Synthesis and characterizations of the Cr-containing anodeCrO42–-NiFe LDH/Cr2O3/NF was obtained by immersing hydrothermally prepared NiFe LDH/NF in an aqueous solution containing sodium chromate (Supplementary Note 1). As depicted in Fig. 1a, Cr species exist in the form of oxidized Cr (i.e., Cr2O3) as well as anionic CrO42– inserted between basal planes. For CrO42–-NiFe LDH/Cr2O3/NF, strong peaks corresponding to the (003) and (006) crystal faces demonstrate that the Cr-containing catalyst is still present as a layered compound (Fig. 1b). While the LDH structure remains mostly intact, the shift of the interlayer space peak (003) suggests an expansion in the layer spacing of 1.13 Å37,38,39,40,41. Such an increase in the interlayer spacing and the Raman peak of CrO42– (890.2 cm–1, Fig. 1c) both imply successful insertions of CrO42–40,41,42,43. Moreover, the Raman peaks at 308.9 cm–1 and 630.9 cm–1 (Fig. 1c) suggest the generation of Cr2O344,45,46. Scanning electron microscopy (SEM) images for CrO42–-NiFe LDH/Cr2O3/NF (Fig. 1d) only confirm the sheet array structure of CrO42–-NiFe LDH/Cr2O3, which is not markedly distinct from the sheet-like NiFe LDH morphology (Supplementary Fig. 1). Besides, SEM images in Fig. 1d do not detect any structures that are possibly Cr2O3, suggesting that the Cr2O3 is present at much smaller sizes. Transmission electron microscopy (TEM) image for CrO42–-NiFe LDH/Cr2O3 (Fig. 1e) displays that a substantial amount of nanoparticles (dotted circles) was produced on the LDH, which may stem from reactions between the CrO42– with Ni2+/Fe2+. According to high-resolution TEM images (Fig. 1f), ultra-small nanoparticles are crystal Cr2O3 (interplanar distance of 0.247 nm for the (110) plane), and these nano-Cr2O3 form a large number of interfaces with NiFe LDH (0.279 nm, the (012) plane), which should contribute to improved eASO activities. High-resolution TEM images prove the good crystallinity of Cr2O3, whereas Cr2O3 cannot be detected by X-ray diffraction (XRD) in Fig. 1b, thereby suggesting a rather low content of Cr2O3. Cr is evenly distributed throughout in NiFe LDH without aggregation (Fig. 1g), albeit at the low Cr level. TEM-element energy dispersive spectroscopy (EDS) mapping analysis reveals that the Cr content in the freshly prepared CrO42–-NiFe LDH/Cr2O3 is about 1.62%. (Supplementary Fig. 2 and Supplementary Table 2). The above characterizations illustrate that the Cr(VI) species are intercalated in the basal planes by electrostatic interaction with NiFe LDH, while the Cr(III) species are also uniformly dispersed on the nanosheet structure by forming an interface with the NiFe LDH.Fig. 1: Material characterizations.a Schematic diagram of synthesizing the electrode. b XRD patterns and (c) Raman spectra of CrO42–-NiFe LDH/Cr2O3/NF and NiFe LDH/NF. d SEM images of CrO42–-NiFe LDH/Cr2O3/NF. e TEM and (f) high-resolution TEM images of CrO42–-NiFe LDH/Cr2O3. g TEM image and corresponding elemental distributions of CrO42–-NiFe LDH/Cr2O3. XPS spectra of CrO42–-NiFe LDH/Cr2O3 and CrO42–-NiFe LDH in the (h) Ni 2p, and (i) Fe 2p regions. j XPS spectra of CrO42–-NiFe LDH/Cr2O3, Cr2O3, CrCl3, and NaCr2O4 in the Cr 2p region.X-ray photoelectron spectroscopy (XPS) survey spectrum of CrO42–-NiFe LDH/Cr2O3 confirms the characteristic Cr peak (Supplementary Fig. 3). XPS spectra in the regions of Ni 2p, Fe 2p, and Cr 2p can reveal surface chemistry changes after introducing Cr species. Ni 2p and Fe 2p peaks shift towards higher binding energy (BE) obviously (Fig. 1h, i), indicating the increased valence states of both Ni and Fe species for CrO42–-NiFe LDH/Cr2O3. Indeed, the ratios of Ni2+/Ni3+ and Fe2+/Fe3+ dropped significantly, from 4.02:1 to 1.75:1 and 4.22:1 to 3.60:1, respectively. Cr 2p XPS spectrum can be easily deconvoluted into four peaks (Fig. 1j): the BE of Cr 2p1/2 at 587.05 eV and Cr 2p3/2 at 577.30 eV that belong to Cr3+, and the BE at 588.95 and 579.65 eV related to the Cr6+47,48,49,50. Besides, the fitting results are in line with standard samples including Cr2O3, CrCl3, and Na2CrO4. XPS spectra in the region of O 1s are also provided (Supplementary Fig. 4).Enhancements in eASO activities with evenly dispersed reaction sitesCrO42–-NiFe LDH/Cr2O3/NF has higher water oxidation activity than NiFe LDH/NF, RuO2/NF, and NF in 1 M KOH solution as well as simulated alkaline seawater (Supplementary Figs. 5 and 6). As expected, CrO42–-NiFe LDH/Cr2O3/NF exhibits the highest eASO activity as well (Fig. 2a). The overall intrinsic activity is boosted with the presence of Cr species (inset of Fig. 2a and Supplementary Fig. 7). We note an increase in the electrochemically active surface area (ECSA) (Supplementary Fig. 8) as well as large decreases in the charge-transfer resistance values (Supplementary Fig. 9 and Supplementary Table 3) for CrO42–-NiFe LDH/Cr2O3/NF in contrast to NiFe LDH/NF, both of which correlate with the improved activity. Even when the current is normalized by ECSA (Supplementary Fig. 10), the Cr-containing electrode is superior to the NiFe LDH electrode at high j. Notably, the Δη/Δlog|j| values of CrO42–-NiFe LDH/Cr2O3/NF, which consistently show the smallest changes with j (Supplementary Fig. 11), imply that this Cr-containing electrode is the easiest to afford high j. Moreover, the Tafel slope for CrO42–-NiFe LDH/Cr2O3/NF only reaches 37.2 mV dec–1, which is significantly lower than those of NF, RuO2/NF, NiFe LDH/NF (Supplementary Fig. 12), and many recent alkaline seawater oxidation electrodes (Fig. 2b and Supplementary Table 4). Such a low Tafel slope again suggests that the introduction of Cr species optimizes the surface micro-environments for making O2. Furthermore, η at various j, which can be used as the measure of whether electrolysis is energy efficient, was also significantly reduced for CrO42–-NiFe LDH/Cr2O3/NF, not only far below those of NF, RuO2/NF, and NiFe LDH/NF, but also below those of some recently reported electrodes (Fig. 2b and Supplementary Table 4).Fig. 2: Remarkable eASO performance.a Evaluation of the eASO activities at a scan rate of 5 mV s−1 with 100% iR correction. Inset shows the TOF values. b Direct comparison of overpotentials at various j as well as Tafel slopes toward eASO. Mapping the eASO activities on (c) NiFe LDH/NF and (d) CrO42–-NiFe LDH/Cr2O3/NF in situ with the aid of infrared thermography. The thermographic data were recorded after chronopotentiometry tests. Electrochemical data (0, 1.5, 1.56, and 1.69 V vs. RHE) were not iR corrected.In addition, we measured the quantity of heat (q) generated on the catalyst surface in that the q as a result of charge transfer is linearly correlated with the current density (q\(\propto\)j)51,52. The recording of the heat generated was carried out before and after chronopotentiometry tests under different applied potentials, and the j of 0 mA cm–2, 50 mA cm–2, 100 mA cm–2, and 500 mA cm–2 are realized at different potentials (i.e., 0 V, 1.5 V, 1.56 V, and 1.69 V vs. RHE) for CrO42–-NiFe LDH/Cr2O3/NF. For NiFe LDH/NF, equal applied potentials were adopted to perform chronopotentiometry tests. Note that the applied potentials for Fig. 2c, d are reported without iR compensation. As shown in Fig. 2c, d, we suspended the electrode on the surface of alkaline seawater and used a Cu wire connected to a piece of CrO42–-NiFe LDH/Cr2O3/NF or NiFe LDH/NF as the working electrode. The dispersion of the charge transfer sites is indirectly reflected by recording the heat distribution of the electrode during electrolysis. At the oxidation j of 100 and 500 mA cm–2, the relatively uniform in situ heat maps for the CrO42–-NiFe LDH/Cr2O3/NF are equivalent to a relatively uniform current density distribution (Fig. 2d). In contrast, the q distribution on the NiFe LDH/NF is not as uniform as that of the Cr-containing electrode (Fig. 2c). This very high state of uniformity in j is a direct result of the even dispersion of Cr species, which proves to be a useful design in safeguarding as many reaction sites as possible.Enhancements in the eASO durability and O2 selectivityWe investigated the improvement in catalyst durability using different electrolyzers. CrO42–-NiFe LDH/Cr2O3/NF was employed directly as the working electrode in a three-electrode configuration for prolonged alkaline seawater oxidation electrolysis lasting up to 2500 h. As shown in Fig. 3a, the 2500-h stability consists of a 1000-h test under a constant j of 1 A cm–2 and a subsequent 1500-h test under 2 A cm–2. The additional 1500-h test under 2 A cm–2 was performed once we observed that the CrO42–-NiFe LDH/Cr2O3/NF was running at an exceptionally stable level at 1 A cm–2, and the results proved that the electrode remains robust. Unlike the Cr-containing electrode, the NiFe LDH/NF deactivates after just 50 h of electrolysis. Ultraviolet-visible (UV-vis) data show that CrO42–-NiFe LDH/Cr2O3/NF makes only a small amount of active chlorine after the initial 1000-h eASO test, whereas NiFe LDH/NF generates much higher levels of active chlorine after only 50 h of electrolysis (inset plot of Fig. 3a and Supplementary Fig. 13). The commercial test strip, which exhibits virtually unchanged color, can serve as a means of cross-validating the UV-vis data (inset digital photo of Fig. 3a). Besides, this Cr-containing electrode demonstrates a high Faradaic efficiency (FE) for synthesized O2 (Supplementary Fig. 14). Therefore, both the near-100% O2 generation FE and trace active chlorine production confirm the good four-electron reaction selectivity in the presence of Cr species. Moreover, the corrosion potential (Ecorr) of the CrO42–-NiFe LDH/Cr2O3/NF becomes more positive compared to that of the Cr-free NiFe LDH/NF counterpart (Supplementary Fig. 15). We further assembled the CrO42–-NiFe LDH/Cr2O3/NF into a closely packed electrolyzer, which consists of Ti end plates, gaskets, Pt/C/NF, ion exchange membrane, and CrO42–-NiFe LDH/Cr2O3/NF. The cathode and anode in membrane electrode assembly (MEA) both share a geometric area of 1 cm2. As shown in Fig. 3b, the CrO42–-NiFe LDH/Cr2O3/NF performs much better than a RuO2/NF counterpart in MEA tests. The MEA electrolyzer based on Pt/C/NF | |CrO42–-NiFe LDH/Cr2O3/NF shows largely improved stability than the Pt/C/NF | |RuO2/NF-based counterpart (Fig. 3b), with reduced estimated costs under different j (Supplementary Fig. 16a, b). The overall seawater electrolysis activity after 200-h electrolysis remains good and still exceeds that of the noble metal counterpart (Supplementary Fig. 16c). As shown in Supplementary Fig. 17a, b, Pt/C/NF | |NiFe LDH/NF electrodes exhibit poor activity and stability, being stable for electrolysis for less than 15 h at the j of 500 mA cm–2. Pt/C/NF | |NiFe LDH/NF electrodes also deactivate severely for only 36 h of operation. Note that with the Cr-containing electrode, the two-electrode MEA electrolyzer achieves a high and rather stable O2 generation efficiency throughout the 200-h operation as well (see the dotted line of pink triangles in Fig. 3b). Pt/C/NF | |NiFe LDH/NF has a two-electrode electrolysis lifespan that is superior to many previously documented two-electrode systems in alkaline seawater (Supplementary Table 5). The transport of CrO42– from the anolyte to the catholyte through the cation exchange membrane under an electric field should be challenging. To verify this, we conducted inductively coupled plasma-optical emission spectrometry (ICP-OES) tests and found almost no CrO42– in the anode solution before and after electrolysis (500 mA cm–2 for 200-h electrolysis) (Supplementary Table 6). The greatly improved selectivity and stability of the CrO42–-NiFe LDH/Cr2O3/NF led us to examine it further in a scaled-up electrolyzer consisting of single cells. We carefully customized the stacks with three pairs of cathodes and anodes (Supplementary Note 2), which have a total electrode area of 50 cm2. Remarkably, even in the electrolysis stacks, CrO42–-NiFe LDH/Cr2O3/NF with a much higher total electrode area still delivers a stable 200-h lifespan (Fig. 3c). We should point out that whether used for the typical three-electrode testing, MEA testing or scaled-up testing in electrolysis stacks, CrO42–-NiFe LDH/Cr2O3/NF reaches the most stable levels for eASO, far exceeding the vast majority of anodic electrodes available today (Fig. 3d and Supplementary Table 5). The following factors should explain why CrO42–-NiFe LDH/Cr2O3/NF is able to perform well over its super-long lifetime. (1) Cl– species are more unlikely to be adsorbed by surface sites due to the electrostatic interaction between Cl– and CrO42–. (2) OH– species occupy catalytic sites with absolute dominance owing to Cr(III) species with small sizes and high dispersion, which can successfully modify the active NiFe sites and favor the enrichment of surface OH– conducive to electrocatalysis. Thus, this surface anion species regulation (i.e., increasing the surface coverage of OH– as well as weakening the competitiveness of Cl– adsorption), achieved by Cr(III) and Cr(VI) species with high-degree dispersity, keeps the catalyst stable. Although it is difficult to directly compare the performance of NiFe LDH/Cr2O3 and CrO42– intercalated NiFe LDH toward the eASO, we tried to evaluate the respective contribution of OH–-enriching Cr(III) sites and Cr(VI) sites (CrO42–) in the CrO42–-NiFe LDH/Cr2O3/NF. Firstly, stability tests of Cr2O3/NF and bare NF in alkaline seawater and stability tests of bare NF in alkaline seawater with 40 mM sodium chromate were performed, respectively. As depicted in Supplementary Fig. 18a, Cr2O3/NF exhibits enhanced performance in 1 M KOH + seawater for OER compared to NF. As shown in Supplementary Fig. 18b, Cr2O3/NF and NF exhibit comparable stability in 1 M KOH + seawater, maintaining stable electrolysis for approximately 80 min. Noticeably, NF in alkaline seawater + 40 mM Na2CrO4 can be electrolyzed stably for over 200 min, indicating significantly enhanced stability. The results above suggest that Cr(III) sites can adsorb OH– and mostly enhance the OER activity, whereas Cr(VI) sites can repel Cl– and mostly enhance the OER stability. Furthermore, theoretical calculations were conducted to gain a better understanding of the contribution of Cr(III) sites and Cr(VI) sites (Supplementary Fig. 19) in terms of *Cl adsorption. The results show that the free energy of *Cl adsorption for NiFe LDH/Cr2O3 and CrO42–-NiFe LDH were 0.25 eV and 0.12 eV, respectively. In addition, we speculate that mechanical stability, which has a direct relationship to the oxygen bubbles’ release, is also one of the causes of the enhanced eASO stability shown by the CrO42–-NiFe LDH/Cr2O3/NF anode. As depicted in Supplementary Fig. 20a, both the NiFe LDH/NF and CrO42–-NiFe LDH/Cr2O3/NF electrodes demonstrate highly hydrophilic properties. This also suggests that it is difficult to tell the difference between their surface wettability to the electrolyte based on the water contact angle tests alone. Therefore, we further investigated the bubble contact angles under water (Supplementary Fig. 20b) and performed the adhesive forces measurements (Supplementary Fig. 20c, d). As observed, the introduction of Cr species leads to a higher air contact angle and a decreased catalyst-bubble interfacial adhesion force, both of which suggest a faster bubble evolution kinetics of CrO42–-NiFe LDH/Cr2O3/NF. The solid-liquid-gas interface theory states that a smaller bubble adhesion force (from 44.3 μN to 35.8 μN in our work) is advantageous for bubble desorption and a greater air contact angle (from 148.4° to 154.8° in our work) is capable of minimizing the contact area of bubbles with electrodes. Thus, faster bubble removal processes would strengthen the mechanical stability of CrO42–-NiFe LDH/Cr2O3/NF under eASO conditions, making the anode more proficient in inhibiting detachment and shedding of catalyst materials.Fig. 3: Electrochemical durability of CrO42-NiFe LDH/Cr2O3/NF and NiFe LDH/NF.a 2500-h electrolysis curves. Inset images show the UV-vis data and commercial test strip results. b Comparison of long-term voltage-time curves in the MEA device (j: 0.5 A cm−2), with O2 FE values. The symbols for the purple square and triangle correspond to electrolysis data of Pt/C/NF | |CrO42–-NiFe LDH/Cr2O3/NF and Pt/C/NF | |RuO2/NF, respectively. The pink triangles correspond to O2 FE values. c Electrolysis durability test for electrolysis stacks at a constant j of 0.5 A cm−2. Insets show the device. The purple squares correspond to electrolysis data of Pt/C/NF | |CrO42–-NiFe LDH/Cr2O3/NF. The pink triangles correspond to O2 FE values. d Overall comparison of electrolysis lifespans in different reaction conditions.Real catalysts changeCrO42–-NiFe LDH/Cr2O3 samples after 50-h eASO catalysis were characterized by TEM. Nano-sized Cr2O3 particles remain well dispersed on the LDH structures (Fig. 4a), and the measurements of lattice spacing further confirm that both NiFe LDH (spacing of 0.28 nm, (111) facet) and Cr2O3 (spacing of 0.25 nm, (111) facet) are still in existence (Fig. 4b), with a good degree of crystallinity. According to more high-resolution TEM images of different CrO42–-NiFe LDH/Cr2O3 samples after alkaline seawater electrolysis under high j of 500 mA cm−2 for 50 h, there is no highly crystalline Cr(OH)3 in the catalyst (Supplementary Fig. 21). Time-of-flight secondary ion mass spectrometry (TOF-SIMS) was employed to identify the Cr(VI) species after eASO tests. The mapping on the electrode following a 50-h testing period clearly indicates the existence of rich chromate species (Fig. 4c), hence validating the (oxy)hydroxides intercalated/adsorbed with CrO42–. TEM and TOF-SIMS results here verify the stable presence of Cr species on the catalyst, but we still detected Ni, Fe, and Cr in the electrolyte after electrolysis, indicating trace levels of metal leaching (Fig. 4d). Furthermore, we used XPS to probe the surface chemistry of CrO42–-NiFe LDH/Cr2O3 samples at different depths (surface and 100 nm) after different electrolysis times (50, 100, and 250 h) (Supplementary Figs. 22‒24). Valence states of Ni and Fe species increase naturally with the extension of electrolysis time, especially the Ni species, with more than half of all Ni species transforming into Ni3+ (Fig. 4e and Supplementary Fig. 22). High-resolution XPS data of Cr 2p do not show any weakening of the peaks (Supplementary Fig. 24), which indicates that the reconstruction process does not contain notable Cr leaching. At the same time, the valence states of Cr at different depths gradually increase with the time of electrolysis (Fig. 4f), which may be caused by strong anode polarization conditions. After a 250-h stability test, the M-OH content of CrO42−-NiFe LDH/Cr2O3/NF increased, as indicated by data in the O 1s spectrum (Supplementary Fig. 25). At the same time, the Cr contents are consistently maintained at around 2%, both on the outer surface and 100-nm deep inside (Supplementary Fig. 26). In situ Raman spectroscopy was performed to detect surface Cr species and obtain an in-depth understanding of catalyst changes. In the process of increasing the applied potentials, typical γ-NiOOH is derived on the NiFe LDH (Fig. 4g), and this transformation is evidenced by the pairwise Raman peak shifts from 455/525 cm–1 to 472/553 cm–153,54,55,56. As for the Cr-containing electrode, such structural transformations (metal hydroxide to metal oxyhydroxides) clearly occur at a more positive potential (Fig. 4h). The changed transformation behaviors of NiFe LDH and CrO42−-NiFe LDH/Cr2O3 imply that the Cr species accelerate the generation of the catalytically favorable γ-phase under eASO conditions. When the anodic potential rises to a potential of 1.7 V or even greater, the Cr-OH signal at ~720 cm−1 appears and shows a tendency to become stronger (Fig. 4h). This means that the electrode surface is always covered with rich *OH species under high j operation. This strong peak at approximately 720 cm−1 indicates that heterogeneous Cr sites do have a strong capacity to adsorb OH−. This may lead to a Cr-OH bond that is too strong for OH spillover or the transfer of *OH to the NiFe site. However, this strong adsorption capacity will draw OH− near to the catalytic surfaces, resulting in NiFe sites with full access to OH−. This is comparable to catalysis in tandem, and our in situ tests support this view. Meanwhile, both CrO42− and Cr2O3 Raman signals are consistently observed (Fig. 4h, i), revealing that both Cr species are present on the electrocatalytic surfaces that function at various potentials and different lengths of time. Furthermore, in time-dependent Raman tests, we reached phenomena consistent with ex situ XPS data, i.e., large increases in Cr6+/Cr3+ ratios with growing electrolysis time. The original Cr(III) sites are converted into Cr-OH sites and anionic Cr6+ sites with a strong repulsive force to repel Cl−. According to the Pourbaix diagram for Cr, under an alkaline testing environment (pH = 14), Cr(III) species do not exist in the form of Cr(OH)3, and CrO42− should be the exclusive Cr(VI) species (Cr2O72− + H2O ↔ 2CrO42− + 2H+)50,56,57. Since the Cr oxides and the derived Cr sites are uniformly dispersed near the catalytic Ni sites (evidences: in situ heat maps and TOF-SIMS mapping), CrO42–-NiFe LDH/Cr2O3 is very stable. In fact, in addition to repelling Cl–, sufficient surface CrO42– would repel negatively charged OH− as well. This is overlooked by a large number of eASO anion-rich anode designs that rely on surface anions (e.g., SO42−, CO32−) to repel Cl−. On the surfaces of CrO42–-NiFe LDH/Cr2O3, since OH− is a harder Lewis base than Cl−, OH− is able to be screened as the dominant adsorbate in the presence of nano-sized Cr oxides as powerful Lewis acid centers. Moreover, with different mapping techniques (Figs. 1g, 2d, and 4c), high-degree dispersity of the Cr species is verified, which ensures that Cl−-repelling and OH−-absorbing capabilities of our catalyst are guaranteed at almost every reaction site. Thus, our work enables the highly selective rejection of Cl– (Fig. 5a, the concept of creating anion-rich catalytic surfaces that selectively repel Cl– and enrich OH–).Fig. 4: Real catalyst structure changes.a TEM and (b) high-resolution TEM images of CrO42–-NiFe LDH/Cr2O3 after the eASO. c TOF-SIMS mapping of CrO42– fragments from the CrO42–-NiFe LDH/Cr2O3/NF surface after testing in alkaline seawater. d Leaching of each element over the reaction time. e Detailed changes in the valence distribution of Ni and Fe species. f Detailed changes in the valence distribution of Cr species. Potential-dependent Raman data recorded on (g) NiFe LDH/NF and (h) CrO42–-NiFe LDH/Cr2O3/NF. i Time-dependent Raman data recorded on the CrO42–-NiFe LDH/Cr2O3/NF. The time-dependent Raman tests were performed at 1.8 V vs. RHE for 60 min. The electrolyte was not changed throughout the time-dependent test.Fig. 5: Theoretical insights.a Concise electrocatalytic diagram. b–e Possible AEM- and LOM-based pathways. Free energy diagrams of (f) AEM and (g) LOM. h Free energy calculations for the adsorption of *Cl with atomic structures. i PDOS. j Polarization curves with ρRHE values. k Polarization curves with Tafel slope values in alkaline seawater with/without TMA+ with 100% iR correction.Mechanism understandingTo gain a deep understanding of experimental results, we conducted comprehensive density functional theory (DFT) calculations to elucidate the catalysis processes on Cr-containing and Cr-free models. The experimental details of the DFT calculations are included in Supplementary Note 3. Inspired by previous studies58, we constructed the structures of NiFe LDH and CrO42–-NiFe LDH/Cr2O3. Interestingly, two materials exhibit significant magnetic moments, mainly resulting from the Fe (about 3.90 μB) and Ni sites (about 1.60 μB), as depicted by Supplementary Table 7. We analyzed the charge density differences. According to the Bader analysis, there is a significant charge redistribution. The Fe atom of NiFe-LDH gathers positive charges (1.47 e), which is larger than that of Fe on CrO42–-NiFe LDH/Cr2O3 (1.57 e), thus indicating that after the introduction of Cr species, the charge carried by Fe becomes more positive (Supplementary Fig. 27). The surface Pourbaix diagrams for CrO42–-NiFe LDH/Cr2O3 and NiFe LDH were established to evaluate their stability under electrochemical conditions (Supplementary Fig. 28). According to the obtained oxidation potentials and the limiting potentials, we found that the limiting potential of OER on CrO42–-NiFe LDH/Cr2O3 and NiFe-LDH is much lower than the oxidation potentials. Thus, the pristine CrO42–-NiFe LDH/Cr2O3 and NiFe LDH surfaces are the most favorable, which suggests that their excellent stabilities are beneficial for their application in electrochemical reactions. In addition, AIMD simulations were also conducted to assess the thermal stability of CrO42–-NiFe LDH/Cr2O3. The results showed that the structural integrity of CrO42–-NiFe LDH/Cr2O3 catalysts can be maintained well at T = 300 K after 10 ps (Supplementary Fig. 29).For the OER on the two catalysts, we considered two general mechanisms, including the adsorbate evolution mechanism (AEM) and lattice oxygen mechanism (LOM) (Fig. 5b–e). According to the free energy profiles computed by the DFT + U method, which has been widely utilized to describe the catalytic performance of transition metal-based catalysts with unfilled d orbitals59, we observed that all elementary reaction steps in the two mechanisms show uphill trends, in which all oxygenated intermediates are adsorbed on the Fe active site. For CrO42–-NiFe LDH/Cr2O3, the rate-determining step (RDS) is the dehydrogenation of *OH to *O via the AEM with the ∆G value of 1.74 eV, whereas O2 formation is the RDS via the LOM and the corresponding ∆G is computed to be 1.64 eV (Fig. 5f, g). Thus, the LOM is the dominant pathway for OER on CrO42–-NiFe LDH/Cr2O3 and the overpotential is 0.41 V according to the CHE model. Conversely, on NiFe LDH, O2 formation is a big challenge via both AEM and LOM due to its energy input of 1.79 and 2.22 eV, indicating that the OER proceeds via the AEM mechanism with an overpotential of 0.58 V. Thus, CrO42–-NiFe LDH/Cr2O3 exhibits a higher OER catalytic performance due to a lower overpotential of 0.41 V via the LOM, well accounting for our experimental results. In addition, we computed the adsorption energies (Eads) of *Cl and *OCl on the two catalysts to evaluate their *Cl resistance. The results showed that the computed Eads values of *Cl and *OCl on CrO42–-NiFe-LDH/Cr2O3 are 0.33 and 0.59 eV, respectively, which are much higher than that on NiFe LDH (0.08 eV and 0.34 eV, Supplementary Table 8), indicating the excellent Cl-resistance CrO42–-NiFe-LDH/Cr2O3 (Fig. 5h). We also considered utilizing Ni in CrO42–-NiFe-LDH/Cr2O3 and NiFe LDH as active sites for *Cl adsorption. As shown in Supplementary energies of 0.44 eV and 0.20 eV for CrO42–-NiFe-LDH/Cr2O3 and NiFe LDH (Supplementary Table 9), respectively. The projected density of states was calculated (Fig. 5i), and the d-band center (ɛd) value is –2.00 eV and –2.26 eV for NiFe LDH and CrO42−-NiFe LDH/Cr2O3, respectively, which is consistent with our research. As shown in Fig. 5j, the effect of increasing pH (from 12.5 to 14) on the eASO activity of CrO42−-NiFe LDH/Cr2O3/NF far exceeds that on the activity of NiFe LDH/NF, indicative of changed proton-electron transfer (PET) steps after introducing the Cr species60,61,62. Indeed, the CrO42−-NiFe LDH/Cr2O3/NF shows a proton reaction order ρRHE of 0.917, whereas the ρRHE of NiFe LDH/NF is only 0.687. This indicates a greater degree of nonconcerted PET (NCPET) steps (i.e., O2-evolution pathways alter) on the Cr-containing electrode surfaces. Except for pH-dependence tests, we used tetramethylammonium cations (TMA+) as the chemical probe to confirm once again whether the surface reaction paths change. After replacing K+ with TMA+ in the alkaline solution, the change in activity of the CrO42−-NiFe LDH/Cr2O3/NF is significantly greater than that of the NiFe LDH/NF (Fig. 5k), which indicates that TMA+ on the CrO42−-NiFe LDH/Cr2O3 surface electrostatically react more strongly with oxygenated intermediates (OIδ−)63,64. Such reaction/binding inhibits lattice oxygen-mediated O2 evolution on CrO42−-NiFe LDH/Cr2O3, thus leading to lower j and higher Tafel slope65.