Fe‐Doping‐Accelerated Deep Reconstruction of KCoF₃ for Enhanced Alkaline Oxygen Evolution Reaction

ABSTRACT

Designing catalysts that can undergo deep reconstruction to form active oxyhydroxides for efficient oxygen evolution reactions (OER) remains urgently needed yet challenging. Herein, a series of KCo_{1‐ x} Fe _x F₃ was rationally synthesized, in which the leaching of K⁺ and F⁻ ions induces self‐reconstruction during alkaline OER. Moreover, the introduction of Fe enhances the ionicity of the M−F bonds, promoting bond dissociation and inducing local structural disordering, strengthening the electrochemical reconstruction. In situ Raman reveals that the Co sites serve as the catalytic centers within the amorphous (Co, Fe)OOH active layer, and Fe incorporation increases the formation of active species and promotes deep reconstruction. Further studies reveal that Fe doping enables the OER mechanism to switch from a dual pathway of the adsorbate evolution mechanism and lattice oxygen‐mediated mechanism (LOM) to a LOM, thus improving the OER performance. The reconstructed KCo_0.7Fe_0.3F₃ delivers a low overpotential of 257 mV at a current density of 10 mA cm⁻² and remains stable for over 100 h at 100 mA cm⁻², superior to that of commercial RuO₂. This work provides a rational method for deep reconstruction of the catalyst and fundamental insights into enhancing OER catalytic performance.

The Fe‐doped KCoF₃ (KCo_0.7Fe_0.3F₃) achieves excellent oxygen evolution reaction (OER) performance, where the leaching of K⁺ and F⁻ triggers self‐reconstruction under alkaline OER conditions. Meanwhile, Fe incorporation boosts the formation of (oxy)hydroxide active species and promotes deep reconstruction, greatly improving the catalytic activity.

1. Introduction

With the energy shortage and growing environmental concerns, the development and utilization of renewable energy sources have become increasingly important [1]. Hydrogen is seen as a crucial sustainable energy carrier, owing to its zero‐carbon emission, abundant sources, and high energy density [2]. Green hydrogen production can be achieved through a viable strategy by using renewable electricity to power electrochemical water splitting. The overall energy conversion efficiency of electrochemical water splitting is fundamentally governed by the anodic oxygen evolution reaction (OER) [3]. Currently, the widespread application of noble‐metal‐based OER electrocatalysts, such as RuO₂ and IrO₂, is hampered by their scarcity, high cost, and poor electrochemical stability [4]. Although non‐noble‐metal‐based oxides and hydroxides show promising OER performance, they still suffer from insufficient active sites exposure and low intrinsic activity, which restricts their practical applications [5, 6, 7, 8].

The formation of active sites is associated with the in situ reconstruction of the catalyst during the OER [9]. It is widely recognized in the field that many catalysts undergo structural reconstruction, yielding active species with catalytic performance far superior to that of their pristine counterparts [5, 10]. Maximizing the degree of structural reconstruction can promote the conversion of a larger proportion of catalyst into active species and induce the formation of defect‐rich, low‐crystallinity structures, increasing the amount of exposed active sites and improving mass transfer kinetics [11, 12]. However, for most reported catalysts, reconstruction remains confined to the near‐surface region. Although the (oxy)hydroxide shell can enhance OER performance, the internal components remain largely underutilized [13, 14]. Various strategies have been explored to promote deep reconstruction, including ionic doping [15], defect engineering [16], and the introduction of leachable components [17]. These approaches aim to increase the specific surface area, induce amorphization, and achieve multicomponent coleaching, facilitating more extensive structural evolution during the OER process [18]. Although these strategies have enhanced the reconstruction to some degree, achieving deep reconstruction remains challenging [19]. Therefore, exploring catalysts that possess both intrinsic reconstruction propensity and compositional tunability has become crucial for overcoming the current bottleneck [9, 20].

Perovskite fluorides (ABF₃) have aroused widespread attention in diverse fields, benefiting from their tunable composition and flexible electronic structure [21]. Alkali metal cations typically occupy the A‐site of ABF₃, whereas transition metal ions (e.g., Co, Fe) reside at the B‐site [22]. Many studies have demonstrated that materials containing alkali metals are conducive to cation leaching, while anions such as F⁻, Cl⁻, P³⁻, and S²⁻ also commonly undergo leaching during the OER process [23, 24, 25]. The coleaching of cation and anion can significantly promote and deepen the structural reconstruction of catalysts [24]. As the most electronegative element, fluorine (F) not only serves as a promising n‐type dopant but also leads to stronger ionicity due to its tendency to form weak M−F bonds [26]. These weak M−F bonds are prone to hydrolytic cleavage under alkaline OER, leading to the dissolution of F⁻ from the lattice and its subsequent occupation by OH⁻ from the electrolyte [27]. This process induces reconstruction of the ABF₃ and finally transforms it into a highly active (oxy)hydroxide phase, which acts as the real catalytic active center for boosting OER performance [28, 29]. This inherent self‐reconstruction behavior makes ABF₃ an excellent candidate for exploring the relationship between reconstruction and enhanced OER performance [30]. Additionally, B‐site doping modification can further regulate the ionicity and bond strength of M−F bonds, optimize the electronic structure, and achieve more thorough deep reconstruction [31].

Motivated by the above considerations, a series of Fe‐doped perovskite fluoride catalysts, KCo_1‐xFe_xF₃ (x = 0, 0.1, 0.2, 0.3, 0.4), was synthesized via a solvothermal method. The leaching of K⁺ and F⁻ ions induces self‐reconstruction, while the introduction of Fe increases the ionicity of the M−F bond, making it easier to dissociate and inducing local structural disorder, accelerating the electrochemical reconstruction process during the OER. In situ Raman spectroscopy demonstrates that Fe doping can remarkably reduce the reconstruction onset potential and promote the formation of an amorphous (Co, Fe)OOH active layer, which acts as the real active species, thus enhancing the OER activity. Consequently, the reconstructed KCo_0.7Fe_0.3F₃ surpasses commercial RuO₂ in electrocatalytic activity and stability. This work demonstrates that ion leaching and Fe doping together promote deep reconstruction of the catalyst, offering new perspectives and effective methods for designing advanced OER electrocatalysts.

2. Results and Discussion

A series of KCo_1‐xFe_xF₃ was successfully synthesized via a solvothermal method (Figure 1a). The x‐ray diffraction (XRD) was used to analyze the phase and structure of the KCo_{1‐ x} Fe _x F₃ (Figure 1b). All diffraction peaks match the cubic perovskite fluoride structure of KCoF₃ (Pm‐3m, PDF#97–001–5425) without detectable impurities, confirming the successful preparation of pure‐phase perovskite fluoride crystals. Notably, with increasing Fe doping content, the characteristic diffraction peaks gradually shift to lower 2θ angles, which is indicative of lattice expansion arising from the incorporation of larger Fe³⁺ ions replacing smaller Co³⁺ ions [32]. Scanning electron microscopy (SEM) was used to characterize the morphologies of the Fe‐doped samples. The KCo_0.7Fe_0.3F₃ (Figure 1c) exhibits an agglomerated morphology that consists of numerous nanoparticles. This morphology was consistently observed for the other samples as well (Figure S1). The microstructure of the KCo_0.7Fe_0.3F₃ was further examined by transmission electron microscopy (TEM) (Figure 1d). Nanoparticles with sizes of approximately 20–50 nm are clearly discernible. High‐resolution TEM (HRTEM) analysis (Figure 1e) displays clear lattice fringes at 0.297 nm, corresponding to the (110) plane of the cubic perovskite structure of KCoF₃. These nanoscale building blocks offer a larger specific surface area and more active sites, while their small size effectively shortens the electron transport pathway, thereby facilitating charge transfer kinetics [33, 34, 35, 36]. In addition, the homogeneous distribution of K, Co, Fe, and F in KCo_0.7Fe_0.3F₃ is demonstrated by high‐angle annular dark‐field scanning transmission electron microscopy (HAADF‐STEM) and corresponding energy‐dispersive X‐ray (EDX) elemental mappings (Figure 1f), confirming the successful incorporation of Fe into the perovskite lattice. EDX analysis confirms that the atomic ratios for all the samples are consistent with the nominal stoichiometries (Figure S2).

FIGURE 1.

(a) Schematic illustration of the preparation of KCo_{1‐ x} Fe _x F₃. (b) XRD profiles of KCoF₃, KCo_0.9Fe_0.1F₃, KCo_0.8Fe_0.2F₃, KCo_0.7Fe_0.3F₃, and KCo_0.6Fe_0.4F₃. (c) SEM image, (d) TEM image, (e) HRTEM image, and (f) HAADF‐STEM image and corresponding elemental mapping images of KCo_0.7Fe_0.3F₃.

To investigate the correlation between surface elemental states and electrocatalytic OER performance of KCoF₃ and KCo_0.7Fe_0.3F₃, X‐ray photoelectron spectroscopy (XPS) was performed. The XPS survey spectra of KCoF₃ and KCo_0.7Fe_0.3F₃ (Figure S3a) reveal the presence of K, Co, and F in the pristine KCoF₃, while an additional Fe 2p signal is detected for KCo_0.7Fe_0.3F₃. The Co 2p spectrum of KCo_0.7Fe_0.3F₃ (Figure 2a) reveals two sets of spin‐orbit doublets, which are assigned to Co³⁺ (778.2/794.3 eV) and Co²⁺ (780.2/795.8 eV), along with characteristic shake‐up satellite features [37]. Notably, the Co valence state in KCo_0.7Fe_0.3F₃ is slightly lower than that in KCoF₃, which is further confirmed by peak area analysis (Table S1). This can be ascribed to the fact that Fe has a lower electronegativity than Co, so partial electrons are transferred from Fe to Co, reducing the Co oxidation state [38]. The Fe 2p spectrum of KCo_0.7Fe_0.3F₃ (Figure 2b) is fitted into two spin‐orbit doublets, alongside two shake‐up satellite signals. The doublets at 710.4/719.2 eV and 712.9/723.4 eV correspond to Fe²⁺ and Fe³⁺ species, respectively [39]. Compared with KFeF₃, the Fe 2p peaks of KCo_0.7Fe_0.3F₃ shift to higher binding energies. The peak area fitting results indicate an increased Fe³⁺ proportion in KCo_0.7Fe_0.3F₃ (Table S2), suggesting that electrons are transferred from Fe to Co. In addition, the F 1s spectrum (Figure S3b) shows that the M–F peak of KCo_0.7Fe_0.3F₃ appears at a lower binding energy than that of KCoF₃, implying that Fe doping enhances the electron density around F atoms and M–F bond ionicity. This promotes the cleavage of the M–F bond during OER, providing favorable conditions for triggering deep reconstruction [40].

FIGURE 2.

(a) Co 2p XPS spectra of KCoF₃ and KCo_0.7Fe_0.3F₃. (b) Fe 2p XPS spectra of KCo_0.7Fe_0.3F₃ and KFeF₃. (c) Normalized XANES spectra at the Co K‐edge, and (d) Fourier‐transformed EXAFS patterns for KCoF₃ and KCo_0.7Fe_0.3F₃. (e) Normalized XANES spectra at the Fe K‐edge, and (f) Fourier‐transformed EXAFS patterns for KCo_0.7Fe_0.3F₃. (g) and (h) Co K‐edge EXAFS data wavelet transforms for KCoF₃ and KCo_0.7Fe_0.3F₃. (i) Fe K‐edge EXAFS data wavelet transform for KCo_0.7Fe_0.3F₃.

The X‐ray absorption spectroscopy (XAS), including x‐ray absorption near‐edge structure (XANES), and extended x‐ray absorption fine structure (EXAFS) spectroscopy, was further employed to investigate the electronic structure and local coordination of KCoF₃ and KCo_0.7Fe_0.3F₃. The Co K‐edge XANES spectra of KCoF₃ and KCo_0.7Fe_0.3F₃ are displayed in Figure 2c. The absorption edges of KCoF₃ and KCo_0.7Fe_0.3F₃ lie between those of CoO and Co₃O₄, suggesting the average Co oxidation state is between +2 and +3. Compared with KCoF₃, the absorption edge shifts to lower energy for KCo_0.7Fe_0.3F₃, which indicates that Fe doping reduces the average oxidation state of Co, in good agreement with the XPS findings. In the Co K‐edge EXAFS Fourier‐transformed spectra (Figure 2d), the peaks of KCoF₃ and KCo_0.7Fe_0.3F₃ at R ≈ 1.6 Å corresponds to the Co–F coordination bond. In the second and higher shell coordination regions, both samples exhibit peaks at R ≈ 3.1 Å and R ≈ 3.8 Å, which are attributed to the Co–F–K and Co–F–M (M = Co/Fe) scattering paths, respectively [31, 41]. Notably, these peaks exhibit weaker intensity in KCo_0.7Fe_0.3F₃ relative to KCoF₃. This arises from Fe‐doping‐induced local structural disorder, which results in a slight reduction in crystallinity [42]. Low crystallinity is accompanied by abundant structural defects, which in turn accelerate the reconstruction process [16, 42]. Figure 2e displays the Fe K‐edge XANES spectra of KCo_0.7Fe_0.3F₃. In the Fe K‐edge EXAFS Fourier‐ transform spectra (Figure 2f), the Fe–F coordination bond appears at R ≈ 1.5 Å in KCo_0.7Fe_0.3F₃, confirming the formation of a stable Fe–F coordination bond [43]. In contrast to the Fe foil reference, which shows a characteristic Fe–Fe metal bond peak at R ≈ 2.2 Å, no such feature is observed for KCo_0.7Fe_0.3F₃, indicating the absence of Fe clusters within the perovskite lattice. Additionally, weak peaks corresponding to Fe–F–K and Fe–F–M scattering paths are observed at R ≈ 3.3 Å and R ≈ 3.8 Å, respectively [44]. The low intensity of these features suggests a disordered local coordination environment around Fe, likely arising from local lattice strain induced by the ionic radius mismatch between Fe³⁺/Fe²⁺ and Co³⁺/Co²⁺ [23, 44, 45].

Wavelet transform (WT) analysis was performed to resolve the backscattering atoms in both k‐space and R‐space (Figure 2g–i). For KCoF₃ (Figure 2g) and KCo_0.7Fe_0.3F₃ at the Co K‐edge (Figure 2h), the WT contour plots show intensity maxima at R ≈ 1.6 Å with k ≈ 4–5 Å⁻¹, corresponding to Co–F coordination. Weaker signals at higher R ≈ 3.1 Å and R ≈ 3.8 Å are attributed to Co–F–K and Co–F–M paths, consistent with the EXAFS analysis [41]. For the Fe K‐edge WT of KCo_0.7Fe_0.3F₃ (Figure 2i), the intensity maximum at R ≈ 1.5–1.6 Å with k ≈ 4–5 Å⁻¹ is assigned to Fe–F coordination [44]. These WT results corroborate the EXAFS findings.

To study the effect of Fe doping on the OER performance of KCoF₃, linear sweep voltammetry (LSV) tests were carried out on a series of Fe‐doped KCoF₃ catalysts (KCoF₃, KCo_0.9Fe_0.1F₃, KCo_0.8Fe_0.2F₃, KCo_0.7Fe_0.3F₃, and KCo_0.6Fe_0.4F₃) in 1.0 M KOH. The results show that Fe incorporation markedly improves the OER activity (Figure 3a). Specifically, KCo_0.7Fe_0.3F₃ exhibits the best performance, achieving overpotentials of only 257 and 312 mV at 10 and 100 mA cm⁻², which are markedly lower than those of KCoF₃ (η ₁₀ = 361 mV, η ₁₀₀ = 426 mV), KCo_0.9Fe_0.1F₃ (η ₁₀ = 292 mV, η ₁₀₀ = 353 mV), KCo_0.8Fe_0.2F₃ (η ₁₀ = 283 mV, η ₁₀₀ = 330 mV), KCo_0.6Fe_0.4F₃ (η ₁₀ = 268 mV, η ₁₀₀ = 318 mV), and RuO₂ (η ₁₀ = 262 mV, η ₁₀₀ = 399 mV). Tafel slopes and electrochemical impedance spectroscopy (EIS) analyses were performed to probe the kinetic effect of Fe doping on the OER process. As shown in Figure 3b, the Tafel slope of KCo_0.7Fe_0.3F₃ is the smallest (50.93 mV dec⁻¹), indicating that it possesses the fastest interfacial reaction kinetics [46]. Furthermore, the EIS profiles in Figure 3d demonstrate that KCo_0.7Fe_0.3F₃ displays the smallest semicircle diameter, with a minimum charge‐transfer resistance (R _ct) of only 11.4 Ω (Table S3). This is further supported by in situ EIS measurements (Figure S4), where KCo_0.7Fe_0.3F₃ exhibits a characteristic frequency that shifts to higher values as the potential increases and also shows a higher characteristic frequency and a lower phase angle than KCoF₃ at all measured potentials. These results confirm that Fe doping enhances the interfacial charge transfer process, thus accelerating the OER kinetics [47]. The CV curves recorded within the non‐Faradaic region (Figure S5) were used to determine the double‐layer capacitance (C _dl) of the catalysts. As shown in Figure 3e, KCo_0.7Fe_0.3F₃ displays the largest C _dl value (65.01 mF cm⁻²), outperforming other samples. From the measured C _dl, the electrochemically active surface area (ECSA) of KCo_0.7Fe_0.3F₃ was calculated to be 319 cm² (Figure S6), suggesting that the KCo_0.7Fe_0.3F₃ exposes more active sites and provides more abundant reaction centers [48]. Figure S7 shows the ECSA‐normalized LSV results, which exhibit a consistent trend with the original LSV curves, demonstrating that Fe doping can expose more active sites and significantly boost the intrinsic catalytic activity.

FIGURE 3.

(a) LSV curves, (b) Tafel slopes, (c) overpotential and Tafel slope columns, (d) EIS Nyquist plots, and (e) C_dl of KCoF₃, KCo_0.9Fe_0.1F₃, KCo_0.8Fe_0.2F₃, KCo_0.7Fe_0.3F₃, and KCo_0.6Fe_0.4F₃. (f) Chronoamperometry curves of KCoF₃, KCo_0.7Fe_0.3F₃, and RuO₂ (The inset shows the initial current rise of KCoF₃ and KCo_0.7Fe_0.3F₃, followed by stabilization, corresponding to the catalyst activation process). (g) Comparisons of the overpotential of KCo_0.7Fe_0.3F₃ with reported electrocatalysts at 10 mA cm⁻². (h) AEMWE device using KCo_0.7Fe_0.3F₃ as an anode and Pt/C as a cathode. (i) Polarization curves of the Pt/C||KCo_0.7Fe_0.3F₃ water electrolyzer. (j) Chronoamperometric durability test of the Pt/C||KCo_0.7Fe_0.3F₃ AEMWE device at a constant cell voltage, with an initial current density of 0.5 A cm⁻² at 25°C. (k) Comparison of experimental and theoretical O₂ production yields at 100 mA cm⁻² and corresponding Faradaic efficiency.

The long‐term stability of KCoF₃ and KCo_0.7Fe_0.3F₃ during the OER process was evaluated by chronoamperometry, with the results shown in Figure 3f. The KCo_0.7Fe_0.3F₃ exhibits a continuous current rise, which is associated with the gradual generation of active (oxy)hydroxide by in situ self‐reconstruction, eventually reaching a steady state. In contrast, the KCoF₃ shows only a slight initial current increase and subsequently remains at a steady state. This indicates that Fe doping can enhance OER activity and promote the formation of the active phase during the reaction [49]. After this activation process, the reconstructed KCo_0.7Fe_0.3F₃ can maintain stable operation for over 100 h at 100 mA cm⁻², outperforming KCoF₃ and commercial RuO₂, verifying that Fe incorporation can boost the electrochemical stability of the catalyst. In addition, KCo_0.7Fe_0.3F₃ exhibits superior OER activity at 10 mA cm⁻² compared with most reported electrocatalysts (Figure 3g and Table S4).

To evaluate the industrial application potential of KCo_0.7Fe_0.3F₃, an anion exchange membrane water electrolysis cell (AEMWE) was constructed with the KCo_0.7Fe_0.3F₃ as an anode catalyst and Pt/C as a cathode catalyst (Figure 3h). The LSV and stability tests show that the Pt/C||KCo_0.7Fe_0.3F₃ electrolyzer achieves a cell voltage of only 2.2 V at 500 mA cm⁻² at 25°C, and it can operate stably for 40 h (Figure 3i–j). Furthermore, the Faradaic efficiency for oxygen evolution was quantified by the water displacement method at 100 mA cm⁻² during alkaline water electrolysis with the Pt/C||KCo_0.7Fe_0.3F₃ electrolyzer (Figure S8). As shown in Figure 3k, the experimentally measured oxygen volume closely aligns with the theoretical yield derived from the passed charge, verifying a high Faradaic efficiency of 97.9%. This confirms the outstanding OER energy conversion efficiency of KCo_0.7Fe_0.3F₃. These results demonstrate that KCo_0.7Fe_0.3F₃ is a high‐performance OER electrocatalyst with great practical application potential in the field of industrial water electrolysis to produce hydrogen.

The post‐OER KCo_0.7Fe_0.3F₃ was characterized to probe its structural evolution. The characteristic diffraction peaks of the pristine KCo_0.7Fe_0.3F₃ vanish after the OER test, replaced by a broad hump indicative of an amorphous phase, with no new crystalline phases detected (Figure 4a). This indicates thorough amorphization and deep reconstruction. The Raman spectrum in Figure 4b reveals distinct peaks at ∼459 and 504 cm⁻¹, which are ascribed to CoOOH, confirming that KCo_0.7Fe_0.3F₃ has been reconstructed into an (oxy)hydroxide phase [50]. Notably, a strong peak at ∼667 cm⁻¹, assigned to the A_1g vibrational mode of Co₃O₄, is also observed. However, as no crystalline Co₃O₄ phase is detected by XRD, this Raman signal is tentatively attributed to a local phase transition from CoOOH to Co₃O₄ induced by laser heating during the measurement [51]. Furthermore, Fourier‐transform infrared (FTIR) spectroscopy was carried out on the KCo_0.7Fe_0.3F₃ before and after the OER (Figure 4c). Before the OER, only a narrow peak attributed to adsorbed water is observed at 3280 cm⁻¹. The post‐OER KCo_0.7Fe_0.3F₃ exhibits a prominent O–H stretching signal spanning 3000–3600 cm⁻¹ region, along with a distinct M–O peak at 570 cm⁻¹ [31, 52, 53]. This further verifies that KCo_0.7Fe_0.3F₃ undergoes transformation into amorphous (oxy)hydroxide. SEM image (Figure S9) reveals that post‐OER KCo_0.7Fe_0.3F₃ exhibits a mixed nanoscale morphology consisting of granular agglomerates and nanosheets. TEM image (Figure 4d) reveals the structural evolution whereby some nanoparticles transform into nanosheets upon reconstruction, with the darker contrast in the central region corresponding to the stacking of nanoscale building blocks. This nanosheet structure remarkably enlarges the specific surface area, thereby enabling abundant catalytic sites to be exposed [43]. HRTEM image (Figure 4e) exhibits only short‐range ordered lattice fringes with no detectable long‐range order, confirming that the KCo_0.7Fe_0.3F₃ has transformed into a predominantly amorphous (Co,Fe)OOH active phase [54]. Multi‐region HRTEM images (Figure S10) further verify the fully amorphous nature of the particles, confirming deep reconstruction. Corresponding elemental mapping images (Figure 4f) demonstrate that K, Co, Fe, F, and O remain dispersed in the electrochemically reconstructed catalyst. Fe is homogeneously distributed with Co and O, without obvious phase separation or agglomeration. As shown in Figures S2d and S11b, EDX analysis confirms a significant enrichment of O, accompanied by a marked depletion of F⁻ and K⁺ in the post‐OER KCo_0.7Fe_0.3F₃. Notably, the Co/Fe atomic ratio shows negligible change before and after OER. As illustrated in Figure S12, Inductively Coupled Plasma‐Mass Spectrometry (ICP‐MS) measurements further reveal a remarkable rise of dissolved K⁺ in the electrolyte after OER. Although the dissolved Fe concentration is slightly higher than that of Co, both concentrations show only minimal increases relative to the blank electrolyte, confirming no significant metal leaching and that Fe remains retained in the reconstructed active phase. During OER, the leaching of F⁻ creates anionic vacancies that are then occupied by OH⁻ from the solution [55, 56, 57]. However, the sluggish kinetics of the OH⁻ filling process leads to the transient accumulation of vacancies, ultimately collapsing the crystalline lattice and resulting in amorphization [57]. To clarify the surface chemical states of the catalyst, XPS measurements were carried out. The survey spectrum (Figure S13) confirms that the post‐OER KCo_0.7Fe_0.3F₃ contains K, Co, Fe, F, and O. Peak deconvolution of Co 2p and Fe 2p spectra (Figure 4g–h) indicates a distinct positive shift, confirming the oxidation states of Co and Fe have increased. Furthermore, the peak fitting results (Tables S5–S6) show that the relative contents of Co²⁺ and Fe²⁺ decrease, while those of Co³⁺ and Fe³⁺ increase correspondingly, which aligns with the shifts in binding energies [37]. This indicates that OER‐induced oxidation of metal centers promotes the generation of high‐oxidation‐state intermediates, which optimizes OER kinetics and contributes to the significantly reduced overpotential. The F 1s XPS spectra (Figure S14) further confirm the leaching of F⁻. The intensity of the M−F peak decreases significantly after OER, indicating substantial leaching of F⁻ from KCo_0.7Fe_0.3F₃. Moreover, a new peak appears after OER, which is attributed to the C−F bond from the Nafion [57]. Figure 4i displays the deconvoluted O 1s spectrum, with three peaks attributed to M–O, M–OH, and adsorbed H₂O (H₂O_ads) [43]. This result further confirms the generation of (oxy)hydroxides, consistent with the Raman and FTIR analyses.

FIGURE 4.

(a) XRD patterns, (b) Raman spectra, and (c) FTIR spectra of KCo_0.7Fe_0.3F₃ before and after OER. (d) TEM image, (e) HRTEM image, and (f) elemental mapping images of KCo_0.7Fe_0.3F₃ after OER. (g) Co 2p, (h) Fe 2p, (i) O 1s XPS spectra of KCo_0.7Fe_0.3F₃ before and after OER.

To clarify the mechanism by which Fe doping accelerates reconstruction during OER, a comparative analysis of the phase evolution of KCo_0.7Fe_0.3F₃ and KCoF₃ before and after OER was performed (Figure 4a, Figure S15). After OER, the original crystalline diffraction peaks of KCoF₃ disappear, indicating that the KCoF₃ undergoes amorphization. Consistent with the behavior of KCo_0.7Fe_0.3F₃, this result demonstrates that amorphization represents a common structural evolution trend for both samples. Raman spectroscopic analysis (Figure S16) reveals that KCoF₃ exhibits a broad signal from 420 to 620 cm⁻¹ after OER, which can be attributed to CoOOH [51]. Meanwhile, a characteristic signal at ∼667 cm⁻¹ corresponds to the A_1g vibrational mode of Co₃O₄. EDX analysis reveals that KCoF₃ and KCo_0.7Fe_0.3F₃ undergo significant depletion of F⁻ and K⁺ after OER (Figure S2a, d, and S11). Compared with KCoF₃, KCo_0.7Fe_0.3F₃ exhibits a more marked depletion of F⁻ and K⁺ and a greater enrichment of O after OER, indicating that Fe doping effectively promotes ion leaching and drives more complete structural reconstruction. ICP‐MS results (Figure S12a) confirm more significant K⁺ dissolution from KCo_0.7Fe_0.3F₃ than from KCoF₃, further corroborating the above conclusions. To probe the dynamic structural reconstruction during OER and clarify the influence of Fe doping on reconstruction kinetics, in situ Raman spectroscopy was performed at 0.1 V intervals from the open circuit potential (OCP) to 1.5 V versus RHE. As shown in Figure 5a, KCoF₃ exhibits no distinct Raman signals below 1.2 V versus RHE. Upon reaching 1.3 V versus RHE, a broad peak attributable to CoOOH appears and subsequently intensifies with increasing potential. Upon reaching 1.5 V versus RHE, the A_1g vibrational mode of Co₃O₄ gives rise to an obvious signal at 682 cm⁻¹ [50, 51]. All Fe‐doped samples exhibit two distinct characteristic peaks of CoOOH. The KCo_0.9Fe_0.1F₃ (Figure S17a) exhibits CoOOH bands at 461 and 524 cm⁻¹ when the potential reaches 1.2 V versus RHE. The CoOOH peaks of KCo_0.8Fe_0.2F₃ (Figure S17b) can be detected at 449 and 517 cm⁻¹, with a reconstruction onset potential of 1.2 V versus RHE. The KCo_0.7Fe_0.3F₃ undergoes reconstruction at 1.1 V versus RHE, with the appearance of CoOOH vibrational features at 446 and 511 cm⁻¹. When the potential reaches 1.4 V versus RHE, a weak signal attributable to the A_1g vibrational mode of Co₃O₄ appears at 680 cm⁻¹ (Figure 5b) [51, 58]. As shown in Figure S17c, when the Fe doping level is further increased, the CoOOH characteristic peaks of KCo_0.6Fe_0.4F₃ further shift to 443 and 508 cm⁻¹, and its reconstruction onset potential increases back to 1.2 V versus RHE. The intensity of the CoOOH characteristic peaks for all Fe‐doped samples increases significantly with increasing potential, and a continuous redshift to lower wavenumbers is observed as Fe doping level rises. These results confirm that appropriate Fe doping significantly lowers the reconstruction onset potential and accelerates structural transformation, thus driving a more complete deep reconstruction. Notably, the CoOOH Raman signals of Fe‐doped catalysts are much narrower and sharper than the single broad CoOOH peak of reconstructed KCoF₃, which confirms that Fe is preserved within the reconstructed active phase [59, 60]. Furthermore, dynamic contact angle measurements reveal that Fe doping improves the surface wettability of the catalysts (Figure S18). With an increasing Fe doping level, the surface wettability of the catalysts first increases and then decreases. Among all samples, KCo_0.7Fe_0.3F₃ exhibits the optimal surface wettability, with an initial contact angle of 12.8° decreasing to 6.9° within 2 s and the droplet spreading completely after 2 s. The other doped samples and KCoF₃ show higher initial contact angles and slower droplet spreading rates. These results demonstrate that an appropriate Fe doping level optimizes the surface wettability of the catalyst in alkaline electrolyte, facilitating sufficient electrode‐electrolyte contact and providing favorable interfacial conditions for the cleavage of M–F bonds and the infiltration of OH⁻, thereby further promoting the deep reconstruction process [61, 62].

FIGURE 5.

In situ Raman spectra of (a) KCoF₃ and (b) KCo_0.7Fe_0.3F₃. (c) LSV curves of KCoF₃ and KCo_0.7Fe_0.3F₃ at different pH values. (d) pH‐dependent logarithmic current density of KCoF₃ and KCo_0.7Fe_0.3F₃ at 1.6 V versus RHE. (e) In situ ATR‐FTIR spectra of KCoF₃ and (f) KCo_0.7Fe_0.3F₃. (g) Schematic diagram of the LOM, (h) Schematic representation illustrating the reconstruction process of KCo_0.7Fe_0.3F₃ during OER.

To elucidate the boosted OER mechanism of KCo_0.7Fe_0.3F₃, the reaction pathways of KCoF₃ and KCo_0.7Fe_0.3F₃ were further investigated. Figure 5c demonstrates that increasing pH enhances the OER activity of KCoF₃, with an even stronger enhancement observed for KCo_0.7Fe_0.3F₃. The pH dependence can be quantified by the proton reaction order on the RHE scale (ρ = ∂log(j)/∂pH), where a ρ value close to 1 signifies a more significant pH dependence [57, 63]. KCo_0.7Fe_0.3F₃ exhibits a ρ value of 0.88 (Figure 5d), which is substantially higher than that of KCoF₃ (0.51), confirming its stronger pH‐dependent OER activity and suggesting a lattice oxygen‐mediated mechanism (LOM) reaction pathway. As shown in Figure S19, the current density of KCo_0.7Fe_0.3F₃ decreases markedly in 1 M tetramethylammonium hydroxide (TMAOH), indicating that the OER process of KCo_0.7Fe_0.3F₃ is dominated by the LOM pathway. In contrast, KCoF₃ exhibits only a slight decay in OER activity in TMAOH, which confirms a mixed reaction pathway involving both the adsorbate evolution mechanism (AEM) and LOM [64]. In situ ATR‐FTIR spectroscopy measurements were used to track surface intermediate evolution during OER. As shown in Figure 5e, KCoF₃ shows no distinct peaks at OCP. At a potential of 1.1 V versus RHE, double peaks emerge at nearly 1050 and 1250 cm⁻¹, which are attributed to the M–O–O and O–O intermediates, respectively [65]. Further increasing the potential to 1.3 V versus RHE, OOH intermediates emerge at approximately 1190 cm⁻¹ [66]. This indicates that KCoF₃ follows a mixed reaction pathway involving both AEM and LOM. Figure 5f shows that KCo_0.7Fe_0.3F₃ exhibits no distinct peaks at OCP, while double peaks emerge at about 1050 and 1250 cm⁻¹ at 1.1 V versus RHE, which correspond to M–O–O and O–O species, respectively [65]. Their intensities increase significantly with increasing potential, demonstrating that KCo_0.7Fe_0.3F₃ follows a predominantly LOM‐dominated reaction pathway [65]. These in situ ATR‐FTIR results provide direct spectroscopic evidence for the participation of lattice oxygen in the OER process of KCo_0.7Fe_0.3F₃, and the corresponding reaction mechanism is schematically illustrated in Figure 5g. Figure 5h illustrates the structural evolution mechanism of KCo_0.7Fe_0.3F₃ during OER. In the initial stage, F⁻ vacancies are generated and immediately occupied by OH⁻ from the electrolyte, thus initiating the subsequent OER reactions. During the reaction process, (Co, Fe)OOH active species are generated on the surface at the Co and Fe sites of KCo_0.7Fe_0.3F₃, accompanied by the leaching of F⁻ and K⁺. This promotes the cleavage of M–F and M–K bonds and structural reconstruction. As the reaction proceeds, the continuous leaching of F⁻ and K⁺ drives more complete structural and chemical transformations, ultimately forming an amorphous (Co, Fe)OOH active phase. Notably, it is widely acknowledged that amorphous (oxy)hydroxides represent the core active components driving alkaline OER activity in Co/Fe‐based materials [67, 68]. This further demonstrates that the essence of Fe‐doping‐driven deep reconstruction lies in the efficient conversion of the KCo_0.7Fe_0.3F₃ precursor into highly active (Co, Fe)OOH phases, thus offering a rational design strategy for the development of advanced alkaline OER electrocatalysts.

3. Conclusion

In summary, the as‐prepared perovskite fluoride KCo_1‐xFe_xF₃ undergoes self‐reconstruction by utilizing the leachable characteristics of K⁺ and F⁻ ions. The introduction of Fe enhances the ionicity of M−F bonds and induces local structural disordering, boosting the electrochemical reconstruction of KCo_0.7Fe_0.3F₃. In situ Raman spectroscopy reveals that Fe doping markedly lowers the reconstruction onset potential and facilitates the formation of an amorphous (Co, Fe)OOH active layer. In situ ATR‐FTIR spectroscopy confirms that Fe doping switches the OER pathway from AEM and LOM to the LOM, greatly improving the OER performance. Consequently, the reconstructed KCo_0.7Fe_0.3F₃ outperforms commercial RuO₂ catalysts in electrocatalytic activity and long‐term durability. This work offers a rational route for designing deeply reconstructed electrocatalysts and deepens the understanding of the role of doping in promoting OER catalysis.

Supporting information

Supporting Information: smll74659‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.