Coupling cation migration with segregation for versatile air electrode in proton-conducting ceramic cells Authors: Junda You, Zhipeng Liu, Yuan Zhang, Suling Shen, Hongxin Yang, Junbiao Li, Haojie Zhu, Shuo Zhai, Nai Shi, Nanqi Li, Yuchen Wu, Jun Zhou, Meijie Yin, Qianling Zhang, Minhua Shao, Meng Ni, Zongping Shao, Bin Chen, Heping Xie Published: 2026 ABSTRACT Proton conducting ceramic cells enable efficient electrocatalytic H2-to-power conversion, yet their perovskite air electrodes demand versatile properties: high bulk proton uptake/conductivity, fast surface oxygen kinetics, and operational stability. Conventional single-phase perovskites usually fail to meet these demands simultaneously. Here, we show a facile strategy that couples bulk cation migration with surface segregation to enhance the versatility of the air electrode. The as-prepared (Ba0.9Ce0.1-α)(CeαFe0.8Ni0.2-β)O3-δ-βNiO (cm-BCFN@NiO) electrode features a coupled bulk–surface architecture where finely dispersed NiO nanoparticles and Ce migration from A-sites to B-sites synergistically optimize surface oxygen kinetics, bulk proton uptake/conductivity, and electronic transport. The optimized electrode exhibits a low area-specific resistance of 0.3 Ω cm2 at 550 °C—a 75% reduction compared to the NiO-free counterpart—enabling a 77.1% increase in maximum power density and over 1200 hours of stable single-cell operation. Beyond protonic ceramic cells, this work presents an approach for designing perovskites by leveraging cation migration and surface segregation at the atomic scale, opening more opportunities for a wide range of electrocatalytic applications. Introduction Protonic ceramic fuel/electrolysis cells (PCFC/ECs) outperform traditional oxygen-conducting solid oxide fuel/electrolysis cells (SOFC/ECs) in the intermediate temperature range (500–600 °C)1–5. However, they still face various bottlenecks towards commercialization, with sluggish air electrode processes being particularly prominent, despite the development of various perovskite materials6–8. Air electrodes must exhibit versatility, functioning as triple (H+/O2-/e-) conductors, while being catalytically active and mechanically/chemically durable. Perovskites with the general formula ABO3 promise substantial compositional flexibility to tune their physicochemical properties; however, achieving all of these requirements in a single-phase perovskite is challenging. For example, achieving triple conductivity in perovskites is often hindered by the competing defect chemistry (e.g., the competition between water uptake and oxygen dissociative absorption9–11). Furthermore, good thermomechanical durability often conflicts with high electrochemical activity, as the latter demands rich defects and a redox-active surface, which often undermines the structural/chemical stability of the lattice4. To address these challenges, the formation of multi-phase composites has emerged as a practical strategy for developing air electrodes, wherein different phases serve distinct functions. Examples include the in-situ-formed air electrodes composed of Ba0.62Sr0.38CoO3−δ and Pr1.44Ba0.11Sr0.45Co1.32Fe0.68O6−δ12; Ba0.5Sr0.5(Co0.7Fe0.3)0.6875W0.3125O3-δ, comprising double and simple perovskite phases5 and Sr0.9Ce0.1Fe0.8Ni0.2O3-δ13-a nanocomposite comprising simple perovskite, Ruddlesden–Popper perovskite, and surface-decorated minor oxides. Methods for fabricating multi-phase electrodes vary, including mechanical mixing, infiltration techniques, and thermal/chemical re-assembly14–16. Notably, the emerging self-assembly techniques—which involve exceeding material tolerance limits of the lattice17,18 or stoichiometry19—have demonstrated effectiveness in enhancing oxygen reduction reaction (ORR) activity in PCFCs. However, the traditional self-assembly methods face limitations, such as poor interfacial stability caused by severe phase segregation, particularly under conditions of high humidity or elevated oxygen partial pressure20,21. In addition, mechanical stability may also be compromised due to boundary separation during assembly. Consequently, further innovations are required to design composite air electrodes for robust and efficient PCFC/ECs. Incorporating fine functional nano-oxides anchored onto the surface of the bulk phase provides a more effective strategy, ensuring robustly anchored particles with integration into the parent material. Building on this concept, we propose a holistic approach to functionalizing air electrodes through an intrinsic lattice relaxation process—coupling the migration of bulk lattice cations within the ABO3 perovskite structure with the segregation of NiO nanoparticles at the perovskite surface22,23. To realize this, we selected BaFexNi1−xO3−δ (BFN) as the parent material, as it is a well-studied cubic perovskite air electrode that provides a suitable platform for A-site doping and subsequent cation migration within the lattice. Furthermore, the Ni at the B-site is intrinsically prone to segregation under lattice perturbations24. Therefore, the coupling of migration and segregation is feasible in the BFN system, enabling the simultaneous tuning of both bulk and surface electrochemical properties, thereby demonstrating the core concept of this work. Through density functional theory (DFT) predictions and experimental observations, we have developed an air electrode with a stable nanocomposite structure: (Ba0.9Ce0.1-α)(Fe0.8CeαNi0.2-β)O3-δ with abundant NiO nanoparticles segregated on the surface (termed as cm-BCFN@NiO). The formation of NiO nanoparticles is considered to be facilitated by the migration of Ce to the B-site due to the oversized ionic radius of Ce3+ (1.34 Å) in the A-site:1\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{{{\rm{Ba}}}}}_{0.9}{{{{\rm{Ce}}}}}_{0.1}{{{{\rm{Fe}}}}}_{0.8}{{{{\rm{Ni}}}}}_{0.2}{{{{\rm{O}}}}}_{3-\delta }\mathop{\longrightarrow }^{{{{\rm{Ce }}}}\atop{{{\rm{migration}}}}}({{{{\rm{Ba}}}}}_{0.9}{{{{\rm{Ce}}}}}_{0.1-\alpha })({{{{\rm{Fe}}}}}_{0.8}{{{{\rm{Ce}}}}}_{\alpha }{{{{\rm{Ni}}}}}_{0.2-\beta }){{{{\rm{O}}}}}_{3-\delta^{\prime} }+\beta {{{\rm{NiO}}}}$$\end{document}Ba0.9Ce0.1Fe0.8Ni0.2O3−δ→Cemigration(Ba0.9Ce0.1−α)(Fe0.8CeαNi0.2−β)O3−δ′+βNiO Characterization results confirm that the well-decorated NiO nanoparticles enhance the oxygen surface exchange kinetics, while the bulk (Ba0.9Ce0.1-α)(Fe0.8CeαNi0.2-β)O3-δ (C-BCFN)—featuring partial migration of Ce from A-site to B-site—provides abundant oxygen vacancies, improved electronic conductivity, and increased proton uptake (hydration) capacity. This synergistically results in a 1.77-fold increase in peak power density in fuel cell mode—1.16 W cm−2 vs. 0.65 W cm−2 of Ba0.9Ce0.1FeO3-δ at 650 °C and enhanced durability (with no degradation over 1200 h). These findings highlight the potential of leveraging cation migration to develop high-performance air electrodes for protonic ceramic cells, paving the way for significant advancements in PCFC/PCECs technology. Results Prediction of cation migration in selected perovskites To demonstrate the cation migration concept, we first need to identify metal cations that can occupy the A- and B-site simultaneously with sufficiently low migration energy barriers. Previous studies have indicated that cations with appropriate ionic radii (0.086–0.105 nm) can occupy both the A- and B-sites of ABO3 perovskite25–33, as highlighted by the two-color atoms in Fig. 1a out of the periodic table. To predict the ease of migration for these cation candidates, we used DFT to calculate their migration energy barriers in an ORR-active cubic perovskite in a model system—Ba0.9X0.1Fe0.8Ni0.2O3-δ (X represents the cation candidates undergoing A-to-B migration, including Ce, Y, Yb, La, Pr, Nd, Sm, and Gd). To simulate the migration, we removed one B-site cation from the pristine supercell model consisting of 8 unit cells, as illustrated in Fig. 1b. Energy barriers values were obtained at six intermediate coordinates (I to VI) along the migration path from the A-site to the B-site (Fig. 1c, Supplementary Fig. S1, Supplementary Table S1 and Supplementary Data 1). These values define the energy differences between each intermediate configuration and the initial A-site position, thus describing the full migration landscape for each cation. Figure 1d summarizes the maximum migration energy barriers extracted from the minimum energy path (MEP) of the seven elements. Among them, Ce exhibits the lowest migration energy barrier (~ 2.583 eV), whereas La shows the highest (~ 4.251 eV). A clear positive correlation is observed between the maximum migration energy barrier and the ionic radius of the migrating cation, confirming that ionic size is the dominant factor governing cation migration in the Ba0.9 X 0.1Fe0.8Ni0.2O3-δ system, in agreement with previous reports34.Fig. 1Prediction of cation migration in perovskites.a Periodic tables highlighting elements that can occupy both A and B-sites of the perovskite. b Schematic illustration of the cation migration process. The red, blue, two-toned, and gray spheres represent A-site, B-site, A/B-site cations and oxygen ions, respectively. c Calculated migration energy barriers for Ba0.9 X 0.1Fe0.8Ni0.2O3-δ (X = Ce, Y, Yb, La, Pr, Nd, Sm, Gd) at six coordinates (Shannon ionic radii74 are used to represent the cation radii, R). d Correlation between the maximum energy barrier and the typical ionic radius (for 3+ or 4+ oxidation states). e Calculated MEP of Ce migration from the A-site to the B-site. Source data for Fig. 1 are provided as a Source Data file. Prediction of cation migration in perovskites. a Periodic tables highlighting elements that can occupy both A and B-sites of the perovskite. b Schematic illustration of the cation migration process. The red, blue, two-toned, and gray spheres represent A-site, B-site, A/B-site cations and oxygen ions, respectively. c Calculated migration energy barriers for Ba0.9 X 0.1Fe0.8Ni0.2O3-δ (X = Ce, Y, Yb, La, Pr, Nd, Sm, Gd) at six coordinates (Shannon ionic radii74 are used to represent the cation radii, R). d Correlation between the maximum energy barrier and the typical ionic radius (for 3+ or 4+ oxidation states). e Calculated MEP of Ce migration from the A-site to the B-site. Source data for Fig. 1 are provided as a Source Data file. To further illustrate the migration process in detail, Fig. 1e presents the MEP for Ce migration from the A-site to the B-site, discretized into six configurations (I–VI). Four transition states connecting these configurations were determined using the standard climbing-image nudged elastic band (CI-NEB) method35. The highest energy point along this path occurs near Configuration IV, defining a migration energy barrier of 2.583 eV. The relatively low migration energy barrier of Ce is attributed to its weaker electrostatic interactions with the surrounding lattice atoms, which facilitate its migration between the two sites36. To further validate the DFT predictions, perovskites with the nominal stoichiometry of Ba0.9 X 0.1Fe0.8Ni0.2O3-δ were synthesized via a sol-gel method followed by calcination at 1000 °C. X-ray diffraction (XRD) patterns of all the synthesized perovskites are shown in Supplementary Fig. S2. Except for Pr, which did not form a well-defined cubic phase, all metal cations (X3+/4+) successfully incorporated into the cubic phase of the parent material without observable peaks of corresponding simple oxides (i.e., X2O3). Furthermore, NiO peaks were observed in the XRD patterns of Ba0.9X0.1Fe0.8Ni0.2O3-δ (where X = Ce, Y, La, Nd and Sm). Analysis of their energy-dispersive X-Ray spectroscopy (EDX) mapping and transmission electron microscopy (TEM) images (Supplementary Figs. S3–10) further indicates that the Ba0.9X0.1Fe0.8Ni0.2O3-δ (X = Ce, Sm, La, Nd, Y) samples possess surface-segregated NiO. Based on the XRD results, TEM observations and predicted migration energy barriers of these perovskite materials, the Ba0.9Ce0.1Fe0.8Ni0.2O3-δ composition shows a high possibility of Ce migration and well-defined NiO nanoparticles. There is likely a synergistic correlation between cation migration and B-site segregation to relax the lattice according to lattice tolerance theory37; sepecifically, the formation of NiO nanoparticles serves to reduce the tolerance factor (t) while the migration of Ce ions in Ba0.9Ce0.1Fe0.8Ni0.2O3-δ increases the t factor (assuming complete NiO segreagation in Ba0.9Ce0.1Fe0.8Ni0.2O3-δ, the t factor exceed 1.09, at which a cubic phase of ABO3 probably cannot be sustained). Therefore, in the following sections, we selected the as-designed Ba0.9Ce0.1Fe0.8Ni0.2O3-δ (cm-BCFN@NiO) as the experimental group to thoroughly verify the coupled cation migration and NiO segregation and investigate their effects on its ORR activity and durability, while Ni-free BCF was selected as the control group. Experimental verification of cation-migration and surface segregation The morphology of cm-BCFN@NiO is observed in the scanning electron microscope (SEM) image shown in Fig. 2a, where the green-highlighted areas indicate the uniformly distributed NiO nanoparticles (100 ~ 150 nm) on the surface. The phase structure of the cm-BCFN@NiO electrode is further verified by high-resolution transmission electron microscopy (HRTEM), as shown in Fig. 2b and Supplementary Fig. S11.Fig. 2Experimental observation of Cation migration and NiO nanoparticles in cm-BCFN@NiO.a SEM micrograph showing the surface morphology of NiO nanoparticles on cm-BCFN@NiO. b High-angle annular dark-field imaging (HAADF) STEM image of cm-BCFN@NiO. c, d Corresponding EDS mapping images, comfirming that cm-BCFN@NiO is composed of C-BCFN and NiO. e, f HRTEM images of the NiO and C-BCFN phases with corresponding FFT. g corresponding STEM EDS maps of the Ni distribution in cm-BCFN@NiO acquired from a TEM lamella. h STEM image of the C-BCFN. i HR-TEM images of the C-BCFN sample and corresponding FFT pattern. j The atomic-resolution STEM image and elemental mapping: Ba (yellow), Ce (purple). k Schematic illustration of cm-BCFN@NiO morphology. l Cation migration and NiO formation in cm-BCFN@NiO. Red, pink, blue, green, and gray spheres represent Ce, Ba, Fe, Ni, and oxide ions (O2−). Source data for Fig. 2 are provided as a Source Data file. Experimental observation of Cation migration and NiO nanoparticles in cm-BCFN@NiO. a SEM micrograph showing the surface morphology of NiO nanoparticles on cm-BCFN@NiO. b High-angle annular dark-field imaging (HAADF) STEM image of cm-BCFN@NiO. c, d Corresponding EDS mapping images, comfirming that cm-BCFN@NiO is composed of C-BCFN and NiO. e, f HRTEM images of the NiO and C-BCFN phases with corresponding FFT. g corresponding STEM EDS maps of the Ni distribution in cm-BCFN@NiO acquired from a TEM lamella. h STEM image of the C-BCFN. i HR-TEM images of the C-BCFN sample and corresponding FFT pattern. j The atomic-resolution STEM image and elemental mapping: Ba (yellow), Ce (purple). k Schematic illustration of cm-BCFN@NiO morphology. l Cation migration and NiO formation in cm-BCFN@NiO. Red, pink, blue, green, and gray spheres represent Ce, Ba, Fe, Ni, and oxide ions (O2−). Source data for Fig. 2 are provided as a Source Data file. The C-BCFN main phase and the NiO decorating phase are observed to be well-connected (Fig. 2e, f) from a selected region of Fig. 2b, which is confirmed by elemental mapping (Fig. 2g, h) and lattice spacings of d(011) = 2.93 Å of C-BCFN and d(200) = 2.07 Å of NiO in Fig. 2e, f and Supplementary Fig. S11, which is consistent with the XRD refinement results (discussed later in the context of Fig. 3). In addition, Ni elemental mapping of the TEM lamella fabricated by focused ion beam (FIB) in Fig. 2g shows well-distributed clusters of Ni, consistent with the SEM observations. The relatively weak EDS signals for Ni throughout the cross-section also indicate that Ni segregation is incomplete, with a certain amount of Ni remaining in the B-site of C-BCFN (consistent with Fig. 2c). The enlarged high-resolution STEM image C-BCFN in Fig. 2h confirms its cubic lattice structure with corresponding fast Fourier transform (FFT) shown in Fig. 2j along the [−1-1-1] axis. Meanwhile, the atom-resolution EDS element mapping of C-BCFN in Fig. 2i clearly indicates distinct positions for the Ba and Ce cations along the [111] zone axis. The incomplete overlap of the Ba and Ce signals suggests Ce partially resides at the B-site, supporting the migration of Ce cations from the A-site to the B-site.Fig. 3Structural characterizations of Ce cation in cm-BCFN@NiO.a The HAADF image of C-BCFN viewed along the (111) zone axis with the path for electron energy-loss line-scan spectra (EELS) line-scan indicated. b EELS profiles from 870 eV to 920 eV across 7 atom layers are indicated in (a). The graph on the right shows corresponding Ce M5 and M4 edge intensity ratios. c Rietveld refinement of the XRD pattern for cm-BCFN@NiO powder. d Fourier transformed of Ce L3-edge XANES spectra for CeO2, cm-BCFN@NiO and BCF. e Ce L3-edge EXAFS spectra of CeO2, cm-BCFN@NiO and BCF. f In situ high-temperature XRD patterns of the cm-BCFN@NiO recorded from 25 °C up to 1000 °C and back to 25 °C. Source data for Fig. 3 are provided as a Source Data file. Structural characterizations of Ce cation in cm-BCFN@NiO. a The HAADF image of C-BCFN viewed along the (111) zone axis with the path for electron energy-loss line-scan spectra (EELS) line-scan indicated. b EELS profiles from 870 eV to 920 eV across 7 atom layers are indicated in (a). The graph on the right shows corresponding Ce M5 and M4 edge intensity ratios. c Rietveld refinement of the XRD pattern for cm-BCFN@NiO powder. d Fourier transformed of Ce L3-edge XANES spectra for CeO2, cm-BCFN@NiO and BCF. e Ce L3-edge EXAFS spectra of CeO2, cm-BCFN@NiO and BCF. f In situ high-temperature XRD patterns of the cm-BCFN@NiO recorded from 25 °C up to 1000 °C and back to 25 °C. Source data for Fig. 3 are provided as a Source Data file. Based on the observations above, the coupled cation migration and NiO formation process in cm-BCFN@NiO are illustrated in Fig. 2k, l. First, the ABO3 perovskite is stoichiometrically designed with Ba and Ce at the A-site and Fe and Ni at the B-site; Next, due to size effects, A-site Ce cations migrate to the B-site, “pushing” the original B-site metal ions out of the lattice due to the lattice relaxation; Meanwhile, A-site cation vacancies emerge in the bulk phase C-BCFN with Ce-migrated, allowing for the stable segregation of NiO at the C-BCFN surface. To further verify cation migration, we probed the Ce electronic structure using scanning transmission electron microscopy-electron energy-loss line-scan spectroscopy (STEM-EELS) line-scans across seven atomic columns along the path of cm-BCFN@NiO in Fig. 3a. The corresponding EELS profiles in Fig. 3b show distinctive peaks at 884.7 eV and 902.3 eV, corresponding to the Ce M5 and M4 edges. Since the intensity ratios between M5 and M4 edges are 1.12 and 0.85 for Ce4+ and Ce3+, respectively38, the M5/M4 intensity ratio of cm-BCFN@NiO can be used to probe the oxidation state of Ce in the lattice. Figure 3b shows significant local variations of the M5/M4 ratio from 1.05 to 0.85 at different atomic sites in Fig. 3a, implying the co-existence of Ce4+ and Ce3+ cations and the atomic-level variation in the coordination environment of Ce from columns in cm-BCFN@NiO39. This can be a good indication of co-occupation of Ce in A- and B-sites due to cation migration. Rietveld refinement of the XRD pattern for cm-BCFN@NiO confirmed a two-phase composite: the main cubic perovskite phase with Ce migrated (Ba0.9Ce0.1-α)(CeαFe0.8Ni0.2-β)O3-δ (C-BCFN, Pm-3m, 93.08 wt%, a = b = c = 4.08269(13) Å) and a secondary minor NiO phase (6.92 wt%) in Fig. 3c, with details in Supplementary Table S2, S3, and there is no other Ce-containing perovskites, simple oxides (i.e., CeO2) or solid solutions identified. Therefore, Ce cations are deemed to only occupy both the A and B-sites of C-BCFN due to the migration process. To confirm the electronic environment of Ce cations in EELS, complementary X-ray absorption fine structure (XAFS) spectra of cm-BCFN@NiO and BCF were compared. The X-ray absorption near-edge structure (XANES) spectra in Fig. 3d indicate a higher proportion of Ce4+ ions in the cm-BCFN@NiO compared to the BCF and standard CeO240. This observation aligns with the tolerance factor principle, as Ce4+ (0.087 nm) is more likely to occupy the B-site due to its smaller ionic radius compared to Ce3+ (0.134 nm). In addition, the coordination environment of Ce was analyzed using its L3-edge Fourier-transformed extended X-ray absorption fine structure (FT-EXAFS) spectra to identify the occupancy sites of Ce. In an ideal perovskite structure, the A-site cations are coordinated with 12 oxygen atoms, while B-site cations are coordinated with 6 oxygen atoms41, the FT-EXAFS fitting results revealed that the average coordination number (CN) of Ce is 6.96 in cm-BCFN@NiO, a value lower than that of BCF (CN = 7.78) and CeO2 (CN = 8), as shown in Fig. 3e and Supplementary Table S4. This confirms the migration of a fraction of Ce cations migrated from the A-site to the B-site in cm-BCFN@NiO. The slightly higher CN of Ce in BCF also indicates reduced migration in BCF, termed as (BaCe0.1-α)(CeαFe)O3-δ, according to the XRD in Supplementary Fig. S12 and Supplementary Table S5, S6, regardless of the absence of B-site Ni segregation. To evaluate the stability of the NiO nanoparticles and cation migration, in-situ high-temperature XRD testing was performed on cm-BCFN@NiO from 25 to 1000 °C, as shown in Fig. 3f. The material exhibited phase stability during heating and cooling, evidenced by the full recovery of the thermal expansion-induced peak shifting. Notably, the minor peaks of NiO, marked with asterisks, at 2θ = 37.24° in the cm-BCFN@NiO XRD pattern show complete recovery of the peak shift, indicating that both the main C-BCFN phase and the NiO phase possess stability when subjected to thermal cycles. Electrochemical performance of functionalized air electrode We then assessed the electrochemical performance of the functionalized cm-BCFN@NiO as the air electrode in PCECs. Electrochemical impedance spectroscopy (EIS) measurements were first performed on a symmetrical cell (cm-BCFN@NiO|BZCYYb|cm-BCFN@NiO) under open-circuit and 3 vol% humidified air conditions at various temperatures (BZCYYb refers to BaZr0.1Ce0.7Y0.1Yb0.1O3-δ). As shown in Fig. 4a, the Arrhenius plots of measured area-specific resistance (ASR) reveal that the BCF air electrode existed an ASR of 1.254 Ω cm² at 550 °C, while cm-BCFN@NiO showed a notably lower value of only 0.30 Ω cm2 at the same temperature, highlighting its performance benefiting from the nano-compositing. A second symmetrical cell serving as a control group was also prepared with a mechanically mixed BCF-NiOmix electrode (mixed by ball milling BCF and NiO with a mass ratio of 1:0.06), which exhibited an impedance value intermediate between those of cm-BCFN@NiO and BCF. This verifies that NiO promotes ORR activity of BCF (even at the microscale, see SEM image in Supplementary Fig. S13), though it requires nanolization and optimized distribution to further enhance performance, highlighting the role of in situ formed NiO nanoparticles in achieving low ASR observed for cm-BCFN@NiO.Fig. 4Electrochemical performance of functionalized air electrode.a Arrhenius plots of area-specific polarization resistance for BCF, BCF-NiOmix, and cm-BCFN@NiO electrode in symmetrical cells (BZCYYb electrolyte) from 450 to 700 °C. b DRT analysis for BCF, BCF-NiOmix and cm-BCFN@NiO air electrodes at 550 °C. c The ASR of cm-BCFN@NiO electrodes (fired at 800 °C) measured at 550 °C with varying humidity (0, 3, and 10 vol% H2O) in a symmetrical cell. d, e Long-term stability of symmetrical cells with BCF and cm-BCFN@NiO electrode at 550 °C in humidified air. Representative EIS Nyquist plots for the (f) BCF and (g) cm-BCFN@NiO symmetric cells after long-time operation. h I-V curves of the single cell with cm-BCFN@NiO and BCF air electrode in fuel cell mode at 650 °C using pure hydrogen for the fuel electrode and synthetic air (3 vol% H2O-air) for the air electrode. i The Galvanostatic stability of cm-BCFN@NiO single cell at 400 mA cm−2 at 550 °C for 1200 (The irregular fluctuations observed are attributed to minor equipment anomalies, specifically intermittent software interruptions requiring manual restarts). All electrochemical performance data represent raw measured values without iR compensation. Source data for Fig. 4 are provided as a Source Data file. Electrochemical performance of functionalized air electrode. a Arrhenius plots of area-specific polarization resistance for BCF, BCF-NiOmix, and cm-BCFN@NiO electrode in symmetrical cells (BZCYYb electrolyte) from 450 to 700 °C. b DRT analysis for BCF, BCF-NiOmix and cm-BCFN@NiO air electrodes at 550 °C. c The ASR of cm-BCFN@NiO electrodes (fired at 800 °C) measured at 550 °C with varying humidity (0, 3, and 10 vol% H2O) in a symmetrical cell. d, e Long-term stability of symmetrical cells with BCF and cm-BCFN@NiO electrode at 550 °C in humidified air. Representative EIS Nyquist plots for the (f) BCF and (g) cm-BCFN@NiO symmetric cells after long-time operation. h I-V curves of the single cell with cm-BCFN@NiO and BCF air electrode in fuel cell mode at 650 °C using pure hydrogen for the fuel electrode and synthetic air (3 vol% H2O-air) for the air electrode. i The Galvanostatic stability of cm-BCFN@NiO single cell at 400 mA cm−2 at 550 °C for 1200 (The irregular fluctuations observed are attributed to minor equipment anomalies, specifically intermittent software interruptions requiring manual restarts). All electrochemical performance data represent raw measured values without iR compensation. Source data for Fig. 4 are provided as a Source Data file. To further explore the enhancement mechanisms, the distribution of relaxation time (DRT) technique was employed to analyze the impedance response of all air electrodes in humidified air at 550 °C. In Fig. 4b, the DRT profiles exhibit three distinct peaks at high-frequency (HF), intermediate-frequency (IF), and low-frequency (LF) ranges. These peaks correspond to charge transfer at the electrolyte-electrode interface, ion migration or oxygen surface exchange within the electrode, and gas-phase (O2/H2O) diffusion, respectively42,43. The cm-BCFN@NiO electrode demonstrates a marked decrease in peak intensity across all frequency ranges compared to the BCF and BCF-NiOmix electrodes, indicating substantial and holistic improvement in charge transfer, surface oxygen exchange kinetics, and gas/H2O diffusion. It is clearly shown that the HF and IF peak areas for the BCF-NiOmix electrode are lower than those of BCF, indicating that additional NiO particles promote ion transfer, ion migration, and surface exchange processes. Prior studies have demonstrated that NiO nanoparticles could effectively increase the number of active sites for O2 adsorption and dissociation and thus reducing the IF resistence44. However, the BCF-NiOmix electrode exhibits an increased LF peak region, indicating a reduced capability for gas or water diffusion. To further explore the electrochemical reaction kinetics of the cm-BCFN@NiO electrode under varying water content, EIS measurements were conducted at 550 °C under different pH2O conditions (Fig. 4c and Supplementary Fig. S14) and analyzed using the DRT technique (Supplementary Fig. S15). Both the impedance spectra and DRT curves exhibited noticeable shifts in response to variations in the pH2O level. Increasing the water content from 0 to 10 vol% in air reduced the Rp of cm-BCFN@NiO from 0.855 Ω cm2 to 0.292 Ω cm2. Under dry conditions, a large LF peak area was observed, which significantly decreased upon the introduction of water, indicating accelerated water vapor diffusion45. Furthermore, higher water content reduced the IF peak area, suggesting that increased external water vapor may accelerate surface water exchange and facilitate proton-related transport processes, thereby enhancing OER activity46. To evaluate the stability of the reconstructed cm-BCFN@NiO electrode, long-term stability testing was conducted using symmetric cells in humidified air at 550°C. As illustrated in Fig. 4d, e, the ASR value of the cm-BCFN@NiO air electrode exhibited a negligible degradation rate of 1.18 × 10−4 Ω cm2 h−1 over 200 h, while the BCF electrode experienced a much higher degradation rate of 0.05 Ω cm2 h−1, leading to a sevenfold increase to 6.993 Ω cm2. The ohmic resistance of cm-BCFN@NiO slightly decreased over time compared to BCF (Fig. 4f, g), likely due to its robust connection of reconstructed phases, which enhances the electrode-electrolyte interface. Post-test HRTEM and elemental mapping (Supplementary Figs. S17, S18) further confirm that the phase structure of cm-BCFN@NiO remains intact, with no new phases detected following the stability test. XRD analysis of cm-BCFN@NiO powder before and after 100 hours of sintering in humidified air (3 vol% H2O) revealed unchanged diffraction peaks (Supplementary Fig. S16), confirming that the coupled cation migration and the formation of NiO nanoparticles structure, maintains stability, with stable ohmic and polarization resistances during long-term testing47. The electrochemical performance of the cm-BCFN@NiO electrode is demonstrated through its application in single-cell applications. In full cells configured as Ni+BZCYYb|BZCYYb|air electrode, cm-BCFN@NiO, BCF-NiOmix, and BCF achieved a peak power density of 1160, 839 and 671 mW cm−2 at 650 °C (Fig. 4h and Supplementary Fig. S19). Furthermore, the cm-BCFN@NiO-equipped single cell delivers competitive performance compared to several widely used cobalt-free air electrodes, exhibiting a lower ASR value, as shown in Supplementary Table S7. In electrolysis mode, cm-BCFN@NiO achieved a current density of 1.96 A cm−2 at 650°C with an applied voltage of 1.4 V in humidified air (3 vol% H2O) (Supplementary Fig. S20), underscoring its potential for water electrolysis in PCECs. To assess the durability of the cm-BCFN@NiO, a long-term stability test at 550 °C was conducted, as shown in Fig. 4j. Under a current density of 400 mA cm⁻2 at 550 °C in fuel-cell mode, the single cell equipped with the cm-BCFN@NiO operated continuously for 1200 h with noticeable degradation, demonstrating long-term stability. Cross-sectional SEM analysis (Supplementary Fig. 21) of the post-test electrode indicated an interfacial robust bond between the cm-BCFN@NiO electrode and the electrolyte, with no notable structural failures. These findings underscore the potential of cm-BCFN@NiO as a highly robust air electrode for protonic ceramic cells. Characterizations and mechanisms for holistic electrochemical enhancement Electron paramagnetic resonance (EPR) spectroscopy was employed to assess the oxygen vacancy concentrations in BCF and cm-BCFN@NiO (Fig. 5a). The cm-BCFN@NiO exhibited a stronger signal at g = 2.003, indicating a higher concentration of oxygen vacancies relative to BCF. The elevated oxygen vacancy concentration is expected to facilitate proton uptake, provide active sites for surface reactions, and improve charge-transfer kinetics, thereby enhancing electrode activity48–53. X-ray photoelectron spectroscopy (XPS) O 1 s spectra (Fig. 5b and Supplementary Table S9) further validates these findings, showing a significantly higher adsorbed-to-lattice oxygen ratio (Oad/Olattice) in cm-BCFN@NiO (~ 2.9) compared to BCF ( ~ 1.8), reflecting the increased oxygen vacancies in cm-BCFN@NiO54, which is consistent with EPR results.Fig. 5Material characterizations and theoretical calculations.a comparison of oxygen vacancy concentration for BCF and cm-BCFN@NiO at 25 °C. b XPS O1s spectra for BCF and cm-BCFN@NiO samples at 25 °C. c The spectra of Fe K-edge XANES for BCF and cm-BCFN@NiO. d Thermogravimetric analysis curves of BCF and cm-BCFN@NiO recorded during the cooling stage under dry and humidified air (pH2O = 0.03 atm). e The calculated proton conductivity and hydrogen permeability of cm-BCFN@NiO and BCF at 650 °C. f Calculated H2O and O2 adsorption energies, hydration and oxygen vacancy formation energies, and O-p-band center for BCF and cm-BCFN@NiO. Source data for Fig. 5 are provided as a Source Data file. Material characterizations and theoretical calculations. a comparison of oxygen vacancy concentration for BCF and cm-BCFN@NiO at 25 °C. b XPS O1s spectra for BCF and cm-BCFN@NiO samples at 25 °C. c The spectra of Fe K-edge XANES for BCF and cm-BCFN@NiO. d Thermogravimetric analysis curves of BCF and cm-BCFN@NiO recorded during the cooling stage under dry and humidified air (pH2O = 0.03 atm). e The calculated proton conductivity and hydrogen permeability of cm-BCFN@NiO and BCF at 650 °C. f Calculated H2O and O2 adsorption energies, hydration and oxygen vacancy formation energies, and O-p-band center for BCF and cm-BCFN@NiO. Source data for Fig. 5 are provided as a Source Data file. To further examine the electronic structure evolution accompanying these defect changes, X-ray absorption near-edge structure (XANES) spectra at the Fe K-edge were analyzed (Fig. 5c), which reveals a higher Fe oxidation state in cm-BCFN@NiO compared to BCF, and XPS further corroborates this finding. According to the XPS results (Supplementary Fig. S22 and Supplementary Table S10), the Fe oxidation state in our system increases in the cm-BCFN@NiO sample. Compared to BCF, which exhibits an average Fe oxidation state of + 2.6741, the cm-BCFN@NiO sample shows a higher Fe oxidation state of + 2.9684, with a difference of 0.2943. This elevated Fe oxidation state suggests a stronger oxygen reduction capability in cm-BCFN@NiO, which may facilitate the observed electrochemical activity. O2-temperature programmed desorption (O2-TPD) analysis (Supplementary Fig. S23) was used to assess oxygen species activity and Fe ion reducibility. cm-BCFN@NiO exhibits a larger desorption peak area than BCF, indicating higher oxygen vacancy concentration and enhanced oxygen exchange activity, facilitating itss oxygen ion mobility and ORR kinetics55. This finding aligns with the EPR results, confirming the enhanced oxygen vacancy concentration in cm-BCFN@NiO. Notably, the integrated XPS, XAFS, O2-TPD, and EPR results reveal concurrent changes in Fe valence state, oxygen vacancy concentration, and hydration behavior in cm-BCFN@NiO. While variations in the Fe valence play a role in regulating defect chemistry, these changes occur simultaneously with Ce migration and in-situ segregation of the secondary NiO phase. Consequently, the observed enhancement in oxygen vacancy concentration and protonic conductivity cannot be attributed to Fe valence variation alone. Instead, these results suggest a synergistic mechanism wherein cation redistribution and secondary-phase formation collectively modulate the local defect environment and catalytic activity of the electrode. To assess the effect of oxygen vacancies on electrode conductivity, electrical conductivity measurements were performed on dense bars under dry air conditions. Supplementary Fig. S24 shows that BCF exhibits typical p-type semiconducting behavior, characterized by increasing conductivity at higher temperatures, whereas cm-BCFN@NiO’s conductivity declines between 500 and 700 °C. This decline, linked to hole depletion due to excessive oxygen vacancies, which disrupts hole hopping along Fe–O–Fe bonds56. Above 500 °C, oxygen desorption and Fe4+ reduction further diminish the hole concentration, leading to a decrease in total electrical conductivity. Despite this transition, cm-BCFN@NiO maintains significantly higher overall conductivity than BCF, confirming its suitability for PCFC applications. In previous EIS tests, water was shown to significantly affect air electrode performance. To further investigate, thermogravimetric analysis (TGA) was used to quantify the mass increase of BCF and cm-BCFN@NiO samples under both humidified and dry air during cooling from 900 °C to 200 °C (Fig. 5d). In dry air, the mass gain arises from oxygen incorporation into vacancies, while in humidified air, both oxygen and water incorporation contribute to the total weight change. The difference in mass change between the two conditions reflects net water uptake, which indicates each material’s hydration capacity57. Hydration, essential for enhancing triple-conducting (H+/O2-/e-) properties in proton-conducting ceramics, relies on oxygen vacancies and typically occurs in humid environments58, as described by the following hydration reaction equation:2\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${{\mbox{H}}}_{2}{\mbox{O}}+{{\mbox{O}}}_{{\mbox{O}}}^{\times }+{{\mbox{V}}}_{{\mbox{O}}}^{\bullet \bullet }\rightleftharpoons {2{\mbox{OH}}}_{{\mbox{O}}}^{\bullet }$$\end{document}H2O+OO×+VO∙∙⇌2OHO∙ During cooling, cm-BCFN@NiO exhibits a significantly larger weight increase of approximately 0.2 wt.%, compared to only 0.05 wt.% for BCF, indicating an enhanced hydration capacity, which enhances the triple-conducting behavior of the cm-BCFN@NiO electrode. This enhancement is attributed to the synergistic effects of cation migration and the presence of NiO nanoparticles, which collectively modulate oxygen vacancy concentration and hydration properties. The influence of Ce migration and NiO segregation on proton migration was visually observed through hydrogen permeation measurement (Fig. 5e). The proton conductivity of cm-BCFN@NiO and BCF was found to be 1.74 × 10−2 and 0.16 × 10−2 S cm−1 at 650 °C, respectively. DFT calculations were conducted to gain deeper insights into the proton transport capability. Futhermore, the proton migration energy profiles (Supplementary Fig. S25) illustrate the migration energy barriers at different coordination sites. The lower migration barrier for cm-BCFN@NiO (0.1631 eV) compared to BCF (0.2785 eV) confirms that proton transport is more favorable within the cm-BCFN@NiO lattice. This enhanced proton mobility may be attributed to the increased oxygen vacancy concentration (\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${V}_{O}^{{{{\rm{\bullet }}}}{{{\rm{\bullet }}}}}$$\end{document}VO∙∙) and enhanced H2O uptake capacity demonstrated by TGA. The coordinated effect, driven by the coupling of cation migration and NiO exsolution, establishes a synergistic mechanism that simultaneously enhances oxygen transport, proton conduction, and electronic conductivity. The thermal expansion coefficient (TEC) of cm-BCFN@NiO (21.99 × 10−6 K−1) is slightly higher than that of BCF (16.53 × 10−6 K−1) between 25 and 800 °C (Supplementary Fig. S26), though it remains lower than those of cobalt-based electrodes. This increase in TEC for cm-BCFN@NiO is attributed to enhanced electrostatic repulsion59 stemming from the higher oxygen vacancy concentration, which arises from the coupling of in-situ NiO phase segregation and A-site cation vacancies60. However, the electrode’s porous structure (Supplementary Fig. S21) could accommodate thermal expansion differences, while the in-situ NiO particles help maintain mechanical connectivity and enhance surface ORR kinetics61. To elucidate the mechanisms behind the enhanced ORR activity and hydration capability in cm-BCFN@NiO, DFT calculations were performed. The water adsorption energy for cm-BCFN@NiO was found to be − 1.1024 eV, which is more exothermic than BCF’s − 0.8242 eV, which is consistent with the TGA results (Fig. 5d). Moreover, the hydration energy of cm-BCFN@NiO (− 1.2032 eV) is significantly more negative that of BCF (− 0.6548 eV), Fig. 5f presents the minimum oxygen vacancy formation energies for cm-BCFN@NiO and BCF at various O sites (see Supplementary Fig. S27 and Table S8), showing that cm-BCFN@NiO has a lower formation energy of − 0.5352 eV versus BCF’s − 0.2562 eV. In addition, Fig. 5f, Supplementary Fig. S28 and Supplementary Data 1 reveal that the O-p band center for cm-BCFN@NiO (− 1.17 eV) is closer to the Fermi level that of BCF (− 1.47 eV). To further investigate the oxygen adsorption properties of BCF and cm-BCFN@NiO, the oxygen adsorption energies were also conducted. In comparison to BCF, the oxygen adsorption energy of cm-BCFN@NiO is more negative, measuring approximately − 0.7512 eV (Fig. 5f), which shows good agreement with the O2-TPD results. While these theoretical results are consistent with our experimental findings, we note that DFT and MD models are idealized; they may not fully capture the structural complexities and interfacial dynamics of the electrode under actual high-temperature catalytic operating conditions. To clarify the role of Ce migration in regulating triple conductivity, DFT calculations were performed on two representative structural models in Supplementary Fig. S29 and Supplementary Data 1. BCFN-A represents the configuration where Ce is confined exclusively to the A-site of the perovskite lattice, while cm-BCFN@NiO corresponds to the cation-migration state wherein Ce occupies both the A and B-sites. The cm-BCFN@NiO model exhibits significantly more exothermic hydration energy (− 1.2032 eV), representing a 54.5% increase in thermodynamic favorability compared to BCFN-A (− 0.7791 eV) in Supplementary Table S11. Concurrently, its H2O adsorption energy also becomes more negative (− 1.0240 eV vs. − 0.7548 eV). Furthermore, the distribution of Ce across both lattice sites reduces the oxygen-vacancy formation energy from − 0.5201 eV (BCFN-A) to − 0.5352 eV(cm-BCFN@NiO), representing a 2.9% increase in vacancy-formation favorability (Supplementary Fig. S30). These theoretical results provide insights into the possible energetic benefits of Ce migration, though we acknowledge that these idealized models may not fully capture the complexities of the material under actual catalytic conditions. Discussion In this work, we propose a holistic strategy to synergistically tune both the surface and bulk electrochemical properties in a perovskite air electrode by harnessing the coupled processes of bulk cation migration and surface segregation. Demonstrated by Ce migration in (Ba, Ce)(Fe, Ni)O3-δ, we successfully developed a high-performance and versatile perovskite-based composite air electrode: cm-BCFN@NiO. This electrode features abundant oxygen vacancies, enhanced hydration, and improved protonic/electronic conductivity in the bulk, while simultaneously exhibiting rich NiO nanoparticles on the surface that enhance surface oxygen kinetics, leading to the boost in ORR activity by four times. Moreover, single cells with the versatile electrode demonstrated stable operation for over 1200 h with obvious degradation. Further optimization of air electrodes may be achieved by carefully selecting migrating cations and their segregation within the “playground” of the ABO3 lattice. Beyond its application in proton-conducting fuel cells, this work provides a framework for tailoring material properties of nanocomposites to meet the demands of diverse energy and environmental technologies. Methods Synthesis of materials Cm-BCFN@NiO and BCF powder were synthesized by a combined EDTA-citrate complexing process. Stoichiometric amounts of Ba(NO3)2-99.5% (Sinopharm), Ce(NO3)3·6H2O-99.95% (Aladdin), Fe(NO3)3·9H2O-99.99%(Aladdin), Ni(NO3)2·6H2O-99%(Xilong scientific), were mixed in a molar ratio of 0.05 for each synthesis process. Subsequently, the molar ratio of total metal ions: ethylene diamine tetraacetic acid (EDTA-99.99%, Aladdin): citric acid (CA-99.5%, Xilong scientific) was 1:1:1.5. Then, NH3·H2O was added to keep a pH between 8 and 9. The solution was evaporated at 90°C until a homogenous gel was achieved. The solid precursor for cm-BCFN@NiO was obtained after placing the obtained gel in an oven at 160 °C for 10 h. The solid precursor was calcined at 1000 °C for 5 h in air to obtain the final product. The electrolyte powder of BaZr0.1Ce0.7Y0.1Yb0.1O3−δ (BZCYYb) was synthesized via the commonly used sol–gel method. Stoichiometric amounts of Ba(NO3)2, ZrO(NO3)3·Xh2O-99.99%(Aladdin), Ce(NO3)3·6H2O, Y(NO3)3·6H2O-99.5%(Aladdin), and Yb(NO3)3·6H2O(Macklin) were dissolved in deionized water under continuous stirring. EDTA and citric acid CA were employed as complexing agents, with a molar ratio of total metal ions: EDTA: CA = 1:1:1.5. An appropriate amount of NH3·H2O was added to adjust the pH value of the solution to approximately 8. The homogeneous solution was heated and stirred at 80 °C for 3 h until a viscous gel formed. The obtained gel was then transferred to an oven and dried at 180 °C to yield a fluffy precursor. Finally, the precursor was calcined in air at 1000 °C for 5 h to obtain phase-pure perovskite BZCYYb powder. BZCYYb was stored in a vacuum desiccator (sealed container) at 25 °C to prevent moisture absorption and ensure material integrity before electrochemical testing. Fabrication of single cells and symmetric cells To fabricate single cells, anode-supported half-cells (NiO+BZCYYb|BZCYYb) were prepared by dry-pressing and a co-sintering process. NiO, BZCYYb, and soluble starch in a weight ratio of 6:4:1 were mixed and pressed at 4 MPa. Then, 0.012 g of BZCYYb electrolyte powder was homogeneously coated onto the surface of the anode pellet, compressed at 8 MPa to form a two-layered subtracted, and then co-sintered at 1450 °C for 5 h. Afterward, the cm-BCFN@NiO slurry was sprayed over the center of the electrolyte surface (with a circular area of 0.27 cm2), followed by sintering at 800 °C for 2 h in air condition. To fabricate symmetrical cells for electrochemical impedance spectroscopy (EIS) measurements, cells with the configuration of cathode|BZCYYb|cathode were prepared. First, the as-synthesized BZCYYb powders were dry-pressed into pellets and calcined at 1450 °C to obtain dense electrolyte disks. Subsequently, the air electrode suspension was prepared by ball milling a mixture containing powder, glycerol, ethylene glycol and isopropyl alcohol for 1 h to ensure homogeneous dispersion. The resulting slurry was symmetrically sprayed onto both sides of the dense BZCYYb electrolyte disks, followed by calcination at 800 °C for 2 h in air to form porous cathode layers. Finally, silver paste was applied onto both cathode surfaces to form current-collecting grids, and silver wires were attached to connect the cell to the electrochemical workstation for EIS measurements. Characterizations The crystal structure of cm-BCFN@NiO powder was determined by XRD (Bruker D8 Advance, Germany) in the range of 10–90°. The morphology and chemical composition of cm-BCFN@NiO were examined using transmission electron microscopy (TEM, American FEI Talos F200s). The weight loss data were acquired using a TGA analyzer. The sample underwent an initial heating phase from 30 °C to 80 °C in dry air, employing a heating rate of 5 °C min−1. Subsequently, it underwent a cooling phase from 800 to 200 °C with a cooling rate of 5 °C min−1 under both dry and humidified air conditions. O2-TPD was obtained by using Hiden Analytica, HPR20. The TEC test was performed using a Netzsch DIL402C with a heating rate of 10 °C min−1 and a cooling rate of 10 °C min−1. The oxidation state of elements in cm-BCFN@NiO was performed by XPS using the PHI 5000 Versaprobe ΙΙΙ instrument at room temperature. The Ce L3-edge, Fe K-edge, and Ni K-edge were performed in fluorescence mode at the TPS 44 A beamline at the Taiwan Photon Source. Data processing of the XANES62 and the EXAFS were carried out using Athena and Artemis modules (Version 0.9.26)63,64. Hydrogen permeation flux was measured by gas chromatography (GC-2014, Shimadzu). BCF and cm-BCFN@NiO membranes with a thickness of ~ 0.65 mm were tested at 650 °C using a feed gas of 20 mL min−1 H2 and 80 mL min−1 N2, while high-purity Ar (80 mL min−1) was used as the sweep gas. Electrochemical tests EIS measurements of single cell and symmetrical cells were performed on the Princeton Electrochemical Workstation in the frequency range of 100 kHz to 0.1 Hz at 30 mV under open circuit potential (auto calibration). The EIS data were fitted by equivalent circuit models using Zview software. The I-V and I-P polarization curves were collected using a Keithley 2460 in a fuel cell test station. During the experiments, the fuel side was exposed to dry H2 gas at a flow rate of 80 mL min−1, while ambient air was supplied to the fuel side at a flow rate of 100 mL min−1. The measurement uncertainty is primarily governed by the instrumental accuracy and calibration limits of the testing systems. Computational method of DFT DFT calculations65,66 were performed using the Vienna Ab Initio Simulation Package (VASP) with the projector augmented wave (PAW) method67,68. Exchange–correlation effects were treated within the generalized gradient approximation (GGA) using the Perdew–Burke–Ernzerhof (PBE) functional69. A plane-wave cutoff energy of 500 eV was employed. For bulk Ba0.9 X0.1Fe0.8Ni0.2O3 (X = Ce, Y, Yb, La, Pr, Nd, Sm, Gd), Brillouin zone integration was carried out with a 3 × 3 × 3 Γ-centered Monkhorst–Pack mesh70. The structures were fully relaxed until the residual forces on each atom were below 0.02 eV/Å and the total energy converged within 10⁻5 eV. Migration energy barriers were determined using the climbing-image nudged elastic band (CI-NEB) method with four intermediate images generated by linear interpolation71. The transition states of cation migration from the A-site to the B-site were investigated using a 2 × 2 × 2 superlattice. To discretize the migration process of cation from the A-site to the B-site, five migration steps were defined within the lattice, resulting in six configurations, including the initial and final states. The four transition states along the Minimum Energy Path (MEP) were determined using the classical Climbing Image–Nudged Elastic Band (CI-NEB) method, a computational algorithm widely employed in materials science and computational chemistry to elucidate reaction pathways and identify transition states35. For BCF and cm-BCFN@NiO structures, a 3 × 3 × 3 Γ-centered Monkhorst-Pack mesh was used for bulk calculations. The DFT + U method was applied to account for the strong correlation of Fe-3d and Ni-3d electrons, with Ueff values of 5.3 eV and 6.2 eV, respectively, consistent with the Materials Project45. The hydration energy was calculated as3\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\Delta {E}_{h{ydr}}={E}_{2O{H}_{O}^{*}}-{E}_{{defect}}+{E}_{{H}_{2}O}({{{\rm{g}}}})$$\end{document}ΔEhydr=E2OHO*−Edefect+EH2O(g)where \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${E}_{2O{H}_{O}^{*}}$$\end{document}E2OHO* is the electronic energy of the perovskite with two protons (i.e., two -OH groups), and \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$${E}_{{H}_{2}O}\,({{{\rm{g}}}})$$\end{document}EH2O(g) is the energy of a water molecule72. The (01-1) plane of BCF and cm-BCFN@NiO was chosen to calculate H2O/O2 adsorption energies. A vacuum layer of 20 Å was introduced along the c-axis normal to the surface, and a 2 × 3 × 1 Γ-centered Monkhorst–Pack mesh was employed. In addition, the CI-NEB method was performed to analyze the reaction barriers of proton migration. All structural models were visualized with VESTA73. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Transparent Peer Review file Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Transparent Peer Review file Source data Source Data Source Data