Geometrically periodic architecture for enhanced mass transport in electrocatalysis Authors: Zenglong Wu, Haimin Zhang, Huajie Yin, Huijun Zhao Published: 2026 ABSTRACT AbstractElectrochemical energy conversion reactions are central to sustainable energy technologies, yet their efficiency at industrially relevant current densities is often constrained by mass-transfer limitations. Here, we introduce a periodic flow field strategy for enhancing mass transfer based on a geometrically periodic microstructure that induces spatially varying velocity fields and wall shear oscillations within the pores, generating a coherent convective contribution to otherwise diffusion-dominated transport. To implement this concept, we design hierarchical ordered interconnected porous nickel electrodes with long-range order, tunable structural periods and scalable centimeter-scale fabrication, serving as a general electrocatalyst support. The resulting architecture alleviates mass-transfer constraints across multiple electrochemical reactions. As a representative example, the NiFe-modified electrode delivers a current density of 7 A cm−2 for the oxygen evolution reaction at 25 °C with an overpotential of 393 mV, while an anion-exchange-membrane water electrolyzer sustains 1 A cm−2 at 1.68 V for 1083 h at 60 °C. We also construct a lightweight flexible device for portable oxygen and hydrogen production with efficient operation under fluctuating solar power.Subject terms: Electrocatalysis, Electrocatalysis, Electrocatalysis Introduction Electrochemical energy conversion reactions (EECRs), including water electrolysis1,2 and electrooxidation of small molecules such as urea, methanol and biomass-derived substrates (e.g., 5-hydroxymethylfurfural)3,4, are central to the sustainable development of future energy systems5–7. At industrially relevant high current densities, however, in addition to intrinsic catalytic activity, the performance of EECRs is frequently constrained by mass transfer limitations (Fig. 1a, b) at the electrode-electrolyte interface, where local reactant depletion, product accumulation, and intermediate retention generate large concentration overpotentials, manifesting as distinct current plateaus8–14. Inspired by microfluidic transport in confined channels, periodic modulation of the flow field has been demonstrated to markedly enhance convection and mixing even under low Reynolds number conditions (Re ≪ 1)15–17. This enhancement primarily arises from spatially periodic velocity profiles (Fig. 1c) and wall shear within the pores that drive transverse mass exchange16,18,19, thereby accelerating the transport of ions and molecules at the electrode interface. Translating this principle into electrode design, a potential strategy for boosting mass transfer is to embed an ordered, periodically modulated geometric microstructure within the electrode, enabling the porous structure to spontaneously sustain periodic flow patterns during operation20–23. Periodic flow fields for overcoming mass-transfer limitations. a Conceptual roles of intrinsic electrode activity and mass transfer at low and high current densities. b Schematic polarization curves showing bending caused by mass-transfer limitations at high current densities. c Oscillation characteristics of periodic flow. d Schematic of flow behavior in ordered interconnected porous electrodes. e Magnified schematic of local periodic flow within a representative region, with red, yellow, and blue indicating high, intermediate, and low velocity, respectively. Realizing electrode materials with precisely ordered and tunable periodic geometric microstructures remains challenging, particularly for the large-area, mechanically robust electrodes required for practical devices24,25. In conventional powder catalyst-coated electrodes, such as gas diffusion electrodes, the catalyst layer is structurally random and disordered, making deterministic control over local mass-transfer pathways essentially impossible11,21,26. Top-down nanofabrication techniques (e.g., two-photon lithography and electron-beam lithography) can generate precise periodic architectures with nanoscale precision and complex three-dimensional fidelity27. For instance, Larrazábal et al. employed photolithographic microfabrication to synthesize ordered Cu microelectrode arrays with uniform spacing and well-defined interfacial architectures28. Yet, such approaches are typically confined to micro- to millimeter-scale regions, and scaling them up to centimeter-scale and self-supporting electrodes is hindered by high complexity, low throughput, and prohibitive cost29,30. By contrast, bottom-up self-assembly strategies are capable of producing ordered structures at the nanoscale via intermolecular interactions and energy minimization, offering an attractive route to periodic electrode architectures31. Colloidal crystal templating, for example, can yield inverse opal or close-packed lattices with interconnected porous channels over relatively large areas and at low cost32–35. However, extending these architectures into macroscopic, free-standing electrodes while preserving long-range order across large areas remains non-trivial24,25. Here, we introduce a periodic flow field strategy to mitigate mass transfer limitations in EECRs. We synthesize a hierarchical ordered interconnected porous nickel (hOPN) via a scalable colloidal templating and electrodeposition route and employ it as a general electrode scaffold. The hOPN comprises a long-range ordered network of uniformly sized macropores interconnected by narrower neck channels, forming a highly open and bicontinuous architecture with well-defined geometric periodicity (Fig. 1d). This periodic architecture induces spatially varying velocity fields and shear oscillations within the pores, generating coherent convective flows that periodically compress and refresh the electrode-electrolyte interface even at Re ≪ 1 (Fig. 1e). The resulting convective-diffusive coupling establishes a purely geometry-driven mechanism for mass transfer enhancement, as validated by both experimental and theoretical analyses. Consequently, hOPN-based electrodes exhibit broadly enhanced mass transfer performance across examined EECRs. A NiFe-modified hOPN electrode (NiFe/hOPN) delivers a high current density of 7 A cm−2 for the oxygen evolution reaction (OER), and reaches 1 A cm−2 at 1.68 V with 1083 h of long-term durability in an anion exchange membrane water electrolyzer (AEMWE). Moreover, a lightweight flexible water electrolysis device based on hOPN for portable oxygen and hydrogen production demonstrates rapid response and stable performance under natural fluctuating solar power. This work establishes a generalizable paradigm for overcoming mass transfer bottlenecks in electrochemical energy conversion via geometry-engineered periodic flow fields. Results Synthesis and characterization of hOPN As illustrated in Fig. 2a, the hOPN was fabricated by combining precise colloidal templating with electrochemical deposition to generate an optimized ordered porous framework24,25,36,37. Briefly, monodisperse polystyrene (PS) spheres (diameters of 247, 504, 965, or 1844 nm) were first assembled into a three-dimensional ordered template by evaporation-induced self-assembly (Supplementary Fig. S1). The as-formed PS colloidal crystal was then thermally treated and resin-stabilized to improve mechanical robustness of the template and interconnectivity of the resulting electrode scaffold36. Nickel was subsequently electrodeposited into the interstitial voids of the PS template (Supplementary Fig. S2), followed by template removal to yield the freestanding hOPN. The resulting hOPN is mechanically flexible, can be produced over an area up to 100 cm2, and displays pronounced structural colors arising from its periodic architecture (Fig. 2b and Supplementary Fig. S3). Fabrication and structural characteristics of hOPN. a Schematic of the fabrication procedure for hOPN based on colloidal templating and electrodeposition. b Optical images of 100 cm2 hOPN film showing structural color. c–e Top-view and cross-sectional SEM images of hOPN, highlighting the ordered macropore array, interconnected neck channels and film thickness; the inset in (c) shows the corresponding FFT pattern, evidencing long-range periodic order. f, g Tunability of hOPN thickness, neck pore diameter and macropore diameter, with (f) corresponding to hOPN prepared from 965 nm PS spheres; the inset in (f) schematically defines the macropore diameter (DOH) and neck pore diameter (dNH). X-ray diffraction (XRD) analysis confirms that the hOPN framework consists of crystalline metallic nickel (Supplementary Fig. S4). Field-emission scanning electron microscopy images reveal that the hOPN comprises a long-range ordered array of uniformly sized macropores (965 ± 10 nm in diameter) extending over a large area, interconnected by narrower neck channels to form a highly open and bicontinuous network (Fig. 2c, d). The fast Fourier transform (FFT) pattern derived from the hOPN Scanning electron microscopy (SEM) image (inset of Fig. 2c) displays a well-defined hexagonal array of diffraction spots forming multiple complete concentric rings, further indicative of long-range periodic order in the pore lattice of hOPN. By contrast, the disordered porous control electrode (in-hOPN) exhibits randomly distributed pores with broad size dispersion, originating from non-uniform packing of PS spheres in the precursor template (Supplementary Fig. S5). Cross-sectional SEM images of hOPN further show vertically aligned macropore channels through the film thickness, with a representative thickness of 19.2 μm (Fig. 2e). The absence of PS residues on both the surface and cross-section indicates that toluene penetrates and removes the template throughout the entire volume, confirming that the ordered interconnected pore network pervades the full thickness rather than being confined to a surface layer. The experimentally derived porosity of hOPN (0.75 ± 0.03) closely matches the theoretical value of 0.80 from geometric modeling (Supplementary Fig. S6), further corroborating the periodic ordered architecture and supporting the use of the idealized geometry in subsequent transport simulations. Compared with reported porous electrodes, bare hOPN combines high specific surface area, ordered porosity, and measurable permeability (Supplementary Figs. S7–S9). The geometric parameters of hOPN can be independently and precisely tuned. The macropore diameter (DOH) and thickness are controlled by the diameter and stacking thickness of the PS spheres, and the neck pore size (dNH) is adjusted via the sintering time (Fig. 2f, g and Supplementary Figs. S10–S12). The well-defined and highly tunable periodic architecture allows for engineering across multiple geometric dimensions and provides an ideal model system for investigating geometry-driven mass transfer in electrochemical reactions. Electrocatalytic performance of NiFe/hOPN We first evaluated the electrode performance using the OER as a model system, with NiFe hydroxide nanosheets serving as the catalytically active phase. As the larger thickness offers clearer visualization to better reveal the ordered architecture and structural connectivity, the 19.2-µm-thick sample (965 ± 10 nm in diameter) is shown in Fig. 2e. Based on its optimal electrochemical performance (Supplementary Figs. S13 and S14), the hOPN with a thickness of 12.0 µm and diameter of 965 ± 10 nm serves as the representative configuration for performance evaluation. NiFe hydroxide nanosheets were electrodeposited onto hOPN to prepare NiFe/hOPN (Fig. 3a and Supplementary Fig. S15). For comparison, NiFe/in-hOPN and NiFe/NF were prepared under identical conditions on the disordered in-hOPN scaffold and commercial nickel foam (NF, thickness of 1 mm, 110 ppi), respectively (Supplementary Fig. S16). The nanosheet thicknesses of the three samples are similar, all being approximately 4.7 nm (Supplementary Fig. S17). The Fe loadings of NiFe/NF, NiFe/in-hOPN, and NiFe/hOPN are determined to be 0.096, 0.097, and 0.106 mg cm−2, respectively (Supplementary Table S4). All samples exhibit only slight differences in their contact angles, indicating similar hydrophilic behavior (Supplementary Fig. S18). SEM imaging shows that NiFe/hOPN preserves the ordered interconnected architecture while being uniformly coated with dense NiFe hydroxide nanosheets (Fig. 3b). X-ray photoelectron spectroscopy (XPS) analysis confirms that the NiFe/hOPN surface consists mainly of Ni2+ and Fe3+ species (Supplementary Fig. S19)38. The OER tests were performed in 1.0 M KOH using a standard three-electrode configuration, unless otherwise specified (Supplementary Fig. S21). The OER polarization curve of NiFe/hOPN (Fig. 3c) displays a high current density of 7 A cm−2 at an overpotential of 393 mV with good reproducibility (Supplementary Fig. S22), without any obvious current plateau over the measured range. This behavior indicates that the apparent catalytic kinetics are not appreciably constrained by mass transport, even at multi-ampere current densities. By contrast, both NiFe/in-hOPN and NiFe/NF exhibit pronounced curvature in their polarization curves once the current density exceeds ~3000 mA cm−2, indicative of emerging mass-transfer limitations. Across the entire range of current densities, the NiFe/hOPN consistently delivers the lowest overpotentials, reaching 100, 1000, and 3000 mA cm−2 at representative overpotentials of 232, 291, and 333 mV, respectively (Fig. 3d). Tafel analysis (Supplementary Fig. S23) shows lower Tafel slopes for NiFe/hOPN (30.9 mV dec−1) than for NiFe/in-hOPN (42.2 mV dec−1) and NiFe/NF (58.2 mV dec−1), which could be consistent with relatively improved kinetics of NiFe/hOPN under the tested conditions. Furthermore, an interval-based apparent slope analysis based on Δη/Δlog∣j∣ shows that NiFe/hOPN exhibits a much lower slope at high current densities (Supplementary Fig. S24), indicating reduced polarization caused by mass-transfer limitations. To directly probe the impact of mass transport, we examined the dependence of the polarization response on scan rate, a sensitive indicator (|E–E0|/E0) of diffusion limitations in high-rate electrolysis, where E denotes the overpotential required to reach 3 A cm−2 at a given scan rate, and E0 is the corresponding overpotential measured at the quasi-steady 5 mV s−1 scan rate39. NiFe/hOPN exhibits the weakest dependence of operating potential on the scan rate among all samples (Fig. 3e and Supplementary Fig. S25). As shown in Fig. 3e, at a fixed current density of 3 A cm−2, the potential variation of NiFe/hOPN is limited to 0.51% when the scan rate increased from 5 to 50 mV s−1, markedly lower than that of NiFe/in-hOPN (12.04%) and NiFe/NF (15.25%). In addition, NiFe/hOPN shows minimal sensitivity to variations in electrolyte concentration, temperature and electrode thickness (Supplementary Figs. S13 and S26). The electrochemically active surface area (ECSA) was systematically evaluated, with relatively small differences among the three electrodes, allowing surface area effects to be excluded as the dominant origin of the observed performance enhancement (Supplementary Fig. S28). Collectively, these results demonstrate that the ordered interconnected porous architecture of NiFe/hOPN affords a markedly enhanced mass-transfer capability, which underpins its competitive high-current OER performance. Electrocatalytic performance and universality of hOPN-based electrodes. a Schematic illustration of the fabrication procedure for NiFe/hOPN. b SEM image of NiFe/hOPN showing the ordered porous scaffold uniformly coated with NiFe hydroxide nanosheets. c OER polarization curves of NiFe/hOPN, NiFe/in-hOPN, and NiFe/NF in 1.0 M KOH at 25 °C, with 95% iR correction. d Comparison of OER overpotentials at selected current densities. e Dependence of operating potential on scan rate at a fixed current density of 3 A cm−2. f Long-term stability of NiFe/hOPN during OER at 1000 mA cm−2 in 1.0 M KOH (25 °C) over 1110 h. g Schematic diagram of the universality of mass transfer enhancement strategy for diverse EECRs. h Polarization curves of hOPN-based electrodes for different reactions in 1.0 M KOH: NiFe/hOPN for UOR, MOR, and OER; NiCo/hOPN for HMFOR; PtNi/hOPN for HER; the inset in (h) shows enlarged portions of the curves before and after HMF addition. The nature of the active phase on NiFe/hOPN was further examined by operando Raman spectroscopy in 1.0 M KOH (Supplementary Fig. S31). The emergence of characteristic Ni3+-O vibrational modes at ~475 and ~550 cm–1 at anodic potentials above 1.38 V versus the reversible hydrogen electrode (RHE) confirms the formation of high-valent Ni oxyhydroxide species, which are widely recognized as the main OER-active phase10. NiFe/in-hOPN and NiFe/NF exhibit similar Raman features (Supplementary Fig. S32). Additionally, the durability of the NiFe/hOPN was evaluated by chronopotentiometry at 1000 mA cm−2 in 1.0 M KOH using a three-electrode configuration. As shown in Fig. 3f, the electrode exhibits limited performance decay over 1110 h of continuous operation, highlighting its long-term operational stability and potential for industrial applications. SEM images after long-term stability testing show that the geometrically periodic architecture of NiFe/hOPN is well preserved, with no structural collapse, while the nanosheets evolve into thicker features accompanied by some nanoparticles (Supplementary Fig. S33), which can be attributed to electrochemical reconstruction and dissolution-redeposition processes under anodic conditions40,41. XRD patterns show no additional crystalline phases are observed after long-term stability testing, with only a slight decrease in peak intensity observed (Supplementary Fig. S34a), further supporting the structural stability of the NiFe/hOPN. In addition, Raman and XPS analyses reveal the formation of high-valence Ni3+-O and Fe3+-O species along with an increased Ni atomic fraction at the surface (Supplementary Figs. S34b and S35), collectively indicating the transformation toward catalytically active oxyhydroxide phases via surface reconstruction. Inductively Coupled Plasma (ICP) analysis shows that the overall amount of dissolved species is minimal, indicating limited metal leaching. Specifically, Fe dissolution is more pronounced in the early stage and gradually stabilizes after 80 h, while Ni dissolution remains consistently low throughout the entire test (Supplementary Fig. S36). This observation is consistent with the increased Ni atomic fraction at the surface, as revealed by XPS analysis. Overall, despite the observed surface evolution and minor metal redistribution, the electrode maintains its structural integrity and compositional stability, demonstrating operational stability under prolonged high-current-density conditions40. Generality of mass transfer enhancement strategy To assess the generality of the mass transfer strategy based on an ordered geometric microstructure, we extended the hOPN architecture to a series of representative EECRs beyond the anodic water oxidation (Supplementary Table S5), using appropriately selected catalyst loadings (Fig. 3g). These reactions include the cathodic hydrogen evolution reaction (HER), the nitrogen-containing small-molecule oxidation reaction (urea oxidation (UOR)), the alcohol oxidation reaction (methanol oxidation (MOR)), and a biomass-derived oxidation reaction (5-hydroxymethylfurfural oxidation (HMFOR)), with the last one being particularly susceptible to mass transfer limitations due to the sluggish diffusion of HMF42. All measurements were conducted in 1.0 M KOH electrolyte with the corresponding reactant introduced. As shown in Fig. 3h, for HER, our PtNi-coated hOPN (PtNi/hOPN) reaches 7 A cm−2 at an overpotential of 570 mV (and 1 A cm−2 at 220 mV) without an apparent current plateau, indicative of rapid reactant/product transport in the multi-ampere regime43. UOR and MOR on NiFe/hOPN display nearly linear polarization behavior deep into the multi-ampere regime (Supplementary Fig. S37). In particular, the UOR polarization curve shows no discernible curvature up to 5.4 A cm−2 before gradually approaching the OER up to 7 A cm−2, indicating that the enhanced mass transfer sustains reactant availability at high current densities11. For HMFOR, NiCo-hydroxides-coated hOPN (NiCo/hOPN) achieves 1.1 A cm−2 before a diffusion-limited plateau emerges (inset of Fig. 3h). Given that ~1 A cm−2 is already considered a high current density for HMFOR, this performance underscores the capability of the ordered porous microstructure to mitigate mass-transfer constraints even for inherently diffusion-challenged substrates. By contrast, electrodes based on the disordered in-hOPN scaffold or on NF exhibit lower current density and pronounced mass-transfer-induced curvature across these reactions (Supplementary Figs. S37–S39). Collectively, these results establish that the ordered geometric microstructure serves as a universal architecture to overcome mass transfer limitations, enabling robust electrocatalytic performance across various EECRs at high current densities. Geometrically periodic architectures driving mass transfer enhancement Given the micro- to nanoscale dimensions of the pores, the Reynolds number under conventional flow velocities is much smaller than unity (Re ≪ 1), such that inertial flow effects are limited and cannot account for the observed performance enhancement44. Starting from the structural geometric periodicity, we therefore examine the fundamental mechanism by which hOPN enhances mass transport in electrochemical systems. The FFT pattern from the NiFe/hOPN SEM image displays a clear hexagonal array of diffraction spots with four complete concentric rings (Fig. 4a and Supplementary Fig. S40), reflecting the preserved long-range periodic order of the underlying hOPN architecture45–47. The corresponding spatial frequency spectrum shows four symmetric harmonic peaks flanking the central point (Fig. 4c), indicating a well-defined long-range periodicity with a period of 916 nm (reflecting the bright–dark intensity periodicity rather than the exact PS diameter) and an order factor (I/FWHM) of 307 derived from the spatial primary frequency (Fig. 4d). By contrast, NiFe/in-hOPN displays only diffuse rings, a single broad spatial-frequency peak and a much lower order factor of 54 (Fig. 4b–d), indicating the loss of long-range coherence. These frequency-domain fingerprints validate that the pore field of hOPN can be reasonably approximated as a sinusoidally modulated structure. Combined with its highly open bicontinuous network structure, where macropores are interconnected through smaller neck channels, the local pore cross-sectional area can be defined as: where A0 is the average pore cross-sectional area, λ is the structural period, and ε is the amplitude, defined as ε = (DOH − dNH)/(DOH + dNH), with DOH and dNH denoting the diameters of the macropore and neck pore, respectively. This periodic variation in cross-section accelerates the fluid at the necks (local constrictions) and decelerates it in the wider pore chambers. Mass continuity then imposes a corresponding periodic modulation of the axial velocity field: where U0 is the mean axial velocity. The spatial modulation of u(x) generates periodic pressure gradients and shear-rate perturbations, which can be expressed as: where μ is the dynamic viscosity and γ̇(x) is the local shear rate. Even under conditions of Re ≪ 1, these velocity, pressure, and shear oscillations periodically compress and renew the diffusion boundary layer, introducing a coherent convective component into the otherwise diffusive-dominated transport process. This establishes a purely geometry-driven mechanism for mass transfer enhancement arising from structural periodicity. Geometric periodicity and mass-transfer enhancement. a, b FFT patterns of NiFe/hOPN (a) and NiFe/in-hOPN (b), highlighting the presence or absence of long-range periodic order. c Spatial frequency spectra of NiFe/hOPN and NiFe/in-hOPN. d Magnified first-order peak of the spatial frequency spectra used to define the order factor (I/FWHM, I represents the peak intensity and FWHM is the full width at half maximum). e EIS Nyquist plots of NiFe/hOPN, NiFe/in-hOPN, and NiFe/NF; the inset in (e) shows the equivalent circuit used for fitting the impedance spectra. f DRT spectra of NiFe/hOPN, NiFe/in-hOPN and NiFe/NF. g Experimentally measured kma values of NiFe/hOPN, NiFe/in-hOPN, and NiFe/NF, together with the simulated kma of NiFe/hOPN, as a function of superficial velocity. h Structural schematic of the periodic NiFe/hOPN architecture used in the finite element model. i Simulated distributions of normalized flow velocity (top), OH− concentration (bottom left) and O2 concentration (bottom right) within the hOPN structure. j Oscillations characteristics of periodic flow in NiFe/hOPN obtained from finite-element simulations, showing local velocity amplification relative to the mean. k Normalized radial concentration profiles from pore center to pore wall for NiFe/hOPN and NiFe/in-hOPN. Electrochemical impedance spectroscopy (EIS) provides direct experimental evidence for this enhanced transport. As shown in Fig. 4e, Nyquist plots reveal that NiFe/hOPN exhibits a pronounced reduced mass transfer impedance, with a characteristic low-frequency resistance of 2.19 Ω, compared with 4.08 Ω for NiFe/in-hOPN and 8.32 Ω for NiFe/NF. Distribution of relaxation times (DRT) analysis further resolves the associated temporal signature: the characteristic relaxation time for NiFe/hOPN (0.148 s) is threefold to fivefold shorter than that for NiFe/in-hOPN (0.451 s) and NiFe/NF (0.697 s), indicating much faster renewal of the electrochemical interface enabled by the coherent convective-diffusive coupling within the periodic scaffold (Fig. 4f). Consistent with these transient results, the effective mass-transfer resistance coefficient (k) extracted from the limiting-current plateau reflecting steady-state mass-transfer performance is also minimal for NiFe/hOPN (k = 8.22 × 10−4) (Supplementary Fig. S41). The mass-transfer performance is directly evaluated by mass-transfer coefficient measurements using a flow electrochemical system with 1 mM K4Fe(CN)6 in 1 M KCl as the redox-probe electrolyte (Fig. 4g and Supplementary Figs. S42–S44)44,48–50. NiFe/hOPN exhibits a much higher mass-transfer coefficient (kma = 48.25u0.68) than NiFe/in-hOPN (kma = 2.62u0.60) and NiFe/NF (kma = 0.78u0.57). At 0.20 cm s−1, the kma of NiFe/hOPN is approximately 17 and 50 times higher than those of NiFe/in-hOPN and NiFe/NF, respectively. Although additional bubble-induced perturbations during OER cannot be fully described by the ferrocyanide-derived kma values, these values provide quantitative supporting evidence for the enhanced mass-transfer capability enabled by the geometric periodicity of hOPN. Taken together, these transient and steady-state transport metrics confirm that the geometric periodicity of hOPN establishes a highly efficient transport network that mitigates mass-transfer limitations and accelerates interfacial reaction processes. Finite element simulations further support the mass transfer enhancement driven by geometric periodicity (Fig. 4h). The simulated local flow velocity within NiFe/hOPN displays a clear periodic oscillation in phase with the structural period (Fig. 4i), reaching values up to six times the inlet velocity within the periodic architecture (Fig. 4j). These velocity oscillations generate pressure fluctuations with an approximate π/2 phase shift, which create alternating high-shear and low-shear zones along the pore channel. The resulting periodic compression and renewal of the diffusion boundary layer are consistent with the theoretical picture described by Eqs. (2)–(4). In contrast, the NiFe/in-hOPN structure develops irregular flow paths with isolated high-velocity channels adjacent to poorly accessed stagnant regions, forming an inefficient transport network with limited mass-transfer capability (Supplementary Fig. S46). Analysis of the simulated concentration fields further highlights the advantage of periodic architecture (Fig. 4i and k). Normalized concentration profiles from the pore wall to the center reveal that, at any radial position, NiFe/hOPN maintains higher OH− concentration and lower O2 concentration than NiFe/in-hOPN. Whereas NiFe/in-hOPN displays nearly linear concentration gradients characteristic of diffusion-limited transport (Supplementary Fig. S46), NiFe/hOPN exhibits distinctly parabolic profiles, providing supports for coherent convective dispersion induced by geometric periodicity. The convection-diffusion transport of ferrocyanide is used to directly calculate the theoretical mass-transfer coefficient (Supplementary Fig. S47)44. The simulated mass-transfer coefficient of NiFe/hOPN is fitted as kma = 136u0.89 (Fig. 4g), remaining within the same order of magnitude as the experimentally measured relationship over the tested velocity range, with a maximum deviation of less than 2.5-fold. This difference is reasonable because the simulation represents an idealized limiting-current model, in which the reactive surfaces are assumed to be fully accessible and perfectly mass-transfer limited. The general agreement between simulation and experiment indicates that the model can reasonably describe the mass-transfer behavior of hOPN, further supporting the proposed mechanism in which geometric periodicity induces coherent convective dispersion and thereby enhances mass transfer. Performance of hOPN-based AEMWE and lightweight flexible devices To further assess the practical relevance of the hOPN architecture, we assembled an AEMWE using NiFe/hOPN as the anode and PtNi/hOPN as the cathode for alkaline water electrolysis under industrially relevant conditions (Supplementary Fig. S48). The AEMWE cell equipped with NiFe/hOPN||PtNi/hOPN electrode pair delivers a current density of 1 A cm−2 at a cell voltage of 1.68 V, showing a lower cell voltage than the benchmark IrO2||Pt/C configuration, as well as the NiFe/in-hOPN||PtNi/hOPN and NiFe/NF||PtNi/hOPN cells (Fig. 5a). Notably, this device-level performance is competitive with that of reported AEMWE systems, consistent with the high current densities observed in three-electrode measurements (Supplementary Table S1 and Supplementary Fig. S49). The device further sustains high current densities, achieving 5 A cm−2 at 2.22 V (Fig. 5a). The Nyquist plot of the electrochemical impedance revealed a low cell resistance of 58 mΩ (Supplementary Fig. S50), indicative of efficient charge transfer through the electrodes based on hOPN. Long-term durability tests demonstrate that the NiFe/hOPN||PtNi/hOPN AEMWE maintains stable operation for 1083 h at 1 A cm−2 (Fig. 5b), with limited performance degradation observed. The high-frequency resistance (HFR) shows only a slight increase from 58 to 70 mΩ after 200 h of operation, indicating relatively stable interfacial and membrane resistance during the initial stage of long-term electrolysis (Supplementary Fig. S51). This combination of high current density, low operating voltage and extended stability underscores the potential of the hOPN architecture as a robust platform for industrial EECRs. Performance of hOPN-based AEMWE and lightweight flexible electrolysis devices. a Polarization curves of AEMWE cells using NiFe/hOPN, NiFe/in-hOPN, and NiFe/NF as anodes and PtNi/hOPN as the cathode, compared with a commercial IrO2||Pt/C system at 60 °C (without iR compensation, shaded regions represent the standard deviation from three independent measurements). b Long-term stability of the NiFe/hOPN||PtNi/hOPN AEMWE at 1 A cm−2 and 60 °C (Data are from one representative device). c Schematic diagram of the lightweight flexible electrolysis device integrated with a thin-film solar cell module. d Photograph of the lightweight and flexible electrolysis device. e Polarization curves of the lightweight flexible device employing NiFe/hOPN and PtNi/hOPN as electrodes at 25 °C without iR correction. f Time-resolved input voltage and current-density profiles for the solar-integrated lightweight flexible electrolysis device during outdoor operation on Days 1–3. g Time-resolved input voltage and current-density profiles recorded during outdoor operation on Days 4–6; E denotes the input voltage and J denotes the current density in (f and g). Conventional water electrolyzers typically rely on heavy and rigid components, such as stainless steel or titanium bipolar plates and current collectors51, resulting in kilogram-scale systems that lack portability and mechanical adaptability even at laboratory scale52. This severely limits their deployment in applications that demand lightweight or adaptable oxygen/hydrogen sources, including high-altitude, underground, or space environments53,54. To address this limitation, we exploited mechanical flexibility, self-supporting nature, and scalable manufacturability of hOPN-based electrodes (NiFe/hOPN and PtNi/hOPN) to construct a lightweight and flexible water electrolysis device integrated with a thin-film solar module for operation under dynamically varying power inputs (Fig. 5c)55. In this configuration, the hOPN-based electrodes and anion exchange membrane are laminated onto thin polyethylene terephthalate (PET) plates and encapsulated with a heat-shrink film (Fig. 5d), yielding a compact, mechanically robust, and bendable device. Notably, even without applying high compression pressure to the cell stack, the device exhibits a low resistance (Supplementary Fig. S52), which we attribute to the high intrinsic conductivity of the nickel framework and the favorable mass-transfer characteristics of the hOPN-based electrodes. As shown in Fig. 5e, this lightweight and flexible electrolyzer achieves current densities of 300 mA cm−2 at 1.92 V and 500 mA cm−2 at 2.08 V in 1.0 M KOH at 25 °C. It maintains stable operation across current density window (50–500 mA cm−2), with minor performance losses under step changes in load (Supplementary Fig. S53 and Fig. 5e). When integrated with a thin-film solar module, the system operated stably for 6 consecutive days under outdoor conditions at Hefei Science Island, successfully adapting to natural fluctuations in solar intensity (Fig. 5f, g). The combination of high efficiency, mechanical flexibility, and compatibility with intermittent solar input highlights the suitability of hOPN-based devices for decentralized, portable and renewable-driven hydrogen and oxygen generation56. Discussion In summary, we demonstrate a geometry-engineered, periodic flow-field strategy for mitigating mass transfer limitations in EECRs. Using colloidal templating and electrodeposition, we synthesize hOPN with long-range periodic order, precisely tunable pore and neck dimensions, and scalable centimeter-scale form factors. The periodically ordered porous network of hOPN induces velocity and shear oscillations, forming coherent convective flow that continuously refreshes the electrode-electrolyte interface. As a representative case, NiFe/hOPN achieves 7 A cm−2 for OER at an overpotential of 393 mV and sustains 1 A cm−2 at 1.68 V for 1083 h in AEMWE. This geometry-driven mechanism also enables efficient mass transport and universal performance enhancement across various reactions, including UOR, MOR, HMFOR, and HER. Moreover, the flexible hOPN-based electrolyzer retains efficient operation under fluctuating solar input, highlighting that the same ordered architecture can be translated into lightweight, mechanically compliant devices. Overall, the geometry-induced transport mechanism realized in hOPN provides a practical and scalable route to designing high-performance electrodes where mass transfer and reaction kinetics are simultaneously optimized. Methods Materials All reagents and solvents were purchased from commercial suppliers and used as received unless otherwise noted. Ethanol (≥99.7%), acetone (≥99.7%), methanol (≥99.8%), toluene (≥99.5%), and isopropanol (≥99.7%) were purchased from Sinopharm Chemical Reagent Co., Ltd. Sodium hydroxide (NaOH, ≥98%), ammonium persulfate (APS, (NH4)2S2O8, ≥98%), polyvinylpyrrolidone (PVP, Mn ≈ 40000), sodium 3-mercapto-1-propanesulfonate, poly(ethylene glycol) diacrylate (PEGDA, Mn ≈ 600), 2,4,6-trimethylbenzoyl-diphenylphosphine oxide (TPO, ≥97%), nickel sulfamate (≥98%), nickel chloride (NiCl2·6H2O, ≥98%), boric acid (H3BO3, ≥99.5%), ferrous sulfate (FeSO4·7H2O, ≥99%), potassium hydroxide (KOH, ACS), hexachloroplatinic acid hexahydrate (H2PtCl6·6H2O, ACS), cobalt nitrate hexahydrate (Co(NO3)2·6H2O, ≥99%), 5-hydroxymethylfurfural (HMF, ≥99.5%), and urea (≥99.5%) were purchased from Aladdin Co., Ltd. Styrene (99%) was purchased from Macklin Co. Ltd. Nickel foam (thickness of 1 mm, 110 ppi) was purchased from Guangzhou Jiayuan New Materials Co., Ltd. ITO-coated glass substrates (1.1 mm thick, 5 Ω sq−1) were purchased from Luoyang Guluo Glass Co., Ltd. Synthesis of the monodisperse colloidal polystyrene (PS) spheres Monodisperse colloidal PS spheres with diameters of 247, 504, 965, and 1844 nm were synthesized by dispersion or emulsion polymerization by using different solvents (isopropanol, ethanol, or water). In a typical procedure, the inhibitor in styrene monomer was removed by repeated washing with 10 mL of 10 wt.% NaOH aqueous solution. The purified styrene was then transferred to a round-bottomed flask and mixed with the selected solvent and PVP. The mixture was purged with N2 for 15 min, heated to 75 °C, and maintained for 15 min under a N2 atmosphere. An APS solution was added to initiate polymerization, which proceeded under continuous stirring under N2 atmosphere. After completion and cooling to 25 °C, a milky emulsion of PS microspheres was obtained. The PS microspheres were collected by centrifugation, washed repeatedly, and redispersed in ultrapure water. A 50% (w/v) PS suspension was prepared by five ultrasonication-assisted redispersion cycles. Detailed reagent amounts, reaction times, and reaction temperatures for PS spheres of different diameters are summarized in Supplementary Table S2. Functionalized substrate surface ITO-coated glass substrates were functionalized to tailor surface charge prior to PS self-assembly. ITO-coated glass substrates (10 × 10 cm2) were sequentially sonicated in methanol, acetone and isopropanol for 15 min each, then dried under a N2 stream. A thin Au layer was deposited by magnetron sputtering. The Au-coated substrates were immersed in a 1 mM sodium 3-mercapto-1-propanesulfonate solution at 25 °C for 24 h. The thiol groups chemisorbed onto Au, grafting terminal sulfonate groups and imparting a negatively charged surface. Synthesis of the hOPN electrode The hOPN was fabricated by integrating precise colloidal templating with electrochemical deposition. The functionalized Au-coated ITO glass slide (10 × 10 cm2) as substrate was rinsed five times with ultrapure water, dried under N2 stream, and placed horizontally. 50 w/v% PS suspension (1 mL) was drop-cast and evenly spread on the substrate. Upon slow evaporation, the PS microspheres self-assembled into face-centered cubic (FCC) colloidal crystal layer on this substrate over 72 h. The resulting template was thermally treated at 95 °C in air for 6 h, impregnated with a UV-curable resin solution (6 v/v% PEGDA in ethanol containing 2 wt% TPO) for 1 min, and subsequently cured under UV light irradiation for 30 min. After wetting with ethanol, the reinforced template was transferred into a nickel sulfamate electrodeposition bath and used as the cathode. The electrodeposition bath was prepared using nickel sulfamate, nickel chloride, boric acid, and water with a mass ratio of 45:3:3:100. An acid-treated nickel foam (NF, 10 × 10 cm2) was used as the anode in a two-electrode electroplating configuration. Electrodeposition was conducted under galvanostatic control using the initial current profile shown in Supplementary Fig. S2, followed by a current of 50 mA until a total charge of 2000C was applied. The as-plated sample was immersed sequentially in toluene and in a resin-etching solution (NaOH/acetone/methanol, 2 g/10 mL/10 mL) for 24 h to remove the PS template and the cured resin, respectively, yielding a centimeter-scale, all-metal, ordered interconnected porous self-supporting electrode (hOPN). Synthesis of the in-hOPN electrode The disordered porous nickel electrode (in-hOPN) was prepared following a procedure analogous to that used for hOPN, except that the PS template was dried rapidly to suppress ordering. Briefly, 1 mL of a 50 w/v% PS suspension was drop-cast onto the functionalized Au-coated ITO substrate and spread uniformly, then dried in a well-ventilated environment within 1 h. Under these conditions, the PS microspheres did not have sufficient time to self-assemble into an ordered colloidal crystal. The subsequent resin reinforcement, nickel electrodeposition, and template/resin removal steps were performed as described for hOPN. After etching, a self-supporting nickel electrode with non-ordered pores was obtained and denoted as in-hOPN. Synthesis of the NiFe/hOPN, NiFe/in-hOPN, and NiFe/NF electrodes NiFe hydroxide coatings were deposited by pulse electrodeposition using the nickel sulfamate plating bath described above, supplemented with FeSO4·7H2O to a final Fe2+ concentration of 50 mM. hOPN, in-hOPN, or acid-treated nickel foam (NF) substrates were used as the working cathodes, with acid-treated NF of identical geometric area serving as the counter anode. Pulse electrodeposition was conducted at a current density of 3 mA cm−2 for 3 s, followed by a 1 s open-circuit rest per cycle. Deposition was terminated when the areal charge reached 1.5 C cm−2. The resulting electrodes are denoted as NiFe/hOPN, NiFe/in-hOPN, and NiFe/NF. Synthesis of the NiCo- and PtNi-modified electrodes NiCo hydroxide-modified electrodes for the HMF oxidation reaction (HMFOR) were prepared by pulse electrodeposition using the same protocol as for NiFe deposition, except that Co(NO3)2·6H2O was added to the electrolyte to afford a final Co2+ concentration of 50 mM. PtNi-modified electrodes were prepared by Pt deposition onto Ni substrates. Briefly, the substrate was pretreated in 2 M HCl for 10 min, rinsed thoroughly with ultrapure water, and then vertically immersed for 30 min in an aqueous solution of hexachloroplatinic acid hexahydrate (H2PtCl6·6H2O, 2 mg mL−1), using 1 mL of solution per square centimeter of electrode area (Pt loading of 0.54 mg cm−2). Electrochemical measurements Electrochemical measurements were performed in 1.0 M KOH solution (pH = 13.89 ± 0.08) at 25 °C using DH7003B electrochemical workstation (Jiangsu Donghua Analytical Instrument Co., Ltd.). KOH electrolytes were prepared from KOH pellets and ultrapure water, stored in sealed polypropylene bottles at 25 °C, and used within one week. Oxygen evolution reaction (OER) tests were conducted in a three-electrode configuration with a graphite plate as counter electrode and a Hg/HgO (1.0 M KOH) reference electrode. Self-supported electrodes (NiFe/hOPN, NiFe/in-hOPN, and NiFe/NF, 0.5 cm2) served as the working electrodes. The measurements were conducted using a temperature-controlled electrochemical system (Supplementary Fig. S21). A relatively large volume of KOH electrolyte (1 L) was stored in a glass bottle immersed in a constant-temperature water bath maintained at 25 °C. The electrolyte was circulated into the electrochemical cell from the bottom inlet and flowed out from the upper outlet using a peristaltic pump, forming continuous electrolyte circulation during the test. A thermometer was placed near the working electrode to monitor the local electrolyte temperature in real time. This configuration helped minimize temperature fluctuations caused by high-current operation and enabled OER measurements under controlled temperature conditions. All tests have the same configuration, unless otherwise noted. Prior to data collection, all electrodes were conditioned by ~20 cyclic voltammetry (CV) scans until stable responses were obtained. Polarization curves were characterized at scanning rates of 5 mV s−1. The solution resistance was determined by EIS at 0.6 V versus Hg/HgO before the electrochemical test, and 95% iR compensation was applied unless otherwise stated. Potentials were converted to the RHE scale using: where EHg/HgO is the measured potential, R is the electrolyte resistance, and i is the current. Long-term stability test was evaluated galvanostatically at 1000 mA cm−2 utilized the same three-electrode setup. The Hg/HgO reference electrode was calibrated in an electrolyte saturated with high-purity H2 using two Pt plates (1 × 1 cm2) as the working and counter electrodes. Cyclic voltammetry was performed at a scan rate of 5 mV s−1. The RHE potential was determined as the average of the anodic and cathodic zero-current potentials and used to convert the measured potentials to the RHE scale. Electrochemical tests for UOR, MOR, HMFOR, and HER were also conducted in 1.0 M KOH using the same configuration and procedures as for OER. The electrolyte was supplemented with 0.33 M urea (UOR), 1.0 M methanol (MOR), or 100 mM 5-hydroxymethylfurfural (HMFOR). HER measurements were performed in 1.0 M KOH without additional reactants. Polarization curves for HMFOR were recorded at 10 mV s−1. The ECSA was determined by CV in a non-faradaic potential region. Measurements were conducted in the same three-electrode configuration as described above, using a Hg/HgO reference electrode in 1.0 M KOH electrolyte. The double-layer capacitance (Cdl) was obtained from the slope of the half current density difference (Δj/2) versus scan rate plots. The ECSA was calculated as ECSA = Cdl/Cs, where Cs is the specific capacitance. A typical value of Cs = 60 μF cm−2 for Ni-based materials was adopted. Specific surface area was calculated from ECSA and volume. All samples were analyzed using the same method to ensure consistent comparison. MEA fabrication and AEM water-electrolyzer evaluation An anion exchange membrane (PiperION® A40, Versogen, 40 μm, 2.0 × 2.0 cm2) was used for AEM water-electrolyzer assembly. Prior to assembly, the membrane was converted to the hydroxide form by ion exchange in 0.5 M KOH solution for 12 h. The NiFe/hOPN and PtNi/hOPN catalysts were cut into 1.0 × 1.0 cm2 squares and employed as the anode and cathode, respectively. The single-cell electrolyzer featured a unidirectional serpentine flow field, with both the anode and cathode flow-field plates fabricated from Pt-coated titanium plates prepared by physical vapor deposition, and an active area of 1.0 cm². The flow-field plates were cleaned with ethanol and acetone to remove residual hydrocarbons and surface contaminants. Given the relatively thin thickness (12 μm) of the hOPN electrodes, a compliant nickel fiber felt interfacial layer (250 μm) was introduced between the electrodes and the flow-field plates to mitigate localized compressive stress during assembly and to preserve the structural integrity of the porous framework. The pretreated membrane, electrodes, nickel felt interlayers, and polytetrafluoroethylene (PTFE) gaskets (for sealing and electrical insulation) were then assembled to form the membrane electrode assembly (MEA). PTFE gaskets with a thickness of 300 μm were employed on both the anode and cathode sides to provide controlled compression, and the MEA was assembled under a torque of 4 N·m. Performance was evaluated using a membrane electrode integrated testing platform (Shengshui Technology). The operating temperature was maintained at 60 °C using a rod heater and a temperature sensor. The flow rate of the 1.0 M KOH electrolyte was controlled at 1 mL min−1 using a peristaltic pump and continuously supplied to the anode during the measurements. The NiFe/in-hOPN||PtNi/hOPN and NiFe/NF||PtNi/hOPN cells were evaluated following the same testing protocol. For comparison, a benchmark IrO2||Pt/C cell was assembled by spray-coating commercial IrO2 and Pt/C catalysts onto the anode and cathode sides of the membrane, respectively, with catalyst loadings of 1 mg cm−2 on both sides. This benchmark cell was tested under the same AEMWE configuration and operating conditions as the NiFe/hOPN||PtNi/hOPN cell. Notably, for the nickel foam-based electrode, the compliant nickel fiber felt interfacial layer was omitted and replaced with a thickness-matched PTFE gasket. Lightweight flexible devices assembly and evaluation The pretreated AEM (PiperION® A40) was placed between the two electrodes (2 × 2 cm2) to form the membrane-electrode assembly (MEA). The MEA was then mounted onto thin PET sheets with rounded protrusions, which served as lightweight support plates and facilitated fluid flow. The device was encapsulated with a heat-shrink film to provide sealing, mechanical integrity, and environmental protection, while maintaining ionic and electrical contact under bending and twisting. For solar-driven operation, the flexible electrolyzer was connected to a thin-film photovoltaic module (3.2 V, 2.4 A, 294 × 231 mm2, 32 g) and evaluated under intermittent or dynamically varying solar power input. The current and voltage were measured and recorded using an ammeter connected in series and a voltmeter connected in parallel, respectively. Porosity calculation The experimental porosity of the fabricated hOPN (fexperimental) was estimated from the measured areal mass loading (mA, g cm−2), film thickness (t, cm), and the bulk density of nickel (ρNi, g cm−3) according to following expression: Here, mA/ρNi converts the areal mass loading (mass per unit area) into the equivalent volume of dense Ni per unit area. Dividing by the geometric film volume per unit area (t × 1 cm2) yields the solid volume fraction, and the porosity is obtained as its complement (void volume fraction). This volumetric method is used to quantify porosity in compositionally uniform metallic porous frameworks. The theoretical porosity of an idealized hOPN architecture (ftheoretical) was calculated from the modeled inverse-opal unit cell as: where Vvoid and Vsolid are the void and solid volumes in the modeled inverse-opal unit cell, respectively, and Vtotal is the total unit-cell volume. In the ideal geometry, these volumes can be calculated analytically based on the FCC-packed spherical template and the prescribed neck structure. ftheoretical therefore reflects the intrinsic porosity dictated solely by lattice geometry, independent of fabrication-related deviations. Permeability measurement The permeability was determined from pressure-drop measurements under laminar flow48. Briefly, electrolyte was pumped through the electrode at controlled superficial velocities, and the pressure drop across the electrode was recorded. The permeability was then calculated using Darcy’s law: where u is the superficial velocity, k is the permeability, μ is the electrolyte viscosity, Δp is the pressure drop, and L is the electrode thickness. Effective mass-transfer resistance coefficient The linear region of the LSV curve at 1–2 A cm−2 was first fitted and used as the reference baseline. The deviation between the measured potential in the polarization-bending region and the extrapolated linear baseline was then extracted as the additional potential loss. This additional potential loss was further divided by the corresponding current density to obtain the increased IR. The evolution of this increased IR with current density was analyzed after logarithmic transformation, i.e., log (increased IR) versus j. An approximately linear relationship was observed between log (increased IR) and j. The slope of this fitted line was defined as the effective mass-transfer resistance coefficient. Mass-transfer coefficient measurement The mass-transfer performance was evaluated by flow electrochemical measurements using 1 mM K4Fe(CN)6 in 1 M KCl as the redox-probe electrolyte, following previous reports48,49. The electrolyte was driven through the porous electrode at controlled superficial velocities ranging from 0.01 to 0.20 cm s−1. Electrochemical measurements were performed in a three-electrode configuration using the porous electrode as the working electrode. Steady-state polarization curves were reconstructed by averaging the stable current response at each applied potential under continuous-flow conditions. The limiting current densities were extracted from the plateau regions and used to calculate the mass-transfer coefficient (kma) according to: where U is the superficial velocity, L is the electrode thickness, Ilim is the limiting current, n is the number of transferred electrons, F is the Faraday constant, Q is the flow rate, and cb is the bulk concentration of ferrocyanide. The obtained kma values were fitted as a function of superficial velocity using a power-law relationship. Characterization X-ray diffraction (XRD) patterns were collected on a X’Pert PRO (PANalytical) diffractometer using Cu Kα radiation (λ = 1.5406 Å) operated at 40 kV and 40 mA. Diffraction data were recorded in the 2θ range of 40–80° using continuous scanning with a step size of 0.033°. SEM images were acquired on SU8020 (Hitachi) field-emission scanning electron microscope operated at an acceleration voltage of 10.0 kV. Contact angle measurements were performed using a contact angle goniometer (Drop Shape Analyzer DSA25) via the sessile drop method. Inductively coupled plasma atomic emission spectrometry results were obtained on an ICP6300 (ThermoFisher Scientific, USA). XPS measurements were performed using a Thermo Scientific K-Alpha spectrometer (Thermo Fisher Scientific, USA). Atomic force microscopy (AFM) measurements were carried out using an XploRA Plus system (Horiba), which is a nanospectroscopic platform integrating Raman spectroscopy and AFM. Operando Raman spectroscopy measurements Operando Raman measurements were conducted on a XploRA Plus (Horiba) micro-Raman spectrometer with a 532 nm solid-state laser as the excitation source. The measurements were carried out in a custom-designed spectroelectrochemical cell equipped with a quartz window, allowing Raman signal collection directly from the working electrode surface. The as-prepared NiFe/hOPN served as the working electrode, with a Pt wire and a saturated Hg/HgO electrode used as the counter and reference electrodes, respectively. During the measurements, 1.0 M KOH electrolyte was continuously delivered to the working electrode surface at a flow rate of 2 mL min−1 using a peristaltic pump. Raman spectra were accumulated from four consecutive acquisitions with an integration time of 10 s per acquisition. Finite element simulation Finite element simulations were performed using the Laminar Flow and Transport of Diluted Species to examine mass-transfer behavior within the hierarchical ordered interconnected porous structure. Fluid motion is governed by the momentum conservation equation: together with the incompressibility condition: where u denotes the fluid velocity vector, ρ and μ are the electrolyte density and dynamic viscosity, respectively, p is the hydrodynamic pressure, I is the identity tensor, K represents the viscous stress tensor, and F accounts for body forces acting on the fluid. For a Newtonian fluid, the viscous stress tensor is defined as: Species transport of reactant i was described by the steady-state advection-diffusion equation: in which ci is the concentration of species i, Ji is the diffusive flux, Di is the molecular diffusion coefficient, and Ri represents the contribution of non-surface (bulk) reactions to the species balance. Reactions were assumed to occur only at the electrode surface, Ri = 0 in the bulk electrolyte and the electrochemical consumption was imposed as a boundary flux condition via Faraday’s law. Electrochemical consumption of reactant i at the electrode surface was implemented through the normal total molar flux condition: where n is the outward unit normal vector pointing from the electrode into the electrolyte, j is the applied current density, n is the number of electrons transferred per molecule of reactant i, and F is the Faraday constant. This boundary condition directly couples the prescribed electrochemical rate (j) to the local mass transport, enabling quantitative evaluation of concentration depletion and mass-transfer limitations within the ordered pore network under a given operating current. A superficial inlet velocity of 0.01 m s−1 (uin) was adopted to emulate electrolyte circulation and gas-evolution-induced hydrodynamic disturbances under relatively high reaction flux (≥ 1 A cm−2). The corresponding Reynolds number was well below unity (Re ≪ 1, ~ 10−3), justifying a laminar (creeping-flow) assumption in which viscous forces dominate and inertial effects are negligible. The inlet velocity (uin) and outlet pressure condition were applied as boundary conditions to simulate through-pore flow. The electrode surface was defined as a reactive boundary, while the remaining walls were treated as non-slip (u = 0) and impermeable. At the outlet boundary, the relative pressure was fixed at 0 Pa, serving as the reference pressure for the simulation and allowing the electrolyte to leave the computational domain without imposing an additional flow constraint. The electrolyte was modeled as 1.0 M KOH with representative physical parameters: ρ = 1000 kg m−3, μ = 1 × 10−3 Pa s, D = 1 × 10−9 m2 s−1. The hOPN geometry was represented as a periodic array of interconnected spherical pores and necks, parameterized by the pore diameter of 965 nm (DOH) and neck size of 241 nm (dNH) consistent with experimental observations. A tetrahedral mesh was employed with local refinement near pore walls to ensure numerical accuracy and convergence. The simulations were performed using a 3D model, with gas-phase effects neglected. To enable direct comparison between the finite-element simulation and the experimentally measured mass-transfer coefficients, an additional convection-diffusion simulation was performed using ferrocyanide as the redox-probe species44. The laminar-flow field was directly coupled with the transport of Fe(CN)64−, and the original computational unit of hOPN was periodically extended along the flow direction to reproduce the actual electrode thickness of approximately 12 μm. The electrolyte was modeled as 1 mM K4Fe(CN)6 in 1 M KCl. Owing to the high concentration of supporting electrolyte, migration of Fe(CN)64− was neglected, and its transport was described by convection and diffusion only. The bulk Fe(CN)64− concentration was set as cb = 1 mol m−3, and the diffusion coefficient was set as DFeCN = 6.3 × 10−10 m2 s−1. The inlet Fe(CN)64− concentration was fixed at cb, and superficial velocities of 0.01–0.20 cm s−1 were applied, matching the experimental flow range used in the limiting-current measurements. The outlet was defined as an outflow boundary for species transport. Under the limiting-current condition, the reactive electrode surface was treated as a zero-concentration boundary for Fe(CN)64−, corresponding to complete depletion of the redox species at the electrode surface. All non-reactive walls were treated as no-flux boundaries for Fe(CN)64−. Supplementary information Supplementary Information Transparent Peer Review file Source data Source Data Acknowledgements Numerical computations were performed on Hefei Advanced Computing Center. Author contributions H.Y. conceived the idea and supervised the project. Z.W. and H.Y. designed the experiments. Z.W. carried out the experiments and performed the calculations. Z.W., H.M.Z., and H.Y. analyzed the data. Z.W., H.Y., H.M.Z., and H.J.Z. wrote and revised the manuscript. Peer review Peer review information Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Funding H.Y. discloses support for the research of this work from the National Natural Science Foundation of China [grant number 52472263]. Data availability All relevant data that support the findings of this study are presented in the manuscript and supplementary information file. The experimental datasets associated with this work are provided in the form of Source Data files. Source data are provided with this paper. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Huajie Yin, Email: yinhj@issp.ac.cn. Huijun Zhao, Email: hjzhao@issp.ac.cn. Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-77030-0. References Associated Data Supplementary Materials Supplementary Information Transparent Peer Review file Source Data Data Availability Statement All relevant data that support the findings of this study are presented in the manuscript and supplementary information file. The experimental datasets associated with this work are provided in the form of Source Data files. Source data are provided with this paper.