The increasingly flourishing hydrogen energy economy promotes the rapid development of water splitting to fabricate high-purity hydrogen^[1,2]. Water splitting can be conducted by employing the electrical energy produced by solar and wind energy, which has the obvious advantages of cleanliness, no pollution and high sustainability^[3,4]. However, the recent progress of water splitting has been severely hindered by the high cost of electrocatalysts. Nowadays, the widely used hydrogen evolution reaction (HER) electrocatalysts are focused on Pt-based precious metal catalysts, bringing forth the high cost for water splitting^[5-9]. Developing low-cost electrocatalysts to replace Pt-based catalysts has been widely recognized goal, which has attracted considerable attention from researchers^[10-12].

Among the various kinds of investigated HER catalysts, phosphides, sulfides, carbides and nitrides have been studied by many groups owing to their excellent HER activity^[13-18]. However, these catalysts are generally synthesized through the high-temperature calcination or the hydrothermal treatment^[19-23]. The complex steps and huge energy consumption increase the production cost of HER catalysts. Therefore, searching for simpler and more available methods of catalyst preparation is urgent for water splitting. Corrosion engineering is a recently developed strategy for achieving electrocatalysts relying on the formed corrosion layers on metal substrates as electrochemical active layers^[24-28]. In general, metal corrosion can spontaneously occur in a high-humidity environment at room temperature. The corrosion layers are always composed of hydroxides, which are also applied as electrocatalytic materials^[29-31]. Based on this recognition, metal corrosion is applied for preparing highly active electrocatalysts with the aim of “turning damage into treasure”. Previous works on corrosion engineering mainly focus on oxygen evolution reaction (OER) catalysts, while the investigations of HER catalysts are still very few. The main problem of HER catalysts based on metal corrosion is the poor HER activity of the hydroxide corrosion layers due to the strong *H adsorption^[32]. Thus, employing an effective strategy to modulate the electronic structures of corrosion layers is necessary to enhance the HER performance^[33]. RuO₂ was thereby selected as the targeted modifier, based on its well-documented near-optimal ΔG_H* value close to 0 eV that can precisely compensate for the sluggish hydrogen desorption kinetics of the hydroxide layer^[34]. While some prior works have adopted corrosion engineering to fabricate RuO₂-modified hydroxide catalysts for boosted HER performance, the dynamic surface structural reconstruction process of the catalyst before and after RuO₂ modification has not been investigated.

Towards simplifying the synthesis procedures of HER catalysts, we proposed a facile corrosion-reconstruction strategy in this work to fabricate the RuO₂/NiFe(OH)_x HER electrocatalysts at room temperature. Actually, the cleaned NiFe foams (NFF) were immersed in the NiCl₂ solution for 9 h to obtain the NiFe hydroxide [NiFe(OH)_x]. In the subsequent step, the prepared catalyst was immersed in a 2 mM RuCl₃ solution for 2 h to achieve surface reconstruction, and the resulting sample was labeled as RuO₂/NiFe(OH)_x. The as-synthesized electrocatalysts display ultrahigh HER activity, achieving current densities of 50 mA cm^-2 and 100 mA cm^-2 recorded at the low overpotentials of merely 111 mV and 149 mV. Moreover, the assembled water-splitting device could stably run under the relevant large current density of 200 mA cm^-2 lasting for 100 h. Density functional theory (DFT) calculations reveal that the reconstructed surface enriched with RuO₂ nanoclusters optimally adsorbs water molecules and induces a shift in *H adsorption from oxygen to ruthenium atoms, facilitating the desorption of *H and promoting efficient H₂ formation. This work supplies a facile method of preparing efficient HER catalysts at room temperature, largely simplifying the synthesis of HER electrocatalysts. We believe the simple corrosion strategy can effectively boost the practical application of HER catalysts for water splitting.

A RuO₂/NiFe(OH)_x electrode was synthesized via a facile corrosion-reconstruction strategy for HER applications, as illustrated in Figure 1A. Initially, a cleaned NFF was immersed in a NiCl₂ solution to facilitate the formation of NiFe(OH)_x layers. Afterward, the already synthesized NiFe(OH)_x catalyst was introduced into a RuCl₃ solution, leading to the generation of RuO₂ nanoclusters that were strategically decorated on the NiFe(OH)_x surface. These decorated RuO₂ clusters serve to modulate the electronic configuration of NiFe(OH)_x, thereby enabling highly efficient hydrogen evolution performance.

A preliminary investigation into the corrosion behaviors of NiFe(OH)_x and NFF was conducted to explore the preparation process of electrodes. The optical images of the synthesized NFF, NiFe(OH)_x, and RuO₂/NiFe(OH)_x electrodes are displayed in the inset of Figure 1B and C. The distinct yellow color of NiFe(OH)_x and the black appearance of RuO₂/NiFe(OH)_x starkly contrast with the pristine silver color of NFF, signifying a corrosion-induced transformation in the composition and structure of NFF. By fitting the corrosion polarization curves depicted in Figure 1B and C, it is observed that the corrosion current density of NFF in a NiCl₂ solution (0.551 mA cm^-2) significantly exceeded that in water (0.045 mA cm^-2). Concurrently, the corrosion potential of NFF in the NiCl₂ solution (-0.635 V) was lower than that in water (-0.454 V). Analogously, NiFe(OH)_x delivered a higher corrosion current density (0.367 mA cm^-2) and a lower corrosion potential (-0.509 V) in a RuCl₃ solution compared to those in water (0.005 mA cm^-2 and -0.439 V, respectively) [Figure 1D]. These findings underscore the accelerated corrosion rates in NiCl₂ and RuCl₃ solutions, attributable to the potent corrosivity of Cl^- ions and the enhanced ionic conductivity of metal salt solutions. Furthermore, the corrosion current density of NiFe(OH)_x in water (0.005 mA cm^-2) was notably lower than that of NFF (0.045 mA cm^-2), which can be ascribed to the protective NiFe(OH)_x layers grown on the NFF surface, impeding further corrosion of the NFF substrate by the aqueous medium. Consequently, by modulating the corrosion behavior of metals, it becomes feasible to engineer surface structural reconstructions in catalysts, thereby optimizing their performance for HER applications.

SEM images were used to investigate the surface morphologies of catalysts. As displayed in Supplementary Figure 1, the pristine NFF displays a smooth surface without any grown materials. After the first corrosion in NiCl₂ solution, uniform flower-like nanosheet arrays were generated over the surface of NiFe(OH)_x. Then, the second corrosion in RuCl₃ solution did not change the nanosheet morphology of the RuO₂/NiFe(OH)_x electrode [Supplementary Figure 2]. Moreover, the TEM images were also used to reveal the structure of RuO₂/NiFe(OH)_x. It is notable that the corrosion layers of RuO₂/NiFe(OH)_x consist of many thin nanosheets [Figure 2A and B], which are beneficial for promoting the exposure of electrochemically active sites. The TEM image shown in Figure 2C indicates the lattice fringes of corrosion layers, in which the lattice fringes with a spacing of 0.626 nm are associated with the (020) plane of FeOOH (JCPDS 73-2326). It is also observed that there are some nanoclusters on the nanosheets, which contain lattice fringes with a spacing of 0.255 nm and 0.315 nm belonging to the (101) and (110) crystalline planes of RuO₂ (JCPDS 88-0323)^[36]. The elemental mapping images presented in Figure 2D-G reveal the uniform elemental dispersion of Ni, Fe, Ru, and O within the nanosheet arrays, confirming the successful loading of Ru species on the NiFe(OH)_x nanosheets. In addition, XRD patterns of the catalysts were recorded to analyze phase composition. As depicted in Supplementary Figure 3, no obvious diffraction peaks were observed except the peaks of NFF substrates (FeNi₃: PDF#38-0419)^[37], demonstrating the low crystallinity of corrosion layers. The above-mentioned results prove that the corrosion layers of the RuO₂/NiFe(OH)_x electrode are composed of the RuO₂ nanoclusters-decorated NiFe(OH)_x nanosheet arrays.

The surface compositions and valence states of these catalysts were investigated through the XPS spectra. The survey XPS spectra in Supplementary Figure 4 prove the co-presence of Fe, Ni and O elements for the NiFe(OH)_x and RuO₂/NiFe(OH)_x samples. It is noteworthy that the Ru elements are detected in RuO₂/NiFe(OH)_x, revealing the available loading of Ru on active layers. Ni 2p spectrum of RuO₂/NiFe(OH)_x in Figure 2H is divided into five peaks. The binding energies located at 855.64 eV and 872.96 eV are attributed to the Ni 2p3/2 and Ni 2p1/2 peaks as well as two satellite peaks centered at 861.58 eV and 879.45 eV, revealing the generation of Ni²⁺ species in the RuO₂/NiFe(OH)_x^[38,39]. Besides, the Ni⁰ peak is observed at 852.19 eV, which is caused by the NFF substrates. Figure 2I depicts the Fe 2p XPS spectrum of RuO₂/NiFe(OH)_x, in which the binding energies centered at 710.69 eV and 723.91 eV are assigned to the Fe²⁺ ions, and the binding energies of 712.89 eV and 726.15 eV belong to the Fe³⁺ ions^[33,40]. The Fe 2p XPS spectrum indicates the co-presence of Fe²⁺/Fe³⁺ species in RuO₂/NiFe(OH)_x. Three O species are detected for RuO₂/NiFe(OH)_x at 529.73, 531.19, and 532.18 eV in O 1s spectrum [Supplementary Figure 5], which correspond to the M-O, M-OH and H₂O bonds, respectively^[41,42]. For the Ru 3p spectrum [Figure 2J], the binding energies at 462.81 eV and 485.10 eV are ascribed to the Ru⁴⁺ ions, further confirming the RuO₂ clusters in the active layers of RuO₂/NiFe(OH)_x^[43]. Moreover, the binding energies of Fe 2p3/2, Ni 2p3/2 and O 1s for RuO₂/NiFe(OH)_x all shift negatively toward lower binding energy in contrast with those of NiFe(OH)_x. In general, this reveals the existence of electron transfer between the RuO₂ cluster and NiFe(OH)_x nanosheets, along with local electron rearrangement on the surface of NiFe(OH)_x induced by RuO₂ nanoclusters.

To clarify the formation mechanism of RuO₂/NiFe(OH)_x, we have provided the time-dependent color evolution and pH change of RuCl₃ solution, and Inductively coupled plasma optical emission spectrometry (ICP-OES) analysis of Ru, Ni, and Fe concentrations before and after immersion. During the immersion process, the initial orange color of RuCl₃ solution gradually turned blue [Supplementary Figure 6]. Usually, the Ru³⁺ solution appears yellow, while the Ru²⁺ solution appears blue. The above color change indicates that a redox reaction has occurred between Ru³⁺ ions and NiFe-substrate of the NiFe(OH)_x catalyst. At longer immersion times, the solution became darker and slightly turbid, which is attributed to the slight detachment of newly formed RuO₂/NiFe(OH)_x corrosion layers. Meanwhile, the pH value increased from 3.10 to 5.00 [Supplementary Table 2], resulting from the consumption of Ru³⁺ ions and partial dissolution of surface Ni/Fe-hydroxide species. ICP results in Supplementary Table 3 show that the Ni and Fe concentrations increase markedly from 0.0949 and 0.3047 mg L^-1 to 68.1262 and 210.2132 mg L^-1 after 2 h immersion, verifying the dissolution and reconstruction of NiFe(OH)_x in the acidic RuCl₃ solution. The Ru concentration decreases from 145.7364 mg L^-1 (1.44 mM) to 19.9406 mg L^-1, indicating the consumption and deposition of Ru species on the electrode surface. Considering the solution volume of 30 mL and the electrode area of 4 cm², the calculated Ru loading on RuO₂/NiFe(OH)ₓ was 0.94 mg cm^-2. Thus, it can be concluded that the redox reaction between Ru³⁺ ions and the NiFe substrate is the main reason for the corrosion-induced reconstruction of NiFe(OH)_x in RuCl₃ solution. The reduced Ru species are unstable and are further oxidized to form the RuO₂ phase, resulting in the construction of RuO₂/NiFe(OH)_x corrosion layers.

The electrochemical hydrogen evolution reaction (HER) performances of the NFF, NiFe(OH)_x, and RuO₂/NiFe(OH)_x electrocatalysts were meticulously evaluated in 1 M KOH electrolyte. As evident from the polarization curves presented in Figure 3A, both the pristine NFF and NiFe(OH)_x electrodes exhibit markedly poor HER activity. However, a notable enhancement in the HER activity of NiFe(OH)_x is observed subsequent to its surface reconstruction via corrosion in a RuCl₃ solution. Notably, the RuO₂/NiFe(OH)_x sample achieves remarkable HER performance, demanding a small overpotential of merely 111 mV to attain a current density of 50 mA cm^-2 [Figure 3B]. This HER performance outperforms that of both NFF (350 mV) and NiFe(OH)_x (276 mV) by a significant margin. We have supplemented the benchmark comparison with commercial Pt/C (20 wt%), as shown in Supplementary Figure 7. The RuO₂/NiFe(OH)_x shows a slightly lower current density than Pt/C at low overpotentials, but significantly outperforms it at high overpotentials, highlighting its prominent advantages. The exceptional HER activity demonstrated by RuO₂/NiFe(OH)_x is primarily ascribed to the electronic interplay between RuO₂ nanoclusters and NiFe(OH)_x, along with the augmented exposure of electrochemically active sites facilitated by the nanosheet arrays.

Besides, the corresponding Tafel slopes were calculated to evaluate the HER mechanism [Figure 3C]. It is found that the Tafel slope of RuO₂/NiFe(OH)_x is 33.70 mV dec^-1, much smaller than those of NFF (261.05 mV dec^-1) and NiFe(OH)_x (183.01 mV dec^-1), revealing a faster HER process. Furthermore, the Nyquist curves of NFF, NiFe(OH)_x and RuO₂/NiFe(OH)_x were also tested to study the charge-transfer resistance (R_ct) [Figure 3D]. It is clearly observed that RuO₂/NiFe(OH)_x displays the lowest R_ct value among these catalysts, indicating quick electron transfer in the electrochemically process. Generally, the C_dl values can be adopted to quantify the ECSA of electrocatalysts. The C_dl values of NFF, NiFe(OH)_x and RuO₂/NiFe(OH)_x were calculated from the CV curves in Supplementary Figure 8. It can be seen from Figure 3E that RuO₂/NiFe(OH)_x contains the highest C_dl of 227.66 mF cm^-2 among the three catalysts. We calculated the ECSA using the obtained C_dl values in Figure 3E according to the equation (ECSA = A_geo*C_dl/C_s, C_s = 40 μF cm⁻², A_geo refers to the effective reaction area). The calculated ECSA values for NFF, NiFe(OH)_x and RuO₂/NiFe(OH)_x are 303.63, 426.44 and 1,422.88 cm², respectively. We have also provided the ECSA-normalized HER polarization curves. As shown in Supplementary Figure 9, RuO₂/NiFe(OH)_x still exhibits superior HER activity after ECSA normalization compared with NiFe(OH)_x and NFF. The obtained result verifies that the improved HER performance is not solely caused by the enlarged surface area, but also originates from the enhanced intrinsic activity of RuO₂/NiFe(OH)_x.

Furthermore, the operating stability of RuO₂/NiFe(OH)_x was detected through the chronopotentiometry method. The RuO₂/NiFe(OH)_x displays slight changes in applied potentials after operation at 200 mA cm^-2 for 100 h [Figure 3F], demonstrating admirable durability. For catalysts fabricated via corrosion engineering, the coupling between corrosion layers and substrates has non-negligible weaknesses: partial exfoliation of corrosion layers occurs under high-current operation. Thus, moderately reducing the corrosion rate during preparation facilitates homogeneous and steady growth of corrosion layers, which can improve the interfacial adhesion stability of catalysts. Moreover, the HER activity of RuO₂/NiFe(OH)_x is compared with recently reported electrocatalysts, as shown in Figure 3G and Supplementary Table 4. It is noteworthy that the overpotential of RuO₂/NiFe(OH)_x at 100 mA cm^-2 is much lower than that of most of the reported catalysts. Thus, it can be inferred that the synthesized RuO₂/NiFe(OH)_x possesses excellent HER activity and durability. SEM images of RuO₂/NiFe(OH)_x after the durability determination reveal substantial reconstruction of the nanosheet architecture, evidenced by the formation of abundant nanoparticles across the surface [Supplementary Figure 10]. XPS spectra in Supplementary Figure 11 verify that distinct Ru species are clearly preserved on the catalyst following prolonged electrolysis. The robust electronic coupling between these residual Ru species and Ni/Fe active sites imparts the catalyst with prominent electrocatalytic activity and exceptional long-term operational stability. To verify the hydrogen generation efficiency, the measurement of Faradaic efficiency for RuO₂/NiFe(OH)_x was conducted using the water-displacement method [Supplementary Figure 12]. The measured amount of H₂ evolution was similar to the theoretical value calculated from Faraday’s law, corresponding to a Faradaic efficiency of 87.5%. This confirms that the RuO₂/NiFe(OH)_x catalyst enables efficient H₂ generation during HER.

To gain a deeper insight into the HER mechanism on RuO₂/NiFe(OH)_x surfaces, we performed a comprehensive mechanistic analysis employing DFT calculations. FeNiOOH was adopted as an idealized model to describe the local coordination environment of the experimentally prepared NiFe(OH)_x catalyst. This model aims to capture the essential interfacial electronic interactions with RuO₂ under HER operating conditions. The differential charge density depicted in Figure 4A, and Bader charge analysis in Supplementary Figures 13-15 and Supplementary Tables 5-7 provide compelling evidence of electron redistribution within FeNiOOH, induced by the incorporation of a RuO₂ nanocluster. These results reveal that RuO₂ donates electrons to NiFe(OH)_x, leading to electron enrichment of the NiFe(OH)_x side and local interfacial charge polarization. Figure 4B shows the calculated reaction free energy changes for the water dissociation-related process on FeNiOOH and RuO₂/FeNiOOH. The pristine FeNiOOH surface exhibits a strong interaction with water-derived intermediates, corresponding to a largely negative reaction free energy change of -0.7206 eV. After the introduction of RuO₂ nanoclusters, this value is moderated to -0.0125 eV, indicating that RuO₂ modulates the thermodynamic favorability of water adsorption and dissociation on FeNiOOH. Furthermore, the introduction of RuO₂ nanoclusters leads to an optimization of the adsorption free energy for hydrogen atoms, as illustrated in Figure 4C. Figure 4D presents the optimized and established water dissociation models on both the FeNiOOH and RuO₂/FeNiOOH surfaces. On the FeNiOOH surface, the preferred adsorption site for the hydrogen intermediate (*H) is located at the oxygen atom, resulting in a robust adsorption interaction that hinders the desorption of *H and subsequent H₂ formation. However, upon the incorporation of RuO₂ nanoclusters, the adsorption site of *H shifts to the ruthenium atom due to the electron transfer effect. This shift is accompanied by a change in the adsorption free energy from -1.7966 eV to 0.5112 eV [Figure 4C], facilitating the desorption of *H and H₂ evolution. The DFT calculations offer robust evidence that the electronic interaction mechanism between RuO₂ nanoclusters and FeNiOOH significantly enhances the HER property of the RuO₂/NiFe(OH)_x catalyst under alkaline conditions.

The overall water-splitting device was constructed to explore the potential applicability of RuO₂/NiFe(OH)_x, as depicted in Figure 5A. The as-prepared NiFe(OH)_x catalyst serves as the anodic electrode, and its OER performance has also been evaluated. The LSV curves and a comparison of overpotentials presented in Figure 5B and C reveal that NiFe(OH)_x can achieve the current densities of 50 mA cm^-2 and 100 mA cm^-2 by supplying the small overpotentials of 196 mV and 257 mV, respectively, much smaller than those of NFF (290 mV and 325 mV). The enhanced OER activity is ascribed to the generation of NiFe(OH)_x corrosion layers on the NFF substrate. For comparison, a water-splitting device utilizing NFF‖NFF has also been assembled. As evident from the polarization curves in Figure 5D, the NiFe(OH)_x‖RuO₂/NiFe(OH)_x-based water-splitting device exhibits superior performance compared to the NFF‖NFF device. Notably, to attain a current density of 50 mA cm^-2, the constructed NiFe(OH)_x‖RuO₂/NiFe(OH)_x device requires a voltage of merely 1.556 V, considerably lower than the 1.909 V demanded by the NFF‖NFF electrolyzer. Moreover, to assess the practical application potential, the durability of the water electrocatalysis device is evaluated. As depicted in Figure 5E, the chronopotentiometry curves of the NiFe(OH)_x‖RuO₂/NiFe(OH)_x device show remarkable operational stability, maintaining stable performance even at high current densities of 100 mA cm^-2 and 200 mA cm^-2 with only a slight increase in potential after the 100 h testing. Furthermore, a comparative analysis of the water-splitting performance of the NiFe(OH)_x‖RuO₂/NiFe(OH)_x system with recently reported devices, as summarized in Supplementary Table 8, confirms that both the NiFe(OH)_x and RuO₂/NiFe(OH)_x possess outstanding durability, making them promising candidates for practical water-splitting applications.

To validate the universality of the corrosion-reconstruction strategy, diverse metal substrates, including IF, NF, CF, and SSM, were employed to synthesize HER electrodes, as shown in Figure 6A. Notably, aside from the inherently corrosion-resistant SSM, the colors of the prepared electrodes have all changed, indicative of corrosion reactions taking place. A comparison of the HER polarization curves between pristine substrates and their corrosion-derived electrodes [Supplementary Figure 16] reveals a remarkable enhancement in HER activity for IF, NF, and CF substrates after the corrosion-reconstruction process. Conversely, the HER performance of SSM remained almost unchanged, attributed to its robust corrosion resistance. As evident in Figure 6B and C, a substantial increase in HER current density is observed for the corrosion-derived electrodes. Notably, the RuO₂/NiFe(OH)_x electrode supported on NFF stands out, realizing a remarkable current density of 181.7 mA cm^-2 at 200 mV, significantly surpassing that of the untreated NFF (10.1 mA cm^-2). To attain a current density of 100 mA cm^-2, RuO₂/NiFe(OH)_x necessitates a mere overpotential of 149 mV [Figure 6D], underscoring its superior HER electrocatalytic activity in comparison to other electrodes. Furthermore, the corrosion-reconstruction approach is conducted at ambient temperature, facilitating straightforward operation and precise control for achieving the desired scaled-up fabrication. Consequently, we have successfully synthesized the large-sized NiFe(OH)_x and RuO₂/NiFe(OH)_x catalysts (8 × 8 cm²), as depicted in Figure 6E-H. SEM images shown in Supplementary Figure 17 illustrate that the large-sized RuO₂/NiFe(OH)_x electrode still retains its nanosheet array morphology. The electrode with a 0.25 cm² area cut from the 8 * 8 cm² electrode was used for testing. The polarization curves presented in Figure 6I indicate that the HER property of the large-size RuO₂/NiFe(OH)_x electrode closely mirrors that of its smaller counterpart (2 × 2 cm²). More importantly, the long-term durability determination of the bulk electrode demonstrates that it can sustain continuous operation for over 100 h at a large current density of 200 mA cm^-2 [Supplementary Figure 18]. These comprehensive findings confirm the remarkable universality of the corrosion-reconstruction strategy in synthesizing high-performance HER electrodes.

In summary, we engineered RuO₂ nanocluster-endowed NiFe hydroxides [NiFe(OH)_x] via a corrosion-reconstruction strategy. Comprehensive characterizations indicate that RuO₂ nanoclusters anchor onto the NiFe(OH)_x surface, forming a highly active reconstructed interface that induces localized electronic rearrangement in the corrosion layers. The resultant RuO₂/NiFe(OH)_x catalyst delivers exceptional HER activity, which only demands 149 mV overpotential to drive the 100 mA cm^-2 current density. An electrolyzer assembled with this catalyst stably maintains 200 mA cm^-2 for 100 h. DFT calculations further reveal that RuO₂ introduction reconstructs active sites through electron transfer effects, shifting *H adsorption sites from oxygen to ruthenium atoms. This synergistically optimizes both water dissociation and *H desorption free energy. Moreover, the corrosion-reconstruction strategy demonstrates scalability to large-area electrodes (8 × 8 cm²) with room-temperature synthesis, highlighting its process simplicity and industrial viability. This work puts forward a promising design approach for efficient non-noble/noble metal hybrid HER catalysts and advances the mechanistic understanding of interfacial electronic modulation in alkaline hydrogen evolution reactions.