Introduction

Hydrogen production through water electrolysis is a cornerstone renewable technology crucial for realizing widespread sustainable and clean energy adoption in the future1,2,3,4. It also plays a vital role as a strategy for storing intermittent energy sources and achieving carbon neutrality. However, the low energy efficiency of water electrolysis primarily arises from sluggish kinetics of oxygen evolution reaction (OER), which underscores the need for highly efficient electrocatalysts to navigate the intricate multiple proton/electron processes5,6,7. Over decades of exploration, researchers have found that transition metal (e.g., Co, Fe, Ni) oxides/hydroxides/oxyhydroxides are favorable candidates to replace noble metal electrocatalysts whose widespread commercial applications are limited by their scarcity and high costs8,9,10. However, due to the insufficient intrinsic activity of 3d transition metals, their electronic structures, and sometimes even the entire reaction pathways must be adjusted in order to meet the requirements of practical applications11,12,13,14.

Doping with inorganic heteroatoms (e.g., sulfur) is a typical strategy to enhance the OER performance of transition metal-based electrocatalysts by effectively introducing long-range lattice strain through atomic size effects, charge compensation mechanisms, or structural reconstruction, thereby modulating the adsorption thermodynamics/kinetics of the oxygen-containing intermediates15,16,17,18. To a certain extent, the adjustment of atomic positions can cause disturbances in the charge density distribution on catalysts19. Nevertheless, the bonding characteristic is dominated by ionic bonds, which exhibit limited electron delocalization capacity and a dynamically tunable donor-acceptor mechanism. This makes it difficult to drastically alter the activation energy barrier through such doping. Recently, organic ligand modulation has also attracted great attentions in electrocatalysis20,21,22. This heightened interest arises from the high diversity and designability inherent in organic molecules, as well as the capacity to significantly regulate the electronic structures of metal sites by completely changing their coordination environments23,24. Moreover, the organic units can introduce supplementary adsorption sites through their weak bonding interactions (e.g., hydrogen bonds, halogen bonds, or van der Waals forces), potentially breaking the linear scaling relation of multiple intermediates, ultimately resulting in enhanced electrocatalytic activity25,26. However, due to their predominant reliance on surface adsorption and inherent molecular flexibility, organic components cannot induce substantial bulk lattice strain effects comparable with inorganic heteroatoms. Obviously, a single modulation strategy struggles to simultaneously optimize the geometric and electronic structures of metal sites (Fig. 1a, b). However, the complementary properties of organic and inorganic components suggest that integrating both could enable dual modulation of these structures (Fig. 1c). Such an approach would overcome the inherent limitations of single-component modulation strategies and provide an alternative design principle for developing advanced electrocatalysts. Undoubtedly, the first challenge to realize this vision will be in synthesis technology. As we all know, the traditional approach to S doping often demands high temperature and pressure, accompanied emission of toxic gaseous sulfur compounds, or necessitates toxic and explosive reducing agents like sodium borohydride, inevitably introducing boride impurities27,28. Besides, the efficient synthesis pathway for atomic-level modification of organic ligands remains elusive.

Fig. 1: Catalyst engineering through distinct doping strategies.
Fig. 1: Catalyst engineering through distinct doping strategies.
Full size image

a Inorganic heteroatom doping induces lattice strain and electronic structure modification. b Organic ligand modification enables strong electronic regulation. c Synergistic effects of organic-inorganic dual-modification for both significantly changing the geometric and electronic structures of active metal atoms. Red and blue shadows denote electron gain and loss, respectively, with color depth indicating the extent of change.

Here, we present an eco-friendly and facile methodology to simultaneously modify Co(OH)2 matrices with organic 2-methylimidazole (MI) molecules and inorganic sulfur (S). Using a cobalt-based metal-organic framework (ZIF-67) as the precursor, a gentle microbial vulcanization (MV) process29,30 triggers the structural transformation of ZIF-67, converting it into MI-modified cobalt hydroxide while incorporating sulfur atoms into the lattice, which resulting in an efficient electrocatalyst, denoted as MV-S/Co(OH)2/MI (MEC-17, MEC: lab of molecule-based energy chemistry). The electrocatalyst exhibits notable activity towards OER, achieving an overpotential of 285.6 ± 1.7 mV at a current density of 1000 mA cm−2 on cobalt foam. Moreover, it can be stably operated for over 300 h at 1000 mA cm−2 in an anion-exchange-membrane water electrolyzer (AEMWE). Experimental and theoretical analyses indicate that the incorporation of sulfur exerts a significant influence on the geometric configuration of the metal sites, leading to a shorter interatomic Co−Co distance for facilitating the O−O radical coupling. On the other hand, the introduction of MI further intensifies the electron-deficiency of the cobalt active center, optimizing the adsorption of the oxygen-containing intermediates. The combined effect of these organic-inorganic dual dopants augments the individual advantages of each dopant and remarkably accelerate the OER kinetics.

Results and discussion

Synthesis and characterization of catalysts

The MV-S/Co(OH)2/MI was synthesized by a facile one-step microbial vulcanization of ZIF-67 precursor (Fig. 2a, b and Supplementary Fig. 1). The crystal structure transformation of ZIF-67 during microbial vulcanization were monitored by powder X-ray diffraction (PXRD). The results illustrated that four diffraction peaks, located at 9.7°, 20.1°, 33.1°, 38.2°, and 58.7°, corresponding to the (001), (002), (102), (104), and (110) crystal facets of Co(OH)2, began to appear after just 1 h vulcanization, then the diffraction peaks of ZIF-67 gradually decreased until disappeared thoroughly after 12 h, suggesting the complete conversion of ZIF-67 to Co(OH)2 (Fig. 2c). With prolonged microbial vulcanization, the morphology of nanoparticles changed from a typical truncated dodecahedron for ZIF-67 into nanosheets, while the color of the catalyst changed from purple to green, and both changes were basically balanced after 24 h of reaction (Fig. 2d). The high-resolution transmission electron microscope (HRTEM) image reveals clear legible lattice fringes with a pitch of ca. 0.272 nm, consistent with the (100) planes of Co(OH)2 (Supplementary Fig. 2). The double aberration-corrected HAADF-STEM image further confirms the formation of Co(OH)2. A lattice point array, corresponding to the (001) facets of Co(OH)2, is clearly visible in MV-S/Co(OH)2/MI (Fig. 2e). Besides, atomic force microscope (AFM) image indicates an average thickness of 3.54 ± 0.12 nm for MV-S/Co(OH)2/MI (Fig. 2f and Supplementary Fig. 6a). The elemental mapping images reveal that Co, S, O, C and N species are homogeneously distributed on MV-S/Co(OH)2/MI, indicating that the product by microbial vulcanization still contains MI ligands and successfully implants S element (Fig. 2g). In addition, MV-S/Co(OH)2/MI can be synthesized simply without high energy consumption and easily for gram-scale preparation even in lab conditions (Fig. 2h, i). It is worth noting that this microbial vulcanization method is also environmentally friendly, and the bacteria after vulcanization operation can still maintain activity for recycling used (Fig. 2j and Supplementary Fig. 3). The lab-estimated cost of MV-S/Co(OH)2/MI is only ca. 673 $ kg−1, which is ~1/227 of commercial RuO2 (Table S1).

Fig. 2: Molecular vulcanization and morphology characterizations.
Fig. 2: Molecular vulcanization and morphology characterizations.
Full size image

a Schematic illustration of the microbial vulcanization of ZIF-67 to synthesize MV-S/Co(OH)2/MI. b SEM image of microbial cells and MV-S/Co(OH)2/MI. c PXRD monitoring of ZIF-67 by microbial vulcanization under different times. d SEM and sample photos of the transitions from ZIF-67 to MV-S/Co(OH)2/MI using different vulcanizing time (the scale bars are 400 nm). e Double aberration-corrected HAADF-STEM image (inset: metal arrays of Co(OH)2 lattice), f AFM (inset: thickness profile), and g elemental mapping images of MV-S/Co(OH)2/MI. h Photograph of MV-S/Co(OH)2/MI laboratory preparation process. i The MV-S/Co(OH)2/MI obtained by one lab-scale preparation. j Plates (LMEC represents Lab of Molecule-based Energy Chemistry) marked with bacteria after vulcanization of ZIF-67. Source data are provided as a Source Data file.

In order to investigate the influence of microbial vulcanization and organic-inorganic dual-dopants on the performance of Co(OH)2, we synthesized single-doped and bacteria-free catalysts as controls. Among them, MV-S/Co(OH)2 without organic ligands is directly prepared by microbial vulcanization of Co(OH)2. Both MV-Co(OH)2/MI and S/ZIF-67 utilized ZIF-67 as a precursor, but lack sulfur source and bacteria cells in the reaction system, respectively (Fig. 3a). Even in the absence of sulfur source, ZIF-67 can still undergo transformation into Co(OH)2 nanosheets through bacterial corrosion, yet the structure and morphology of ZIF-67 remain intact when bacteria are not introduced (Supplementary Figs. 4, 5). Interestingly, the diffraction peak of (001) plane for MV-S/Co(OH)2/MI and MV-Co(OH)2/MI shifted to lower degree, suggesting that the interlayer spacing is expanded by the intercalation of MI (Supplementary Fig. 4). The MV-S/Co(OH)2 and MV-Co(OH)2/MI nanosheets have thickness of 3.51 ± 0.05 nm and 3.62 ± 0.07 nm, respectively. These values are both similar to that of MV-S/Co(OH)2/MI (Supplementary Figs. 5, 6). Additionally, the uniform distribution of S or N elements either in MV-S/Co(OH)2 or MV-Co(OH)2/MI, indicates the homogeneous doping of S and MI in the respective catalyst (Supplementary Fig. 7). The BET specific surface area of MV-S/Co(OH)2/MI (116.9 m2 g−1) and MV-Co(OH)2/MI (39.3 m2 g−1) are significantly higher than that of MV-S/Co(OH)2 (19.2 m2 g−1), which is contributed to the large number of micropores that have been retained from the ZIF-67 precursor (Supplementary Figs. 8, 9).

Fig. 3: Catalyst components and structures.
Fig. 3: Catalyst components and structures.
Full size image

a Comparison of synthesis conditions and sample colors, b the MS and 1H-NMR spectra, c Raman spectra, d the Co, S and N contents, and e FT-IR spectra of MV-S/Co(OH)2/MI, MV-S/Co(OH)2, MV-Co(OH)2/MI, and S/ZIF-67, respectively. f TG-MS curves for the thermal decomposition of MV-S/Co(OH)2/MI. g Co 2p XPS spectra of catalysts. Source data are provided as a Source Data file.

A series of spectroscopic measurements were conducted to confirm the catalyst structures. The mass spectra (MS) of MV-S/Co(OH)2/MI, MV-Co(OH)2/MI, and S/ZIF-67 all exhibit prominent peaks at 83.06 m/z (C4H7N2+), indicating the presence of MI. This was further supported by the 1H-nuclear magnetic resonance (NMR) spectra, which display the distinct resonance at δ = 6.62 corresponding to protons on the ring carbons, while sharp peak at δ = 2.13 identifies the methyl group (Fig. 3b)31. Raman spectra further confirm the organic coordination in the catalysts. Both MV-S/Co(OH)2/MI, MV-Co(OH)2/MI, and S/ZIF-67 exhibit bands at 186 and 677 cm−1, which are assigned to the Co−N deformation vibration and C−CH3 stretching vibration originating from MI, respectively (Fig. 3c)32. Moreover, the FT-IR spectra of MV-S/Co(OH)2/MI, MV-S/Co(OH)2 and MV-Co(OH)2/MI show that three peaks at 1647 cm−1, 1383 cm−1 and 515 cm−1 attributed to Co−OH bending vibration from Co(OH)2, and all of MV-S/Co(OH)2/MI, MV-Co(OH)2/MI and S/ZIF-67 exhibit a characteristic peak at 425 cm−1 corresponding to the stretching vibration of Co−N (Fig. 3e and Supplementary Fig. 10)33. Based on the elemental analysis, significant sulfur is contained in MV-S/Co(OH)2/MI (6.16 wt%) and MV-S/Co(OH)2 (1.47 wt%), as well as nitrogen contained in MV-S/Co(OH)2/MI (3.49 wt%) and MV-Co(OH)2/MI (5.88 wt%), indicating that MV-S/Co(OH)2 and MV-Co(OH)2/MI are doped with sulfur and MI, respectively, whereas MV-S/Co(OH)2/MI contains both of dopants (Fig. 3d and Supplementary Fig. 11). The small amount of nitrogen detected in MV-S/Co(OH)2 should come from the hexamethylenetetramine (HMT) used in the synthesis of Co(OH)2 substrate. Besides, the thermogravimetric analysis-mass spectrometry (TG-MS) technology reveals that the primary gases emitted during the thermal degradation of MV-S/Co(OH)2/MI in an oxidative atmosphere are carbon dioxide and sulfur dioxide (m/z = 44 and 64), further verified the co-doping of organic and inorganic components in MV-S/Co(OH)2/MI (Fig. 3f). The chemical states of elements were characterized using X-ray photoelectron spectroscopy (XPS). The high-resolution spectra of S 2p and N 1 s also show that dual-dopants in the MV-S/Co(OH)2/MI (Supplementary Figs. 12–14)34,35. In the high-resolution Co 2p spectrum of MV-S/Co(OH)2/MI, the main signal can be divided into peaks of Co2+ (782.9 2p3/2 and 798.3 eV 2p1/2) and Co3+ (781.1 2p3/2 and 796.8 eV 2p1/2)36,37, which reflects the coexistence of Co2+ and Co3+. In addition, the binding energy of Co in MV-S/Co(OH)2/MI is the highest, indicating a more electron-deficient characteristic in this material, which should be beneficial to catalytic performance (Fig. 3g).

Electrocatalytic OER performances

By using a standard three-electrode system in O2-saturated 1.0 M KOH on glassy carbon electrode (GCE), we normalized the current density using the geometric area and electrochemically active specific surface area (ECSA), denoted as jgeo and jECSA, respectively. MV-S/Co(OH)2/MI exhibits good catalytic activity (Fig. 4a and Supplementary Figs. 15–19), with an overpotentials of 265.6 ± 2.1 mV at 10 mA cm−2geo, 303.0 ± 2.6 mV at 0.1 mA cm−2ECSA, and a Tafel slope of 58.4 ± 1.2 mV dec−1. These values are lower to those of MV-S/Co(OH)2 (ηgeo10 = 287.6 ± 0.6 mV, ηECSA0.1 = 305.3 ± 0.6 mV, Tafel slope = 79.1 ± 2.1 mV dec−1), MV-Co(OH)2/MI (ηgeo10 = 306.9 ± 1.2 mV, ηECSA0.1 = 343.7 ± 1.5 mV, Tafel slope = 114.5 ± 1.9 mV dec−1), Co(OH)2 (η10 = 345.7 ± 2.1 mV, ηECSA0.1 = 370.7 ± 9.5 mV, Tafel slope = 117.2 ± 2.8 mV dec−1), and the commercial RuO2 (ηgeo10 = 305.0 ± 1.0 mV, Tafel = 122.7 ± 3.8 mV dec−1). Furthermore, MV-S/Co(OH)2/MI demonstrates the fastest electron transfer (Rct = 39 Ω), the largest exposed active specific area (Cdl = 2.73 mF cm−2), and the highest mass activity of 119.0 ± 10.8 A g−1 when compared to other tested catalysts (Fig. 4b and Supplementary Figs. 20–22). Additionally, the Faradaic efficiency analysis reveals a near-unity OER efficiency of 98.99 ± 0.22 % with no detectable side products, confirming that the measured current is primarily attributed to water oxidation (Supplementary Fig. 23). To further investigate the advantages of microbial vulcanization on catalytic properties, we also synthesized vulcanized ZIF-67 using hydrothermal and pyrolytic methods (denoted as St/ZIF-67 and Sp/ZIF-67, respectively) for comparison. Notably, their electrocatalytic performances were found to be significantly inferior to that of MV-S/Co(OH)2/MI (Supplementary Figs. 24, 25).

Fig. 4: Electrocatalytic activity evaluation.
Fig. 4: Electrocatalytic activity evaluation.
Full size image

a OER polarization curves of catalysts on GCE in O2-saturated 1.0 M KOH (pH = 13.6 ± 0.2) under 25 °C at a scan rate of 5 mV s−1. b Comparison of overpotentials, Tafel slopes, Cdl, Rct and mass activities of catalysts. c LSVgeo curves of MV-S/Co(OH)2/MI on different substrates (inset: summary of overpotentials of MV-S/Co(OH)2/MI on different substrates at 0.1, 0.5, and 1.0 A cm−2. Error bars represent standard deviation, n = 3 independent replicates.). d The AEMWE activities of Pt/C||MV-S/Co(OH)2/MI and commercial Pt/C||IrO2 catalysts. Polarization curves of the electrolyzer without iR-compensation 1.0 M KOH (pH = 13.6 ± 0.2) solution at 25 °C. e Overpotential comparison of MV-S/Co(OH)2/MI and reported advanced Co-based electrocatalysts on different substrates at 10 mA cm−2 in 1.0 M KOH. f Consecutive chronoamperometric test of the AEMWE using Pt/C||MV-S/Co(OH)2/MI were conducted at a potential of 2.0 V in 1.0 M KOH (inset: photograph and assembly schematic of the AEMWE device). Source data are provided as a Source Data file.

To investigate the substrate effect, we evaluated the OER performance of MV-S/Co(OH)2/MI on various substrates, including carbon cloth (CC), Co foam, and Ni foam (Fig. 4c and Supplementary Figs. 26, 27). Specifically, on Co foam, MV-S/Co(OH)2/MI requires only 235.6 ± 6.2 mV, 273.6 ± 1.4 mV, and 285.6 ± 1.7 mV to impel current densities of 0.1 A cm−2, 0.5 A cm−2, 1.0 A cm−2, respectively. Moreover, as the current density increases, the performance of MV-S/Co(OH)2/MI exhibits a more favorable advantage compared with that of commercial RuO2 (Supplementary Fig. 28). In particular, MV-S/Co(OH)2/MI exhibits competitive performance among reported Co-based electrocatalysts, regardless of whether it is loaded on GEC, CC, or Ni foam perform (Fig. 4e and Tables S2–4). Moreover, MV-S/Co(OH)2/MI exhibits good stability with only an 8.1 mV overpotential rise at 350 mA cm−2 after 10000 CV cycles. Besides, we also conducted a long-term stability test on MV-S/Co(OH)2/MI at 100 mA cm−2 by chronopotentiometry, and the potential only increased by 3% after 200 h (Supplementary Fig. 29). Considering the high potential for industrial application, we constructed an AEMWE using MV-S/Co(OH)2/MI as the anode and Pt/C as the cathode. The Pt/C(−)||MV-S/Co(OH)2/MI(+) electrolyzer achieves a current density of 2.0 A cm−2 with a potential requirement of 2.12 V. Conversely, at the same potential, the Pt/C(−)||IrO2(+) electrolyzer shows a current density of only 650 mA cm−2, which is about one third of Pt/C(−)||MV-S/Co(OH)2/MI(+) (Fig. 4d). The AEMWE also maintains stable operation for over 300 h under industrial current density of 1.0 A cm−2 (Fig. 4f). When the Pt/C(−)||MV-S/Co(OH)2/MI(+) device is combined with a solar panel, hydrogen can be continuously generated with just 1.46 V under sunlight illumination (Supplementary Fig. 30).

Insights into OER mechanism

A series of operando spectroscopic techniques and density functional theory (DFT) calculations were employed to deeply understand the catalytic mechanism of MV-S/Co(OH)2/MI. The real active structure of the catalysts was firstly characterized by PXRD and operando Raman. The PXRD patterns reveal that after undergoing the OER process, MV-S/Co(OH)2/MI, MV-S/Co(OH)2, MV-Co(OH)2/MI, and Co(OH)2 exhibit distinct diffraction peaks related to CoOOH (JCPDS card no. 07-0169), signifying a structure change from Co(OH)2 to CoOOH (Supplementary Fig. 31). The operando Raman spectra exhibit the bands at 448 and 518 cm−1 before electrolysis, which are owed to the A1 and A2 mode characteristics of Co(OH)2. As the applied potential increased from 0.925 to 0.985 V, the peaks of Co(OH)2 shifted to 461 and 573 cm−1. These peaks were ascribed to the Eg and A1g modes typical for CoOOH, being consistent with the result of PXRD (Fig. 5a)38. In addition, XPS analysis also demonstrated the structural transformation of catalysts through the OER process (Supplementary Figs. 32–36). The binding energy of Co species in all catalysts exhibit a notable increase post-electrolysis. Notably, the binding energy of Co in MV-S/Co(OH)2/MI after OER (denoted as MV-S/Co(OH)2/MI-OER) remains the highest compared with other controls (Supplementary Fig. 33). Besides, the S 2p fine spectrum reveals that the peak corresponding to Co−S vanishes post-reaction in both MV-S/Co(OH)2/MI and MV-S/Co(OH)2, indicating that the doped S in catalyst is oxidized to SOx2− via the OER process, which has been widely observed in S-doped electrocatalysts (Supplementary Fig. 36)39. Moreover, HRTEM images of MV-S/Co(OH)2/MI-OER also provided evidence for the reconstruction, as the lattice fringes of the core structure clearly correspond to the (011) lattice plane of CoOOH (Supplementary Fig. 37). Besides, the uniform distribution of C, N and S in MV-S/Co(OH)2/MI-OER was further confirmed by EDS mapping (Supplementary Fig. 38). To evaluate the stability of MI ligands under OER conditions, we performed 1H NMR and MS analyses on both MV-S/Co(OH)2/MI-OER and the electrolyte after prolonged testing. The NMR spectrum of MV-S/Co(OH)2/MI-OER shows a distinct signal at δ = 2.13 ppm, which is attributed to the methyl group of MI. In contrast, the electrolyte exhibited no detectable signals associated with free MI or any oxidative degradation products (Supplementary Fig. 39). Correspondingly, the MS spectrum of MV-S/Co(OH)2/MI-OER displays a prominent peak at m/z = 83.05, being assigned to the C4H7N2+, without observable impurities/extra peaks, also indicating the preservation of MI ligand within the catalyst and the absence of detectable degradation products (Supplementary Fig. 40). Moreover, the structural integrity of MI within MV-S/Co(OH)2/MI-OER was further confirmed by FT-IR and Raman spectroscopy. Both spectra clearly showed the preservation of characteristic vibrational modes associated with MI, indicating its stability under operational conditions (Supplementary Figs. 41, 42). These results suggest that Co(OH)2 in MV-S/Co(OH)2/MI-OER was oxidized to CoOOH, while maintaining co-doping of MI and S components. Furthermore, geometric phase analysis (GPA) reveals distinct strain distribution patterns across the catalysts (Supplementary Fig. 43). The pristine CoOOH exhibits relatively homogeneous stress distribution, while the introduction of both MI and S dopants brings about pronounced lattice distortions, as evidenced by significant color variations in the strain field mapping. Particularly, S doping induces greater lattice strain compared with MI doping, and the dual-doping system demonstrates the highest strain level, suggesting synergistic effect of co-doping in modifying the local atomic structure.

Fig. 5: Investigation of OER catalytic pathway.
Fig. 5: Investigation of OER catalytic pathway.
Full size image

a Operando Raman spectra of MV-S/Co(OH)2/MI during OER electrocatalysis. b Co K-edge XANES profiles, and c Fourier transform EXAFS of MV-S/Co(OH)2/MI, MV-S/Co(OH)2, MV-Co(OH)2/MI, and Co(OH)2 after OER. d Potential-dependent operando ATR-SEIRAS of MV-S/Co(OH)2/MI. e DEMS signals of 36O2 products for 18O-surface-labeled MV-S/Co(OH)2/MI and Co(OH)2 tested in the electrolyte using H216O as solvent. f Polarization curves of MV-S/Co(OH)2/MI with or without TMA+. g Polarization curves of MV-S/Co(OH)2/MI in the electrolytes utilizing H2O and D2O as solvents, respectively. Inset shows the dependence of KIEs value on overpotential. Source data are provided as a Source Data file.

The electronic structure and coordination environment of Co centers in catalysts were further investigated by X-ray absorption fine spectroscopy (XAFS). In the Co K-edge X-ray absorption near edge structure (XANES) spectra, the absorption edge position of MV-S/Co(OH)2/MI-OER is the highest among the catalysts, while those of MV-S/Co(OH)2-OER and MV-Co(OH)2/MI-OER are higher than Co(OH)2-OER, aligning with the XPS results (Fig. 5b). These findings indicate that either individual doping of MI or S can enhance the valence state of Co. Notably, the electronic perturbation effect becomes more pronounced when MI and S are co-existence. After OER process, all catalysts exhibit a cobalt valence state close to +3, being consistent with the structural transformation as observed above (Supplementary Fig. 44). Extended X-ray absorption fine structure (EXAFS) analysis and fitting show that all catalysts exhibit similar prominent peaks located at 1.44 Å and 2.45 Å, which are in good agreement with the Co−O and Co−Co scattering paths typically observed in cobalt oxyhydroxide (Fig. 5c, Supplementary Figs. 45, 46 and Table S5)40.

To investigate the adsorption behaviors of key reaction intermediates during OER, operando attenuated-total-reflection surface enhanced infrared adsorption spectroscopy (ATR-SEIRAS) was performed. As the applied potential increases, two distinct absorption peaks are observed at 1035 and 1492 cm−1. The peak at 1035 cm−1 can be attributed to the oxygen bridges between two metal sites (*O−O*), while that at 1492 cm−1 corresponds to linearly bonded superoxol species (O−O*), both of which are crucial intermediates in the oxide path mechanism (OPM) pathway41,42. It is noteworthy that an enhancement of the peak at 1181 cm−1, which corresponds to the increase of SO42−, aligns with the results of XPS analysis (Fig. 5d)43. For MV-S/Co(OH)2, its spectral changes are similar to those of MV-S/Co(OH)2/MI, confirming it also follows the OPM pathway. In contrast, Co(OH)2 is primarily characterized by a peak at 1230 cm−1, assigned to the O−O stretching vibration of *OOH intermediates (crucial intermediates in the adsorbate evolution mechanism, AEM)44. These results demonstrate that S doping induces a transition from the AEM to OPM in Co(OH)2 (Supplementary Fig. 47). Besides, operando differential electrochemical mass spectrometry (DEMS) was conducted with isotope labeling experiments. In general, if the OER follows the ordinary adsorbate evolution mechanism (AEM), no 36O2 by coupling will be produced when using H218O-labeled catalyst tested in H216O. It is intriguing that both 18O-surface-labeled MV-S/Co(OH)2/MI and MV-S/Co(OH)2 exhibit a significant enhancement of 36O2 signal compared to the baseline background, when measured in an electrolyte with H216O, whereas no such enhancement is observed for Co(OH)2 under the same conditions (Fig. 5e and Supplementary Figs. 48–50). These results indicate that the occurrence of OER on MV-S/Co(OH)2/MI and MV-S/Co(OH)2 proceeds via O−O coupling through the OPM pathway, whereas Co(OH)2 follows the AEM pathway45. To rule out the lattice oxygen oxidation mechanism (LOM), the polarization curve is nearly identical with or without tetramethylammonium (TMA+), which easily bind to the *O22− species (intermediates of LOM). This observation could exclude the possibility of LOM on MV-S/Co(OH)2/MI (Fig. 5f). Furthermore, deuterium kinetic isotope effects (KIEs) experiments were conducted to investigate the involvement of proton transfer in the rate determining step (RDS). The KIEH/D values of MV-S/Co(OH)2/MI were found to be <1.5 across the catalytic overpotential range from 0.30 to 0.35 V, indicating that proton transfer is not the dominant factor in the RDS, which is consistent with the characteristic of OPM pathway (Fig. 5g)46.

To gain preliminary insights into the OPM pathway of MV-S/Co(OH)2/MI and to explore the possible effects of MI and S co-doping on the intrinsic catalytic activity, a series of DFT calculations were performed. According to the actual active structures illustrated above, we constructed periodic surface models of CoOOH, MI-doped CoOOH, SO42−-doped CoOOH, and MI/SO42− co-doped CoOOH to simulate the catalytic structures of Co(OH)2, MV-Co(OH)2/MI, MV-S/Co(OH)2, and MV-S/Co(OH)2/MI, respectively. Atomic coordinates for all optimized models are provided in Supplementary Data 1. To highlight the local catalytic environment, only the reactive region is displayed in the figures (Fig. 6a, b and Supplementary Figs. 51–53). The calculated RDS energy barrier for classic Co(OH)2 along the AEM pathway is 1.68 eV, which is lower than that of 1.92 eV for the OPM pathway. In contrast, the RDS for MV-S/Co(OH)2/MI along the AEM pathway is 1.50 eV, being higher than that of 1.45 eV followed the OPM pathway. Moreover, the RDS barrier for the LOM pathway on MV-S/Co(OH)2/MI is even higher, reaching 1.71 eV (Supplementary Fig. 51). These results suggest that AEM may be more favorable on pristine Co(OH)2, while the OPM could be more advantageous on MV-S/Co(OH)2/MI, which appears consistent with the trends observed in operando experiments and KIEs studies. Notably, even following the AEM pathway, MV-S/Co(OH)2/MI still exhibits the lowest energy barrier of RDS when compared with other catalysts (Supplementary Fig. 54). For the OPM pathway, the O−O coupling is prohibited for Co(OH)2 and MV-Co(OH)2/MI with rather high RDS energy barrier of 1.92 eV and 1.85 eV, respectively. However, for S-doped catalysts of MV-S/Co(OH)2 and MV-S/Co(OH)2/MI, the RDS energy barriers significantly decrease to 1.54 eV and 1.45 eV, respectively (Fig. 6c and Supplementary Fig. 54). This suggests that S doping could facilitate the occurrence of O−O coupling, which is generally consistent with the findings from operando spectroscopy and DEMS with isotope labeling. After analyzing the relation between the calculated d-band center of Co and the ΔGO*−O*, we found that incorporation of MI dopants has a more pronounced effect on the electronic structure, as reflected by a greater shift in the d-band center. Co-doping with both MI and sulfur resulted in Co(OH)2 exhibiting optimal adsorption strength for oxygen species (Fig. 6d and Supplementary Fig. 55). This electronic modulation is further supported by electron density difference plots, which show that both MI and SO42− may seize electrons from Co to increase electron depletion on Co sites, being consistent with the results of XPS and XAFS (Fig. 6e). In contrast, sulfur seems to play a more critical role in optimizing ΔGO*−O* and reducing the O−O coupling barrier. Notably, the interatomic Co−Co distances in MV-S/Co(OH)2/MI (2.668 Å) and MV-S/Co(OH)2 (2.681 Å) are significantly shorter than those in MV-Co(OH)2/MI (2.758 Å) and Co(OH)2 (2.879 Å). This structural contraction appears to correlate with strain analysis results and may help facilitate O−O coupling during the OER (Fig. 6f). Therefore, unlike the significant contribution of organic MI in altering the electronic structure of the Co sites, the inorganic anion primarily influences the geometric structure of adjacent Co atoms, which might be associated with a shift in the OER pathway from AEM to OPM.

Fig. 6: Synergistic mechanism of the organic-inorganic dual dopants.
Fig. 6: Synergistic mechanism of the organic-inorganic dual dopants.
Full size image

Optimal conformations of a Co(OH)2 and, b MV-S/Co(OH)2/MI for key intermediates in the reaction pathways of OPM (green arrow) and AEM (blue arrow). c Comparisons of the energy barriers at the OPM pathway for the catalysts. d Quantitative relation between ΔGO*−O* and the d-band center of Co in different catalysts. The area of the circle represents the current density of different catalysts at an overpotential of 0.3 V. e Electron density difference fields for the S doped, MI doped, and S, MI co-doped Co(OH)2. The yellow and blue regions indicate electron depletion and accumulation, respectively, isovalue = 0.04. f The interatomic Co−Co distances in Co(OH)2 (I), MV-Co(OH)2/MI (II), MV-S/Co(OH)2 (III), and MV-S/Co(OH)2/MI (IV). Color codes: dark blue, Co; pink, O; yellow, S; gray, C; green, N; white, H. Source data are provided as a Source Data file.

In summary, we have developed an efficient organic-inorganic hybrid electrocatalyst, MEC-17, through an eco-friendly microbial-mediated vulcanization method. This electrocatalyst demonstrates good OER performance, with an overpotential of 285.6 ± 1.7 mV and sustains operation for over 300 h at high current density of 1000 mA cm−2. Our carefully designed experiments and theoretical calculations reveal that the incorporating both MI and sulfur dopants into the classic Co(OH)2 structure leads to a synergistic enhancement of its electrocatalytic activity. On one hand, sulfur doping mainly contributes to shorten the interatomic Co−Co distance, promoting O−O radical coupling, and on the other hand, MI molecule plays a more critical role by modulating the electronic structure of the metal centers, which optimizes the adsorption of reaction intermediates to ensure the efficient execution of the OPM pathway. This study demonstrates the potential of the organic-inorganic dual-modulation strategy for improving the efficiency and efficacy of catalytic systems, providing a feasible perspective for upgrading the metal-based electrocatalysts with environment and economic benefits.

Methods

Material preparation

2-Methylimidazole (98%), Co(NO3)2·6H2O (AR, 99%), RuO2 (99.9%), 4-(2-hydroxyerhyl)piperazine-1-erhanesulfonic acid (HEPES, 99.5%), Thiourea (AR, 99%), Sodium DL-lactate solution (73% in water), Thiamine hydrochloride (AR, 99%), Sublimed sulfur (99.95%), Nafion 117 solution (~5% in a mixture of lower aliphatic alcohols and water), potassium hydroxide (KOH, semiconductor grade, 99.99%), and Magnesium Sulfate (99%) were all bought from Shanghai Aladdin Bio-Chem Technology Co., Ltd. CoCl2·6H2O (AR, 99%) and Thioacetamide (99%) were purchased from Beijing Energy Engineering Technologies Co., Ltd. CaCl2·2H2O (99%) and Sodiumthiosulfate (99%) were obtained from Beijing Innochem Science & Technology Co., Ltd. Luria-Bertani (LB) Agar Powder (FMB Grade) and LB Broth Powder (FMB Grade) were offered by Sangon Biotech (Shanghai) Co., Ltd. All the reagents and chemicals were used directly without further purification. For KIEs experiments, D2O (99.9%) was purchased from Anergy Reagents Co., LTD. Carbon cloth (CC), Ni foam and Co foam were all procured from Shanghai Aladdin Bio-Chem Technology Co., Ltd. Before utilization, these self-supporting materials underwent a meticulous pretreatment process to eliminate surface contaminants such as oils and oxides. In particular, CC was subjected to treatment in concentrated HNO3 at 90 °C for 3 h, followed by thorough rinsing with ultrapure water. As for the Ni foam and Co foam, they were initially cleansed in an acetone solution using ultrasonication for 30 min to remove oil, and subsequently purified in a 3.0 M HCl solution using ultrasonication for 20 min each. Lastly, they were rinsed multiple times with ultrapure water and ethanol, respectively, and then dried overnight at 60 °C. The counter electrode, reference electrode, and glassy carbon working electrode were purchased from Gaoshi Ruilian Science and Technology Co., Ltd.

Material synthesis

Preparation of ZIF-67

The synthesis method of ZIF-67 was based on the literature47. Co(NO3)2·6H2O (7.276 g, 20 mmol) and 2-methylimidazole (8.21 g, 100 mmol) were dissolved in 200 mL methanol solution, respectively. Then, the metal salt solution was poured into the organic ligand solution and stirred for 24 h. The solid product was separated by centrifugation and washed 3 times with methanol. Finally, ZIF-67 powder was obtained after 24 h of freeze-drying.

Preparation of MV-S/Co(OH)2/MI

S. oneidensis MR-1 was obtained from The Global Bioresource Center (ATCC700550) and incubated in LB medium at 30 °C for 12 h, with constant shaking at 200 rpm, until reaching a final optical density (OD600) of 0.5. Then, the bacterial cells were concentrated by centrifugation at 3820 × g for 5 min and washed three times with sterile deionized water. The S. oneidensis MR-1 cells (450 mg wet weight) and ZIF-67 (500 mg) were re-suspended in 45 mL sterilized buffer [pH 7.0, HEPES (50 mmol L−1), MgSO4 (2 mmol L−1), CaCl2 (0.1 mmol L−1), Vitamin B1(0.01%) and Sodium DL-lactate (20 mmol L−1)] with Na2S2O3 (2.6 mL, 1 mol L−1) was added. Prior to the addition of bacteria, the above solution was bubbled with N2 for 45 min to remove the dissolved O2. Then, the mixture was incubated at 37 °C in an anaerobic environment for 24 h with shaking at 200 rpm. The solid product and bacterial cells were collected by centrifugation at 3000 rpm and 8000 rpm, respectively. Afterwards, solid products were washed 3 times with sterile water, and then freeze-dried for 24 h to obtain MV-S/Co(OH)2/MI. The synthesis of MV-S/Co(OH)2/MI-Air followed the same procedure as that of MV-S/Co(OH)2/MI, except that the reaction was conducted under ambient conditions without anaerobic protection.

Preparation of the MV-Co(OH)2/MI, MV-S/Co(OH)2 and S/ZIF-67

The synthesis of MV-Co(OH)2/MI, MV-S/Co(OH)2, and S/ZIF-67 were similar with that of MV-S/Co(OH)2/MI. The synthesis of MV-Co(OH)2/MI and S/ZIF-67 was not added Na2S2O3 and S. oneidensis MR-1 cells, respectively, while the synthesis of MV-S/Co(OH)2 by using α-Co(OH)2 instead of ZIF-67 (yield: 75%). Precursor α-Co(OH)2 was synthesized by hydrothermal method. In a typical procedure, 1.682 g (60 mM) hexamethylenetetramine (HMT) and 200 mg poly-(vinylpyrrolidone) (PVP) were sequentially added to 196 mL mixed water/ethanol middle solution (V/V = 9:1) under magnetic stirring. Then, 584.4 mg (50 mM) NaCl and 475.86 mg (10 mM) CoCl2·6H2O were dissolved in the above solution and heated at 90 °C for 2 h. After the reaction, the α-Co(OH)2 was collected by centrifugation and washed 3 times with deionized water, and freeze-dried overnight.

Preparation of Sp/ZIF-67 and St/ZIF-67

Sulfur powder (4 g) and ZIF-67 (0.3 g) were placed upstream and downstream of the tube furnace, respectively. After heated under N2 atmosphere at 900 °C for 2 h at a heating rate of 5 °C min−1, the vulcanized sample was collected and denoted as Sp/ZIF-67, yield: 30%. Apart from this, St/ZIF-67 was synthesis by added ZIF-67 (1.5 mmol) and thioacetamide (4.5 mmol) to 75 mL mixed water/ethylene glycol solution (V/V = 1:1). After vigorous stirred for 30 min, the mixture was transferred into a 100 mL autoclave and heated at 180 °C for 24 h. The solid product was by collected centrifugation and washed with deionized water and ethanol for three times, respectively. Finally, St/ZIF-67 was obtained after dried at 60 °C overnight.

Material characterizations

The crystalline phase date was characterized by Powder X-ray diffraction (PXRD, Rigaku MiniFlex600 X-ray diffractometer) using CuKα radiation at a scanning speed of 15 degrees per minute. The morphology of catalysts was acquired via scanning electron microscopy (SEM, Carl Zeiss EVO 10 microscope) and transmission electron microscopy (TEM, JEM2100). S. oneidensis MR-1 cells for SEM analysis were fixed by soaking in glutaraldehyde solution for 3 h, and then dehydrated in a gradient of low to high concentration aqueous ethanol for 15 min each time. The thickness of sample was measured by Atomic force microscopy (AFM, SPM9700). The content of Co and S was determined on Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES, Agilent 730), and the content of S and N was performed on elemental analysis (EA, UNICUBE). The Fourier transform infrared spectroscopy (FT-IR), confocal Raman spectroscopy, and X-ray photoelectron spectroscopy (XPS) date were acquired on Nicolet 6700, LabRAM HR (632 nm excitation wavelength) and Thermo ESCALAB 250Xi, respectively. Operando Raman spectra were recorded on LabRAM HR Evolution with an excitation of wavelength 532 nm (Supplementary Fig. 56). Operando ATR-SEIRAS measurements were performed using a Nicolet iS50 FT-IR spectrometer with potentials ranging from 1.2 to 1.65 V versus the reversible hydrogen electrode (RHE, Supplementary Fig. 56). Mass spectrometry was performed on a TripleTOF 6600 system. For solid powder analysis, the sample was dissolved in 5% (v/v) hydrochloric acid. Proton nuclear magnetic resonance (1H NMR) were acquired on an AVANCE III HD 400 spectrometer. For NMR analysis, approximately 10 mg of the sample was dissolved in 500 µL of 1 M sulfuric acid via ultrasonication, after which 500 µL of D2O was added. Electrolysis was carried out using 1 M KOD as the electrolyte, with the catalyst loaded onto a titanium mesh. The electrolyte was collected after electrolysis for further analysis. Thermogravimetric analysis (TGA) was carried out on a TGA 55 at a heating rate of 10 °C min−1 under an argon atmosphere. Nitrogen sorption isotherms at 77 K were measured on Micrometrics ASAP 2020 Plus after heated under high vacuum at 150 °C to remove the remnant guest molecules. The specific surface areas and pore size distribution were calculated using the BET and the Barrett-Joyner-Halenda method, respectively. The X-ray Absorption Fine Structure (XAFS) spectroscopy was recorded at the 5S1 X-ray Absorption Beamline of the Aichi Synchrotron Radiation Center. This beamline utilizes a double-bounce channel-cut Si (111) monochromator for single-beam X-ray absorption spectroscopy. The end-station is equipped with three ionization chambers and a seven-element Silicon Drift Detector (SDD) positioned after the focusing Rh mirror and Au mirror, facilitating both transmission and fluorescence mode X-ray absorption spectroscopy.

Electrochemical measurements

All electrochemical measurements were performed on an CHI760E electrochemical workstation (CHI Instrument, Shanghai, China) via a typical three-electrode system in O2-saturated 1 M KOH (Volume: 50 mL). The data were automatically collected by the instrument’s software and processed with origin for plotting and analysis. For each experiment, we prepared a fresh KOH electrolyte solution (pH = 13.6 ± 0.2) by dissolving 5.6 g of semiconductor-grade potassium hydroxide (KOH, 99.99% purity) in ultrapure water and bringing the solution to a final volume of 100 mL. The electrolyte was stored in a clean polyethylene bottle to avoid contamination. Platinum sheet and saturated Hg/HgO electrode were used as counter electrode and reference electrode (EHg/HgO = 0.098 V). The Hg/HgO reference electrode was calibrated versus a saturated Ag/AgCl reference electrode (in saturated KCl) by monitoring the open-circuit potential in argon-purged 1.0 M KOH solution for at least 1 h. 5 mg of each catalyst, 10 μL of Nafion, 210 μL of isopropyl alcohol and 90 μL of deionized water were mixed by ultrasonic for 30 min to obtain catalyst ink. 3.6 μL and 40 µL of catalyst ink was evenly dropped on the cleaned glassy carbon electrode (GCE, diameter of 3 mm) and CC (0.2 cm2) as the working electrode, respectively. Besides, we also dropped 60 µL of catalyst ink on Co foam (0.2 cm2) and Ni foam (0.2 cm2) as working electrode, respectively. The samples were then allowed to air-dry completely at 25 °C, resulting in final catalyst loadings of 0.29 mg cm−2 (GCE), 3.23 mg cm−2 (CC), 4.84 mg·cm−2 (CF), and 4.84 mg cm−2 (NF), respectively. Prior to the measurements, cyclic voltammetry (CV) was carried out on each electrode until a steady state was achieved. Subsequently, linear sweep voltammetry (LSV) was performed at a scan rate of 5 mV s⁻1 with 95% iR correction on GCEs and 85% iR correction on self-supporting electrodes. For the AEMWE tests, the electrolyzer employed a Fumasep FAA-3-PK-130 membrane (2.0 × 2.0 cm, 130 μm thickness), which was pretreated by soaking in 1.0 M KOH solution for 24 h. Commercial IrO2 or MV-S/Co(OH)2/MI-coated CF was used as the anode catalyst, while Pt/C-coated CF served as the cathode catalyst. The catalyst loadings were of 4.84 mg cm⁻2 for MV-S/Co(OH)2/MI, 2.0 mg cm⁻2 for IrO2, and 2.0 mg cm⁻2 for Pt/C. The components were assembled into an AEMWE flow cell operated at 25 °C with a continuous flow of 1.0 M KOH. Chronoamperometry tests were conducted at an applied voltage of 2.0 V. All measured potentials were converted to the reversible hydrogen electrode (RHE) based on the formula:

$${E}_{{{{\rm{RHE}}}}}=E+{E}_{{{{\rm{Hg}}}}/{{{\rm{HgO}}}}}+0.0591\times {{{\rm{pH}}}}$$
(1)

The Tafel slope of the catalyst was determined using the formula presented below:

$$\eta=a+b\,\log i$$
(2)

Where η represents the overpotential, i is the current, a signifies the Tafel constant, and b indicates the Tafel slope. The electrochemical impedance spectroscopy (EIS) was measured in a frequency scan range of 105–0.01 Hz at 1.50 V vs. RHE on GCEs. All data were fitted using ZView software to extract the relevant parameters. Through measured the electrochemical double-layer capacitance (Cdl), the electrochemical surface area (ECSA) was calculated according to the following equation:

$${{{\rm{ECSA}}}}=\frac{{C}_{{{{\rm{dl}}}}}}{{C}_{{{{\rm{s}}}}}}$$
(3)

Where Cdl was evaluated by CV measurements from 0.924 to 1.024 V at the scan rates of 20, 40, 60, 80, 100, 120, 140, and 160 mV s−1. Cs is the specific capacitance of Co (0.035 mF cm−2). The Faraday efficiency was determined using the following equation during constant voltage electrolysis (1.53 V vs. RHE):

$${{{\rm{Faraday}}}}\,{{{\rm{efficiency}}}}=\frac{nzF}{Q}$$
(4)

Where n represents the number of moles of O2 produced, as measured by gas chromatography (SHIMADZU, Nexis GC-7820), z denotes the number of electrons transferred (which is 4 for O2), F is the Faraday constant (96485 C mol−1), and Q is the total charge accumulated at the working electrode under the applied potential. The mass activity was calculated using the following equation:

$${{{\rm{Mass}}}}\,{{{\rm{activity}}}}=\frac{J}{m}$$
(5)

where J represents the current density measured at overpotentials of 300 mV, and m denotes the mass of Co loaded onto the electrode. Operando DEMS testing with isotope labeling measurements were performed on an QAS 100 device (detection limit: 3 × 10−15 mbar, sensitivity: 5 × 10−4 A mbar−1, electron impact ionization) utilizing a similar approach reported by He et al.42 The DEMS setup combines a high-vacuum mass spectrometer with an ambient-pressure electrochemical cell through a medium-vacuum capillary interface, enabling real-time detection of gaseous products via pressure-driven sampling. The working electrode was fabricated via sequential magnetron sputtering of a thin Au layer onto a porous PTFE membrane (Linglu Co., China; ≥50% porosity, ≤20 nm average pore size) and controlled drop-casting of the catalyst. We employed a conventional three-electrode configuration consisting of the prepared working electrode, a Pt wire counter electrode and an Ag/AgCl reference electrode, all immersed in Ar-saturated 1.0 M KOH electrolyte solution. To investigate isotopic labeling, a 1.0 M KOH solution was prepared using H218O as the solvent. Prior to the electrochemical measurements, the electrolyte was purged with high-purity argon gas to eliminate any dissolved oxygen. Each catalyst was subjected to continuous CV scans within a potential window of 1.1–1.8 V vs. RHE at a scan rate of 5 mV s−1. Concurrently, the mass spectra of different molecular weight gas species, including 32O2, 34O2, and 36O2 were meticulously recorded. Each catalyst was initially electrolyzed in a 1.0 M KOH solution prepared using H216O as the solvent to establish a baseline background, with corresponding gas signals being collected. Subsequently, isotopic labeling was performed using H218O as the electrolyte solvent. After labeling the catalyst surfaces with 18O, the samples were rinsed with H216O to eliminate any physically adsorbed H218O molecules. The catalyst was then evaluated in a 1.0 M KOH electrolyte prepared with H216O, while also monitoring the gas products of 32O2, 34O2, and 36O2 (Supplementary Fig. 56). The KIEs experiments were carried out in 1.0 M KOH/H2O and 1 M KOD/D2O electrolytes via electrochemical methods46,48. LSV was performed at a scan rate of 5 mV s−1 with 95% iR compensation. The overpotential in 1.0 M KOH was calculated utilizing the following equation:

$${\eta }^{{{{{\rm{H}}}}}_{2}{{{\rm{O}}}}}={E}_{({{{\rm{vs}}}}.{{{\rm{Hg}}}}/{{{\rm{HgO}}}})}+{{E}_{{{{\rm{H}}}}}}_{({{{\rm{Hg}}}}/{{{\rm{HgO}}}})}+0.059\times {{{\rm{pH}}}}-1{{{{\rm{.229V}}}}}_{{{{\rm{RHE}}}}}$$
(6)

Similarly, the overpotential in 1.0 M NaOH was determined using the equation:

$${\eta }^{{{{{\rm{D}}}}}_{2}{{{\rm{O}}}}}={E}_{({{{\rm{vs}}}}.{{{\rm{Hg}}}}/{{{\rm{HgO}}}})}+{{E}_{{{{\rm{D}}}}}}_{({{{\rm{Hg}}}}/{{{\rm{HgO}}}})}+0.059\times {{{\rm{pD}}}}-1{{{{\rm{.262V}}}}}_{{{{\rm{RHE}}}}}$$
(7)

Where pH = 14 and pD = 14.84, the values of EH (Hg/HgO) and ED (Hg/HgO) were 0.0977 V and 0.0847 V, respectively. The KIEs were calculated using the following equation:

$${{{\rm{KIEs}}}}={\left|\frac{{j}_{{{{{\rm{H}}}}}_{2}{{{\rm{O}}}}}}{{j}_{{{{{\rm{D}}}}}_{2}{{{\rm{O}}}}}}\right|}_{\eta }$$
(8)

where jH2O and jD2O represent the respective current densities at a given overpotential η in H₂O and D₂O solutions.

Theoretical calculation

All the geometry optimization and electron density difference of intermediate models were calculated using Cambridge Sequential Total Energy Package (CASTEP) module in Materials Studio software. The lattice parameters of models without intermediates were optimized with TPSD algorithm before further calculations. The following geometry optimizations are adopted the BFGS algorithm. The calculations were performed using spin-polarized DFT with plane wave pseudo-potential method (OTFG ultrasoft). The Perdew–Burke–Ernzerhof (PBE) method from the generalized gradient approximation (GGA) gradient-corrected functional was used to describe the DFT exchange–correlation potential during calculation. The plane-wave energy cut-off was set as 450 eV, and convergence tolerance for the energy, forces, displacement, and stress were set to be 1 × 10−5 eV atom−1, 0.03 eV Å−1, 0.001 Å, and 0.05 GPa, respectively.