The synthesis of CuO-MoS₂/MoO₃ samples followed a three-step protocol. Initially, commercial MoS₂ (2 g, Aladdin) and Cu powders (1 g, Sigma-Aldrich) were soaked in liquid nitrogen for 8 h. Subsequently, the cryo-treated powders were dispersed in a 200 mL mixed solvent composed of isopropyl alcohol (Sinopharm Chemical Reagent Co., Ltd.) and deionized water (v/v ratio of 1:1) via ultrasound liquid phase exfoliation conducted at 180 W for 4 h. The resulting exfoliated solution was then centrifuged at 8,000 rpm for 30 min to collect the composite of Cu nanodots and MoS₂ nanosheets. Finally, the as-obtained Cu-MoS₂ powder was transformed into the CuO-MoS₂/MoO₃ sample through annealing in air at 500 °C for 2 h.

Morphological features of the catalysts were examined by transmission electron microscopy (TEM, JEM-2100F) and spherical aberration correction scanning transmission electron microscopy (AC-STEM, JEM-ARM200F). X-ray diffraction (XRD) patterns were acquired on a PANalytical X-Pert PRO MPD diffractometer employing a Cu Kα source operating at 40 kV and 40 mA. X-ray photoelectron spectroscopy (XPS) measurements were performed with an ESCALAB 250Xi spectrometer. Raman spectra were recorded using a LabRAM HR Evolution spectroscopy. In situ FTIR spectroscopy was carried out on a Nicolet iS50 spectrometer coupled with an H-type electrochemical cell (Gaossunion Technology Co., Ltd.) to monitor the reaction intermediates.

All electrochemical tests were conducted on a CHI760E workstation in 1.0 M KOH alkaline electrolyte, using a standard H-type three-electrode cell with a saturated calomel electrode (SCE) as the reference electrode and a carbon rod as the counter electrode. The as-prepared CuO-MoS₂/MoO₃ sample deposited on the carbon cloth served as the working electrode. Polarization curves for both HER and OER were obtained at a scan rate of 5 mV s^-1 with 90% iR-compensation. All measured potentials were converted to the reversible hydrogen electrode (RHE) scale according to the equation: E_RHE = E_SCE + 0.059 pH + 0.241 V^[19]. Electrochemical impedance spectroscopy (EIS) was recorded over a frequency range of 0.01 to 10⁵ Hz, with applied potentials of -0.3 V vs. RHE for HER and 1.2 V vs. RHE for OER. The electrochemically active surface area (ECSA) was estimated from the double-layer capacitance (C_dl), derived from cyclic voltammetry (CV) curves at scan rates varying from 10 to 200 mV s^-1. The non-Faradic potential windows of 0-0.1 V vs. RHE and 1.0-1.1 V vs. RHE were selected for HER and OER, respectively. The electrocatalytic overall water splitting was evaluated on a membrane electrode assembly (MEA) configuration, employing a reconstructed Cu-MoS₂ cathode and a CuO-MoS₂/MoO₃ anode separated by an anion-exchange membrane (AEM).

All DFT calculations were performed using the CP2K package with the Quickstep module^[36,37]. The input file was generated using Multiwfn 3.8^[38,39]. Periodic boundary conditions were imposed along the x and y directions, with a vacuum layer of 15-20 Å introduced along the z-axis to avoid periodic interactions. The Perdew-Burke-Ernzerhof (PBE) functional incorporating Grimme’s DFT-D3 (BJ) dispersion correction was adopted within the generalized gradient approximation (GGA) to treat exchange-correlation and van der Waals forces^[40,41]. Core electrons were described by norm-conserving Goedecker-Teter-Hutter (GTH) pseudopotentials, while valence electrons were expanded with the DZVP-MOLOPT-SR-GTH basis set^[42]. A plane-wave cutoff of 400 Ry and a relative cutoff of 55 Ry were set. The DFT + U method was applied to Mo atoms to better capture the localized d-electron behavior. Geometry optimization was performed using the Broyden-Fletcher-Goldfarb-Shanno (BFGS) algorithm, with convergence criteria of 4.5 × 10^-4 Ha/Bohr for maximum force and 3.0 × 10^-3 Bohr for maximum displacement. During optimization, the bottom two atomic layers were kept fixed to simulate surface constraints. SCF convergence was accelerated by the direct inversion in the iterative subspace (DIIS) minimizer and a full preconditioner, with an energy tolerance of 1.0 × 10^-5.

The CuO-MoS₂/MoO₃ nanosheets were obtained via controlled oxidation of the exfoliated Cu-MoS₂ precursor. XRD analysis [Figure 1A] verified the coexistence of diffraction peaks ascribed to MoO₃, MoS₂, and CuO phases, confirming the successful formation of a ternary composite. This assignment was further substantiated by Raman spectroscopy [Figure 1B], which distinctly captured the characteristic vibrational signatures of all three constituents, offering complementary structural confirmation. Transmission electron microscopy [Figure 1C] disclosed typical nanosheet morphologies decorated with CuO nanodots that were evenly spread over the MoS₂/MoO₃ support. This homogeneous distribution was more clearly visualized in spherical aberration-corrected STEM images [Figure 1D], where the embedded CuO nanodots appeared as bright spots exhibiting strong contrast against the underlying substrate. High-resolution TEM analysis [Figure 1E] resolved clear lattice fringes with measured spacings of 0.270, 0.232, and 0.227 nm, indexed respectively to the (101) plane of MoO₃, the (111) plane of CuO, and the (103) plane of MoS₂. The presence of these three distinct spacings within the same region attests to the intimate interfacial contact among the different phases. Additionally, elemental mapping [Figure 1F] confirmed the uniform spatial distribution of O, Cu, Mo, and S throughout the composite architecture, indicating that CuO nanodots were well incorporated into the MoS₂/MoO₃ matrix and that a thoroughly mixed ternary material had been formed.

XPS analysis was conducted to probe the surface chemical states of CuO-MoS₂/MoO₃ in parallel with MoS₂/MoO₃ for reference. The Cu 2p spectrum [Figure 1G] exhibited distinct peaks at 934.5 eV for Cu 2p_3/2 and 954.6 eV for Cu 2p_1/2, accompanied by pronounced shake-up satellite features, which are characteristic of Cu²⁺ oxidation state^[43]. The high-resolution Mo 3d spectrum of CuO-MoS₂/MoO₃ [Figure 1H] displayed two characteristic peaks at 233.3 and 236.4 eV, corresponding to Mo⁴⁺ and Mo⁶⁺, respectively^[19,44]. A notable positive shift in binding energy was observed for these Mo 3d peaks compared to those in MoS₂/MoO₃ (232.9 and 236.0 eV, Supplementary Figure 1), suggesting a modified electronic environment around Mo centers due to electron transfer toward CuO species. In the S 2p region [Figure 1I], the doublet located at 168.7 eV for S 2p_3/2 and 169.9 eV for S 2p_1/2 further confirmed the presence of SO_x species in the MoS₂/MoO₃ phase^[20]. Moreover, the O 1s spectrum [Figure 1J] was well-fitted with three components, including lattice oxygen in metal oxides (Mo) at 530.9 eV, surface hydroxyl groups (M-OH) at 532.2 eV, and adsorbed water molecules at 535.4 eV^[45]. Collectively, these results confirm the successful synthesis of CuO-MoS₂/MoO₃ precatalyst.

The as-prepared CuO-MoS₂/MoO₃ precatalyst was subjected to appreciable electrochemical reconstruction during electrolytic operation. To track this transformation and clarify the true active species under working conditions, we performed in situ electrochemical XRD to monitor phase evolution in real time for both HER and OER, given the composition dynamics induced by applied potentials. For the HER process [Figure 2A and B], the CuO-MoS₂/MoO₃ precatalyst initially showed coexisting phases of CuO, MoO₃, and MoS₂ at open circuit potential (OCP). Upon applying a cathodic potential of -0.1 V vs. RHE, the CuO and MoO₃ peaks gradually diminished, while metallic Cu phases emerged and intensified with increasing cathodic polarization. Quantitative analysis revealed that the CuO (111) peak at 2θ = 35.5° decreased by 92% and became undetectable at -0.3 V vs. RHE, confirming complete reduction to metallic Cu, with new peaks emerging at 2θ = 43.3°, 50.4°, and 74.1° corresponding to the (111), (200), and (220) planes of Cu (JCPDS No. 04-0836). Similarly, the MoO₃ (101) peak intensity diminished by 87% and became undetectable at -0.3 V vs. RHE. In contrast, the MoS₂ phase maintained its structural integrity throughout the HER process without detectable formation of other Mo-containing species. These findings point to the electrochemical reduction of CuO to metallic Cu coupled with preservation of MoS₂, suggesting that the reconstructed Cu-MoS₂ constitutes the genuine HER-active phase. Turning to the OER side [Figure 2C and D], the catalyst exhibited distinct reconstruction behavior. The CuO and MoS₂ phases remained stable across the potential range from OCP to 1.50 V vs. RHE, while the MoO₃ phase underwent partial reconstruction at 1.10-1.15 V vs. RHE, as evidenced by reduced diffraction peak intensity. Notably, the MoO₃ (101) peak shows a maximum intensity decrease of only 38% at 1.15 V vs. RHE, which subsequently recovers to 85% of its original intensity at 1.50 V vs. RHE, confirming partial and reversible reconstruction. The CuO (111) peak maintains 95% of its original intensity throughout the OER potential range, indicating structural stability under anodic conditions. These quantitative differences substantiate the complete reconstruction of CuO-MoS₂/MoO₃ into Cu-MoS₂ for HER, while maintaining an optimized CuO-MoS₂/MoO₃ structure for OER.

Post-reaction AC-STEM characterization [Figure 2E and F] provided deeper structural insights into the reconstructed catalysts. After HER testing, finely dispersed Cu clusters were distinctly observed on the MoS₂ (103) plane, corroborating the formation of Cu-MoS₂. Following OER testing, distinct lattice spacings of 0.227 and 0.270 nm were observed, corresponding to the (103) plane of MoS₂ and (101) plane of MoO₃, respectively, with CuO clusters embedded in the MoS₂/MoO₃ matrix. Notably, both Cu and CuO clusters in the reconstructed catalysts exhibited significantly reduced sizes compared to the original CuO nanodots, indicating that electrochemical reconstruction generates finer, highly dispersed active species with potentially enhanced catalytic properties. XPS measurements were conducted to analyze the surface chemical states of the reconstructed catalysts [Figure 2G-J]. The reconstructed CuO-MoS₂/MoO₃ after OER still showed characteristic peaks of Cu²⁺ species, whereas Cu-MoS₂ gave distinct signals attributable to Cu⁰ or Cu⁺ species, confirming the reduction of pristine CuO during HER operation. The appearance of Cu²⁺ species in the air-exposed Cu-MoS₂ sample was attributed to surface reoxidation. Similar surface oxidation phenomena were observed for Mo and S elements in Cu-MoS₂, manifested as Mo⁶⁺ and SO_x species^[46,47]. By contrast, the reconstructed CuO-MoS₂/MoO₃ catalyst showed enhanced peak intensities for Mo⁶⁺ and SO_x species, together with the emergence of metal-oxygen (M-O) bonds in the O 1s spectrum, collectively demonstrating the coexistence of MoS₂ and MoO₃ phases in this OER-active material^[47]. This comprehensive characterization provides compelling evidence that the electrochemical reconstruction process generates distinct, optimized active structures for HER and OER, with the transformed interfacial chemistry playing a crucial role in enhancing catalytic performance.

The HER activity of the reconstructed Cu-MoS₂ electrode was assessed in a 1.0 M KOH within a standard three-electrode cell. For comparison, pure Cu, MoS₂, and commercial 20% Pt/C electrodes were also measured under identical conditions. As revealed by the polarization curves in Figure 3A, the Cu-MoS₂ electrode exhibited a sharp increase in cathodic current with increasing overpotential, demonstrating significantly enhanced HER activity compared to the reference materials. Quantitative analysis of the polarization data [Figure 3B] shows that the Cu-MoS₂ electrode required remarkably low overpotentials of only 35, 54, 112, and 287 mV to deliver current densities of 10, 50, 100, and 500 mA cm^-2, respectively. These values are substantially lower than those required by pure Cu, MoS₂, and even the Pt/C benchmark, particularly at higher current densities relevant to industrial applications. Further investigation of the HER kinetics through Tafel slope analysis, which was obtained from the linear regions of the Tafel plots with 90% iR compensation. As shown in Figure 3C, the Cu-MoS₂ electrode displayed a favorable slope of 168.9 mV dec^-1, significantly smaller than those of Cu (252.7 mV dec^-1) and MoS₂ (191.6 mV dec^-1). These values, while higher than those typically reported in acidic media, were consistent with alkaline water splitting, where the Volmer step (water dissociation) introduced an additional kinetic barrier. For comparison, commercial Pt/C in 1.0 M KOH exhibited a Tafel slope of 170.6 mV dec^-1 under identical conditions, confirming that the measured values were reasonable for alkaline electrolytes. The HER Tafel slope suggested a Volmer-Heyrovsky mechanism with the electrochemical adsorption step being rate-limiting^[48].

EIS measurements were employed to probe the interfacial charge transfer kinetics. The Nyquist plots [Figure 3D] show that the Cu-MoS₂ electrode exhibited the smallest charge-transfer resistance (R_ct = 3.2 Ω) among all samples, signifying markedly accelerated electron transfer during the HER process. To evaluate the ECSA, C_dl was determined from CV scans in the non-Faradaic potential region [Supplementary Figure 2]. The Cu-MoS₂ electrode displayed the highest C_dl value of 5.73 mF cm^-2 [Figure 3E], confirming a substantial increase in the number of accessible active sites compared to the reference materials. To evaluate the intrinsic activity, we normalized the catalytic current by the ECSA using the measured C_dl values [Supplementary Figure 3]. At an overpotential of 100 mV, the ECSA-normalized HER current density for Cu-MoS₂ reached 0.611 mA cm^-2_ECSA, which is 3.94 and 3.99 times higher than those of MoS₂ (0.155 mA cm^-2_ECSA) and Cu (0.153 mA cm^-2_ECSA), respectively. The corresponding turnover frequency (TOF) at the same overpotential, calculated based on the ECSA-derived active site density (assuming 10¹⁵ sites cm^-2, i.e., 1.66 × 10^-9 mol cm^-2), was 1.91 s^-1 for Cu-MoS₂ [Supplementary Table 1], significantly exceeding that of MoS₂ (0.48 s^-1) and Cu (0.48 s^-1). The mass activity of Cu-MoS₂ reached 175.2 A g^-1 at 100 mV, substantially higher than MoS₂ (34.0 A g^-1) and Cu (23.8 A g^-1). These results demonstrate that the enhanced catalytic performance originates from improved intrinsic activity of the active sites, not merely from increased surface area. Moreover, the Cu-MoS₂ catalyst demonstrated outstanding operational stability, maintaining a constant current density of 500 mA cm^-2 for 100 h without apparent potential decay [Figure 3F]. This exceptional stability, combined with superior activity and favorable kinetics, positions the Cu-MoS₂ electrode as a highly promising catalyst for practical hydrogen production applications.

The OER performance of the CuO-MoS₂/MoO₃ electrode was rigorously evaluated in 1.0 M KOH electrolyte using a standard three-electrode system. As clearly depicted in the polarization curves [Figure 3G], the CuO-MoS₂/MoO₃ electrode demonstrated a substantially steeper increase in anodic current density with applied potential compared to all control electrodes, including CuO, MoS₂/MoO₃, and RuO₂. This pronounced activity enhancement is quantitatively reflected in the exceptionally low overpotentials needed to attain 10, 50, 100, and 500 mA cm^-2, which were measured at only 124, 143, 185, and 355 mV, respectively [Figure 3H]. These values represent significant improvements over the corresponding overpotentials of the reference catalysts, particularly at higher current densities relevant to industrial applications.

Further insight into the OER kinetics was obtained through Tafel analysis [Figure 3I]. The CuO-MoS₂/MoO₃ electrode exhibited a Tafel slope of 277.4 mV dec^-1, which is notably lower than those of MoS₂/MoO₃ (298.4 mV dec^-1), CuO (322.2 mV dec^-1), and RuO₂ (286.9 mV dec^-1). The OER Tafel slope likely reflected surface coverage effects or a chemical rate-determining step (RDS), yet the CuO-MoS₂/MoO₃ catalyst still exhibited the lowest Tafel slope among all tested samples, demonstrating superior reaction kinetics^[49]. EIS measurements provided additional evidence for enhanced reaction kinetics, showing the smallest charge-transfer resistance (R_ct = 4.0 Ω) for CuO-MoS₂/MoO₃ among all samples [Figure 3J], suggesting significantly facilitated charge transfer during the OER process^[50]. In addition, the CuO-MoS₂/MoO₃ electrode displayed the highest C_dl value of 8.17 mF cm^-2 [Figure 3K and Supplementary Figure 4], substantially larger than those of the counterpart materials, indicating a greater number of accessible active sites contributing to the improved OER activity^[51]. At an overpotential of 200 mV, the CuO-MoS₂/MoO₃ catalyst exhibited an ECSA-normalized current density of 0.612 mA cm^-2_ECSA with a TOF of 0.956 s^-1 and a mass activity of 250.0 A g^-1 [Supplementary Figure 5 and Supplementary Table 2]. At 500 mV, the corresponding values increased to 5.981 mA cm^-2_ECSA, 9.33 s^-1, and 2,442.8 A g^-1, respectively. The mass activity of CuO-MoS₂/MoO₃ at 500 mV is 1.74 times higher than that of MoS₂/MoO₃ (1,406.8 A g^-1) and 3.92 times higher than CuO (623.7 A g^-1). It is noteworthy that the promotional effect of Cu species exhibits distinct characteristics for HER and OER: for HER, Cu incorporation significantly enhances the intrinsic activity per active site, while for OER, CuO decoration primarily increases the number of accessible active sites. These distinct yet complementary mechanisms collectively demonstrate the versatility of Cu/CuO modification in tailoring catalytic properties for different half-reactions. These results unequivocally demonstrate that the enhanced catalytic performance originates from improved intrinsic activity of the active sites, not merely from increased surface area, confirming the critical role of Cu incorporation in promoting the electrocatalytic activity of the Mo-based catalyst system. Moreover, the catalyst demonstrated exceptional operational stability, maintaining nearly constant performance over 100 h of continuous operation at 500 mA cm^-2 with negligible activity loss [Figure 3L]. These results indicate that the incorporation of CuO nanodots into MoS₂/MoO₃ nanosheets markedly enhances both HER and OER performance. This improvement can be attributed to the formation of highly active Cu-Mo interfacial sites, which synergistically promote water dissociation, facilitate charge transfer kinetics, and increase the density of ECSAs, positioning this material as a promising alternative to noble metal-based catalysts for practical water splitting applications.

To probe deeper into the water splitting mechanism, we employed in situ FTIR spectroscopy to examine the behavior of interfacial water molecules on Cu-MoS₂ and CuO-MoS₂/MoO₃ surfaces. The in situ FTIR spectra [Figure 4A and B, and Supplementary Figure 6] revealed two characteristic vibrational modes of adsorbed water molecules at 1,640 and 2,800~3,700 cm^-1, corresponding to the H-O-H bending and O-H stretching band, respectively^[52,53]. Notably, both vibrational bands exhibited significantly enhanced intensity on Cu-MoS₂ during HER compared to MoS₂ [Supplementary Figure 7], demonstrating substantially improved water adsorption capability resulting from Cu incorporation. A parallel enhancement was observed for CuO-MoS₂/MoO₃ relative to MoS₂/MoO₃ under OER conditions [Supplementary Figure 8]. Time-dependent in situ FTIR measurements [Supplementary Figure 9] further captured the progressive accumulation of water-related signals during electrolysis, confirming continuous and efficient water activation on the catalyst surfaces throughout the electrochemical processes.

For a deeper understanding of the interfacial water structure, we performed detailed Gaussian fitting analysis of the O-H stretching region using a multi-peak fitting routine. The fitting was conducted with peak positions allowed to vary within ±10 cm^-1 of the literature values, while the full width at half maximum (FWHM) was constrained to 150-250 cm^-1 and a linear baseline correction was applied. All fits achieved R² values exceeding 0.99, confirming the high reliability of the deconvolution. This procedure resolved three distinct components centered at approximately 3,230, 3,420, and 3,550 cm^-1 [Figure 4C and D, Supplementary Figure 10], which were assigned to symmetric strongly hydrogen-bonded water, asymmetric hydrogen-bonded water, and weakly hydrogen-bonded (isolated) water, respectively^[54,55]. Peak area ratios were calculated from integrated areas after normalization to the total O-H stretching intensity. Quantitative analysis revealed a remarkable increase in the proportion of symmetric strongly hydrogen-bonded water on both Cu-MoS₂ and CuO-MoS₂/MoO₃ compared with their unmodified counterparts [Figure 4E and F]. This specific water configuration is of particular significance, as it facilitates efficient proton transfer via a well-structured hydrogen-bonding network, indicating that Cu-modified surfaces actively promote the organization of interfacial water into a more ordered and strongly bonded structure^[56].

Further potential-dependent measurements provided additional mechanistic clues. The O-H stretching frequency of symmetric hydrogen-bonded water displayed a pronounced potential-dependent shift characteristic of the Stark effect [Figure 4G]. The calculated Stark slopes for Cu-MoS₂ (44.5 cm^-1 V^-1) and CuO-MoS₂/MoO₃ (51.2 cm^-1 V^-1) were substantially higher than those for MoS₂ (11.8 cm^-1 V^-1) and MoS₂/MoO₃ (15.9 cm^-1 V^-1). These markedly higher Stark slopes imply that interfacial water molecules on Cu-modified catalysts possess greater flexibility and enhanced sensitivity to applied potential, resulting in more pronounced vibrational frequency shifts under electrochemical conditions^[57]. Taken together, these findings offer strong evidence that both Cu-MoS₂ and CuO-MoS₂/MoO₃ catalysts effectively promote the formation of strongly hydrogen-bonded water networks at the electrode-electrolyte interface. These structured water networks serve as efficient proton transfer pathways, thereby accelerating water dissociation and proton/hydroxide transport processes, which fundamentally enhance the kinetics of both HER and OER.

DFT calculations were systematically performed to obtain atomic-level insights into the catalytic mechanisms of both HER and OER. Well-optimized structural models were constructed to represent the catalytic systems, with MoS₂ and Cu-MoS₂ serving as HER catalysts, while MoS₂/MoO₃ and CuO-MoS₂/MoO₃ models were developed for OER investigations [Supplementary Figure 11]. Differential charge density analysis [Figure 5A] revealed substantial electron redistribution at the heterointerfaces, demonstrating distinct charge transfer patterns. Specifically, Cu-MoS₂ exhibited significant electron transfer from Cu clusters to the MoS₂ substrate with a total charge transfer of 0.36 e^-, creating an electron-rich MoS₂ surface. In contrast, CuO-MoS₂/MoO₃ showed reverse electron transfer from the MoS₂/MoO₃ substrate to CuO clusters, amounting to 0.15 e^-. This pronounced interfacial charge redistribution effectively modulates the d-band centers of the catalytic sites, consequently optimizing their adsorption properties toward reaction intermediates^[58]. The partial density of states (PDOS) analysis provided further evidence of electronic structure modulation^[59]. As illustrated in Figure 5B, the d-band center of Mo 3d orbitals exhibited a notable upward shift from -1.69 eV in MoS₂ to -1.02 eV in Cu-MoS₂, and similarly from -0.38 eV in MoS₂/MoO₃ to -0.15 eV in CuO-MoS₂/MoO₃. A parallel electronic modulation was observed for S 2p orbitals [Figure 5C], confirming that the incorporation of Cu/CuO clusters effectively tunes the electronic structure of both Mo and S sites, thereby enhancing the water adsorption capability and optimizing the binding strength of reaction intermediates.

The reaction energetics were quantitatively evaluated through Gibbs free energy calculations [Figure 5D and E], with detailed reaction pathways schematically illustrated in Figure 5F and G, Supplementary Figures 12 and 13. For HER on Cu-MoS₂, the RDS shifted from water dissociation on pristine MoS₂ (1.06 eV) to hydrogen desorption on Cu-MoS₂. For Cu clusters alone, the hydrogen desorption barrier was 0.71 eV [Supplementary Figure 14]; notably, when considering the Cu-Mo interfacial site as the true catalytic center, this barrier was further reduced to only 0.45 eV, confirming that the interface between Cu and MoS₂ serves as the most active site for HER. In the case of OER, the RDS energy barrier was dramatically decreased from 2.35 eV for the *O → *OOH transformation on MoS₂/MoO₃ to 1.85 eV for the *OH → *O conversion on CuO clusters. More importantly, at the CuO-MoS₂/MoO₃ interfacial site, the RDS barrier was further lowered to 1.68 eV, demonstrating that the interfacial sites, rather than individual CuO clusters, are the genuine catalytic centers responsible for the enhanced OER performance^[60,61]. These computational results clearly demonstrate that Cu incorporation, particularly at the Cu-Mo and CuO-MoS₂/MoO₃ interfaces, effectively optimizes the reaction pathways and reduces the energy barriers for both HER and OER. Integrating these computational findings with in situ FTIR spectroscopic results, we establish a comprehensive mechanism where both Cu-MoS₂ and CuO-MoS₂/MoO₃ catalysts synergistically optimize the adsorption/desorption behavior of key reaction intermediates while facilitating the formation of a strongly hydrogen-bonded water network at the electrode-electrolyte interface.

For a thorough assessment of the practical viability of our catalysts in industrial water electrolysis, we assembled an AEMWE featuring the reconstructed Cu-MoS₂ as the cathode and CuO-MoS₂/MoO₃ as the anode. As a benchmark, a reference cell was constructed using commercial Pt/C (20 wt% Pt) and RuO₂ as the cathode and anode, respectively, representing the state-of-the-art noble metal-based configuration. Schematic and photographic views of the AEMWE setup are presented in Figure 6A and B. Initial performance evaluation through polarization curves measured at 25 °C [Figure 6C] revealed the exceptional efficiency of our catalyst system. The Cu-MoS₂||CuO-MoS₂/MoO₃ electrolyzer required only 1.82 and 1.98 V to achieve current densities of 0.5 and 1.0 A cm^-2, respectively, significantly outperforming the Pt/C||RuO₂ benchmark system, which required 1.91 and 2.06 V for the same current densities. This performance advantage became even more pronounced when the operating temperature was elevated to 80 °C [Figure 6D], where the voltage required to maintain 1.0 A cm^-2 decreased substantially from 1.98 to 1.77 V, demonstrating the system's excellent temperature responsiveness and potential for high-efficiency operation under practical industrial conditions.

The long-term stability test was performed at 1.0 A cm^-2 and 80 °C using a closed-loop circulation system with 1.0 M KOH electrolyte circulated at 100 mL min^-1 through both compartments. The electrolyte reservoir volume was 500 mL and was not replenished during the 500-h test. Impressively, the Cu-MoS₂||CuO-MoS₂/MoO₃ electrolyzer exhibited remarkable durability throughout the 500-h testing period [Figure 6E], maintaining stable voltage output with a minimal degradation rate of merely 23.62 µV h^-1. This performance stands in stark contrast to the commercial Pt/C||RuO₂ benchmark system, which suffered rapid performance degradation with a voltage increase rate of 1.03 mV h^-1 within just 160 h of operation, representing a degradation rate approximately 44 times faster than our system. To further validate the practical applicability, we evaluated the hydrogen production power consumption and energy conversion efficiency without iR compensation. The Cu-MoS₂||CuO-MoS₂/MoO₃ electrolyzer achieved outstanding power consumption values of only 3.70 and 4.23 kWh Nm^-3 at 0.5 and 1.0 A cm^-2, respectively, corresponding to remarkable energy conversion efficiencies of 79.35% and 69.49%. These efficiency metrics, combined with the comprehensive performance comparison summarized in Figure 6F and Supplementary Table 3, demonstrate that our electrolyzer surpasses previously reported AEMWE systems. These outstanding results unequivocally highlight the strong potential of our catalyst system for practical implementation in industrial-level water electrolysis applications, positioning it as a viable and efficient alternative to conventional noble metal-based technologies.