Spin‐Dominated Reconstruction Kinetics in Prussian Blue Analog Enables Efficient Seawater Splitting

ABSTRACT

The spin state of metal centers is identified as a pivotal descriptor dominating the electrochemical reconstruction kinetics of Prussian blue analogs (PBAs). We demonstrate that constructing an interfacial electric field in MnFeO_x/PBA pre‐catalyst triggers a partial high spin configuration of Fe centers, introducing partial t_2g occupancy and thereby substantially enhancing metal to ligand π backdonation. This electronic perturbation destabilizes the cyanide ligand field, rendering the entire coordination framework susceptible to rapid electrochemically driven transformation into active (oxy)hydroxide phases under OER conditions. The reconstructed catalyst (R‐MnFeO_x/PBA) exhibits markedly accelerated reconstruction kinetics and delivers ultralow overpotential of 201 mV to reach 10 mA cm^‒2 in simulated seawater. Crucially, R‐MnFeO_x/PBA enabled alkaline membrane electrode assemblies (MEAs) affords Faraday efficiencies of up to 97.6% and can operate stably for 1150 h at 1000 mA cm^‒2. This work establishes spin‐state engineering as a unique pre‐catalyst design strategy that bridges electronic structure manipulation with reconstruction‐enhanced electrocatalysis.

This work presents a spin‐state engineering, activated by built‐in electric fields (BIEF) in heterojunction pre‐catalysts, as a powerful and universal method to accelerate electrochemical reconstruction, resulting in breakthrough performance and stability for seawater electrolysis.

1. Introduction

Electrochemical overall seawater splitting powered by intermittent renewable energy is a pivotal pathway toward zero‐carbon energy systems. It converts sustainable power into hydrogen, an energy carrier capable of seasonal storage, long‐distance transportation, and the decarbonization of hard‐to‐electrify industries [1, 2]. Among the emerging electrolysis technologies, alkaline membrane electrode assemblies (MEAs) are especially promising. Their alkaline operating environment enables the use of cost‐effective non‐precious catalysts, while still achieving the high current densities and efficient gas separation typically associated with proton exchange membrane systems [3, 4]. Nevertheless, the sluggish kinetics of the oxygen evolution reaction (OER) arise from its complex four‐electron proton‐coupled transfers, resulting in high overpotentials that severely undermine energy conversion efficiency. Overcoming this bottleneck calls for electrocatalysts that integrate exceptional intrinsic activity with industrial‐grade durability [5, 6]. While noble metal oxides exhibit benchmark OER activity, their scarcity and structural degradation under industrial conditions drive the exploration of transition metal‐based alternatives [7, 8].

Transition metal‐based catalysts demonstrate particular advantages owing to their earth abundance, tunable d‐band electronic configurations, and excellent stability in alkaline media [9, 10, 11]. Crucially, transition metal‐based catalysts undergo surface transformations under anodic polarization in alkaline electrolytes, evolving into species with enhanced water oxidation activity [12, 13]. Due to this inherent electrochemical reconstruction under operational OER potentials, the initial materials are aptly termed “pre‐catalysts” [14]. This dynamic and irreversible reconfiguration process typically culminates in the formation of transition metal oxyhydroxides (MOOH), which serve as the true catalytically active species [15].

Prussian blue analogues (PBAs) are considered a promising class of pre‐catalyst materials, owing to their open metal organic framework (MOF) structure in which nitrogen‐coordinated metal cations and carbon‐coordinated Fe sites are linked by cyanide (CN^‒) bridges [16]. These cyanido‐bridged bimetallic three‐dimensional networks, commonly denoted as M_x[Fe(CN)₆]·nH₂O, consist of repeating Fe^III‒C≡N‒Fe^II units (Figure 1a) [17]. The cyanide ligands in PBAs undergo cleavage and transform into active (oxy)hydroxide phases under OER conditions [18, 19]. Previous studies have identified the catalytic active sites in PBAs, confirming that the amorphous hydroxide formed after OER exhibits catalytic activity comparable to that of the metal oxide involved in OER [20, 21]. However, in such conventional PBA units, the low‑spin Fe^III (t_2g5) and Fe^II (t_2g6) centers possess nearly fully occupied t_2g orbitals, resulting in extremely limited π* backdonation capability, rendering the cyano group robust and making the entire framework resistant to electrochemically driven dissociation and recombination. Accordingly, accelerating the inherently sluggish reconstruction kinetics imposed by strong Fe─CN bonds, which is crucial for timely active phase generation, remains a critical challenge.

FIGURE 1.

(a) Cyanide localized structure and crystal structure. (b) Schematic illustration of the spin state regulation (from LS state into IS state). (c) Bond length of Fe–C of PBA and Mn_0.4Fe_1.6O₃/PBA. (d) The differential charge density of Mn_0.4Fe_1.6O₃/PBA. (e) e_g and t_2g orbital density of states (DOS) of PBA and Mn_0.4Fe_1.6O₃/PBA. (f) Schematic diagram illustrating the reduced t_2g orbital energy levels combined with the increased π* orbital occupancy.

Herein, we establish that the spin state of metal centers serves as a decisive descriptor for controlling the reconstruction kinetics of PBAs. By constructing an interfacial electric field in a MnFeO_x/PBA heterostructure, we induce partial high spin Fe^III sites, which introduce partial occupancy into the t_2g orbitals and substantially enhance metal to ligand π backdonation (Figure 1b). This electronic perturbation destabilizes the cyanide ligand field, rendering the coordination framework susceptible to rapid electrochemical transformation into a highly active amorphous FeOOH phase under OER conditions. Consequently, the reconstructed catalyst (R‐MnFeO_x/PBA) delivers exceptional performance, requiring only 201 mV overpotential at 10 mA cm^‒2 in simulated seawater and demonstrating outstanding long‐term stability exceeding 1150 h at 1000 mA cm^‒2 within an assembled alkaline membrane electrode (MEAs). Most importantly, this work elucidates spin state engineering driven by interfacial electric fields, which can effectively accelerate the electrochemical reconstruction kinetics of PBA pre‐catalysts toward active (oxygen) hydroxide phases. The demonstrated strategy for harnessing surface reconstruction provides a theoretical basis for the design of pre‐catalysts with controllable reconstruction for electrochemical seawater oxidation.

2. Results and Discussion

To elucidate the electronic origin of this interfacial‐field‐induced spin state conversion and its impact on cyanide bonding, we performed density functional theory (DFT) calculations. We constructed three slab models: pristine PBA, Fe₂O₃/PBA heterostructure, and Mn_0.4Fe_1.6O₃/PBA heterostructure. The Fe₂O₃/PBA interface serves as a reference to isolate the effect of band alignment‐induced electric field, while Mn_0.4Fe_1.6O₃/PBA represents the targeted pre‐catalyst design. Upon heterojunction formation, the Fe─C bond length in PBA exhibits a marked elongation from 1.86 Å in pristine PBA to 2.21 Å in Mn_0.4Fe_1.6O₃/PBA (Figure 1c). This significant increase directly indicates a weakened Fe–C≡N coordination, suggesting that the cyanide ligand field has been destabilized. Differential charge density analysis reveals pronounced electron depletion at the PBA side and accumulation at the Mn_0.4Fe_1.6O₃ side, demonstrating that electrons spontaneously transfer from the PBA framework toward the oxide component (Figures 1d and S1). Bader charge analysis quantifies this transfer: compared to pristine PBA, the Fe centers in Mn_0.4Fe_1.6O₃/PBA lose ≈1.8 e, while a smaller loss is observed in Fe₂O₃/PBA (Figure S2). The directional electron withdrawal effectively reduces the electron density at Fe 3d orbitals, which is the initial trigger for spin‐state conversion.

Projected density of states (PDOS) analysis provides deeper insight into the electronic consequences of this charge transfer [22]. In Mn_0.4Fe_1.6O₃/PBA, the t_2g band of Fe 3d shifts downward and broadens compared to that in pristine PBA or Fe₂O₃/PBA (Figure 1e). This downshift lowers the crystal field splitting energy (Δ₀), and, more importantly, introduces partial occupancy into the t_2g orbitals, which is a hallmark of the transition from low‐spin to high‐spin configuration [23]. The partially filled t_2g orbitals exhibit enhanced spatial overlap with the π* antibonding orbitals of the cyanide ligands, thereby strengthening metal‐to‐ligand π backdonation [24]. Elevated π* occupancy injects additional electron density into the antibonding manifold of C≡N, which directly weakens the Fe─C bond. Consistently, the PDOS also shows reduced hybridization between Fe 3d and C 2p states in Mn_0.4Fe_1.6O₃/PBA, corroborating the observed Fe–C bond elongation. The projected d band center of the Fe t_2g orbitals shifts from –1.23 eV in pristine PBA to –1.29 eV in Mn_0.4Fe_1.6O₃/PBA (Figure S3), reflecting an energetically optimized alignment with the π* orbitals of cyanide ligands that underpins the enhanced π backdonation [25]. This ligand‐field destabilization renders the entire PBA unit susceptible to rapid anodic dissociation and in situ conversion into the active FeOOH phase, thereby accelerating the reconstruction kinetics that underpin the outstanding OER performance of R‐MnFeO_x/PBA (Figure 1f).

Inspired by the aforementioned theoretical prediction, we synthesized MnFeO_x/PBA pre‐catalyst to verify the proposed models (Figure S4). Briefly, MnFeO_x nanorods were grown on nickel foam and subsequently calcined. PBA shell was then deposited onto the MnFeO_x nanorods through a hydrothermal process. Raman spectrum confirmed PBA successfully loaded on MnFeO_x nanorods, evidenced by characteristic C─N stretching vibrations between 2000 and 2200 cm^‒1 (Figure S5) [26]. The XRD pattern of MnFeO_x/PBA shows the characteristic peaks located at 24.12°, 49.41°, and 75.40°, which can be attributed to the (012), (024), and (220) crystal planes of Fe₂O₃ (JCPDS 72‐0469) (Figure 2a). Notably, these diffraction peaks shift to the lower angles relative to pure Fe₂O₃, indicating lattice expansion consistent with Mn‐doping (Figure S6) [27]. XPS spectrum analysis confirms that the Mn doping amount is consistent with the Mn content in the theoretical calculation model. (Table S1) Concurrently, the characteristic peaks located at 17.49°, 24.83°, and 39.76° are indexed to the (200), (220), and (420) crystal planes of Fe₄(Fe(CN)₆)₃ (JCPDS 73‐0687) respectively, and the heterostructure interfaces of MnFeO_x/PBA have been successfully established. Scanning electron microscopy (SEM) shows PBA wrapped‐MnFeO_x nanorods with an average height of 3‐4 µm are uniformly distributed on nickel foam, forming multi‐dimensionally assembled nanostructure (Figures 2b and S7). Transmission Electron Microscope (TEM) images further reveal a distinct ≈9 nm PBA layer coating the MnFeO_x nanorods, contrasting with the pristine MnFeO_x surface (Figure 2c). High‐resolution TEM (HRTEM) image of the interface region resolves lattice fringes with spacings of 2.97 Å and 2.55 Å, corresponding to the PBA (222) and MnFeO_x (110) planes, respectively, verifying the successful establishment of the interface of the heterogeneous structure of MnFeO_x/PBA (Figure 2d). The results can be further confirmed by the fast Fourier transformation (FFT) patterns. The inverse fast Fourier transform (IFFT) analysis and atomic structure simulations provide additional evidence for the well‐defined heterointerface (Figure S8). Finally, these images suggest epitaxial growth of the PBA (220) plane onto MnFeO_x facets, facilitated by compatible lattice parameters. The energy dispersive spectrometer (EDS) elemental mapping analysis of MnFeO_x/PBA reveals the homogeneity of the heterogeneous structure by the uniform distribution of the Mn, Fe, O, and N elements (Figure S9). The selected area electron diffraction (SAED) patterns of MnFeO_x/PBA exhibit the Debye‐Scherrer rings corresponding to the crystal planes (222) of the PBA and (110) of MnFeO_x, respectively (Figure S10) [28].

FIGURE 2.

(a) XRD patterns of MnFeO_x/PBA. (b) SEM image of MnFeO_x/PBA. (c) TEM image of MnFeO_x/PBA. (d) HR‐TEM image of MnFeO_x/PBA heterostructure. (e) UPS spectra of MnFeO_x and PBA. (f) Magnetic susceptibility of MnFeO_x and MnFeO_x/PBA. (g) Normalized Fe K‐edge XANES spectra of Fe₂O₃/PBA and MnFeO_x/PBA. (h) Fe K‐edge FT‐EXAFS spectra of MnFeO_x and MnFeO_x/PBA. (i) Wavelet‐transformed k³‐weighted EXAFS spectra of MnFeO_x and MnFeO_x/PBA.

Generally speaking, the electron transfer behavior is closely related to the work function difference (ΔΦ) in semiconductor heterostructure [29]. Ultraviolet photoelectron spectroscopy (UPS) determines work functions (Φ) of 5.14 eV for MnFeO_x and 4.57 eV for PBA (Figures 2e and S11), yielding a significant interfacial difference (ΔΦ = 0.57 eV). The larger Φ of MnFeO_x facilitates localized charge redistribution at the heterointerface, creating electrophilic regions on PBA that are beneficial for producing more high‐valent Fe. Tauc plot analysis confirms that an increased band gap (E _g) of MnFeO_x and PBA are 1.54 eV (Fe₂O₃) to 1.72 eV (MnFeO_x) (Figure S12) [30]. Schottky curve analysis reveals that MnFeO_x exhibits a charge density increase of one order‐of‐magnitude compared to the pristine Fe₂O₃, confirming Mn doping‐induced energy level restructuring (Figure S13). Furthermore, the space charge region width (W) of the MnFeO_x/PBA hybrids is calculated (Figure S14) [31]. The enhanced interfacial electron exchange within the MnFeO_x/PBA heterojunction, which increases the built‐in electric field intensity by 164%, as quantified via surface voltage and accumulated electron density (Figure S15). The derived schematic illustration of band structure further demonstrates that the large ΔΦ between MnFeO_x and PBA offers a strong thermodynamic driving force for electron transfer from PBA to MnFeO_x, spontaneously establishing a directional BIEF. This directed BIEF promotes charge separation and the formation of electron‐rich MnFeO_x regions and electron‐deficient PBA regions.

Based on the magnetization‐versus‐temperature plots measured under zero‐field‐cooling conditions with an applied magnetic field of 300 Oe, we conduct further analysis to examine the changes in spin states after the formation of the heterojunction. The Néel temperature varied continuously from 2 to 200 K confirming the enhanced exchange interaction within the samples [32]. By fitting the 1/χm versus T curve (Figure 2f), the unpaired electron numbers can be obtained according to the following equation:

$${\mu }_{eff}=\sqrt{n\left(n+2\right)}$$

where χm is the magnetic susceptibility, µ_eff is effective magnetic moment (µ_B), n is unpaired d‐electrons, and T is absolute temperature (K). The effective magnetic moments of Fe in the PBA and MnFeO_x/PBA samples are measured as 1.62 and 2.55 µ_B, respectively, and thus the unpaired d‐electron number (n) is calculated as 0.91 for PBA and 1.74 for MnFeO_x/PBA (Figure S16). The increased unpaired electrons in MnFeO_x/PBA directly evidence the transition from low‐spin to high‐spin Fe centers, confirming the consistency with the results predicted by DFT. X‐ray photoelectron spectroscopy (XPS) was further carried out to analyze the chemical valence states and elemental compositions of MnFeO_x/PBA, MnFeO_x, and PBA. The MnFeO_x/PBA heterostructure exhibits a positive binding energy shift for Fe 2p and a negative shift for Mn 2p compared to pristine MnFeO_x, confirming interfacial electron transfer from PBA to MnFeO_x (Figure S17) [33]. In summary, the substantial work function difference between MnFeO_x and PBA drives interfacial electron transfer, which in turn modulates the ligand field and spin state of Fe ions. This electronic and spin restructuring is critical for optimizing the adsorption energetics of reaction intermediates, thereby contributing to the enhanced electrocatalytic performance of the MnFeO_x/PBA heterostructure.

The electronic states of MnFeO_x/PBA was further investigated using x‐ray absorption fine structure (XAFS) spectroscopy. X‐ray absorption near edge structure (XANES) analyses of MnFeO_x and MnFeO_x/PBA are compared with control samples of Fe foil and Fe₂O₃ (Figure 2g). The XANES shows that the Fe K‐edge absorption energies of MnFeO_x/PBA are closer to Fe₂O₃, signifying the average valence state of Fe is close to +3. Specifically, the oxidation states of Fe in MnFeO_x and MnFeO_x/PBA are estimated to be approximately +2.4 and +2.8, respectively, based on the linear relationship between oxidation states and the Fe K‐edge white line area (Figure S18). This indicates that the MnFeO_x/PBA heterogeneous interfaces facilitate the oxidation of Fe sites, which may stem from the charge redistribution at the heterogeneous interface. To gain deeper insight into the atomic‐scale coordination, extended x‐ray absorption fine structure (EXAFS) analysis was performed. The Fourier‐transformed EXAFS spectra of MnFeO_x exhibit a main peak at ∼1.41 Å, attributed to the first‐shell Fe–O coordination (Figure 2h). Upon forming the MnFeO_x/PBA heterostructure, this Fe–O peak shifts slightly to ∼1.44 Å, indicating a modest elongation of the Fe─O bond (Figures S19, S20). Concurrently, a new peak emerges at ∼1.61 Å, corresponding to Fe─N bonds from the PBA framework, confirming the successful interfacial bonding between MnFeO_x and PBA. The EXAFS fitting results further quantify these changes: the coordination number (CN) of Fe–O decreases from 5.2 in MnFeO_x to 4.8 in MnFeO_x/PBA while Fe–N coordination appears with a CN of ∼1.2 (Tables S2, S3). This local coordination adjustment implies electron redistribution toward Fe sites, potentially creating electron‐rich Fe centers that are beneficial for adsorbing reaction intermediates. The wavelet‐transform (WT) of the EXAFS spectra provides a resolution‐enhanced visualization of the coordination shells (Figure 2i). MnFeO_x shows a single intensity maximum at ∼4.2 Å^‒1 in k‐space, corresponding to Fe─O scattering. In contrast, MnFeO_x/PBA exhibits an additional contour at ∼5.5 Å^‒1, assigned to Fe–N/Fe–C paths, further confirming the coexistence of oxide and cyanide coordination environments. The shift of the Fe–O WT maximum to a lower k‐value in MnFeO_x/PBA aligns with the elongated Fe─O bond observed in the EXAFS fitting, reinforcing the modified electronic environment induced by heterojunction formation.

The MnFeO_x/PBA electrode underwent in situ reconstruction via 50 cyclic voltammetry (CV) cycles in 1.0 M KOH medium at 50 mV s^‒1 from 0.90 to 1.80 V vs. reversible hydrogen electrode (RHE), yielding R‐MnFeO_x/PBA (Figure 3a). The area of oxidation peak and the OER current increase with the cycling numbers. R‐MnFeO_x/PBA exhibits significantly enhanced OER activity, requiring 81.7 mV lower overpotentials than MnFeO_x/PBA to achieve 100 mA cm^‒2 current density (Figures 3b–c). Morphological analysis confirms that R‐MnFeO_x/PBA retains the nanorod framework, but the edge portion is transformed into nanosheets (Figure S21). The determination of the contact angle (CA) reveals that the contact angle of the reconstructed sample R‐MnFeO_x/PBA (43.6°) is significantly lower than MnFeO_x/PBA (76.1°), indicating that the R‐MnFeO_x/PBA catalyst has significant hydrophilicity (Figure S22). XRD analysis shows complete disappearance of PBA diffraction peaks while Fe₂O₃ reflections persisted (Figure S23), confirming transformation of PBA into an indeterminate phase during reconstruction.

FIGURE 3.

(a) CV curves of MnFeO_x/PBA recorded at a scan rate of 50 mV s^‒1 in 1 M KOH during first 50 cycles. (b) Polarization curves before and after reconstruction. (c) Comparison of overpotential values of MnFeO_x/PBA at different current densities before and after reconstruction. (d) OER polarization curves of MnFeO_x/PBA, R‐Fe₂O₃/PBA, R‐MnFeO_x/PBA and RuO₂. (e) Tafel slopes comparison of various electrocatalysts. (f) Comparison of the overpotentials and Tafel slopes with those reported non‐precious metal electrocatalysts. (g) Chronoamperometry measurement of the R‐MnFeO_x/PBA at 1.5 and ‒0.2 V versus RHE for 400 h.

The electrochemical performance of the as‐prepared catalysts was evaluated by a standard three‐electrode system in simulated seawater electrolyte (1.0 M NaOH + 0.5 M NaCl). The R‐Fe₂O₃/PBA was prepared from pre‐catalyst Fe₂O₃/PBA by the same reconstruction method as MnFeO_x/PBA (Figure S24). And the R‐Fe₂O₃/PBA, MnFeO_x/PBA and RuO₂ are adopted as the reference. Linear sweep voltammetry (LSV) curves (Figure 3d) reveal a dramatic performance enhancement upon reconstruction: R‐MnFeO_x/PBA requires ultralow overpotentials of only 201 mV and 246 mV to reach 10 mA cm^‒2 and 100 mA cm^‒2, respectively, significantly lower than its unreconstructed precursor MnFeO_x/PBA, demonstrating the necessity of the reconstruction process for achieving high activity. Importantly, the superiority of R‐MnFeO_x/PBA over the R‐Fe₂O₃/PBA (243 mV and 312 mV) directly attributes the exceptional activity to the Mn‐doping induced 164% BIEF enhancement, which steers the reconstruction pathway towards a more active phase. This kinetic advantage is further confirmed by Tafel analysis (Figure 3e), where R‐MnFeO_x/PBA exhibits the lowest slope (48.9 mV dec^‒1), substantially outperforming R‐Fe₂O₃/PBA (128.2 mV dec^‒1), MnFeO_x/PBA (106.7 mV dec^‒1), and RuO₂ (95.2 mV dec^‒1), signifying accelerated OER kinetics enabled by the BIEF‐optimized reconstruction. In addition, the exceptional OER activity of R‐MnFeO_x/PBA is overall better than that of RuO₂ noble metal catalysts and superior to advanced reported non‐precious metal catalysts (Figure 3f and Table S4) [34]. Electrochemical impedance spectroscopy (EIS) corroborates the superior interfacial charge transfer (Figure S25), showing R‐MnFeO_x/PBA possesses the smallest charge transfer resistance (R _ct, 0.13 Ω), significantly lower than R‐Fe₂O₃/PBA (0.68 Ω), MnFeO_x/PBA (0.76 Ω), and RuO₂ (0.45 Ω), reflecting the highly efficient charge exchange facilitated by the BIEF‐driven reconstructed active surface.

The electrochemical active surface area (ECSA) was evaluated using the electric double‐layer capacitance (C _dl) collected by successive CV scans at different scan rates in the non‐Faradaic regions (Figure S26). The R‐MnFeO_x/PBA affords the largest non‐Faradaic double‐layer capacitance (C _dl, 5.24 mF cm^‒2), noticeably larger than those of R‐Fe₂O₃/PBA (3.72 mF cm^‒2), MnFeO_x/PBA (3.62 mF cm^‒2), and RuO₂ (1.65 mF cm^‒2), which means that the presence of restructured heterointerface can provide more accessible active sites for improved OER performance (Figure S27). The ECSA of R‐MnFeO_x/PBA and R‐Fe₂O₃/PBA are further calculated through electrochemical double‐layer capacitance (EDLC), and the results show that the active specific surface area of R‐MnFeO_x/PBA with 131.0 cm^‒2 exceeds that of R‐Fe₂O₃/PBA (93.0 cm^‒2), MnFeO_x/PBA (90.5 cm^‒2), and RuO₂ (41.3 cm^‒2), indicating the enhanced inherent activity of R‐MnFeO_x/PBA (Figure S28). The large C _dl and small R _ct illustrate the more increased active sites, further underpinning the remarkable OER activities in LSV results [35]. Moreover, the R‐MnFeO_x/PBA sample exhibits excellent stability at different potentials of 1.43 and 1.47 V vs. RHE, and negligible activity attenuation for up to 200 h in simulated seawater electrolyte (Figure S29). In particular, the durability testing of the R‐MnFeO_x/PBA electrode at a current density of 100 mA cm^‒2 in simulated alkaline seawater for OER and HER show <3 % potential degradation over 400 h, which is much superior to those of Pt/C and RuO₂ (Figure 3g). Additionally, the turnover frequency (TOF) of R‐MnFeO_x/PBA at 1.5 V vs. RHE are calculated to be 16.5 s^‒1. These values are significantly higher than those of the R‐Fe₂O₃/PBA (3.31 s^‒1) and RuO₂ (2.57 s^‒1) samples (Figure S30). Radar chart analysis integrating key metrics (η₁₀, η₁₀₀, Tafel slope, EIS, ECSA, TOF) confirm that R‐MnFeO_x/PBA superior overall OER performance (Figure S31).

R‐MnFeO_x/PBA also delivers outstanding HER performance, requiring only 121 mV overpotential at 10 mA cm^‒2 (Figure S32 and Table S5), outperforming R‐Fe₂O₃/PBA (173 mV) and MnFeO_x/PBA (200 mV). Moreover, the Tafel slope for R‐MnFeO_x/PBA (64.7 mV dec^‒1) is smaller than that for R‐Fe₂O₃/PBA (98.9 mV dec^‒1), which indicates that fast reaction kinetics (Figure S33). The R _ct value of R‐MnFeO_x/PBA is ≈1.50 Ω during HER, which is smaller than that of R‐Fe₂O₃/PBA (4.93 Ω) and MnFeO_x/PBA (6.39 Ω), verifying the fast electron transport (Figure S34). As expected, R‐MnFeO_x/PBA shows exceptional stability, with minimal degradation of its initial potential after hundreds of hours of long‐term chronoamperometry at ‒0.13 and ‒0.17 V versus RHE (Figure S35). The cycling stability of the R‐MnFeO_x/PBA catalyst was further tested with continuous CV scanning. The polarization curve after 5000 cycles nearly overlaps with the initial one, demonstrating outstanding cyclability (Figure S36). Furthermore, R‐MnFeO_x/PBA exhibits a TOF similar to that of Pt/C during the HER, confirming its superior intrinsic activity (Figure S37) [36]. Based on the coverage area in the radar chart and stability test, the R‐MnFeO_x/PBA electrocatalyst exhibits the most comprehensive HER performance (Figure S38). While its catalytic activity remains comparable to that of Pt/C, it significantly improves both the utilization efficiency of precious metals and long‐term stability. After 100 h stability test, the electrocatalyst could still maintain its basic structure and surface chemical state, further demonstrating its stability (Figure S39, S40). Meanwhile, we compared the performance indicators of the R‐MnFeO_x/PBA catalyst in simulated seawater and 1.0 M KOH. The performance of the catalyst remains almost unchanged, still maintaining bifunctional activity in the OER and HER, indicating that R‐MnFeO_x/PBA has excellent OER selectivity as well as resistance to chlorine interference (Figures S41–S44).

In situ Raman spectroscopy elucidates the BlEF‐dependent electrochemical reconstruction dynamics of R‐Fe₂O₃/PBA and R‐MnFeO_x/PBA under anodic potentials (Figure S45). For both catalysts, the Raman bands at 338 and 446 cm^‒1 correspond to Fe‒O stretching vibrations in the crystalline iron oxide phase [37]. Compared to R‐Fe₂O₃/PBA, the Fe‒O band of R‐MnFeO_x/PBA exhibits a slight weakening and shift, attributable to Mn introduction modifying the Fe‒O coordination environment. As the applied anodic potential increases, two additional Raman vibrations emerge at 269 and 708 cm^‒1, assigned to Fe─O bonds in α‐FeOOH and β‐FeOOH, respectively (Figures 4a and S46) [38].

FIGURE 4.

In situ Raman spectra of (a) R‐Fe₂O₃/PBA, (b) R‐MnFeO_x/PBA and (c) R‐MnFeO_x/PBA in the range of 2000‐2200 cm^‒1. (d) Normalized Fe K‐edge XANES of R‐MnFeO_x/PBA, R‐Fe₂O₃/PBA, MnFeO_x/PBA and Fe foil, respectively. (e) Fe K‐edge FT‐EXAFS spectra of R‐MnFeO_x/PBA, R‐Fe₂O₃/PBA and Fe foil, respectively. (f) Wavelet‐transformed k³‐weighted EXAFS spectra of R‐Fe₂O₃/PBA and R‐MnFeO_x/PBA. g) Calculated Gibbs free energy diagrams for OER. (h) Calculated H*adsorption Gibbs free energy diagrams for HER.

Notably, the Fe‒O intensity in the β‐FeOOH characteristic band of R‐MnFeO_x/PBA appears at lower potentials (<1.47 V vs. RHE) compared to R‐Fe₂O₃/PBA, indicating reduced reconfiguration polarization for the Mn‐modified pre‐catalyst (Figure 4b). Concurrently, the Fe‒CN‒Fe vibration band (2050‒2200 cm^‒1) progressively diminishes and disappears, confirming complete destruction of the PBA crystal structure and transformation to FeOOH (Figure 4c). In situ electrochemical impedance spectroscopy was used to study the interfacial electrochemical kinetics during catalysis. When the potential reaches 1.42 V versus RHE, the phase angle of R‐MnFeO_x/PBA decreases significantly compared to R‐Fe₂O₃/PBA (1.62 V vs. RHE). Moreover, the OER phase angle of R‐MnFeO_x/PBA is much smaller than that of R‐ Fe₂O₃/PBA at the same potential (Figure S47). This indicates that there is a faster reaction kinetics during the reaction of R‐MnFeO_x/PBA electrode, and is in agreement with the results of in situ Raman spectral analysis. The Fe K‐edge XANES spectra indicate a broadening of the pre‐edge peak (≈7110 eV) for both R‐Fe₂O₃/PBA and R‐MnFeO_x/PBA relative to Fe₂O₃, consistent with increased structural disorder in the Fe─O bonds following reconstruction (Figure 4d) [39, 40]. The R‐MnFeO_x/PBA catalyst exhibits markedly different k‐space oscillations compared to R‐Fe₂O₃/PBA and Fe₂O₃. This indicates that electrochemical reconfiguration and the formation of FeOOH significantly modify the local atomic environment of Fe sites in R‐MnFeO_x/PBA (Figures S48, S49). The extended EXAFS spectrum of R‐MnFeO_x/PBA shows a prominent peak at 1.51 Å, assigned to Fe‒O scattering. This distance is longer than that observed in R‐Fe₂O₃/PBA (1.44 Å) (Figure 4e and Table S6, S7). Similarly, the WT contour plot for the k³‐weighted Fe K‐edge of R‐MnFeO_x/PBA displays a single intensity maximum at a longer bond length compared to R‐Fe₂O₃/PBA, further confirming the altered Fe‒O coordination (Figure 4f). Linear fitting of the Fe K‐edge XANES first‐derivative spectra reveals higher average Fe oxidation states in R‐MnFeO_x/PBA (+4.8) compared to R‐Fe₂O₃/PBA (+4.2), confirming more radical electrochemical reconstruction in R‐MnFeO_x/PBA (Figure S50) [41].

DFT calculations were performed to elucidate the electronic structure of R‐MnFeO_x/PBA and its role in enhancing HER/OER activities. Based on HRTEM results, we modeled the heterostructure of Fe₂O₃/FeOOH and MnFeO_x/FeOOH (Figures S51–S54). Plane‐averaged charge density difference calculations confirmed electron transfer from MnFeO_x to FeOOH creates an electron‐enriched space charge region (Figure S55). Charge density analysis revealed significant interfacial charge transfer, consistent with the work function difference between MnFeO_x and PBA that drives built‐in electric field (BIEF) formation upon heterojunction formation. Density of states (DOS) shows R‐MnFeO_x/FeOOH exhibits higher electronic states near the Fermi level than Fe₂O₃/FeOOH, indicating enhanced charge transfer capability (Figure S56). The d‐band center of active Fe sites in MnFeO_x/FeOOH (–1.20 eV) shifts closer to the Fermi level compared to Fe₂O₃/FeOOH (–1.39 eV), facilitating intermediate adsorption through antibonding orbital modulation (Figure S57). The upward shift of the d‐band center will cause its antibonding orbitals to transition to higher energy levels, which is conducive to enhancing the adsorption between the catalyst and the intermediate product, reducing the activation energy barrier [42, 43]. The catalytic activity is evaluated by calculating the detailed pathways of OER and HER. The primitive steps of the OER process on the optimized models including *OH, *O, and *OOH intermediates are outlined (Figure S58). The water dissociation step (from *O to *OOH) is the rate‐determining step (RDS) for two heterojunction catalysts. The reaction energy barriers involved in RDS for MnFeO_x/FeOOH heterostructure under alkaline conditions is only 1.70 eV, being much lower than that of Fe₂O₃/FeOOH (2.07 eV), suggesting that MnFeO_x/FeOOH is more thermodynamically favorable for OER (Figure 4g). It is well established that the Gibbs free energy of hydrogen adsorption (ΔG_H*) proves to be a reliable descriptor for the intrinsic HER activity evaluation, where a near‐zero ΔGH* value could balance the hydrogen adsorption/desorption processes to maximize the HER performance [44, 45]. The optimization models for *H adsorption on Fe₂O₃/FeOOH and MnFeO_x/FeOOH were established (Figure S59). The ΔG_H* of MnFeO_x/FeOOH (0.73 eV) is lower than Fe₂O₃/FeOOH (0.82 eV), which indicates the favorable H* adsorption kinetics during HER process, aligning well with the experimental results (Figure 4h). These computational results confirm that the BIEF‐mediated oxidative reconstruction of MnFeO_x/PBA optimizes the electronic configuration of MnFeO_x/FeOOH, synergistically enhancing both OER and HER activities through facilitated charge transfer and optimized intermediate adsorption energetics.

To evaluate the practical application potential of R‐MnFeO_x/PBA in alkaline membrane electrode assemblies (MEAs), we assembled catalysts‐coated membrane (CCM) and constructed a flow‐type AEM with simulated seawater electrolyzer. This CCM employed R‐MnFeO_x/PBA as both the cathode and anode catalyst, paired with an anion exchange membrane (FAA‐3‐50) and titanium fixtures (Figure 5a and Table S8). The polarization curves were tested in the flow cell and benchmarked against MEAs using commercial Pt/C cathode and RuO₂ anode (Figure 5b). In simulated seawater at 60 °C, the R‐MnFeO_x/PBA‐based MEA achieves current densities of 500, 1000, and 2000 mA cm^‒2 at applied voltages of 1.75, 1.92, and 2.09 V, respectively, exceeding that of most reported metal‐based electrocatalysts (Figure 5c). In contrast, the Pt/C||RuO₂ system requires significantly higher voltages to achieve the same current densities. Furthermore, the Faraday efficiency (FE) of R‐MnFeO_x/PBA for O₂ reaches 97.6% (Figure S60). Furthermore, the R‐MnFeO_x/PBA||R‐MnFeO_x/PBA cell demonstrates exceptional stability, operating continuously for over 1150 h at 1000 mA cm^‒2 without significant performance decay (Figure 5d). R‐MnFeO_x/PBA enabled flow‐type AEM shows superior stability and high operation current densities compared that of based almost reported‐catalysts (Figure 5e and Table S9). Replacing the oxygen evolution reaction (OER) with the urea oxidation reaction (UOR) further improves the electrolysis efficiency. In simulated urea wastewater, the R‐MnFeO_x/PBA||R‐MnFeO_x/PBA system under a UOR||HER configuration reduces the cell voltage by 0.3 V relative to conventional seawater splitting, reflecting improved thermodynamic favorability (Figures 5f and S61). And the stable operation persists for over 95 h at 1.49 V without degradation (Figure 5g).

FIGURE 5.

(a) Schematic diagram of CCM. (b) Polarization curves of the AEMWE device based on the R‐MnFeO_x/PBA. (c) Cell voltages for simulated seawater electrolysis of R‐MnFeO_x/PBA at 500, 1000, and 2000 mA cm^‒2. (d) Stability curve of R‐MnFeO_x/PBA in flow cell at 1000 mA cm^‒2. (e) Comparison of cell voltage at 1000 mA cm^‒2 and stability of R‐MnFeO_x/PBA||R‐MnFeO_x/PBA with reported catalysts. (f) Polarization curve of urea oxidation replacing OER. (g) Stability curve of hydrogen production by electrolyzing seawater with UOR coupled with HER.

3. Conclusion

In conclusion, we identify the Fe spin state as a fundamental descriptor dominating the electrochemical reconstruction kinetics of prussian blue analog (PBA)‐based pre‐catalysts. The interfacial electric field in MnFeO_x/PBA induces partial high‑spin Fe^III sites with partially occupied t_2g orbitals, substantially enhancing metal‑to‑ligand π backdonation. Quantitatively, this electronic perturbation is reflected in a downshifted t_2g d‑band center (from –1.23 to –1.29 eV) and reduced Fe–C hybridization, which collectively destabilize the cyanide ligand field and render the Fe–C≡N framework kinetically labile. Consequently, the pre‐catalyst undergoes rapid in situ conversion into an amorphous FeOOH active phase, yielding a reconstructed catalyst (R‐MnFeO_x/PBA) that achieves an ultralow overpotential of 201 mV at 10 mA cm⁻² in simulated seawater and maintains stable operation for over 1150 h at 1000 mA cm^‒2 in alkaline membrane electrode assemblies (MEAs). By elucidating the electronic origin of reconstruction kinetics through DFT calculation, this work establishes interfacial electric field triggered spin state engineering as a mechanistically grounded and generalizable strategy for pre‐catalyst design, offering a foundational blueprint for sustainable seawater electrolysis.

Author Contributions

Zi‐Qi Ge: methodology, software, data curation, formal analysis, validation, investigation, writing – original draft. Shaokuan Chen: methodology, software, data curation, formal analysis, validation. Hui‐Jian Zhang: investigation, visualization, validation, data curation. Chunbo Liu: methodology, resources, conceptualization, funding acquisition. Xiao‐Tong Wang: writing – original draft, formal analysis, investigation, conceptualization. Zhao‐Qing Liu: conceptualization, writing – review and editing, funding acquisition, project administration, resources, supervision.

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File: anie72624‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that supports the findings of this study are available in the supplementary material of this article