Electronic Modulation and Surface Reconstruction of NiS₂ for Enhanced Alkaline Oxygen Evolution Reaction Activity and Durability

ABSTRACT

Practical anion exchange membrane water electrolysis (AEMWE) requires non‐noble‐metal oxygen evolution reaction (OER) anodes that deliver high activity while maintaining structural integrity under alkaline operation. Here, Fe‐substituted Ni disulfide (Ni_1‐xFe_xS₂, NFS‐x) is designed as a conductive Fe‐modulated sulfide framework and combined with alkaline pretreatment (AT) to pre‐form a NiFe (oxy)hydroxide‐rich surface before electrolysis. DFT calculations show that Fe substitution narrows the NiS₂ band gap by ∼0.44 eV and lowers the free‐energy barrier for the rate‐determining ^*O → ^*OOH step by 0.59 eV, while spectroscopic analysis indicates an increased average Ni oxidation state. The optimized NFS‐15 achieves an overpotential of 286 mV at 10 mA cm⁻². AT produces a largely amorphous/poorly crystalline NiFe (oxy)hydroxide‐rich outer layer on the retained sulfide core, confining subsequent reconstruction, suppressing catalyst detachment, and enabling superhydrophilicity‐assisted bubble release. In an AEMWE single cell, the NFS‐15‐AT/NF anode reaches 1 A cm⁻² at 1.65 V, outperforms commercial IrO₂ under identical conditions, and maintains stable operation for >200 h. This work demonstrates that integrating Fe‐induced electronic modulation of the sulfide framework with surface‐confined pre‐reconstruction provides an effective route to durable sulfide‐derived anodes for practical AEMWE.

A simple alkaline pretreatment converts Fe‐doped nickel disulfide into a sulfide‐derived core–shell anode with a NiFe (oxy)hydroxide‐rich outer layer for anion exchange membrane water electrolysis. Iron tuning optimizes oxygen‐evolution intermediate binding, while the reconstructed outer layer confines further surface changes, suppresses metal and anion leaching, and improves wettability for rapid bubble release. These synergistic effects deliver high activity and durable operation.

1. Introduction

The transition toward sustainable and clean energy systems requires green hydrogen production technologies that can convert and store intermittent renewable energy without carbon emissions [1, 2]. Anion exchange membrane water electrolysis (AEMWE) is an electrolysis platform that integrates the merits of alkaline water electrolysis and proton exchange membrane water electrolysis. Based on a zero‐gap architecture, AEMWE can achieve high gas purity and energy efficiency. In addition, its operation in an alkaline environment alleviates the corrosion burden associated with proton exchange membrane systems and enables the replacement of platinum‐group‐metal catalysts, drawing considerable attention as a promising route for cost‐effective hydrogen production [3, 4]. Nevertheless, the intrinsically sluggish four‐electron oxygen evolution reaction (OER) at the anode still requires large overpotentials, which remains a primary factor limiting the overall energy efficiency of AEMWE systems [5, 6]. Therefore, for the commercialization of AEMWE, it is essential to develop non‐noble metal‐based OER catalysts that simultaneously deliver efficient oxygen evolution and long‐term stable operation under practical operating conditions [7, 8].

Recently, compounds based on earth‐abundant transition metals, such as transition metal sulfides, selenides, phosphides, and nitrides, have been widely investigated as non‐noble‐metal catalysts for alkaline OER because of their high electrical conductivity and tunable electronic structures [9, 10]. Among them, nickel disulfide (NiS₂) has been considered a promising candidate due to its low cost, elemental abundance, and appreciable OER activity [11, 12]. However, pristine NiS₂ still shows limited performance because the electronic states of its active sites are not optimized for the adsorption and desorption of OER intermediates. To address this limitation, multicomponent structural design through heterometal doping has been explored as an activity‐enhancement strategy for Ni‐based catalysts [13, 14]. In particular, Fe incorporation and NiFe‐based compositions have been widely investigated in Ni‐based oxides, hydroxides, and oxyhydroxides as effective motifs for enhancing OER activity, because Fe can modulate the electronic structure and redox behavior around Ni centers [15, 16]. However, in a disulfide lattice, the metal–sulfur coordination environment can induce charge redistribution and intermediate‐binding behavior distinct from those in oxide/hydroxide systems [17, 18]. Therefore, in Fe‐doped NiS₂, the Fe‐induced activity enhancement should be rationalized primarily in terms of Ni‐centered electronic modulation and its influence on OER intermediate binding within the sulfide phase [19, 20].

During alkaline OER, particularly under AEMWE‐relevant anodic conditions, Ni‐based sulfides, along with other transition metal chalcogenides, phosphides, and nitrides, undergo surface oxidation and structural reconstruction to form an (oxy)hydroxide phase, which is widely regarded as the catalytically active phase for OER [21, 22]. Zhong et al. reported that NiS₂, NiSe₂, and Ni₅P₄ all transform into Ni (oxy)hydroxide under OER conditions; however, differences in the initial crystal phase and anion chemistry lead to distinct degrees of NiO₆ octahedral distortion and local electronic structure, resulting in different OER activities, with NiS₂‐based catalysts showing the best performance among them [23]. Although the activity mechanism of the reconstructed (oxy)hydroxide phase has been extensively studied [24, 25], reconstruction‐associated degradation pathways, including core‐phase collapse, catalyst detachment, and the dissolution of metal and anionic species, have been comparatively less addressed despite their impact on the long‐term durability of full‐cell components such as the anode, cathode, and anion exchange membrane [26]. Thus, the rational design of sulfide‐derived AEMWE catalysts should combine sulfide‐phase electronic tuning with controlled surface pre‐reconstruction, enabling the formation of an OER‐active (oxy)hydroxide‐rich surface while retaining the conductive sulfide core and suppressing reconstruction‐induced degradation [27].

In this study, we develop a Fe‐substituted NiS₂ OER catalyst that couples sulfide‐phase electronic modulation with controlled alkaline pre‐reconstruction. Fe‐substituted NiS₂ (Ni_1‐xFe_xS₂, NFS‐x) is first designed as an electronically tunable sulfide framework, in which Fe incorporation modulates the band structure, shifts Ni toward a higher average oxidation state, and lowers the free‐energy barrier of the rate‐determining ^*O → ^*OOH step by 0.59 eV. As a result, the optimized NFS‐15 catalyst exhibits the highest intrinsic OER activity, requiring an overpotential of 286 mV at 10 mA cm⁻². Before electrochemical operation, NFS‐15 is subjected to alkaline pretreatment to stabilize the reconstructed surface, yielding NFS‐15‐AT with a predominantly amorphous or poorly crystalline NiFe (oxy)hydroxide‐rich outer layer on a retained sulfide‐rich core. In this surface/core structure, the Fe‐modulated sulfide core provides a conductive framework, while the pre‐formed NiFe (oxy)hydroxide‐rich surface serves as the OER‐active interface and confines further reconstruction to the outer region. Consequently, in AEMWE single‐cell tests, the NFS‐15‐AT/NF anode achieved a current density of 1 A cm⁻² at only 1.65 V, surpassing the performance of commercial IrO₂, and demonstrated long‐term durability with minimal degradation after 200 h of continuous operation at 1 A cm⁻².

2. Results and Discussion

2.1. Structural and Morphological Characterization of NFS‐x

Figure 1a schematically illustrates the overall synthesis procedure for NFS‐x, which involves the solvothermal formation of spherical NiFe‐glycerate precursors followed by sulfidation. The optimization of sulfur content and annealing temperature to achieve a pure NiS₂ phase is presented in Figure S1. XRD patterns of samples with varying Fe‐doping levels exhibit only the characteristic peaks of the cubic pyrite‐type NiS₂ phase (PDF 88–1709), with no detectable Fe‐related secondary phases (Figure 1b) [28]. As the Fe content increases, a systematic shift of the NiS₂ diffraction peaks toward higher angles is observed, likely due to lattice contraction induced by the substitution of smaller Fe²⁺ ions (0.75 Å) for Ni²⁺ ions (0.83 Å) within the octahedral coordination of the NiS₂ lattice [29]. Raman spectra display the characteristic vibrational modes of NiS₂, corresponding to the Eg (∼279 cm⁻¹) and Ag (∼476 cm⁻¹) modes (Figure 1c). With increasing Fe‐doping, both peaks exhibit a blue shift and noticeable broadening. The blue shift is attributed to the decrease in reduced mass upon Fe doping, whereas the peak broadening is associated with lattice distortion at higher doping concentrations [30, 31, 32]. Elemental analysis by ICP‐OES confirmed that the Fe contents in all samples closely matched the target compositions, demonstrating precise compositional control (Table S1).

FIGURE 1.

(a) Schematic illustration of the synthesis procedure for NFS‐x. (b) XRD patterns, (c) Raman spectra, and (d) SEM images with corresponding EDS elemental mappings of NFS‐x as a function of Fe doping content.

SEM images (Figure 1d) reveals that all samples exhibited uniform spherical morphologies, with a gradual decrease in average particle size as the Fe content increased. This trend can be attributed to an increased number of nucleation sites induced by heteroatom incorporation [33]. However, the NFS‐20 sample showed an increase in particle size and partial agglomeration, likely due to lattice instability and secondary aggregation caused by excessive doping [34]. EDS mapping confirmed the homogeneous spatial distribution of Ni, Fe, and S throughout the particles. BET surface area measurements yielded values of 82.68, 92.68, 99.45, and 95.97 m² g⁻¹ for NFS‐0, NFS‐10, NFS‐15, and NFS‐20, respectively, consistent with the particle size evolution observed in the SEM analysis (Figure S2).

To further examine Fe incorporation, representative HRTEM and line‐scan EDS analyses were performed for NFS‐0 and NFS‐15. HRTEM images show clear lattice fringes with interplanar spacings of ∼0.290 nm for NFS‐0 and ∼0.288 nm for NFS‐15, which are assignable to the (200) plane of cubic NiS₂ (Figure S3). The slightly reduced lattice spacing in NFS‐15 is consistent with the XRD peak shift toward higher angles, supporting lattice contraction upon Fe incorporation. In addition, line‐scan EDS analysis of NFS‐15 shows that Fe is distributed throughout the particle rather than localized as a separate Fe‐rich phase (Figure S4). These results further support the successful incorporation of Fe into the NiS₂ lattice.

2.2. Electronic Modulation by Fe Doping

XPS was performed to investigate Fe‐doping‐induced modulation of the electronic structure in the NFS catalysts. In the Fe 2p region, no Fe signal was detected for NFS‐0, with the Ni LMM Auger peak observed within the same binding‐energy region [35]. After Fe incorporation, distinct Fe 2p_3/2 and Fe 2p_1/2 peaks appeared at 707.4 and 720.1 eV, respectively, confirming the formation of Fe─S bonds (Figure 2a) [36]. In the Ni 2p spectra, the Ni 2p_3/2 and Ni 2p_1/2 peaks corresponding to Ni─S bonds were observed at 853.5 and 871.0 eV for NFS‐0 (Figure 2b) [37]. Additional peaks at 855.3 and 873.9 eV were assigned to Ni─O species, consistent with partial surface oxidation upon air exposure [38]. For the Fe‐doped samples, both Ni─S peaks shifted toward higher binding energies, with NFS‐15 exhibiting the largest positive shift of +0.8 eV, indicating a decrease in the effective electron density around the Ni centers [39]. A similar trend was observed in the S 2p spectra, where the S 2p_3/2 and S 2p_1/2 peaks shifted from 162.5 and 163.7 eV in NFS‐0 to 162.8 and 164.0 eV in NFS‐15, respectively, indicating a positive shift of 0.3 eV (Figure 2c) [40]. These observations reflect changes in the metal–sulfur bonding environment and the local electronic configuration of the sulfur ligands [38].

FIGURE 2.

XPS spectra of the synthesized NFS‐x: (a) Fe 2p, (b) Ni 2p, and (c) S 2p regions.

The binding‐energy shifts suggest that Fe²⁺ (d⁶) acts as a relatively electron‐deficient dopant compared to Ni²⁺ (d⁸), inducing local charge redistribution within the lattice and leading to an increase in the average oxidation state of Ni from Ni^{2 +} to Ni^(2+δ). From an electronic‐structure perspective, octahedrally coordinated Ni²⁺ (Oh, d⁸) corresponds to a $t_{2g}^6{e}_g^2$ configuration, whereas a higher Ni oxidation state (e.g., Ni³⁺ (Oh, d⁷)) corresponds to a $t_{2g}^6{e}_g^1$. Thus, the XPS derived increase in Ni oxidation state suggests that the e_g occupancy approached unity (Figure S5) [23], which has been widely correlated with optimized adsorption‐desorption energetics of OER intermediates (^*OH, ^*O, and ^*OOH) and enhanced OER activity, supporting effective electronic‐structure tuning by Fe doping [41, 42].

2.3. OER Activity of NFS‐x

The OER performance of the synthesized samples was evaluated in a three‐electrode cell using 1 m KOH as the electrolyte. As shown in the LSV polarization curves (Figure 3a), NFS‐15 exhibited the lowest overpotential and thus the best OER activity; at 10 mA cm⁻², the overpotentials of NFS‐0, NFS‐10, NFS‐15, and NFS‐20 were 354, 294, 286, and 290 mV, respectively. This enhanced performance is consistent with the preceding XPS analysis, which suggests that Fe doping modulates the average oxidation state of Ni and e_g occupancy, thereby optimizing the adsorption‐desorption strength toward OER intermediates [42]. Tafel slope analysis (Figure 3b) further supports this trend: the slopes were 203, 127, 87, and 99 mV dec⁻¹ for NFS‐0, NFS‐10, NFS‐15, and NFS‐20, respectively, confirming the most favorable OER kinetics for NFS‐15. In addition, the Nyquist plots from the EIS measurements (Figure 3c) shows that NFS‐15 exhibits the smallest semicircle, corresponding to the lowest charge‐transfer resistance, indicating enhanced interfacial charge‐transfer kinetics upon Fe doping. ECSA was estimated from the C_dl values obtained from CV measurements in the non‐Faradaic region at various scan rates (Figure S6). NFS‐15 exhibited the highest C_dl value of 0.2112 mF cm⁻², indicating that it has the largest ECSA among the synthesized samples (Figure 3d). Even after normalizing the LSV curves by ECSA to compare intrinsic catalytic activities, NFS‐15 still demonstrated the highest intrinsic OER activity (Figure 3e). As shown in Figure 3g, a similar activity trend was observed for the catalysts loaded on NF, confirming that NFS‐15 remains the most active across different electrode substrates. Durability of the most active NFS‐15/NF electrode was evaluated by a CP test at 50 mA cm⁻² for 50 h (Figure 3h). After CP test, the overpotential at 50 mA cm⁻² increased from 283 to 359 mV, indicating pronounced degradation (Figure 3i). This result suggests that although the catalyst exhibits promising activity, further strategies are needed to improve its long‐term durability [24]. The observed degradation is likely associated with surface reconstruction during OER under alkaline conditions. It is well established that conductive compounds such as sulfides, selenides, phosphides, and nitrides undergo surface reconstruction to form (oxy)hydroxide phases (e.g., M─OOH) during OER, which serve as active sites that facilitate reaction intermediates and accelerate the overall reaction kinetics [43]. However, these dynamic surface transformations can simultaneously lead to cation leaching, catalyst delamination, and membrane degradation within the cell, thereby compromising durability. Therefore, developing strategies to control or mitigate surface reconstruction during operation is essential to achieve both high activity and long‐term stability.

FIGURE 3.

(a) iR‐corrected LSV curves for OER. (b) Tafel plots derived from the LSV curves. (c) EIS Nyquist plots. (d) ECSA calculated from the C_dl. (e) LSV curves normalized by ECSA. (f) Comparison of overpotentials and R_ct values. (g) LSV curves of NFS‐x/NF electrodes. (h) CP curve of the NFS‐15/NF electrode at 50 mA cm⁻². (i) LSV curves of NFS‐15/NF before and after the stability test.

2.4. Pre‐Reconstruction by Alkaline Pretreatment

The AT‐induced surface reconstruction of NFS‐15 was first probed by XRD and Raman analyses. The XRD pattern of NFS‐15‐AT exhibited a notable decrease in the relative intensity of the NiS₂ diffraction peaks, accompanied by pronounced peak broadening, leaving only weak reflections at 31.6°, 39.1°, and 53.9° (Figure S9). Concurrently, a broad feature attributable to NiFe layered double hydroxide (LDH) emerged, indicating the formation of an oxyhydroxide‐related surface phase after AT [44]. Raman spectra (Figure S10) further supported this structural evolution, showing significant broadening of the Ni─S vibrational modes and the appearance of a new band at ∼550 cm⁻¹, which is assignable to a Ni‐OOH vibration mode [45]. These results indicate that AT promotes partial reconstruction of the sulfide surface into a NiFe (oxy)hydroxide‐related phase while retaining residual NiS₂ crystallinity. Accordingly, the AT duration was optimized to satisfy two criteria: (i) the appearance of NiFe LDH‐related features in XRD while preserving partial NiS₂ crystallinity without complete phase disappearance, and (ii) the retention of the initial OER activity in a three‐electrode configuration without significant degradation (Figure S8).

TEM and HRTEM analyses further revealed the morphological evolution of NFS‐15 after AT. TEM images showed that the overall spherical particle morphology was preserved after AT, indicating that the treatment did not induce severe particle collapse or bulk morphological degradation (Figure 4a). However, HRTEM images of the surface regions revealed a distinct difference in near‐surface crystallinity between NFS‐15 and NFS‐15‐AT (Figure S11). Pristine NFS‐15 retained relatively clear lattice fringes up to the outer region, whereas NFS‐15‐AT exhibited markedly weakened lattice visibility near the surface. This contrast indicates that the outer region of NFS‐15 was transformed into a predominantly amorphous or poorly crystalline reconstructed layer after AT, while the overall spherical particle framework was maintained.

FIGURE 4.

(a) TEM images of NFS‐15 and NFS‐15‐AT particles. (b) Surface EDS elemental mappings of NFS‐15 and NFS‐15‐AT. XPS spectra of (c) Ni 2p and (d) S 2p for NFS‐15 and NFS‐15‐AT.

The elemental redistribution associated with the reconstructed surface was further examined using EDS mapping, HAADF‐STEM/EDS line‐scan analysis, and XPS. EDS mapping confirmed that Ni, Fe, and S were uniformly distributed throughout the pristine NFS‐15 particles, with only a minor O signal (Figure 4b). In contrast, NFS‐15‐AT exhibited clear elemental redistribution near the particle surface: the S signal gradually weakened toward the surface and became markedly attenuated in the outer region, whereas the O signal increased significantly and was detected together with Ni and Fe. This result indicates that the reconstructed outer region has a Ni─Fe─O─rich composition distinct from the original sulfide phase. HAADF‐STEM/EDS line‐scan profiles further resolved this surface‐to‐core compositional gradient. In the whole‐particle line scan of NFS‐15‐AT, the S signal was relatively weak near the particle edge and became stronger toward the interior (Figure S13). More directly, the representative near‐surface line‐scan profile showed an O‐rich and S‐depleted reconstructed region extending approximately 30–40 nm from the particle surface in the analyzed region, followed by a pronounced increase in the S signal toward the inner region (Figure S14). In contrast, untreated NFS‐15 did not show such a distinct O‐rich/S‐depleted outer region. XPS also supported this compositional evolution: the relative fraction of the Ni─O component in the Ni 2p_3/2 spectra increased from 0.31 for NFS‐15 to 0.50 for NFS‐15‐AT, and the oxidized sulfur component in the S 2p spectra increased from 0.30 to 0.53 after AT (Figure 4c,d and Table S2). These results indicate that AT induces pronounced surface oxidation without complete conversion of the sulfide phase, producing a Ni─Fe─O─rich reconstructed outer region while retaining a Ni─Fe─S─rich sulfide core underneath. The Fe 2p region was not quantitatively interpreted after AT because the Fe‐related signal strongly overlaps with the Ni LMM Auger feature, making reliable independent deconvolution of Fe chemical states difficult (Figure S15).

Collectively, these analyses show that AT converts the surface of NFS‐15 into a predominantly amorphous or poorly crystalline NiFe (oxy)hydroxide‐rich reconstructed layer while preserving the sulfide‐rich interior. The XRD/Raman results identify the emergence of oxyhydroxide‐related surface features, HRTEM confirms the formation of a poorly crystalline outer region, and EDS/XPS reveal an O─rich/Ni─Fe─O‐enriched surface with an S‐retaining/Ni─Fe─S─rich inner region. This surface‐confined reconstruction forms a NiFe (oxy)hydroxide‐rich surface layer on a retained sulfide core without complete conversion of the sulfide framework.

2.5. Durability under OER

To evaluate the long‐term stability of the NFS‐15‐AT/NF electrode, CP was conducted at a constant current density of 50 mA cm⁻² for 50 h (Figure 5a). Although the initial potential of the NFS‐15‐AT/NF electrode was slightly higher than that of the NFS‐15/NF electrode, the potential increase during the CP operation was more gradual, indicating a lower degradation rate. This trend was consistent with the post‐CP LSV results at 50 mA cm⁻². The overpotential of NFS‐15‐AT/NF increased from 304 to 330 mV, indicating a reduced increase in overpotential compared with NFS‐15/NF (Figure 5b). These results indicate that the NiFe (oxy)hydroxide layer formed via AT mitigates degradation under simplified three‐electrode OER conditions, although it does not completely eliminate it.

FIGURE 5.

(a) CP curves of NFS‐15/NF and NFS‐15‐AT/NF electrodes at 50 mA cm⁻² for 50 h. (b) LSV curves before and after CP test. (c) Comparison of OER overpotentials before and after durability testing. (d) Amounts of detached Ni and Fe quantified by ICP‐OES from the solids collected after CP operation. (e) Wettability measurements and in situ gas bubble evolution images of the electrodes.

To quantitatively assess catalyst detachment during durability testing, the electrolyte was collected after CP and filtered to recover detached catalyst particles, which were subsequently analyzed by ICP‐OES [46]. Based on three independent measurements, the total amount of detached Ni and Fe decreased from 4.78 ± 0.04 µg cm⁻² h⁻¹ for NFS‐15/NF to 4.16 ± 0.09 µg cm⁻² h⁻¹ for NFS‐15‐AT/NF (Figure 5d), indicating suppressed catalyst detachment after AT. In addition, a full‐cell‐like NFS‐15‐AT/NF electrode prepared using the same spray‐coating procedure and catalyst loading as the single‐cell anode showed similar initial OER activity but improved CP stability (Figure S16), suggesting that the larger apparent degradation observed in the simplified NF half‐cell test is strongly influenced by electrode‐level catalyst detachment associated with electrode fabrication and catalyst loading.

Enhanced electrode hydrophilicity can facilitate gas bubble removal during OER, accelerate electrolyte replenishment, and alleviate bubble accumulation and active site blockage [47]. To verify this effect, the surface wettability and bubble evolution were examined. Contact‐angle measurements revealed that NFS‐15/NF exhibited hydrophobic behavior with a contact angle of 128.3 ± 1.2°, whereas NFS‐15‐AT/NF showed superhydrophilic behavior, with water droplets spreading instantly. Direct visualization during OER indicated that large bubbles accumulated on NFS‐15/NF, whereas much smaller bubbles rapidly detached from NFS‐15‐AT/NF (Figure 5e). This surface hydrophilization and rapid bubble release likely contributed to the enhanced stability of the NFS‐15‐AT/NF electrode.

2.6. AEMWE Performance and Operating Modes

To assess the applicability of NFS‐15‐AT in AEMWE, an anode with an active area of 4.8 cm² was fabricated and evaluated in a single‐cell configuration (Figure 6a). To identify optimal operating conditions, cell performance was compared under both “both‐wet” and “dry‐cathode” conditions. The dry‐cathode mode is commonly adopted in industrial systems because it simplifies balance‐of‐plant operations and facilitates differential‐pressure operation [48]. The two modes yielded similar performance between 1 and 2 A cm⁻²; however, a pronounced difference emerged at current densities above 2 A cm⁻² (Figure 6b). To investigate the origin of this discrepancy, the HFR was measured using GEIS. Under the both‐wet mode, the HFR increased sharply above 2 A cm⁻², whereas under the dry‐cathode mode, HFR remained stable even at high current densities. Because HFR is dominated by ohmic resistance, improvements in membrane hydration and ion‐conduction pathways can effectively reduce HFR in zero‐gap configurations [49]. In dry‐cathode mode, water molecules generated at the anode via OER effectively back‐diffuse across the AEM toward the cathode driven by a steep water‐concentration gradient, thereby enhancing membrane hydration and ensuring stable ion‐conduction pathways. In contrast, the smaller concentration gradient in the both‐wet mode limits further membrane hydration, and excessive electrolyte supply may hinder gas removal and increase interfacial contact resistance. Furthermore, the transport of OER‐generated water from the anode to the cathode may mitigate local pH dilution at the anode, thereby positively influencing OER kinetics [50].

FIGURE 6.

(a) Schematic diagram of the AEMWE single‐cell configuration. (b) Cell polarization curves and HFR values obtained by GEIS for NFS‐15‐AT under dry‐cathode and both‐wet operating conditions. (c) Comparison of cell polarization curves using NFS‐15‐AT and commercial IrO₂ anodes. (d) Long‐term durability test performed at 1 A cm⁻² for 200 h using CP. (e) Comparison of cell voltages at 1 A cm⁻² for NFS‐15‐AT and previously reported anode catalysts.

The AEMWE single cell using the NFS‐15‐AT/NF anode required only 1.65 V to achieve 1 A cm⁻², outperforming commercial IrO₂ under identical electrode‐fabrication conditions and exhibiting competitive activity relative to other state‐of‐the‐art non‐precious metal catalysts reported in the literature (Figure 6c,e and Table S4). Remarkably, the catalyst maintained stable performance during 200 h of continuous operation at 1 A cm⁻², and even after the durability test, the cell required only 1.68 V to reach 1 A cm⁻², indicating minimal degradation (Figure 6d and Figure S21). The polarization behavior before and after the 200 h test was largely preserved, suggesting limited permanent degradation under practical single‐cell operating conditions. Although the post‐mortem XRD pattern of the used electrode was dominated by the strong diffraction peaks of the Ni foam substrate, limiting direct identification of subtle catalyst‐phase changes (Figure S22), post‐test SEM images and EDS elemental maps confirmed that the catalyst layer remained well retained on the NF substrate after the 200 h single‐cell test (Figure S23). In addition, HAADF‐STEM/EDS line‐scan analysis of catalyst particles recovered from the used electrode showed that residual S was still detected in the particle interior, whereas the near‐surface region remained enriched in O (Figure S24). This post‐operation elemental distribution is consistent with the surface/core structure identified before operation, indicating that the sulfide‐rich core is partially retained while the outer region undergoes further oxidation/reconstruction during long‐term AEMWE operation. These results highlight the excellent activity and durability of the sulfide‐based catalyst and demonstrate its practical potential as an anode material for industrial hydrogen production.

2.7. Computational Insights into Fe‐Induced Electronic Modulation of the NiS₂ Phase

Because the catalytically relevant surface under OER conditions is the reconstructed NiFe (oxy)hydroxide layer rather than the pristine sulfide itself, the DFT calculations were used here to examine how Fe substitution modifies the electronic structure of the parent NiS₂ precursor, rather than to directly model the operando active interface. As confirmed by the electrochemical analyses, Fe doping significantly enhanced the OER performance of NiS₂. Accordingly, DFT calculations were performed using a pyrite‐type NiS₂ supercell, where ∼15% of the Ni sites were substituted with Fe, to evaluate Fe‐induced electronic modulation in the precursor phase. Among the four Ni₁₃Fe₃S₃₂ configurations constructed (Figure S26), the model with the lowest total energy was selected as the representative structure for NFS‐15. Charge‐density difference analysis (Figure 7a), where blue and yellow regions denote charge depletion and accumulation, respectively, revealed pronounced electron accumulation around Fe sites accompanied by electron depletion at adjacent Ni sites. This Fe‐induced charge redistribution is consistent with the increased Ni oxidation tendency inferred from XPS. Concomitantly, Fe substitution modified the local bonding environment: the newly formed Fe─S bond length was ∼2.250 Å, while the neighboring Ni─S bond length slightly decreased from 2.302 to 2.286 Å (Figure 7b). These changes suggest strengthened metal‐sulfur interactions and local structural rearrangement around the Fe‐substituted site.

FIGURE 7.

(a) Charge density difference map of NFS‐15. (b) Bond lengths of NFS‐0 and NFS‐15 models. (c) DOS analysis for NFS‐0 and NFS‐15. (d) Schematic illustration of the relationship between the d‐band center and intermediate adsorption strength. (e) Free‐energy diagrams of NFS‐0 and NFS‐15 for the OER pathway.

The density of states (DOS) calculations (Figure 7c) further supported the experimentally observed electronic modulation. Fe 3d states appear near the Fermi level (E_F), reducing the band gap of NFS‐15 to 0.34 eV, which is 0.44 eV smaller than that of NFS‐0. This reduction implies more facile electronic transport in NFS‐15, which can be beneficial for electrochemical activity [51]. In addition, the d‐band center (ε_d) shifted upward from −2.27 eV in NFS‐0 to −2.15 eV in NFS‐15. The upward shift of ε_d toward E_F raises the antibonding‐state energy and reduces antibonding‐state occupancy (Figure 7d), thereby strengthening the adsorption of OER intermediates [52, 53]. To further compare the effect of Fe substitution on the precursor phase, simplified adsorption models of OER intermediates are provided in Figure S27, and the corresponding comparative free‐energy diagrams are shown in Figure 7e. Although these models do not represent the fully reconstructed NiFe (oxy)hydroxide active interface under operando OER conditions, they provide a simplified comparison of how Fe substitution alters the relative adsorption energetics of the parent NiS₂ phase. From the free‐energy diagrams, the rate‐determining step for NFS‐0 is the ^*O → ^*OOH transition with a maximum free‐energy barrier (ΔG) of 2.31 eV. In contrast, NFS‐15 exhibited a markedly reduced ΔG of 1.72 eV for the same step. This decrease in ΔG supports that Fe substitution renders the parent NiS₂ precursor electronically more favorable prior to reconstruction, which is consistent with the superior OER performance observed for NFS‐15.

3. Conclusion

In this study, a Fe‐substituted NiS₂ OER catalyst was developed by coupling sulfide‐phase electronic modulation with controlled alkaline pre‐reconstruction for durable alkaline OER and AEMWE operation. Fe incorporation modulated the Ni‐centered electronic structure of NiS₂, lowering the free‐energy barrier of the rate‐determining ^*O → ^*OOH step by 0.59 eV and enabling the optimized NFS‐15 catalyst to achieve an overpotential of 286 mV at 10 mA cm⁻². Alkaline pretreatment transformed the NFS‐15 surface into a largely amorphous or poorly crystalline NiFe (oxy)hydroxide‐rich reconstructed outer layer while retaining a sulfide‐rich core, thereby providing an OER‐active interface and preserving the Fe‐modulated sulfide core as a conductive framework. This surface‐confined pre‐reconstruction mitigated excessive reconstruction and catalyst detachment under alkaline OER conditions. As a result, NFS‐15‐AT/NF achieved 1 A cm⁻² at only 1.65 V in an AEMWE single cell, outperformed commercial IrO₂ under identical conditions, and maintained stable operation for over 200 h at 1 A cm⁻² with minimal degradation. These results demonstrate that integrating Fe‐induced electronic modulation of the sulfide framework with surface‐confined pre‐reconstruction is an effective strategy for designing durable sulfide‐derived OER catalysts and practical AEMWE anodes.

4. Experimental Section

4.1. Materials

Nickel(II) nitrate hexahydrate (Ni(NO₃)₂∙6H₂O) was purchased from Sigma–Aldrich, and iron(III) nitrate nonahydrate (Fe(NO₃)₃∙9H₂O) was obtained from Alfa Aesar. Isopropanol (99.5%) and 5 m potassium hydroxide (KOH) solution were supplied by SAMCHUN, and absolute ethanol (C₂H₆O, 99.9%) was purchased from Fisher Scientific Co. A glycerol solution (83.5–89.5%) and sulfur powder (99.98%) were obtained from Sigma–Aldrich. All chemicals were used as received without further purification.

4.2. Catalyst Synthesis

4.2.1. Synthesis of NiFe Glycerate Precursors

Spherical Fe‐doped Ni‐glycerate precursors were synthesized as follows. Ni(NO₃)₂∙6H₂O and Fe(NO₃)₃∙9H₂O were dissolved in 45 mL of isopropanol with stirring for 1 h to ensure complete dissolution. The total molar amount of metal ions (Ni+Fe) was fixed at 0.7 mmol, and the Fe content was adjusted to 0, 10, 15, and 20 mol% (Fe/(Ni+Fe)×100). After dissolution, 9 mL of glycerol was added, and the mixture was stirred for an additional 30 min. The homogeneous solution was transferred into a Teflon‐lined stainless‐steel autoclave and heated at 180°C for 3 h in a box furnace. The resulting product was repeatedly washed with ethanol and separated by centrifugation at 12 000 rpm and dried overnight at 50°C in an oven to obtain the NiFe glycerate precursors.

4.2.2. Synthesis of NFS Catalysts

The dried NiFe glycerate precursor and sulfur powder were thoroughly mixed at a molar ratio of (Ni+Fe):S = 1:8. The mixture was placed in an alumina boat and heated to 300°C at 10°C min⁻¹ under an N₂ atmosphere and held at 300°C for 30 min to complete the sulfidation. The product was then ground to obtain the NFS catalysts. The resulting samples were designated as NFS‐0, NFS‐10, NFS‐15, and NFS‐20, where the number denotes the Fe content in the precursor.

4.2.3. Synthesis of NFS‐15‐AT Catalyst

A total of 100 mg of NFS‐15 powder was ultrasonically dispersed in 200 mL of 5 m KOH solution for 10 min and then stirred at room temperature for 1 h. The precipitate was collected using a 0.45 µm membrane filter (Hyundai Micro), washed several times with deionized (DI) water and ethanol, and dried overnight at 50°C to yield the NFS‐15‐AT catalyst.

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File: smll73968‐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.