefore emerged as a promising strategy to alleviate the alkaline NO₃RR “^*H mismatch” by spatially coupling H₂O activation sites with nitrate/adsorbed nitrogen oxides (^*NO_x) hydrogenation sites, thereby improving the probability that water‐derived ^*H species are consumed by NO₃RR rather than accumulated for HER [13, 14, 15, 16, 17, 18, 19]. This principle is well illustrated by Pt/np–Co₂P, where ^*H is transferred to the Co₂P substrate to hydrogenate key intermediates, enabling a 600 mV operating window with NH₃ Faradaic efficiency (FE) > 90% and stable performance across −0.3 to −0.7 V [20].

However, realizing the full potential of this strategy remains challenging. Existing efforts have largely focused on thermodynamic modulation of the solid catalyst via support engineering (e.g., reducible oxides) [21, 22], bimetallic alloying [23, 24, 25, 26, 27], or spatial distance control [15] (Scheme 1). These solid‐phase strategies mainly regulate adsorption thermodynamics or average site distances, while the local kinetics of ^*H generation, migration, and consumption at the solid–liquid interface remain difficult to synchronize. Addressing these kinetic “blind spots” requires looking beyond the solid catalyst to the solid–liquid interface. Hydrogen‐spillover efficiency is not dictated by catalyst composition alone, but is also strongly influenced by the interfacial water structure in the electrical double layer, where water coverage, hydrogen–bond motifs, and cation–hydration environments jointly determine the kinetics of water activation and the local ^*H relay [28, 29, 30]. In this scenario, interfacial confinement created by adjacent nanoscale Ag and CoNiS domains offers a means to program hydrogen spillover: it can shorten the relay distance between ^*H‐supply and NO_x‐consumption sites, reorganize interfacial water into more reactive weakly hydrogen‐bonded and ion‐hydrated water species, and thereby balance ^*H generation and utilization [31, 32]. Consistently, Ru single atoms anchored in a nanoporous Ni₃B framework have demonstrated that nanoconfinement–regulated water microenvironments (e.g., enriched K⁺–H₂O) correlate with enhanced water dissociation and efficient atomic–scale spillover, inspiring the design of confined heterointerfaces [33].

SCHEME 1.

Transition from conventional hydrogen spillover to interfacial confinement‐programmed hydrogen spillover in NO₃RR.

Herein, we construct Ag/CoNiS heterostructures with precisely tuned Ag loading density (Ag_L/CoNiS, Ag_M/CoNiS, and Ag_D/CoNiS denote low, medium, and dense Ag loadings, respectively) to establish an interfacial‐confinement‐regulated hydrogen spillover paradigm for alkaline NO₃RR. At −0.6 V versus the reversible hydrogen electrode (RHE), Ag_M/CoNiS delivers an NH₃ yield of 22.31 mg h⁻¹ cm⁻² with a FE of 99.13%, corresponding to 1.41× and 1.69× higher NH₃ yield than Ag_L/CoNiS (FE 75.97%) and Ag_D/CoNiS (FE 85.65%), respectively, while maintaining excellent stability (30 cycles). Optimizing Ag density creates confined Ag–CoNiS interfacial regions that remodel the water microenvironment and balance weakly hydrogen‐bonded/ion‐hydrated water species, thereby promoting productive hydrogen spillover while suppressing HER. Electron paramagnetic resonance (EPR), cyclic voltammetry (CV), hydrogen/deuterium (H/D) isotope experiments, in situ spectroscopy, and DFT calculations collectively support that this density‐dependent interface establishes a balanced ^*H supply–consumption regime, accelerating nitrate‐to‐ammonia conversion. This interfacial‐regulation strategy was further integrated into a paired electrosynthesis system coupling NO₃RR with sulfide oxidation for simultaneous pollutant remediation and energy‐efficient chemical production.

2. Results and Discussion

2.1. The Synthesis and Characterization

The Ag/CoNiS heterostructures were synthesized through a hydrothermal–sulfurization process followed by ligand‐assisted Ag anchoring (Figure 1a). CoNiOH precursors with sheet‐like morphology were first obtained and then converted into CoNiS nanosheets after sulfurization, as confirmed by scanning electron microscopy (SEM), transmission electron microscopy (TEM) observations, and elemental mapping (Figures S1–S3). Subsequently, Ag nanoparticles were anchored onto CoNiS using H₂mba (thiosalicylic acid) as the regulating ligand. The nanosheet morphology is well preserved after Ag introduction (Figure S4), and high‐angle annular dark‐field scanning transmission electron microscopy (HAADF–STEM) mapping verifies the homogeneous distribution of Ag, Co, Ni, and S over Ag_M/CoNiS (Figure S5). Atomic force microscopy (AFM) further reveals the ultrathin nature of Ag_M/CoNiS with a thickness of ca. 10 nm (Figure 1b), while Kelvin probe force microscopy (KPFM) shows local surface‐potential variations of ca. 730–750 mV (Figure 1c), indicating electronic heterogeneity at the Ag/CoNiS interface. The local interfacial structure was resolved by high‐resolution transmission electron microscopy (HRTEM) and aberration‐corrected high‐angle annular dark‐field scanning transmission electron microscopy (AC‐HAADF‐STEM). HRTEM shows lattice spacings corresponding to Ag(200) and Co_0.5Ni_0.5S₂(200), confirming intimate Ag–CoNiS contact (Figure 1d). The selected‐area electron diffraction (SAED) patterns further support the crystalline features of CoNiS and Ag_M/CoNiS (Figures S6 and S7). AC‐HAADF‐STEM directly visualizes neighboring Ag and CoNiS domains at the heterointerface (Figure 1e,f). Fast Fourier transform/inverse fast Fourier transform (FFT/IFFT) analyses of the selected regions identify lattice spacings of 0.204 and 0.281 nm, assignable to Ag(200) and Co_0.5Ni_0.5S₂(200), respectively (Figure 1g,h). Additional AC‐HAADF‐STEM analysis gives consistent interfacial lattice features (Figure S8), further validating the construction of a well‐defined Ag/CoNiS heterointerface.

FIGURE 1.

Synthesis and structural characterization of Ag/CoNiS catalysts. (a) Schematic synthesis of Ag/CoNiS. (b) AFM height profile and (c) KPFM surface‐potential map of Ag_M/CoNiS. (d) HRTEM image of the Ag/CoNiS interface. (e) AC‐HAADF‐STEM image of the Ag/CoNiS heterointerface in Ag_M/CoNiS. (f) Magnified interfacial region selected from the boxed area in e. (g,h) FFT/IFFT analyses and lattice‐spacing profiles of Zone A and Zone B. Same‐region correlative STEM image sets of (i) Ag_L/CoNiS, (j) Ag_M/CoNiS, and (k) Ag_D/CoNiS, respectively, including bright‐field STEM (BF‐STEM) and secondary‐electron imaging composite mode (SEI‐composite) images collected from identical representative regions. (l) Ag particle‐size/density statistics and (m) ICP‐OES metal contents. (n) Schematic illustration of Ag‐loading‐programmed interfacial confinement and hydrogen spillover.

To regulate interfacial confinement, Ag_L/CoNiS, Ag_M/CoNiS, and Ag_D/CoNiS were prepared by tuning the Ag feeding amount. TEM images first confirm that the nanosheet morphology is retained after Ag deposition, with the yellow‐boxed regions highlighting the loading‐dependent surface contrast from Ag_L/CoNiS to Ag_D/CoNiS (Figure S9). Same‐region correlative STEM imaging further visualizes the Ag‐loading‐dependent contrast evolution, where bright‐field scanning transmission electron microscopy (BF‐STEM) and secondary‐electron imaging composite (SEI‐composite) images collected from identical representative regions show a progressive increase in Ag surface population while retaining the nanosheet framework (Figure 1i–k and Figure S10). HAADF–STEM coupled with energy‐dispersive X‐ray spectroscopy (EDS) mapping further confirms the dispersed Ag distribution across different loading densities (Figure S11). Statistical analysis confirms a consistent Ag particle size of ∼5 nm, whereas the Ag particle density increases markedly with Ag loading (Figure 1l). Inductively coupled plasma optical emission spectroscopy (ICP–OES) results further confirm the systematic increase in Ag content (Figure 1m), indicating that Ag density, rather than particle size, is the dominant structural variable. Together with the porosity and thermal‐stability analyses by Brunauer–Emmett–Teller (BET) and thermogravimetric analysis (TGA) (Figure S12), these results establish Ag/CoNiS as a density‐programmed interfacial platform, where medium Ag loading creates optimized confined Ag–CoNiS regions for balancing water activation, interfacial ^*H spillover, and nitrate hydrogenation (Figure 1n).

The X‐ray diffraction (XRD) patterns confirm the formation of CoNiS and the successful introduction of metallic Ag in the Ag/CoNiS samples (Figure 2a). X‐ray photoelectron spectroscopy (XPS) spectra (Figure 2b–c and Figure S13) further indicate pronounced interfacial electronic coupling between Ag and CoNiS. Metallic Ag⁰ is identified, accompanied by distinct negative shifts of the Co 2p_3/2 and Ni 2p_3/2 peaks by ≈0.4 eV relative to pristine CoNiS, indicating increased electron density around the Co/Ni centers. In contrast, the S 2p peaks exhibit a positive shift of ≈0.7 eV, suggesting electron depletion around S species and charge redistribution within the Co/Ni─S framework [34, 35, 36, 37]. These spectroscopic features are consistent with an electronic metal–support interaction (EMSI) at the Ag/CoNiS interface. Notably, the S 2p shift decreases to ∼0.3 eV for Ag_D/CoNiS, implying that excessive Ag loading attenuates the optimal interfacial charge redistribution. This trend is corroborated by X‐ray absorption fine structure (XAFS): Co and Ni K–edge X‐ray absorption near–edge structure (XANES) (Figure S14) place the absorption edges of Ag/CoNiS between those of the corresponding foils and oxides (CoO/NiO), and their lower‐energy shift upon Ag incorporation supports electron injection from Ag into the CoNiS lattice [38]. Consistently, Fourier transformation (FT) of extended X‐ray absorption fine structure (EXAFS) (Figure 2d–e) and fitting results (Tables S1 and S2) show no Co─Co or Ni─Ni contributions, excluding metal aggregation, while revealing bond–length expansion upon Ag incorporation (Co─S: 2.23 → 2.25 Å; Ni─S: 2.23 → 2.28 Å), in line with electron filling into metal–centered orbitals that increases ionic radii and elongates M─S bonds [39]. Such charge redistribution also reshapes the thermodynamic landscape as derived from ultraviolet photoelectron spectroscopy (UPS) and ultraviolet–visible diffuse reflectance spectroscopy (UV–vis DRS) analyses (Figure 2f–h), with Ag_M/CoNiS exhibiting the lowest work function (4.34 eV) and the most negative conduction–band position, indicative of maximized surface electron density and reducing capability favorable for nitrate–reduction intermediates. Finally, an inter‐sample correlation chord diagram (Figure 2i) visualizes a volcano‐type dependence of electronic modulation intensity (XPS shifts and UPS ΔΦ) on Ag loading, where Ag_M/CoNiS displays the broadest chords, especially for the S 2p shift and ΔΦ, highlighting that the interfacial electronic reconstruction is maximized at the optimal Ag density.

FIGURE 2.

Structural and electronic characterization of Ag/CoNiS catalysts. (a) XRD patterns, (b) Ag 3d and (c) S 2p XPS spectra, (d) Co and (e) Ni K–edge EXAFS spectra in R space, (f) Tauc plots derived from UV–vis DRS, (g) UPS spectra, and (h) schematic energy band alignments of CoNiS and Ag/CoNiS samples. (i) Inter–sample chord diagram visualizing the intensity of XPS (binding energy shifts for S, Ni, Co) and UPS (work function reduction compared to CoNiS, ΔΦ) induced by varying Ag loadings.

2.2. Electrocatalytic NO₃RR Performance

NO₃RR tests were carried out in a standard three–electrode H–cell using 1 m KOH containing 100 mm KNO₃. Linear sweep voltammetry (LSV) measurements (Figure 3a) reveal a pronounced Ag–loading–dependent activity, with Ag_M/CoNiS delivering the highest current density among the Ag/CoNiS series (Ag_M/CoNiS > Ag_L/CoNiS > Ag_D/CoNiS), indicating that Ag contributes to nitrate adsorption/activation but its effectiveness is strongly governed by loading density [40]. Pure Ag showed limited NH₃ yield and FE but pronounced NO₂ ⁻ accumulation (Figure S15), indicating that Ag alone mainly promotes the initial nitrate‐to‐nitrite conversion and cannot efficiently drive deep hydrogenation to NH₃ without the CoNiS‐derived ^*H supply. Consistently, the electrochemical active surface area (ECSA) estimated from the double‐layer capacitance (C_dl) shows that Ag_M/CoNiS exposes the largest accessible surface area (Figure 3b and Figure S16), suggesting that a medium loading best balances Ag‐site exposure while preserving the synergistic interface, and electrochemical impedance spectroscopy (EIS) further confirms the most favorable charge‐transfer kinetics for Ag_M/CoNiS, as evidenced by the smallest charge transfer resistance (R_ct) (Figure 3c) [41]. Accordingly, the NH₃ yield rate follows Ag_M/CoNiS > Ag_L/CoNiS > Ag_D/CoNiS > CoNiS (Figure 3d), while the Faradaic efficiency ranks as Ag_M/CoNiS > Ag_D/CoNiS > Ag_L/CoNiS > CoNiS (Figure 3e): although Ag_L/CoNiS attains a relatively high yield due to its large current, its FE is penalized by an oversupply of ^*H that cannot be timely consumed by nitrate reduction and instead recombines to H₂; in contrast, Ag_D/CoNiS exhibits a reduced NH₃ yield and increased NO₂ ⁻ selectivity during NO₃RR (Figure S17), indicating hindered deep hydrogenation under excessive Ag coverage. Meanwhile, the high intrinsic NO₂RR activity of CoNiS implies that its poor NO₃RR performance is mainly limited by the initial nitrate adsorption/activation step rather than hydrogenation capacity (Figure S18). As a result, Ag_M/CoNiS achieves the optimal trade‐off, delivering an NH₃ yield rate of 22.31 mg h⁻¹ cm⁻² with a near‐unity FE of 99.13% at −0.6 V vs. RHE. Beyond peak activity, Ag_M/CoNiS maintains FE > 80% over a broad nitrate–concentration window, and systematic optimization of reaction duration, initial pH, and nitrate concentration identifies 1 h, pH 14, and 100 mm as conditions that simultaneously maximize NH₃ yield and sustain high FE (Figure 3f,g). Durability evaluation further shows stable NH₃ yield and FE over 30 consecutive cycles (Figure 3h), accompanied by negligible morphological/structural changes after cycling, as verified by SEM, XRD, and XPS analyses (Figures S19–S21). Quantification of NH₃ by UV–vis is independently validated by ¹H nuclear magnetic resonance (¹H NMR) using dimethyl sulfoxide (DMSO) as an internal standard (Figures S22–S24), with excellent agreement between the two methods (Figure S25), confirming the robustness of the product analysis.

FIGURE 3.

Electrochemical NO₃RR performance of Ag/CoNiS catalysts. (a) LSV curves of different samples in 1 m KOH with and without 100 mm KNO₃. (b) C_dl and (c) Nyquist plots for CoNiS and Ag/CoNiS catalysts measured at −0.6 V vs. RHE in 1 m KOH with 100 mm KNO₃. The inset shows the equivalent circuit model used for fitting. (d) NH₃ yield and (e) FE of CoNiS and Ag/CoNiS catalysts at various applied potentials in 1 m KOH + 100 mm KNO₃. (f) NH₃ yield and FE of Ag_M/CoNiS catalyst at −0.6 V vs. RHE under different nitrate concentrations. (g) Dependence of NH₃ yield and FE for Ag_M/CoNiS on time and initial pH at −0.6 V vs. RHE. h) Stability performance of Ag_M/CoNiS for NO₃RR at −0.6 V vs. RHE in 1 m KOH + 100 mm KNO₃.

2.3. Hydrogen Spillover Mechanism

To directly clarify the origin and interfacial transfer behavior of active hydrogen species, electron paramagnetic resonance (EPR) measurements were first conducted in a nitrate‐free alkaline electrolyte [42, 43]. CoNiS exhibits a much stronger hydrogen‐related EPR response than pure Ag, confirming that CoNiS is the primary water‐activation/^*H‐generation domain, whereas Ag alone contributes negligibly to ^*H generation (Figure 4a). After Ag anchoring, the Ag/CoNiS series shows an Ag‐density‐dependent hydrogen‐related signal, indicating that Ag loading modulates the accessibility and accumulation of CoNiS‐derived ^*H species. Consistently, cyclic voltammetry (CV) recorded in 1 m KOH displays clear anodic hydrogen desorption/oxidation features (Figure S26) [44, 45]. The enlarged ^*H‐desorption region extracted from the CV curves at 100 mV s⁻¹ further confirms the Ag‐loading‐dependent regulation of electrochemically detectable hydrogen‐related species (Figure 4b). Here, the CV‐derived H‐desorption feature is used as a relative descriptor rather than an absolute ^*H surface coverage [46, 47, 48].

FIGURE 4.

Experimental evidence for interfacial‐confinement‐enabled hydrogen spillover in NO₃RR. (a) EPR spectra collected in nitrate‐free alkaline electrolyte for CoNiS, pure Ag, and Ag/CoNiS catalysts. (b) Enlarged anodic hydrogen desorption/oxidation region extracted from CV curves recorded at 100 mV s⁻¹ in 1 m KOH. (c) H/D kinetic isotope effect of CoNiS, Ag/CoNiS catalysts, and the physical mixture. (d) Schematic illustration of the D‐preload → H‐NO₃RR isotope‐switching protocol. (e) D‐preload → H‐NO₃RR DEMS signals of the physical mixture and Ag_M/CoNiS. (f) Schematic illustration of the H‐preload anodic‐CV protocol for probing stored‐hydrogen desorption and NO₃RR utilization. (g) First anodic CV curves after H‐preload in KOH and after transferring the H‐preloaded electrode to KOH + KNO₃, together with the extracted H‐desorption charge (Q_des) loss and residual ratio.

The kinetic relevance of interfacial H/D‐transfer kinetics was then evaluated by H/D kinetic isotope effect (KIE) measurements, a commonly used probe for water‐derived proton/hydrogen transfer kinetics in electrocatalytic reactions [49]. All catalysts exhibit apparent KIE values larger than unity, confirming that water‐derived H/D transfer participates in the rate‐relevant NO₃RR process (Figure 4c), building on isotope‐tracing strategies used to distinguish hydrogen‐transfer pathways in NO₃RR [42, 50]. Notably, Ag_M/CoNiS shows the smallest isotope penalty among CoNiS, the Ag/CoNiS series, and the physical mixture, indicating accelerated H/D‐transfer kinetics at the optimized Ag–0CoNiS interface. To further track donor‐derived H/D relay, a D‐preload → H‐NO₃RR isotope‐switching differential electrochemical mass spectrometry (DEMS) protocol was designed (Figure 4d). After D preloading in a deuterated alkaline electrolyte, the electrode was transferred to a proton‐rich nitrate‐containing electrolyte for H‐NO₃RR. Compared with the physical mixture, Ag_M/CoNiS retains more sustained D‐containing responses during consecutive cycles (Figure 4e). The normalized evolution of the summed D‐containing signal and individual m/z = 18, 19, and 20 responses further reveals slower decay over Ag_M/CoNiS (Figure S27). Considering possible contributions from water isotopologues and ammonia‐related isotope fragments, these mass channels are interpreted as D‐containing NO₃RR‐related responses rather than exclusive assignments to specific isotopologues. Nevertheless, their slower decay clearly supports more sustained coupling of pre‐accumulated donor‐derived D/H species to the nitrate‐containing interfacial reaction environment on Ag_M/CoNiS.

H‐preload anodic‐CV measurements were further performed to distinguish nitrate‐coupled ^*H utilization from simple residual ^*H storage (Figure 4f), following the logic that nitrate addition can consume pre‐accumulated surface hydrogen species during subsequent hydrogenation [47]. After H preloading in nitrate‐free KOH, the first anodic scan records the residual ^*H‐desorption charge. When the H‐preloaded electrode is transferred to KOH + KNO₃, nitrate‐involved interfacial hydrogenation consumes part of the pre‐accumulated hydrogen‐related species, thereby decreasing the remaining desorption charge. Ag_M/CoNiS shows a larger Q_des loss and a lower residual ratio than the physical mixture (Figure 4g), demonstrating more efficient nitrate‐associated ^*H utilization at the tightly coupled Ag/CoNiS interface. Collectively, the EPR, CV, KIE, DEMS, and H‐preload/transfer CV results support an interfacial‐confinement‐enabled hydrogen‐spillover model: CoNiS supplies active hydrogen species, Ag‐associated sites participate in nitrate/NO_x activation and hydrogenation, and the optimized Ag_M/CoNiS interface enables efficient short‐range ^*H relay and utilization during NO₃RR.

To further clarify the kinetic origin of the Ag‐density‐dependent NO₃RR behavior, EIS‐derived distribution of relaxation times (DRT) analysis was performed at −0.6 V versus RHE in 1 m KOH with 100 mm KNO₃. The DRT spectra deconvolute the overall response into a high‐frequency Zone 1, associated with fast interfacial charge/proton‐transfer processes, and a low‐frequency Zone 2, related to mass transport and intermediate diffusion (Figure 5a). Ag_M/CoNiS exhibits the smallest contributions in both zones, indicative of more favorable interfacial reaction kinetics in which CoNiS‐derived ^*H supply is synchronized with nitrate reduction on adjacent Ag sites. By contrast, Ag_L/CoNiS displays a retarded Zone 2 response, consistent with inefficient ^*H utilization and enhanced HER‐related transport interference, whereas Ag_D/CoNiS shows an enlarged Zone 1 contribution, reflecting hindered water activation and interfacial ^*H relay under excessive Ag coverage [18, 30, 51]. This structure–kinetics correlation is further summarized by mapping the optical bandgap (E_g) against the DRT‐derived kinetic resistance, where Ag_M/CoNiS occupies the optimal regime (Figure 5b).

FIGURE 5.

Origin of interfacial‐confinement‐programmed hydrogen spillover. (a) DRT spectra of CoNiS and Ag/CoNiS catalysts at −0.6 V vs. RHE in 1 m KOH + 100 mm KNO₃. (b) Correlation between bandgap and DRT‐derived kinetic resistance. (c) Potential‐dependent in situ Raman spectra in the O─H stretching region for Ag/CoNiS catalysts. (d) Relative proportions of interfacial water species derived from Raman deconvolution. (e) Normalized comparison of NH₃ FE, Raman‐derived K⁺–H₂O fraction, and CV‐derived H‐desorption charge. (f) TBA scavenger tests showing NH₃ FE and yield rate under different TBA concentrations. (g) Schematic illustration of Ag‐loading‐programmed interfacial confinement and hydrogen spillover.

Molecular‐level insights into the interfacial water environment were obtained from potential‐dependent in situ Raman spectra in the O─H stretching region (Figure 5c). Spectral deconvolution resolves three representative water configurations, including strongly hydrogen‐bonded water (4‐HB·H₂O), weakly hydrogen‐bonded water (2‐HB·H₂O), and ion‐hydrated active water (K⁺–H₂O) (Figure 5d) [52]. Although K⁺–H₂O generally increases at more negative potentials due to electric‐field‐driven electrical double layer restructuring, Ag_L/CoNiS maintains an excessively high K⁺–H₂O fraction, consistent with surplus ^*H generation and parasitic HER. In contrast, Ag_D/CoNiS is dominated by rigid 4‐HB·H₂O with suppressed K⁺–H₂O, indicating sluggish water dissociation and ^*H starvation. Importantly, Ag_M/CoNiS sustains an intermediate and more balanced population of 2‐HB·H₂O/K⁺–H₂O, suggesting that the confined Ag–CoNiS interface remodels interfacial water to promote ^*H generation without triggering uncontrolled HER [52, 53, 54, 55]. The normalized comparison of NH₃ FE, Raman‐derived K⁺–H₂O fraction, and CV‐derived H‐desorption charge further highlights the balanced interfacial ^*H supply–consumption state of Ag_M/CoNiS (Figure 5e).

Finally, tert‐butanol (TBA) perturbation experiments support the involvement of active hydrogen species in NO₃RR (Figure 5f). Increasing TBA concentration suppresses NH₃ yield for all catalysts, confirming that hydrogen‐related species are essential for deep nitrate hydrogenation. Ag_L/CoNiS shows a slight FE increase at low TBA concentration, consistent with partial scavenging of surplus ^*H species that otherwise favor HER, whereas CoNiS and Ag_D/CoNiS display more pronounced FE losses due to aggravated ^*H deficiency. Ag_M/CoNiS remains the most resilient toward TBA perturbation, further supporting its optimized ^*H relay and utilization capability. Collectively, these kinetic and spectroscopic results support an interfacial‐confinement‐programmed hydrogen spillover mechanism (Figure 5g): low Ag loading provides insufficient confined Ag–CoNiS interfaces, leading to inefficient ^*H utilization and H₂ formation; medium Ag loading creates optimized confined interfacial regions that couple CoNiS‐derived ^*H generation with Ag‐associated nitrate hydrogenation; and excessive Ag loading blocks water‐accessible CoNiS sites, suppressing ^*H generation and deep hydrogenation.

2.4. Pathway Studies and Theoretical Simulations

¹⁵N–labeling experiments unequivocally verified the nitrogen origin of the produced NH₃ (Figure S28): chronoamperometry at −0.6 V vs. RHE in K¹⁴NO₃ and K¹⁵NO₃ electrolytes yielded the characteristic ¹⁴NH₄ ⁺ triplet and ¹⁵NH₄ ⁺ doublet in ¹H NMR, respectively, while delivering comparable NH₃ yields and Faradaic efficiencies (Figures S29 and S30), confirming nitrate as the exclusive N source [56]. To interrogate the reaction pathway, in situ attenuated total reflectance Fourier‐transform infrared spectroscopy (ATR–FTIR) was performed on CoNiS and Ag_M/CoNiS (Figures S31 and S32), where Ag_M/CoNiS shows markedly intensified ^*NO₂ bands at 1250 cm⁻¹ and enhanced ^*NH₂ and ^*NH₃ features at 1434 and 1100 cm⁻¹ respectively, indicating that Ag accelerates both the initial nitrate activation/hydrogenation and the deep hydrogenation step; notably, the strengthened H₂O band at 1630 cm⁻¹ further suggests an elevated interfacial water population that thermodynamically and kinetically favors successive hydrogenation [57, 58, 59]. Operando differential electrochemical mass spectrometry (DEMS) was further used to monitor volatile/interfacial NO₃RR‐related intermediates. For Ag_M/CoNiS, H‐NO₃RR produces clear time‐resolved responses associated with N/NH_x fragments at m/z = 14, 15, 16, and 17, together with NO‐related m/z = 30 signals (Figure 6a). Additional monitoring of m/z = 31, 32, and 33 captures NHO‐, NHOH‐, and NH₂OH‐related responses during nitrate reduction (Figure 6b). Compared with CoNiS (Figure S33), Ag_M/CoNiS shows intensified downstream hydrogenation‐related responses, supporting more efficient conversion of NO_x intermediates toward NH₃. When the electrolyte was replaced by a deuterated system, D‐containing responses appeared under D‐NO₃RR conditions (Figure 6c), further confirming that solvent‐derived H/D species participate in the hydrogenation network. Together, the ATR–FTIR and H/D‐DEMS results support a dominant pathway of ^*NO₃ → ^*NO₂ → ^*NO → ^*N → ^*NH → ^*NH₂ → ^*NH₃ [60].

FIGURE 6.

NO₃RR pathway and theoretical simulations. (a) H‐NO₃RR DEMS profiles of Ag_M/CoNiS. (b) DEMS monitoring of NHO, NHOH, and NH₂OH‐related signals on Ag_M/CoNiS. (c) D‐NO₃RR DEMS profiles of Ag_M/CoNiS. (d) NO₃RR Gibbs free energy diagrams on CoNiS and Ag_M/CoNiS surfaces. Bader charge transfer from the catalyst slab to adsorbed NO₃ ⁻ and corresponding charge‐density‐difference plots for e) CoNiS and (f) Ag_M/CoNiS. The isosurface value is 0.00089 e Å⁻³. (g) Computational models simulating different Ag loading densities constructed by fixing the Ag cluster size while varying the surface area of the CoNiS substrate. (h) Projected density of states (PDOS) of Co and Ni sites with d–band center analysis (ε_d). (i) Gibbs free energy profiles for water dissociation on surfaces with different Ag loadings.

DFT calculations were then performed to rationalize the activity origin. Periodic models of CoNiS(200) and Ag_M(200)/CoNiS(200) were constructed for the NO₃RR pathway analysis (Figure S34). The calculated free‐energy diagrams show that pristine CoNiS suffers a pronounced early bottleneck, with ^*HNO₃ → ^*NO₂ as the largest uphill step (RDS = 0.827 eV; Figure 6d and Tables S3 and S4). In contrast, Ag_M/CoNiS lowers this barrier and shifts the RDS downstream to ^*NO₂H → ^*NO with a smaller energy penalty of 0.472 eV, consistent with Ag‐enabled stabilization and activation of early NO_x intermediates. The subsequent hydrogenation cascade becomes more energetically balanced on Ag_M/CoNiS, supporting an interfacial‐confinement‐enabled ^*H relay in which CoNiS‐derived ^*H species hydrogenate Ag‐associated NO_x intermediates toward NH₃. Charge‐density‐difference and Bader charge analyses further reveal stronger interfacial charge redistribution on Ag_M/CoNiS than on CoNiS (Figure 6e,f), with catalyst‐to‐NO₃ ⁻ electron transfer increasing from 0.62 e⁻ on CoNiS to 0.80 e⁻ on Ag_M/CoNiS. This enhanced electron donation implies stronger electronic coupling and more effective N─O bond activation at the heterointerface.

Finally, to model Ag–loading regulation, we fixed the Ag cluster size while varying the CoNiS surface area to construct Ag_L/CoNiS, Ag_M/CoNiS, and Ag_D/CoNiS (Figure 6g and Figure S35), which exhibit a progressive downshift of Co/Ni d–band centers with increasing Ag coverage (εd(Co) = −1.58 → −1.76 → −2.00 eV and εd(Ni) = −1.86 → −2.08 → −2.31 eV; Figure 6h), indicating weakened adsorbate interaction and thus diminished water–activation capability on CoNiS; accordingly, free–energy profiles for H₂O dissociation (Volmer step) show the RDS for all models but with a monotonic increase in ΔG _RDS from Ag_L/CoNiS (0.511 eV) to Ag_M/CoNiS (0.891 eV) and Ag_D/CoNiS (1.06 eV) (Figure 6i, Figure S36 and Tables S5–S7). These results indicate that increasing Ag density progressively screens CoNiS water‐activation motifs and raises the O─H cleavage barrier. Therefore, Ag_M/CoNiS achieves an optimal balance: sufficient Ag sites stabilize nitrate‐derived NO_x intermediates, while accessible CoNiS sites provide interfacial ^*H species for deep hydrogenation.

2.5. Bifunctional Electrocatalyst for NO₃RR and SOR

Given the exceptional NO₃RR performance of Ag_M/CoNiS enabled by optimized interfacial hydrogen spillover, we further evaluated its potential as a bifunctional electrocatalyst for coupled cathodic nitrate reduction and anodic sulfide oxidation. Replacing the sluggish oxygen evolution reaction (OER, E⁰ = 1.23 V vs. RHE) with the sulfide oxidation reaction (SOR, E⁰ = −0.48 V vs. RHE) can substantially reduce the anodic energy barrier while converting toxic sulfide (S²⁻) into value‐added sulfur, offering an energy‐saving strategy for ammonia electrosynthesis [61]. In a three‐electrode configuration, Ag_M/CoNiS exhibits a pronounced kinetic advantage for SOR over OER (Figure 7a), delivering 50 mA cm⁻² at only 0.42 V vs. RHE in 0.5 m Na₂S, far below the 1.66 V required for OER in 1.0 m KOH. This corresponds to a 1.24 V potential saving and also outperforms the CoNiS substrate (0.50 V at 50 mA cm⁻²), indicating that Ag modulation benefits anodic sulfide adsorption/conversion in addition to tuning cathodic ^*H supply. UV–vis analysis of electrolytes collected after different electrolysis durations reveals a clear color change from colorless to yellow and the emergence of characteristic absorption bands at ∼300 and ∼370 nm (Figure 7b), consistent with the formation of polysulfide species (S_x ²⁻, such as S₂ ²⁻ and S₄ ²⁻) during the S²⁻ → S_x ²⁻ → S pathway [62]. Meanwhile, the monotonic increase of anodic current with Na₂S concentration (0.1–1.0 M; Figure 7c) suggests diffusion‐influenced SOR kinetics, where improved sulfide mass transport accelerates anodic oxidation.

FIGURE 7.

Demonstration of electrochemical evaluation and economic viability of NO₃RR||SOR system. (a) LSV curves of Ag_M/CoNiS in 1 m KOH without or with 0.5 m Na₂S. (b) UV–vis spectra and photographs of 150–times diluted electrolytes at different durations. (c) LSV curves of Ag_M/CoNiS in 1 m KOH with varying Na₂S concentrations. (d) Polarization curves of H–cell two–electrode system for NO₃RR||OER and NO₃RR||SOR. (e) Durability measurements of H–cell three‐electrode system for SOR. (f) Durability measurements of H–cell two–electrode system for NO₃RR||SOR. (g) Schematic of the co‐production system. (h) Chronopotentiometric curves of MEA system for NO₃RR||OER and NO₃RR||SOR. (i) XRD and ¹H NMR spectra of recovered products, with corresponding optical photographs shown in the insets. (j) Production yields. (k) Techno‐economic analysis (TEA) comparison of net profits.

Encouraged by the enhanced half–reactions, a two–electrode hybrid electrolyzer using Ag_M/CoNiS as both NO₃RR cathode and SOR anode (NO₃RR||SOR) achieves markedly reduced cell voltages relative to the conventional NO₃RR||OER couple (Figure 7d): at an industrially relevant 50 mA cm⁻², only 0.70 V is required, 1.16 V lower than NO₃RR||OER (1.86 V), translating into substantial energy savings; meanwhile, robust stability is verified by steady SOR operation over 10 h (Figure 7e) and durable full–cell electrolysis without discernible voltage drift, indicating strong resistance to sulfide poisoning (Figure 7f). For practical validation, we integrated the catalyst into a membrane electrode assembly (MEA) flow cell (Figure 7g), where NO₃RR||SOR sustains 200 mA cm⁻² at ∼1.8 V, significantly below the ∼2.74 V required for NO₃RR||OER (Figure 7h); after electrolysis, acidifying the anolyte yields a yellow precipitate identified as elemental sulfur (S₈) by XRD, while cathodic products are recovered as NH₄Cl and (NH₄)₂SO₄, corroborated by XRD and ¹H NMR (Figure 7i and Figure S37), together delivering high production efficiencies for both sulfur and ammonia (Figure 7j) [63, 64, 65]. Finally, techno–economic analysis (TEA) highlights the economic advantage of the NO₃RR||SOR scheme, projecting a higher net profit (US $ 1.42 kg⁻¹ NH₃) than the OER–coupled benchmark due to the lowered electricity demand and the co–generation of two value–added products (Figure 7k), underscoring Ag_M/CoNiS as an efficient platform for coupled nitrate/sulfide remediation and “waste–to–value” electrosynthesis with substantially reduced energy input.

3. Conclusion

In summary, we demonstrate the rational design of Ag/CoNiS heterostructures via a ligand‐assisted strategy to achieve precise Ag loading density, establishing a high‐performance platform for NO₃RR. The optimized Ag_M/CoNiS creates confined Ag–CoNiS regions that program hydrogen spillover from CoNiS water‐activation domains to Ag‐associated nitrate/NO_x hydrogenation sites. This interfacial‐confinement‐programmed hydrogen spillover remodels the water microenvironment, balances ^*H supply–consumption, suppresses competitive HER, and accelerates deep nitrate hydrogenation. As a result, Ag_M/CoNiS achieves an NH₃ yield of 22.31 mg h⁻¹ cm⁻² with a Faradaic efficiency of 99.13%. Furthermore, the bifunctional NO₃RR||SOR system enables low‐energy co‐production of value‐added ammonia and elemental sulfur, highlighting interfacial confinement as an effective strategy for programming hydrogen spillover in waste‐to‐value electrosynthesis.

4. Experimental Section

4.1. Chemicals

Cobalt(II) acetate tetrahydrate (Co(CH₃COO)₂·4H₂O, AR), ethylene glycol (HOCH₂CH₂OH, CP), hexamethylenetetramine ((CH₂)₆N₄, CP), anhydrous ethanol (CH₃CH₂OH, AR), sublimed sulfur (S₈, AR) and silver nitrate (AgNO₃, AR) were purchased from Sinopharm Chemical Reagent Co., Ltd. Nickel(II) acetate tetrahydrate (Ni(CH₃COO)₂·4H₂O, AR) was purchased from Shanghai Yi'en Chemical Technology Co., Ltd. Thiosalicylic acid (H₂mba, C₇H₆O₂S, 98%) was purchased from Adamas‐beta. Poly(ethylene oxide)–poly(propylene oxide)–poly(ethylene oxide) triblock copolymer (PEO₂₀–PPO₇₀–PEO₂₀, Pluronic P123) was purchased from Aladdin Ltd. All the chemicals were utilized without any subsequent purification. Deionized water (DIW) with a resistivity of 18.2 MΩ·cm was employed for all experimental operations.

4.2. Synthesis of CoNiS

In a typical synthesis, 200 mg of P123 was dissolved in a mixed solvent of ethanol (16 mL), ethylene glycol (12 mL), and deionized water (1 mL). Subsequently, 65 mg of Co(CH₃COO)₂·4H₂O and 65 mg of Ni(CH₃COO)₂·4H₂O were added and dissolved via ultrasonication, followed by the addition of 70 mg of hexamethylenetetramine (HMTA) with another 30 min of sonication. The homogeneous solution was transferred into a Teflon‐lined autoclave and maintained at 170°C for 5 h. The resulting green CoNiOH precursors were collected by centrifugation, washed with water and ethanol, and dried at 60°C. To convert CoNiOH to CoNiS, the precursor and sulfur powder were placed in the downstream and upstream zones of a horizontal tube furnace, respectively. The sulfurization was conducted at 300°C for 2 h with a heating rate of 5°C min⁻¹ under an Ar atmosphere.

4.3. Synthesis of Ag/CoNiS

First, a silver precursor solution was prepared by dispersing 155 mg of H₂mba and 170 mg of AgNO₃ in 6 mL of DI water. After 5 min of sonication, 500 µL of ammonia solution was added dropwise, followed by continuous sonication and stirring until a transparent yellow solution was obtained.

To synthesize Ag_M/CoNiS, 15 mg of the as‐prepared CoNiS was dispersed in 800 µL of DI water, followed by the addition of 200 µL of the Ag precursor solution. The mixture was sonicated and shaken for 30 min to ensure uniform loading. The resulting product was collected by centrifugation, washed once with water, and freeze‐dried. Subsequently, the obtained powder was annealed at 350°C under an Ar atmosphere to obtain the final Ag_M/CoNiS catalyst. For comparison, samples with different Ag densities were prepared by adjusting the volume ratio of the Ag precursor to the CoNiS suspension. Ag_L/CoNiS and Ag_D/CoNiS were obtained using 50 and 500 µL of the Ag precursor solution mixed with 950 and 500 µL of the CoNiS suspension, respectively. Pure Ag was synthesized using the same protocol as Ag_M/CoNiS, except that the CoNiS suspension was replaced with DI water.

4.4. Materials Characterization

The crystallographic structure of the synthesized samples was analyzed by X‐ray diffraction (XRD) using a Bruker D8 Advance diffractometer with a Co Kα radiation source (35 kV, 40 mA). The scanning was performed over a 2θ range of 5° to 80° at a speed of 5°/min. Morphological features were examined using a field‐emission scanning electron microscope (SEM, TESCAN/MAIA3). Detailed microstructures, lattice fringes, and elemental distributions were characterized using transmission electron microscopy (TEM), high‐resolution TEM (HRTEM), and high‐angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM) coupled with energy‐dispersive X‐ray spectroscopy (EDS) mapping on an FEI Talos F200S instrument operating at 200 kV. Same‐region correlative STEM characterization was performed on a JEM‐ARM200F NEOARM atomic‐resolution transmission electron microscope (JEOL) by recording identical representative regions under bright‐field STEM (BF‐STEM) and secondary‐electron imaging composite (SEI‐composite) modes. Aberration‐corrected HAADF‐STEM images were acquired on the same JEM‐ARM200F NEOARM equipped with a cold field‐emission electron gun and an ASCOR probe spherical aberration corrector, operated at 80–200 kV with nominal STEM resolutions of 78 pm at 200 kV and 110 pm at 80 kV. The morphology (height profile) and surface potential distribution of the catalysts were simultaneously characterized using a commercial atomic force microscope (Dimension Icon, Bruker) operating in Kelvin Probe Force Microscopy (KPFM) mode. The surface area and pore structure were determined via nitrogen adsorption–desorption isotherms collected on a Quantachrome Autosorb–IQ instrument. Prior to measurement, samples were degassed at 180°C under vacuum for 6 h. The specific surface area and pore size distribution were calculated using the Brunauer–Emmett–Teller (BET) method and Barrett–Joyner–Halenda (BJH) model, respectively. Thermal stability was assessed using a Thermogravimetric Analyzer (TGA, Q5000IR) under an air atmosphere from room temperature to 800°C with a heating rate of 10°C/min. The actual metal loading (Ag, Co, Ni) in the catalysts was quantified by inductively coupled plasma optical emission spectroscopy (ICP–OES) using a Leeman Labs Prodigy 7 system. The ultraviolet–visible diffuse reflectance spectra (UV–vis DRS) was recorded with a spectrophotometer (UV3600 i Plus). To investigate the electronic structure and band alignment, X‐ray photoelectron spectroscopy (XPS) and Ultraviolet Photoelectron Spectroscopy (UPS) were conducted on a Thermo Scientific Escalab 250Xi system. XPS measurements utilized Al Kα radiation (hν = 1486.6 eV), with binding energies calibrated to the C 1s peak at 284.8 eV. UPS spectra were acquired using a He(I) excitation source (hν = 21.22 eV) to determine the work function and valence band edge. X‐ray absorption near‐edge structure (XANES) spectra (Co and Ni K‐edge) were collected at the Shanghai Synchrotron Radiation Facility (SSRF) to probe the local coordination environment and oxidation states. The optical bandgap was derived from UV–vis diffuse reflectance spectra (UV–vis DRS) using the Kubelka–Munk function and a direct‐allowed Tauc model. Specifically, the bandgap was obtained by extrapolating the linear region of the (F(R)hν)² versus hν plot to the energy axis, where F(R) is the Kubelka–Munk function. For product quantification and mechanistic studies, ¹H Nuclear Magnetic Resonance (¹H NMR) spectroscopy was performed on a Bruker Advance 3 HD 600 MHz spectrometer using dimethyl sulfoxide (DMSO) as an internal standard. Isotopic labeling experiments using ¹⁵NO₃ ⁻ were conducted to verify the nitrogen source. In situ attenuated total reflectance Fourier‐transform infrared (ATR–FTIR) spectroscopy was carried out on a PerkinElmer Spectrum 3 to monitor reaction intermediates. The interfacial water environment was probed by in situ Raman spectroscopy using a Renishaw inVia–Reflex system with a 532 nm laser. Real‐time gaseous and volatile products were detected using Differential Electrochemical Mass Spectrometry (DEMS) on a Linglu QAS 100 apparatus equipped with a semipermeable membrane inlet system.

4.5. Preparation of Working Electrode

To prepare the working electrode, 5 mg of the catalyst powder was dispersed into a mixed solution consisting of 600 µL of ethanol, 350 µL of isopropanol, and 50 µL of Nafion solution (5 wt.%). The mixture was ultrasonicated for 15 min to form a homogeneous catalyst ink. Subsequently, 200 µL of the as‐prepared ink was drop‐cast onto a piece of carbon paper (CP) with a defined surface area of 1 × 1 cm² and dried naturally at room temperature.

4.6. Electrochemical Tests

Electrochemical measurements were carried out on a Biologic VMP3 electrochemical workstation using a standard three–electrode H–type cell separated by a Nafion 117 membrane. Prior to use, the Nafion membrane was pretreated by successively boiling in 5% H₂O₂ aqueous solution at 80°C for 1 h and then in deionized water at 80°C for another 1 h to remove impurities.

The as‐prepared electrode (1 × 1 cm²) served as the working electrode. A Pt foil and a Hg/HgO electrode (filled with 1.0 m KOH) were employed as the counter and reference electrodes, respectively. All measured potentials were converted to the Reversible Hydrogen Electrode (RHE) scale according to the Nernst equation:

$$E_{\mathrm{RHE}}=E_{\mathrm{Hg}/\mathrm{HgO}}+0.059\times pH+0.098V$$

The electrochemical double‐layer capacitance (C_dl) was determined from CV curves collected in a non‐Faradaic potential region at scan rates of 10, 20, 40, 60, 80, and 100 mV s⁻¹. The C_dl value was obtained from the slope of the linear relationship between the capacitive current density difference and the scan rate. The electrochemically active surface area (ECSA) was calculated according to:

$$\mathrm{ECSA}=C_{\mathrm{dl}}/C_{\mathrm{s}}$$

where C_s was taken as 40 µF cm⁻².

For relative hydrogen‐desorption analysis, CV measurements were performed in the potential window covering the hydrogen adsorption/desorption region. The anodic hydrogen‐desorption/oxidation feature was integrated after baseline correction, and the obtained charge was denoted as Q _des:

$$Q_{\mathrm{des}}=\frac{1}{v}{\int }_{E_1}^{E_2}\left[I\left(E\right)-I_{\mathrm{base}}\left(E\right)\right]dE$$

where I(E) is the measured anodic current, I _base(E) is the baseline current, v is the scan rate, and E ₁ −E ₂ is the integration window of the hydrogen‐desorption/oxidation feature. The same integration window and baseline‐correction procedure were applied to all samples. Therefore, Q _deswas used only as a relative descriptor of electrochemically detectable hydrogen‐related species, rather than as an absolute ^*H surface coverage.

The nitrate‐induced change in hydrogen‐desorption charge and residual ratio were calculated as:

$$\mathrm{\Delta }{Q}_{\mathrm{des}}=Q_{\mathrm{des}}\left(\mathrm{preload}\right)-Q_{\mathrm{des}}\left(\mathrm{preload}\to {\mathrm{NO}}_3^{-}\right)$$

$$\mathrm{Residual}\mathrm{ratio}=\frac{Q_{\mathrm{des}}\left(\mathrm{preload}\to {\mathrm{NO}}_3^{-}\right)}{Q_{\mathrm{des}}\left(\mathrm{preload}\right)}$$

where Q _des(preload) is the H‐desorption charge recorded after H preloading in nitrate‐free KOH, and $Q_{\mathrm{des}}(\mathrm{preload}\to {\mathrm{NO}}_3^{-})$ is the residual H‐desorption charge recorded after transferring the H‐preloaded electrode to a nitrate‐containing electrolyte.

Kinetic isotope effect (KIE) experiments were performed by comparing NO₃RR activity in protonic and deuterated alkaline electrolytes. The protonic electrolyte was 1 M KOH containing 100 mm KNO₃ prepared with H₂O, while the deuterated electrolyte was D₂O‐based alkaline electrolyte. Electrolysis was conducted under identical potential, reaction time, and electrode‐loading conditions. The apparent KIE value was calculated as:

$$\mathrm{KIE}=r_{\mathrm{H}}/r_{\mathrm{D}}$$

where r _H and r _D are the NH₃ yield rates measured in H₂O/KOH and D₂O/KOH electrolytes, respectively.

EIS measurements were performed at a potential of −0.6 V vs. RHE over a frequency range from 100 kHz to 0.01 Hz with an AC amplitude of 5 mV to investigate the charge transfer kinetics. The Distribution of Relaxation Times (DRT) analysis was further employed to deconvolve the impedance data and identify specific kinetic processes.

The sulfide oxidation reaction (SOR) performance was first evaluated in a standard three–electrode H–type cell. The CoNiS or Ag_M/CoNiS catalyst served as the working electrode (anode), while a Pt foil and a Hg/HgO electrode acted as the counter and reference electrodes, respectively. The anode compartment was filled with 1.0 m KOH containing 0.5 m Na₂S (or pure 1.0 m KOH for OER comparison), while the cathode compartment contained 1.0 m KOH + 100 mm KNO₃. For the coupled NO₃RR||SOR electrolysis, a two–electrode H–cell system was assembled using CoNiS or Ag_M/CoNiS as both the cathode and the anode. The catholyte was composed of 1.0 m KOH + 100 mm KNO₃, and the anolyte was 1.0 m KOH containing 0.5 m Na₂S (or pure 1.0 m KOH for the NO₃RR||OER control).To assess the potential for practical applications, a membrane electrode assembly (MEA) flow cell was constructed, employing Ag_M/CoNiS as the bifunctional catalyst for both the cathode and anode. The cell compartments were separated by a Nafion 117 membrane. The catholyte (1.0 m KOH + 100 mm KNO₃) and anolyte (1.0 m KOH + 0.5 m Na₂S) were circulated through their respective chambers using a peristaltic pump (Kamoer, DKCP–S10) at a constant flow rate of 40 mL min⁻¹.

Prior to all the measurements, the electrolytes were purged with Ar gas to maintain an inert environment and prevent side reactions.

4.7. UV–vis Analysis for Ionic Concentration Determination in Solution (Determination of NH₄ ⁺)

The concentration of produced ammonia was quantitatively determined using the indophenol blue method. The detection reagents were prepared as follows:

Reagent A: 5 g of salicylic acid and 5 g of sodium citrate were dissolved in 100 mL of 1 m NaOH solution.

Reagent B: 0.5 mL of NaClO solution (available chlorine ≈ 7.5%) was added to 19.5 mL of deionized water.

Reagent C: 0.5 g of sodium nitroferricyanide (C₅FeN₆Na₂O) was dissolved in 50 mL of deionized water.

In a typical measurement, 2 mL of the diluted catholyte sample (or standard solution) was collected and sequentially mixed with 2 mL of Reagent A, 1 mL of Reagent B, and 0.2 mL of Reagent C. The mixture was allowed to react in the dark at room temperature for 2 h to ensure complete color development. The absorbance was subsequently measured at a wavelength of 655 nm using a UV–vis spectrophotometer. The concentration of NH₄ ⁺ in the electrolyte was calculated based on the linear relationship between the absorbance and the standard NH₄ ⁺ concentration.

4.8. UV–Vis Analysis for Ionic Concentration Determination in Solution (Determination of NO₂ ⁻)

The concentration of the nitrite byproduct was quantitatively determined using the typical Griess method. The chromogenic reagent was prepared as follows:

4.9. Griess Reagent

2.0 g of p‐aminobenzene sulfonamide was first dissolved in a mixture of 25 mL of deionized water and 5 mL of phosphoric acid. Subsequently, 0.1 g of N–(1–naphthyl)–ethylenediamine dihydrochloride was dissolved in the above solution. The resulting mixture was transferred to a 50 mL volumetric flask and diluted to the mark with deionized water.

In a typical measurement, 5 mL of the diluted catholyte sample (or standard solution) was prepared. Then, 0.1 mL of the prepared Griess reagent was added to the solution. The mixture was shaken thoroughly and allowed to react at room temperature for 15 min to ensure stable color development. The absorbance was measured at a wavelength of 540 nm using a UV–vis spectrophotometer. The concentration of NO₂ ⁻ in the electrolyte was calculated based on the linear relationship between the absorbance and the standard NO₂ ⁻ concentration.

4.10. Isotope Labeling Experiments by ¹H NMR

The yields of electrolytic ammonia (¹⁴NH₄ ⁺ and ¹⁵NH₄ ⁺) were quantitatively determined using ¹H NMR spectroscopy. Calibration curves were constructed using standard ¹⁵NH₄Cl solutions (¹⁵N abundance: ≥ 99 atom%) with dimethyl sulfoxide (DMSO) serving as the internal standard. In a typical procedure, the pH of the collected electrolyte or standard solution (2 mL) was adjusted to approximately 2.0 by adding concentrated sulfuric acid (4 m H₂SO₄). Subsequently, 800 µL of the acidified solution was mixed with 0.3 mL of deuterium oxide (D₂O) containing DMSO. Finally, the prepared samples were analyzed on a 600 MHz NMR spectrometer employing the water suppression method to minimize solvent interference.

4.11. Calculation of the Yield Rate and the Faradaic Efficiency of NH₃

The ammonia yield and FE were calculated according to the following equations:

$$Yield=\frac{c\times V}{t\times S}$$

$$\mathrm{FE}\left(\%\right)=\frac{n\times F\times c\times V}{M\times Q}\times 100$$

In these equations, c denotes the measured mass concentration of ammonia in the electrolyte. V and t correspond to the volume of the electrolyte (30 mL) and the duration of the electrolysis (1 h), respectively. S represents the geometric surface area of the working electrode (1 cm²). F is the Faraday constant (96485 C mol⁻¹), and Q signifies the total electric charge passed through the circuit, derived from the integration of the i–t curve. M represents the molar mass of one molecule (e.g., NH₃ 17 g mol⁻¹), and n indicates the number of electrons transferred for the reduction of one nitrate/nitrite molecule to ammonia.

4.12. EPR Experiments

Electron paramagnetic resonance (EPR) spectroscopy was employed to detect electrochemically generated hydrogen‐related species using DMPO as the spin‐trapping reagent. The measurements were performed on a Bruker A300 EPR spectrometer (Germany). In a typical experiment, the catalyst‐loaded working electrode was polarized at a given potential in Ar‐saturated NO₃ ⁻‐free KOH electrolyte for 600 s. The electrolyte was deoxygenated by Ar bubbling before the test to remove dissolved oxygen. After polarization, 1–2 mL of the electrolyte near the electrode surface was rapidly collected and immediately mixed with 10 µL of 5,5‐dimethyl‐1‐pyrroline N‐oxide (DMPO). The obtained solution was then transferred into a capillary tube for EPR measurement. All samples were tested under identical electrolyte, polarization, and collection conditions.

4.13. In Situ Raman Spectroscopy

In situ Raman measurements were conducted using a custom‐designed polytetrafluoroethylene (PTFE) cell equipped with a quartz window. The spectra were acquired on a Dilor LabRam–113 micro–Raman spectrometer (France) (532 nm laser) coupled with a CHI 660 electrochemical workstation. The prepared electrode served as the working electrode, positioned perpendicular to the laser beam. A Pt wire and a Hg/HgO electrode were employed as the counter and reference electrodes, respectively. The experiments were performed in an electrolyte of 1.0 m KOH containing 100 mm KNO₃. To investigate the potential‐dependent surface species, chronoamperometry was applied to hold the potential constant for 200 s at each step before spectral acquisition.

4.14. In Situ Differential Electrochemical Mass Spectrometry (DEMS) Analysis

In situ DEMS measurements were performed on a QAS 100 system (Linglu Instrument) using a cell equipped with a hydrophobic breathable membrane. High‐purity Ar was used as the carrier gas. After establishing a stable baseline, the test was conducted in 0.1 m KNO₃ at a constant potential of −0.6 V vs. RHE for 10 s. The mass signals of volatile products were recorded simultaneously with the electrochemical data.

4.15. D‐Labeling DEMS Analysis

D‐labeling DEMS measurements were performed on a Linglu QAS 100 system equipped with a semipermeable membrane inlet. High‐purity Ar was used as the carrier gas. The electrolysis was conducted at the selected potential under the same conditions as those used for H‐NO₃RR DEMS measurements, and time‐resolved mass signals were continuously recorded during NO₃RR.

For the D‐preload → H‐NO₃RR switching experiment, the catalyst electrode was first polarized at −0.6 V vs. RHE in a deuterated alkaline electrolyte to preload interfacial D‐related species. After D preloading, the electrode was carefully rinsed to remove residual deuterated electrolyte and then rapidly transferred into protonic 1 M KOH containing 100 mm KNO₃ for subsequent H‐NO3RR measurements. Consecutive H‐NO₃RR electrolysis cycles were then performed at −0.6 V vs. RHE, while the mass signals at m/z = 18, 19, and 20 were continuously monitored to track the evolution and decay of D‐containing responses. The signals were normalized when necessary for comparison among different samples. Because these mass channels may contain contributions from water isotopologues and ammonia‐related isotope fragments, they were used as comparative D‐containing response channels rather than being exclusively assigned to a single molecular species.

4.16. In Situ FT−IR Experiments

In situ ATR–FTIR measurements were performed using a Thermo Nicolet Nexus 670 spectrometer equipped with a Pike Veemax III accessory (single‐bounce Si crystal coated with an Au membrane) and coupled with a CHI 660 electrochemical workstation. A specialized spectro–electrochemical cell (Shanghai Yuanfang Technology) was employed for the tests. The catalyst ink was prepared by dispersing 5 mg of the catalyst in a mixture of isopropanol (25 µL), water (25 µL), and 5 wt.% Nafion solution (10 µL) via ultrasonication for 30 min. Subsequently, 30 µL of the homogeneous ink was drop‐cast onto the surface of the Si‐supported Au membrane, serving as the working electrode. A Pt wire and a Hg/HgO electrode were used as the counter and reference electrodes, respectively. The measurements were conducted in an electrolyte of 1.0 m KOH containing 100 mm KNO₃. Background spectra were collected at the open‐circuit potential (OCP). Subsequently, time‐resolved infrared spectra were recorded in the wavenumber range of 1000–4000 cm⁻¹ under potentiostatic conditions ranging from −0.3 to −0.7 V vs. RHE, with a holding time of 300 s at each potential.

4.17. DFT Method

Density functional theory (DFT) calculations were performed using the Vienna ab initio Simulation Package (VASP) with the projector augmented wave (PAW) method [66, 67, 68]. Exchange–correlation interactions were treated within the generalized gradient approximation (GGA) using the Perdew–Burke–Ernzerhof (PBE) functional [69]. The plane‐wave kinetic energy cutoff was set to 450 eV. The electronic self‐consistent loop was converged to 1 × 10⁻⁵ eV, and the structures were considered converged when the residual forces were below 0.03 eV Å⁻¹. Dispersion interactions were included using Grimme's DFT–D3 scheme [70]. Surface slab models of CoNiS(200), and Ag_M(200)/CoNiS(200) (Ag supported on the CoNiS(200) surface) were constructed using four‐layer slabs, where the bottom two layers were fixed, and the top two layers were fully relaxed. A vacuum region of 15 Å was added along the surface normal direction to avoid interactions between periodic images. The Brillouin zone was sampled using Monkhorst–Pack k–point meshes [71]. For interface models with different in–plane supercell sizes (corresponding to different Ag loadings/coverages), a constant k–point spacing of KSPACING = 0.040 Å⁻¹ was adopted to ensure comparable k–space sampling density, which resulted in 2 × 2 × 1 meshes for the high– and medium–density models and a 1 × 1 × 1 mesh for the low–density model (a single k–point was used along the surface normal direction). For smaller unit cells (e.g., clean surface calculations), a denser 3 × 3 × 1 mesh was used when applicable. The Gibbs free energy changes (ΔG) of elementary steps were evaluated within the computational hydrogen electrode (CHE) model, where (H⁺+ e⁻) is referenced to 1/2 H₂ at 0 V vs. RHE [72]. Free energies were obtained at 298.15 K as the following equation: [72]

$$G=E_{\mathrm{DFT}}+\mathrm{ZPE}-TS$$

In the Bader analysis, the reported charge‐transfer value denotes electron transfer from the catalyst slab to the adsorbed NO₃ ⁻ species. To avoid direct treatment of the charged NO₃ ⁻ species, gaseous HNO₃ was used as a reference to construct the thermodynamic cycle for nitrate adsorption and related steps, following a reported approach [73].

4.18. Purification and Collection of Ammonia Products

After electrolysis, the catholyte was collected and transferred to a flask, then heated to 70°C while Ar was bubbled through the solution for 24 h to drive dissolved NH₃ out of the electrolyte. The outgoing gas stream was routed to different capture media to obtain three purified ammonia products: directing the gas into 2.0 m HCl immobilized NH₃ as NH₄Cl, after which the absorber was concentrated by rotary evaporation and the residue was dried overnight at 80°C to yield the solid salt; using 1.0 m H₂SO₄ instead fixed NH₃ as (NH₄)₂SO₄, which was likewise isolated by solvent removal (rotary evaporation) followed by overnight drying at 80°C; alternatively, bubbling the gas directly into deionized water afforded NH₃·H₂O as an aqueous product, and its ammonia content can be quantified by ¹H NMR.

4.19. Efficiency Calculations for Ammonia Purification and Product Collection

The Faradaic efficiency toward ammonia in the MEA configuration was obtained from the total produced ammonia amount and the passed charge:

$$\mathrm{F}{\mathrm{E}}_{\mathrm{NH}3}\left(\%\right)=\frac{8F{n}_{\mathrm{NH}3,\mathrm{prod}}}{Q}\times 100$$

where Q is the total charge, and F is the Faraday constant.

The NH₃ removal efficiency reflects how much ammonia was transferred out of the catholyte during Ar stripping:

$${\eta }_{\mathrm{removal}}\left(\%\right)=\frac{n_{\mathrm{cat},0}-n_{\mathrm{cat},\mathrm{res}}}{n_{\mathrm{cat},0}}\times 100$$

where n _cat,0 and n _cat,res are the ammonia amounts in the catholyte before and after stripping, respectively.

The NH₃ absorption efficiency describes the fraction of stripped ammonia captured by the absorber (acid or water):

$${\eta }_{\mathrm{abs}}\left(\%\right)=\frac{n_{\mathrm{abs}}}{n_{\mathrm{strip}}}\times 100,n_{\mathrm{strip}}=n_{\mathrm{cat},0}-n_{\mathrm{cat},\mathrm{res}}$$

For solid product collection, the captured ammonia was converted into salts and isolated after solvent removal and drying. The collection efficiency compares the ammonia contained in the isolated solid to that initially absorbed:

For NH₄Cl:

$${\eta }_{\mathrm{col},\mathrm{NH}4\mathrm{Cl}}\left(\%\right)=\frac{\left(m_{\mathrm{NH}4\mathrm{Cl}}/M_{\mathrm{NH}4\mathrm{Cl}}\right)}{n_{\mathrm{abs}}}\times 100$$

For (NH₄)₂SO₄:

$${\eta }_{\mathrm{col},\left(\mathrm{NH}4\right)2\mathrm{SO}4}\left(\%\right)=\frac{2\left(m_{\left(NH4\right)2SO4}/M_{\left(NH4\right)2SO4}\right)}{n_{\mathrm{abs}}}\times 100$$

For NH₃·H₂O (aq) collected in water, the collection efficiency was evaluated by the ammonia quantified in the final absorption solution relative to the absorbed amount:

$${\eta }_{\mathrm{col},\mathrm{NH}3·\mathrm{H}2\mathrm{O}}\left(\%\right)=\frac{n_{\mathrm{aq}}}{n_{\mathrm{abs}}}\times 100$$

(where n _aq is determined from the measured concentration and volume of the absorption solution).

Finally, the S₈ collection efficiency was defined as the ratio between the experimentally recovered S₈ amount (from collected mass) and the theoretical amount expected from charge (or from the corresponding Faradaic yield when applicable):

$${\eta }_{\mathrm{col},\mathrm{S}8}\left(\%\right)=\frac{n_{\mathrm{S}8,\mathrm{col}}}{n_{\mathrm{S}8,\mathrm{theo}}}\times 100$$

4.20. Techno‐Economic Assessment (TEA) and Calculation Details

A techno‐economic analysis was conducted to compare the economic viability of the proposed NO₃RR||SOR system against the traditional NO₃RR||OER system. The assessment considers both grid‐electricity scenarios and renewable‐energy‐driven scenarios.

4.21. Energy Consumption (E_input )

The electrical energy consumption per kilogram of ammonia produced ($kWh{\mathrm{kg}}_{N{H}_3}^{-1}$) is calculated as:

$$E_{input}=\frac{U_{cell}\times n\times F}{3600\times M_{N{H}_3}\times F{E}_{N{H}_3}}$$

where U_cell is the average cell voltage recorded at 200 mA cm⁻², n  =  8 is the number of electrons required for NO₃ ⁻‐to‐NH₃ conversion, F is the Faraday constant, M _NH3 is the molar mass of NH₃, and FE _NH3 is the Faradaic efficiency for NH₃.

4.22. Techno‐Economic Assessment (TEA) Details

A comparative techno‐economic assessment (TEA) was conducted to evaluate the economic viability of the proposed NO₃RR||SOR system against the traditional NO₃RR||OER benchmark, considering both grid (US $0.03 kWh⁻¹ and renewable electricity (US $0.015 kWh⁻¹ scenarios [74]. The levelized cost of ammonia (LCOA) accounts for electricity consumption (derived from the average cell voltage at 200 mA cm⁻² and 95% FE) and non‐electric costs (US $0.14 kg_NH3 ⁻¹, covering maintenance and materials) [74]. On the revenue side, the cathodic product is valued as either standard anhydrous ammonia (US $0.74 kg⁻¹) or high‐value ammonium hydroxide (US $1.60 kg⁻¹, adjusted for water costs), with an additional 10% premium applied to renewable‐driven “green ammonia” [75]. Anodic revenue integrates the high‐value recovery of solid sulfur (US $0.22 kg⁻¹ with 76% recovery efficiency) for the SOR system, contrasting with the low economic value of oxygen (US $0.05 kg⁻¹) in the OER system [74].

Author Contributions

F. T. Xie designed and wrote the study; Z. Y. Wu and X. X. Kang performed the DFT simulation; H. L. Zhu and Z. T. Li helped with the synthesis and analysis; Z.T. Li, Z. Y. Wu, L. Wang, and J. P. Yang revised the manuscript.

Funding

This work was supported by the National Natural Science Foundation of China (Nos. 52172291 and 52473294), China Postdoctoral Science Foundation (2023M740584), Shanghai Pujiang Program (No. 23PJD001), Key R&D Project in Shaanxi Province (2025CY‐YBXM‐469).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: adma73885‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.