MXene-mediated reconfiguration induces robust nickel–iron catalysts for industrial-grade water oxidation

Significance

We report a significant conceptual advance on stabilizing nickel–iron catalysts based on Fe coordination modification via MXene-mediated reconfiguration. The hydrophilic and conductive MXene is used to navigate coordination reconfiguration at the atomic level via dissolution-incorporation dynamics under electro-oxidative condition. The available interaction between NiFeO_xH_y and MXene through covalent Ni/Fe-O bonds not only facilitates WOR activity but also mitigates Fe leakage, enabling the nickel–iron catalyst capable of long-term industrial-grade WOR.

Abstract

Nickel–iron oxy/hydroxides (NiFeO_xH_y) emerge as an attractive type of electrocatalysts for alkaline water oxidation reaction (WOR), but which encounter a huge challenge in stability, especially at industrial-grade large current density due to uncontrollable Fe leakage. Here, we tailor the Fe coordination by a MXene-mediated reconfiguration strategy for the resultant NiFeO_xH_y catalyst to alleviate Fe leakage and thus reinforce the WOR stability. The introduction of ultrafine MXene with surface dangling bonds in the electrochemical reconfiguration over Ni-Fe Prussian blue analogue induces the covalent hybridization of NiFeO_xH_y/MXene, which not only accelerates WOR kinetics but also improves Fe oxidation resistance against segregation. As a result, the NiFeO_xH_y coupled with MXene exhibits an extraordinary durability at ampere-level current density over 1,000 h for alkaline WOR with an ultralow overpotential of only 307 mV. This work provides a broad avenue and mechanistic insights for the development of nickel–iron catalysts toward industrial applications.

Hydrogen fuel promises as a desirable energy carrier for mankind to realize low-carbon community (1, 2). Water electrolysis powered by renewable electricity is an efficient and economical technology of green hydrogen production (3–5), which also represents an appealing solution for renewable energy conversion and storage. Electrochemical water splitting involves two half-reactions that proceed at the anode and the cathode, respectively. Generally, the anodic water oxidation reaction (WOR) manifests as a bottleneck due to its sluggish four-electron reaction kinetics and consequent high overpotential (6–10), which urgently needs the development of high-performance WOR catalysts.

Nickel–iron oxy/hydroxides (NiFeO_xH_y) emerge as one type of the most active catalysts for the WOR in alkaline media (11–13). It is widely recognized that the synergy between Fe and Ni sites is critical for the optimal WOR activity (14–21), though the magical role of Fe remains debated. However, Görlin et al. found a significant degradation of WOR activity on NiFeO_xH_y in electrolysis at a constant current density of 10 mA cm⁻² (22). Speck et al. discovered the leaching of FeO₄²⁻ species from NiFeO_xH_y during the WOR process, the rate of which was related to pH values and current densities (23). With more and more observations on Fe segregation accompanied by deactivation (24–27), highly active NiFeO_xH_y catalysts face with a serious challenge of poor stability in the sustained electrolysis. To address this issue, Obata et al. employed a permselective CeO_x coating on NiFeO_xH_y catalysts to effectively alleviate the Fe dissolution, achieving an enhanced durability for nearly 100 h at 20 mA cm⁻² (28). Chung et al. reported a dynamic compensation for the Fe dissolved over NiFeO_xH_y via the ionic Fe addition in the electrolyte and whereby proposed a “dynamically stable Fe active sites” concept to understand the maintainable WOR at around 30 mA cm⁻² (24). Further, Kuai et al. developed an intermittent reduction approach for mediating reversible Fe segregation over NiFeO_xH_y to revivify the catalytic activity, indirectly achieving the durability over 60 h at 200 mA cm⁻² (25). Recently, some strategies of bond engineering have been developed to improve radically the structural fastness of Fe coordination in NiFeO_xH_y catalysts (26, 29). For example, Peng et al. introduced atomic cation vacancies to strengthen the Fe-O binding energy in NiFeO_xH_y, resulting in the prolonged WOR stability for 100 h at around 110 mA cm⁻² against Fe dissolution (26). Liao et al. reported an oxyanion-inserted strategy to suppress Fe segregation in NiFeO_xH_y catalysts, thereby affording the durable WOR at around 150 mA cm⁻² within 90 h (29).

Note that the stability evaluation with low current densities less than industrially required 500 mA cm⁻² for tens of hours cannot comply with the long-term durability requirement for practical alkaline water electrolysis (30, 31). That said, despite many successful cases to stabilize NiFeO_xH_y catalysts, their availability still remains uncertain toward industrial applications. Notably, the durability at 500 mA cm⁻² for 110 h on NiFeO_xH_y catalyst was recently realized by Kang et al. through surface functionalization with tetraphenylporphyrin (32), which acted as a protective layer to alleviate the Fe dissolution. This prompts the search for desirable interface layers that endow NiFeO_xH_y catalysts with both high activity and robust stability competent for industrial-scale water electrolysis.

Two-dimensional (2D) layered MXene with surface-terminal electronegative groups is inclined to conjugate with metal cations via the chemical coordination (33), which may provide a possibility to anchor Fe sites. Moreover, the metallic and hydrophilic natures make MXene materials fascinating supports to tune the charge/mass transport capacities for catalysts (34–38). In this work, we construct a chemically bonded NiFeO_xH_y/MXene hybrid by MXene-mediated reconfiguration over NiFe Prussian blue analogue (PBA) to alleviate Fe leakage, which shows exceptional WOR stability even at ampere-level current densities. The MXene coating decelerates the dissolution of [Fe(CN)₆] groups in NiFe-PBA under electro-oxidation, which endows the discrete Fe cations with a chance to bond with MXene via Fe–O bonds, forming the Fe-fixed nodes favorable for both WOR activity and stability. As a result, NiFeO_xH_y/MXene catalyst manifests long-term WOR with 1,000 mA cm⁻² at only 307 mV steady for over 1,000 h in alkaline media, rendering an outstanding activity–stability performance for Ni-Fe catalysts.

Results

Design and Synthesis Scheme.

A multicomponent NiMoO₄/NiFe-PBA/MXene precursor with hierarchical structure was designed to regulate the electrochemical reconfiguration toward the targeted NiFeO_xH_y/MXene catalyst, which is schematically shown in Fig. 1A. Briefly, NiMoO₄ nanorod arrays grown on nickel foam (NF) were used as a three-dimensional (3D) architectural template that can enlarge the specific surface area to facilitate active site exposure and mass/charge transfer for high-current-density catalysis. Meanwhile, the dissoluble [MoO₄] group provides a prerequisite for deep reconfiguration under electro-oxidation (39). In order to introduce Fe into subsequent reconfiguration, an ion exchange procedure was conducted by immersing NiMoO₄ nanorods in K₃[Fe(CN)₆] solution dispersed with ultrathin Ti₃C₂T_x MXene nanosheets to obtain hierarchical NiMoO₄/KNi[Fe(CN)₆]/Ti₃C₂T_x core-shell nanorods, which is labeled as NiMoO/PBA/MXene hereafter. For comparison, another sample was also prepared by similar ion exchange without MXene involving and labeled as NiMoO/PBA. Because the valuable Fe would lose in the reconfiguration of NiMoO/PBA into metal (oxy)hydroxides, evidenced by direct proofs hereafter, the MXene that features with electronegative groups on surface was designed as a functional layer for bonding with Fe cations dissolved in dynamic reconfiguration. The actual photos and scanning electron microscopy (SEM) images of the samples at different stages in the above synthesis procedure are shown in SI Appendix, Fig. S1. Few-layer ultrathin MXene nanosheets came from MAX-phase Ti₃AlC₂ through etching and exfoliation treatments, and the ultrathin texture and terminal –OH and –F groups on surface were validated by the transmission electron microscope (TEM) and Fourier transform infrared (FTIR) spectroscopy (SI Appendix, Fig. S2). Finally, the electrochemical self-reconfiguration over the elaborate NiMoO/PBA/MXene precursor was expected to in situ induce the Fe-stabilized NiFeO_xH_y/MXene catalyst durable for high-current-density WOR.

Fig. 1.

Design scheme and structural characterization. (A) Schematic illustration for the synthetic procedure of R-NiFeO_xH_y/MXene catalyst. (B and C) SEM, (D) TEM, and (E–G) HRTEM images of NiMoO/PBA/MXene. (H) EDS elemental maps displaying the hierarchical distribution of Mo, Ni, Fe, C, N, O, and Ti elements in NiMoO/PBA/MXene.

Morphology and Structure.

NiMoO₄ nanorods grown on porous NF via one-pot hydrothermal process are observed by the field-emission SEM to have smooth surfaces and uniform dimensions with an average length of ~8 µm and an average diameter of ~200 nm (SI Appendix, Fig. S3). After ion exchange treatment, the converted NiMoO/PBA/MXene nanorods maintain the 3D array morphology, but their surfaces become rough due to the formation of PBA cubes around rods (Fig. 1B). The magnified SEM image (Fig. 1C) reveals that they are actually composed of many interlinked cubes. The TEM image further shows a transformed nanorod with jagged edges (Fig. 1D), and a magnified TEM image at typical boundary area discloses the hierarchical structure with different diffractive contrast (Inset). The ultrathin texture and surface electronegativity of MXene nanosheets offer structural flexibility and electrostatic force for conformal curling along uneven PBA edges (SI Appendix, Figs. S2 and S4). The high-resolution TEM (HRTEM) image at a selected area by blue wireframe shows two sets of distinct lattice stripes separated by clear boundary (Fig. 1E), which correspond to NiMoO₄ (Fig. 1F) and the (220) plane of NiFe-PBA (Fig. 1G), respectively. The X-ray diffraction (XRD) pattern of NiMoO/PBA/MXene also displays the characteristic diffraction peaks that can be assigned to NiMoO₄ and PBA crystals (SI Appendix, Fig. S5), without diffraction signals on MXene due to its ultrathin texture. The energy-dispersive X-ray spectroscopy (EDS) spectrum discloses the presence of Ni, Mo, Fe, Ti, O, C, and N elements in single NiMoO/PBA/MXene nanorod, and moreover, the linear sweep EDS analysis along the radial direction shows that the distribution widths of metal elements follow the order: Ti > Fe > Ni > Mo (SI Appendix, Fig. S6). Further, the EDS mapping visualizes the distribution areas of different elements in single nanorod (Fig. 1H), in which the innermost Mo and outermost Ti locations further illustrate the hierarchical structure of NiMoO/PBA/MXene.

Reconfiguration Dynamics.

Cyclic voltammetry (CV) measurements were carried out to probe the reconfiguration dynamics over NiMoO/PBA/MXene precursor. The closed areas of the CV curves increase gradually with the cycle numbers (Fig. 2A), indicating a noticeable reconfiguration process with enhanced charge-discharge capacity. The redox peaks gradually saturate in the 8th to 12th cycles (SI Appendix, Fig. S7), which suggests that the electrode reaches a steady state after full thermodynamic reconfiguration. Typically, the overpotential to deliver an anodic current density of 1,000 mA cm⁻² is decreased gradually (Fig. 2B), and a significant reduction of 133 mV is achieved in the 12th cycle comparing with the initial one. Electrochemical impedance spectra (EIS) record the change in charge transfer resistance ( $R_{\mathrm{c}\mathrm{t}}$) from 12.17 to 6.05 ohm (Fig. 2C), indicating the stepwise optimization of charge transport. In sum, a significant reconfiguration was identified with more active speciation, accompanied by the upgrades in capacitance and conductance, resulting in a favorable OER catalyst.

Fig. 2.

Electrochemical reconfiguration dynamics. (A) Continuous CV curves at 50 mV s⁻¹ over NiMoO/PBA/MXene in 1 M KOH solution, (B) the overpotentials at 500 mA cm⁻², and (C) EIS spectra with different CV cycles (Inset: equivalent circuit model). (D) In situ Raman spectra of NiMoO/PBA/MXene under anodic reconfiguration. (E) SEM image, (F and G) TEM and (H and I) HRTEM images, and (J) EDS mapping images of R-NiFeO_xH_y/MXene.

To gain insights into the dynamic reconfiguration mechanism, in situ electrochemically coupled Raman spectroscopy was performed to analyze the molecular structure evolution with controlled anodic potentials (Fig. 2D). As electro-oxidation potential increases, the characteristic bands of Mo-O-Ni and Fe-CN-Ni vibrations (27, 40) at 926, and 2,050 to 2,200 cm⁻¹ gradually decline and then disappear, indicating structural destruction of NiMoO₄ and PBA crystals. More importantly, the emergence of two characteristic bands at 484 and 568 cm⁻¹, assigned to the $E_{\mathrm{g}}$ and $A_{1\mathrm{g}}$ modes of N³⁺-O vibration (41), uncovers the generation of γ-NiOOH species. Ex situ technologies were further employed to investigate the physical structure variations after reconfiguration. According to morphology characterizations on SEM (Fig. 2E) and TEM (Fig. 2F), the reconstructed catalyst still maintains the 3D morphology of nanorod arrays covered on NF but the individuals have transformed to be porous flocculent texture. The component leaching in NiMoO/PBA/MXene precursor makes structure loose, allowing electrolyte penetration for deep reconstruction. The XRD pattern (SI Appendix, Fig. S8) shows that the diffraction peaks from NiMoO₄ and PBA phases completely disappear and meanwhile new small peaks emerge characteristic of NiFeOOH and NiFe layered double hydroxides (NiFe-LDH), which indicates the collapse of original structure and sequential reconfiguration upon metal hydroxylation. The almost complete loss of Mo element is also evidenced by EDS survey spectrum (SI Appendix, Fig. S9A), where Ni, Fe, and Ti elements are found survival in the reconstructed nanorods. In the HRTEM image (Fig. 2 H and I), the observed lattice domains can be assigned to the (104) plane of NiFe-LDH as well as the (111) and (210) planes of NiFeOOH, which also coincide with the diffraction rings in the selected area electron diffraction (SAED) pattern (SI Appendix, Fig. S9B). The EDS mapping images show the elements in the resultant nanorods are uniformly distributed (Fig. 2J) and no longer stratified in the radial direction like original NiMoO/PBA/MXene nanorods (Fig. 1H), indicating the hybridization of MXene with NiFeO_xH_y species in the reconstructed nanorods. Hence, the resultant MXene-modified NiFeO_xH_y is considered as a really catalytic species for OER, and thus, the reconstructed catalyst is denoted by R-NiFeO_xH_y/MXene.

Fourier transform infrared (FTIR) spectroscopy and X-ray photoelectron spectroscopy (XPS) were employed to further explore the chemical coordination before and after reconfiguration. The FTIR spectra in Fig. 3A show that NiMoO/PBA/MXene has mainly four bands at 3,372, 2,073, 730.3, 1,608, and 503 cm⁻¹, which belong to –OH absorbate, Fe-CN-Ni vibration in PBA (40), MoO₄²⁻ vibration in NiMoO₄ (39), as well as C=O and Ti-O vibrations in MXene (37), respectively. After reconfiguration, the Fe-CN-Ni vibration at 2,073 cm⁻¹ almost disappears while the newly emerging peaks at 869 and 690 cm⁻¹ correspond to Fe-O-H and Ni-O-H vibration modes (42, 43), respectively, indicating the crack of PBA and synchronous formation of active NiFeO_xH_y species. The peaks assigned to C=O and Ti–O vibrations in MXene shifts to a lower wave number after reconfiguration, which may be associated with its coordination environment. Fig. 3 B–E show the high-resolution XPS spectra of metal Ni 2p, Fe 2p, Mo 3d, and Ti 2p cores. For NiMoO/PBA/MXene, the Ni 2p_1/2 and Ni 2p_3/2 peaks can be deconvoluted into four peaks, assigned to Ni²⁺ at 855.2 and 873.0 eV as well as Ni³⁺ at 856.8 and 874.8 eV, respectively (Fig. 3B). Similarly, Ni still exists in R-NiFeO_xH_y/MXene in mixed valence states, but its relative content increases due to the loss of other specific elements. The Fe 2p spectrum of NiMoO/PBA/MXene (Fig. 3C) displays the 2p_1/2 and 2p_3/2 peaks located at 721.2 and 708.3 eV, mostly dominated by Fe²⁺ state upon deconvolution, which is consistent with Fe form in PBA. While after reconfiguration, the blue-shift and broadening of two Fe 2p peaks indicate an increase of Fe³⁺ with respect to Fe²⁺, suggesting the reconfiguration of Fe species. Note that Fe is not completely lost despite PBA dissolving, and instead its portion is incorporated in R-NiFeO_xH_y/MXene. Such an Fe trapping capacity depends primarily on the presence of MXene, which will be discussed further below. The Mo 3d spectrum for NiMoO/PBA/MXene (Fig. 3D) exhibits two peaks at 232.0 and 235.2 eV, respectively, corresponding to the Mo⁶⁺ state in NiMoO₄; whereas, they become undetectable in R-NiFeO_xH_y/MXene due to the complete dissolution of [MoO₄] radicals during reconfiguration. The Ti 2p spectra are deconvoluted into four pairs within 2p_3/2 and 2p_1/2 doublets (Fig. 3E), including Ti-C (454.0 and 460.3 eV), Ti²⁺ (455.6 and 461.6 eV), Ti³⁺ (457.4 and 463.0 eV), and Ti-O (458.5 and 464.2 eV) (38). The slight variations on Ti 2p peaks before and after reconfiguration indicates that most MXene is hybridized into resultant product with coordination modulation. The XPS analyses on nonmetal atoms (SI Appendix, Fig. S10 A–C) reveal the complete N leakage, partial C loss, and growth of O-metal (O-M) coordination after reconfiguration. In addition, ex situ Raman spectra (SI Appendix, Fig. S10D) further show that the vibration bands from Mo-O, Mo-O-Ni, and Fe-CN-Ni bonds disappear due to the structural destruction of NiMoO₄ and PBA after reconfiguration, while other two characteristic bands emerge at 342 and 480 to 640 cm^–1, assigned to Fe-O and Ni-O vibrations (41, 44), indicating that Fe–O and Ni–O bonds survive in R-NiFeO_xH_y/MXene.

Fig. 3.

Chemical state and coordination structure. (A) FTIR spectra of NiMoO/PBA/MXene and R-NiFeO_xH_y/MXene. High-resolution XPS spectra at (B) Ni 2p, (C) Fe 2p, (D) Mo 3d, and (E) Ti 2p regions of NiMoO/PBA/MXene and R-NiFeO_xH_y/MXene. (F) Schematic diagram for dynamic reconfiguration mechanism.

To highlight the role of MXene in mediating reconfiguration, a NiMoO₄-transformed NiFe PBA precursor without MXene involvement was also subjected to CV reconstruction under the same condition, accompanied by morphology, structure, and coordination characterizations (SI Appendix, Fig. S11). In the absence of MXene, the reconstruction with PBA collapse faster proceeds at the end of the fifth cycle; 1D rod-like morphology cannot be maintained; moreover, Fe is completely lost to form flocculent NiO_xH_y as the end product (denoted by R-NiO_xH_y). Therefore, it can be included that MXene mitigates reconfiguration and simultaneous surface-suspended electronegative groups (likely –OH) anchor partial Fe to involve in dynamic reconfiguration. Because the magical Fe in hydr(oxy)oxides is very beneficial for OER performance, here the Fe-anchored effect by MXene-mediated reconfiguration is worthy of expectation for the development of high-performance NiFeO_xH_y catalysts.

Based on the above results, the reconfiguration mechanism is proposed and illustrated in Fig. 3F, including the collapse of NiMoO₄ and PBA crystals with [MoO₄] and [Fe-CN-Ni] dissolution, binding of MXene to discrete Fe ions, and formation of NiFeO_xH_y under OH^– contact and electrooxidation conditions. The MXene with electronegative groups on surface is favorable to bond with unhooked Fe cations, which thus alleviates the Fe leaching into the electrolyte. Nevertheless, MXene-free NMO/PBA fails at Fe harvesting and thus causes a complete Fe segregation.

Electrocatalytic OER Performance.

The OER performance of R-NiFeO_xH_y/MXene and R-NiO_xH_y products by controlled reconstruction with and without MXene was examined by line sweep voltammetry (LSV) scans at 50 mV s^–1 in 1.0 M KOH at room temperature (~ 25 °C), with bare NF and commercial IrO₂ as two benchmarks. All polarization curves are in finally stable state, and the dynamic properties of the precursors before and within reconstruction are excluded for performance evaluation. As shown in Fig. 4A, R-NiFeO_xH_y/MXene has a significantly higher OER activity than R-NiO_xH_y counterpart, apparently outperforming bare NF and even IrO₂ catalyst. For instance, R-NiFeO_xH_y/MXene needs lower overpotentials of 233 and 271 mV to reach typical current densities of 10 and 100 mA cm⁻² with respect to R-NiO_xH_y (275 and 356 mV) as well as IrO₂ (310 and 408 mV) (Fig. 4B). Particularly, the current density of R-NiFeO_xH_y/MXene can achieve 500 and 1000 mA cm⁻² at overpotentials of only 301 and 307 mV, which are much lower than those of R-NiO_xH_y and IrO₂. Accordingly, R-NiFeO_xH_y/MXene exhibits the lowest Tafel slope of 15.9 mV dec⁻¹ compared with R-NiO_xH_y (34.7 mV dec⁻¹), IrO₂ (85.3 mV dec⁻¹), and NF (115.8 mV dec⁻¹), indicating superior OER kinetics (Fig. 4C). Moreover, compared with the overpotentials reported by other NiFe-based catalysts in literatures (SI Appendix, Table S1), R-NiFeO_xH_y/MXene is among the forefront; in particular, an ability to deliever high current densities up to 1,000 mA cm⁻² at low overpotentials is very desirable for practical applications.

Fig. 4.

Electrocatalytic OER performance. (A) OER polarization curves of R-NiFeO_xH_y/MXene and R-NiO_xH_y, with bare NF and commercial IrO₂ as two references. (B) Comparison on overpotentials of contrast catalysts at typical 10, 100, 500, and 1,000 mA cm⁻². (C) Tafel slopes and (D) $\mathrm{\Delta }\eta /\mathrm{\Delta }\mathrm{l}\mathrm{o}\mathrm{g}\mathrm{j}$ ratios of contrast catalysts. (E) CP stability test curves of R-NiFeO_xH_y/MXene at 1,000 mA cm⁻². (F) Comparison on activity–stability performance of R-NiFeO_xH_y/MXene with recently reported NiFe-based catalysts.

Further exploration on the physicochemical properties of R-NiFeO_xH_y/MXene was carried out to understand the causes of its high OER activity. Double-layer capacitances (DLC) were estimated by non-Faradaic CV measurements (SI Appendix, Fig. S12), in which R-NiFeO_xH_y/MXene has a higher value of 4.8 mF cm⁻² with respect to R-NiO_xH_y (3.7 mF cm⁻²), indicating larger electrochemically active surface area (ECSA). An analysis on the DLC and OER dynamics decoupled by in situ EIS spectra (SI Appendix, Fig. S13) indicates that R-NiFeO_xH_y/MXene affords larger DLC by –OH accumulation at the same potentials compared to R-NiO_xH_y, transfusing high polarization force with strong –OH affinity, and then R-NiFeO_xH_y/MXene triggers OER at 1.4 V earlier than R-NiO_xH_y (at 1.45 V). Meanwhile, R-NiFeO_xH_y/MXene exhibits much smaller $R_{\mathrm{c}\mathrm{t}}$ with respect to R-NiO_xH_y and bare NF in EIS Nyquist plots (SI Appendix, Fig. S14), revealing more favorable charge transfer. In addition, R-NiFeO_xH_y/MXene exhibits an increase in turnover frequency (TOF) in comparison with R-NiO_xH_y (SI Appendix, Fig. S15), verifying the superior intrinsic activity of R-NiFeO_xH_y/MXene. Since single MXene itself has no detectable OER activity, the enhanced intrinsic activity might result from the synergistic effect of Ni-Fe sites, which was widely accepted to understand the advantages of NiFe-based catalysts. We further used the conventional electrodeposition method to prepare the NiFeO_xH_y counterpart (E-NiFeO_xH_y) with the same loading mass (SI Appendix, Fig. S16) and found that its $C_{\mathrm{d}\mathrm{l}}$, TOF, EIS, and OER activity are not as good as R-NiFeO_xH_y/MXene. This comparison suggests that the hybridization of MXene also contributes to the increase of OER performance because the conductive and hydrophilic MXene can serve as an effective kinetics-favorable component. Furthermore, the physical resistivity of R-NiFeO_xH_y/MXene is measured as low as 14.07 × 10⁻⁴ Ω m (SI Appendix, Fig. S17), much smaller than those of R-NiO_xH_y and E-NiFeO_xH_y, indicating the enhanced electrical conductivity induced by MXene hybridization. Dynamic surface droplet contacting images (SI Appendix, Fig. S18) manifest the ultrafast surface wetting of R-NiFeO_xH_y/MXene in comparison with R-NiO_xH_y, indicating the responsibility of MXene for superb hydrophilicity. These results illustrate the excellent charge/mass transfer capability of R-NiFeO_xH_y/MXene, which is especially useful at high current densities. The ratios of $\mathrm{\Delta }\eta /\mathrm{\Delta }\mathrm{l}\mathrm{o}\mathrm{g}\mathrm{j}$ were calculated to assess the OER kinetics at different current density regions (45–47), in which only R-NiFeO_xH_y/MXene can keep small ratio less than 50 mV dec⁻¹ as the current density increases to 1,000 mA cm⁻² (Fig. 4D), indicating that vigorous OER kinetics upon ultrafast charge/mass transfer is competent for high-current-density catalysis.

The most critical issue to be answered here is whether R-NiFeO_xH_y/ MXene can anchor Fe against leakage, especially at high current densities. To our best knowledge, the durability for hundreds of hours at industrially required current densities over 500 mA cm⁻² has hardly been reported for NiFeO_xH_y catalysts (48). Note that Fe leakage is dependent on current densities (23), and some Fe-stabilized strategies proposed at low current densities, such as dynamic dissolution/deposition equilibrium (24), intermittent reduction recovery (25), vacancy or surface engineering (26–28), may be unsuitable for high current densities or industrial applications. Here, R-NiFeO_xH_y/MXene catalyst that underwent full reconfiguration with Fe-locked efficacy by MXene should be more robust to improve catalytic stability for industrial OER. As expected, R-NiFeO_xH_y/MXene exhibits an extraordinary stability with only ~1% decay of the initial potential in long-term chronoamperometry (CP) test over 1,000 h at an ultrahigh current density of 1,000 mA cm⁻² (Fig. 4E). The in situ inductively coupled plasma mass spectrometry (ICP-MS) measurement show no detectable Fe leakage into the electrolyte (SI Appendix, Fig. S19). Poststructural characterizations show that R-NiFeO_xH_y/MXene maintains its original morphology, structure, and composition unchanged after the stability test (SI Appendix, Fig. S20). The Faraday efficiency (FE) of oxygen evolving at the anode was evaluated to be close to 100% (SI Appendix, Fig. S21), indicating a stable selectivity of R-NiFeO_xH_y/MXene to OER without any side reactions. As a control experiment, E-NiFeO_xH_y catalyst prepared by electrodeposition method was also subjected to the CP stability test at 50 mA cm⁻² (SI Appendix, Fig. S22). The activity of NiFeO_xH_y drops gradually within 75 h of observation, particularly accompanied by synchronous Fe leakage. These results manifest the Fe-stabilized effect in R-NiFeO_xH_y/MXene on improving high-current-density OER durability, which is attributed to the robust conjugation established by MXene-mediated reconfiguration. We further used a 3D coordinate frame in terms of current density, overpotential, and duration time to evaluate practical application potential (Fig. 4F and SI Appendix, Table S2). Compared to other state-of-the-art catalysts, our R-NiFeO_xH_y/MXene catalyst shows a significant advantage in enduringly operating high current density at low potential, suggesting an outstanding activity-stability balance for industrial applications. Further studies on practical potencies in alkaline and anion exchange membrane electrolyzers are ongoing in our group.

Discussion

In summary, we successfully develop a robust R-NiFeO_xH_y/MXene catalyst through MXene-mediated reconfiguration, which is highly active and durable for alkaline WOR even at industrial-grade high current densities. Besides the role of ultrathin MXene component in accelerating charge/mass transfer favorable for WOR kinetics due to its high conductivity and hydrophilicity, an intriguing function for alleviating Fe leakage is achieved by surface-hanging electronegative groups to bond with unhooked Fe cations in reconfiguration process. As a result, R-NiFeO_xH_y/MXene shows a record-high activity-stability balance over 1,000 h at 1,000 mA cm⁻² for Ni-Fe-based WOR catalysts, taking a solid step toward practical applications. This work presents an effective method to hamper Fe leakage and the relevant insights for the development of durable Ni-Fe catalysts for water oxidation.

Materials and Methods

Synthesis of MXene Nanosheets.

Two grams of Ti₃AlC₂ powder was added slowly to the mixed solution of 40 mL 9 M of HCl and 3.2 g of LiF, which was stirred continuously at 45 °C for 48 h for selective Al component removing. Then, the stripped MXene was extracted and washed with distilled water by centrifugation several times until the pH of dispersion solution reaches neutral (≥6). The precipitate was collected and dried, resulting in MXene powder for reserve. Finally, 0.5 g of MXene powder was dissolved in 200 mL of deionized water with ultrasound, and then, the supernatant containing ultrathin MXene was collected using centrifugation to remove precipitation.

Synthesis of NiMoO₄ Nanorods.

First, a piece of NF with an area of 1 × 2 cm² and a thickness of 1 mm was ultrasonically cleaned in 1 M hydrochloric acid aqueous solution and then washed with ethanol and distilled water. Second, 0.02 mmol of Na₂MoO₄·2H₂O and 0.04 mmol of Ni(NO₃)₂·6H₂O were dissolved into 50 mL distilled water with magnetic stirring for 10 min to form a uniform solution. Third, the mixture was transferred to a Teflon-lined autoclave with NF, sealed, and heated at 150 °C for 6 h. After the reaction, the NF was washed several times with ethanol and deionized water and then completely dried at 60 °C for 24 h.

Synthesis of NiMoO/PBA/MXene and NiMoO/PBA.

The prepared NiMoO₄ was immersed in 50 mL K₃[Fe(CN)₆] solution (8 mg mL⁻¹) at 90 °C for 2-h ion exchange reaction to obtain NiMoO/PBA. Similarly, the NiMoO/PBA/MXene was prepared by soaking NiMoO₄ in the 50 mL K₃[Fe(CN)₆] solution mixed with 10 mL MXene dispersion at 90 °C for 2 h.

Synthesis of E-NiFeO_xH_y.

The trimmed NF (1 cm × 2 cm, thickness: 1 mm) was sequentially immersed in acetone, 8 wt.% H₂SO₄, and deionized water and treated with ultrasound for 10 min. Electrochemical deposition was carried out in a standard three-electrode setup. Initially, a mixture solution of 0.15 M Ni(NO₃)₂⋅6H₂O and 0.15 M FeSO₄⋅7H₂O was stirred uniformly. NF, Pt foil, and Ag/AgCl were used as the working electrode, counter electrode, and reference electrode, respectively. Constant current deposition was performed at a current density of 50 mA cm⁻² for 120 s, resulting in the formation of a hydroxide coating on the surface of NF.

Material Characterizations.

XRD was performed on a Shimadzu XRD-7000 diffractometer with Cu Kα radiation (λ = 1.54 Å) at a scanning rate of 2° min⁻¹. SEM was taken on a Hitachi S-4800II instrument at 5 kV acceleration voltage. TEM was carried out on a Tecnai F30 instrument at 300 kV acceleration voltage, with EDS spectroscopy. FTIR was recorded on a Varian 670-IR spectrometer with a spectral range of 400 to 4,000 cm⁻¹. In situ Raman was measured via a self-built system assembled with a three-electrode cell and RPB4 spectrometer using a 532-nm laser excitation source. XPS was recorded on an ESCALAB250Xi spectrometer using Al Kα as an excitation source at a power of 150 W, and all binding energies were referenced to the C 1 s peak at 284.8 eV. ICP-MS spectrometry was taken on Elan DRC-e (PerkinElmer).

Electrochemical Measurements.

All electrochemical tests were conducted on a CHI660E system in 1 M KOH electrolyte using a standard three-electrode installation. Self-supported catalysts on NF substrates were directly used as the working electrode, while a graphite rod and a Hg/HgO electrode were applied as the counter and reference electrodes, respectively. The LSV curves were recorded with a scanning rate of 5 mV s⁻¹. Faradic-free CV curves were recorded with a potential range of 1.054 to 1.194 V vs. RHE at different scan rates. Operando EIS was conducted with the frequency range of 0.01 to 100 k Hz. All reported potentials were calibrated via 95% iR compensation and referenced vs. RHE unless otherwise specified, with standard formulas as follows:

$$E_{\text{calibrated}}=E_{\text{measured}}-iR,$$ [1]

$$E_{\text{RHE}}=E_{\text{Hg}/\text{HgO}}+0.098+0.059\times \text{pH}.$$ [2]

The $C_{\mathrm{dl}}$ values were obtained by plotting differences in charging currents $\left(∆j/2\right)$ against scan rates ( $v$) in non-Faradic CV curves with the following formula:

$$C_{\text{dl}}=\frac{\mathrm{\Delta }j}{2v}.$$ [3]

The mole number of active sites per square meter could then be estimated with the Faraday constant ( $F$), $C_{\mathrm{dl}}$ and voltage range ( $\mathrm{\Delta }V$) of CV scanning by the following formula:

$$n=\frac{C_{\mathrm{d}\mathrm{l}}\times \mathrm{\Delta }V}{2F}.$$ [4]

So, the TOF was calculated by the following formula:

$$\mathrm{T}\mathrm{O}\mathrm{F}=\frac{j}{4Fn}.$$ [5]

The $\mathrm{FE}$ was calculated by the comparison of measured and theoretical oxygen production with the following formula:

$$\text{FE}=\frac{4\times M\times N_{\mathit{A}}}{j\times s\times t÷e}\times 100\%,$$ [6]

where $M$ is the measured oxygen amount, ( $j\times s\times t÷e)$ is calculated as the transferred electron numbers within catalytic time, and $N_A$ is the avogadro constant.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All sequence data generated inthis study have been deposited at the OSFHOME data (https://osf.io/ecqns/?view_only=6f3c0aad5b3948ed909e8e31a24ff3f7) (49). All other data are included in the article and/or SI Appendix.