Synthesis and characterizations of A-Fe2S1N5/SNC. A continuous two-step pyrolysis and carbonization approach, as shown in Fig. 1, was proven effective for gradually introducing numerous defects and high non-metal heteroatoms into the A-Fe2S1N5/SNC. In detail, a certain mass ratio of 2-benzimidazolethiol and melamine was first ground into a homogeneous mixture, which suffered confined pyrolysis in the flowing inert atmosphere. The thiol group on 2-benzimidazole thiol was easily extracted by hydrogen to form free radicals due to its high activity. Through free radical reactions, the active C atom on melamine probably covalently bonded with the thiol group to form C-S-C bonds. At high temperatures, melamine transformed into a carbon nitride structure (NC), and accompanied by the polymerization of NC frameworks, numerous defects were spontaneously formed.16 During the thermal polymerization of 2-benzimidazolethiol, NC was encapsulated through a C-S-C bond linkage, and the constructed carbon framework was converted to S, N-co-doped carbon nanosheet (SNC) after in situ restricted pyrolysis.17 Then, the abundant defects generated by confinement pyrolysis and the doped sulfur atoms could firmly anchor double Fe atoms through defect capturing and non-metal heteroatoms tethering effects.18 Finally, the diatomic A-Fe2S1N5/SNC catalysts were obtained at selected temperatures through carbonization reduction under an N2 atmosphere. The mass percentages of Fe in the A-Fe2S1N5/SNC catalysts are calculated to be 6.72 wt% through the inductively coupled plasma-optical emission (ICP-OES) test.
Visualization of scanning electronic microscopy (SEM) revealed the curved and wrinkled nanosheet morphology of A-Fe2S1N5/SNC (Fig. 2a), which is inherited from the SNC nanosheets (Supplementary Fig. 1 and Supplementary Fig. 2). The thickness of the nanosheet was calculated to be 4.5 nm (Supplementary Fig. 3). A-Fe2S1N5/SNC is defect-rich thin nanosheets with plenty of edge activity sites (Fig. 2b, Supplementary Fig. 4). The surface/edge defects can easier anchor the metal, further changing the metal charge via atomic modulation strategies and affecting the electrocatalytic activity of the materials.19 The energy-dispersive X-ray spectroscopy (EDS) images and the electron energy loss spectroscopy (EELS) mapping of A-Fe2S1N5/SNC at high magnification confirm the uniformly distributed Fe, S, N, and C elements over the entire A-Fe2S1N5/SNC (Fig. 2c and Supplementary Fig. 5). Furthermore, no obvious metal particles/clusters are observed in the high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image (Supplementary Fig. 6), indicating the atomically dispersed Fe in the SNC matrix.20 Additionally, the morphology of Fe2N6/NC and Fe1N4/NC is consistent with that of A-Fe2S1N5/SNC (Supplementary Figs. 7–8). In Fig. 2d, the aberration-corrected HAADF-STEM distinctly detects the high-density and paired bright dots at the atomic level (circled by red dashed lines), corresponding to the atomically dispersed Fe-Fe atom-pair. The ratio of Fe diatomic pairs is conducted using the computational statistics from the STEM images (histograms in Supplementary Fig. 9), which reaches 81.7%. The intensity distribution along the X-Y in region A indicates that the spacing between the diatomic pair Fe is about 0.25 nm (Fig. 2e), which closely matches the effective diameter of Fe atoms in A-Fe2S1N5/SNC (Supplementary Fig. 10). The EELS spectrum of small pixels of the annular dark field (red rectangle in Supplementary Fig. 5a) result portrayed visible signals of Fe, S, N, and C, authenticating the densely populated FeS1N2-FeN3 sites (Fig. 2f). Additionally, the relevant X-ray diffraction (XRD) and Raman characterization were depicted in Supplementary Figs. 11–12.
The electronic and atomic structural information for Fe species in the samples were further examined by X-ray absorption spectroscopy (XAS). Figure 2g and Supplementary Figs. 13 involve the Fe K-edge X-ray absorption near-edge structure (XANES) spectra of A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC. The adsorption threshold position of A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC was both located between FeS2 and Fe2O3, which indicates the average metal valence states of Fe species in A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC were situated between them.21 From the XANES spectra shown in Supplementary Figs. 14, the average oxidation state of Fe in A-Fe2S1N5/SNC obtained by fitting was 2.78. The Fourier-transformed k3-weighted extended X-ray absorption fine structure (EXAFS) of A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC were displayed in Fig. 2h and Supplementary Fig. 15. The intense peak at 1.47 Å confirms the existence of Fe-N scattering path in A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC. Concurrently, the Fe-Fe interactions around 2.27 Å are found both in A-Fe2S1N5/SNC and Fe2N6/NC, suggesting that double iron atoms exist in the form of direct bonding.22,23 Notably, the peak appears at 1.84 Å in A-Fe2S1N5/SNC indicating the presence of a Fe-S scattering. Ulteriorly, the atomic configurations of A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC are then studied by the Fe K-edge wavelet transform (WT)-EXAFS (Supplementary Fig. 16), demonstrating the isolated characteristic of Fe species.24 The quantitative EXAFS fitting is studied to authenticate the local coordination configuration of Fe in A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC (Fig. 2i and Supplementary Figs. 17–20). The fitting results revealed the coordination configuration in A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC are FeS1N2-FeN3, Fe2N6, and FeN4 moieties, respectively. The first shell of A-Fe2S1N5/SNC was fitted with Fe-N, Fe-S, and Fe-Fe back-scattering path. A coordination number of approximately 2.5, 0.5, and 1.0 was determined using Fe-N, Fe-S, and Fe-Fe in A-Fe2S1N5/SNC, with the average bond lengths of 1.90 Å, 2.30 Å and 2.56 Å, respectively (inset of Fig. 2i and Supplementary Table 1). Furthermore, the atomic structure in A-Fe2S1N5/SNC was evaluated by the comparison of experimental spectra with the simulated XANES spectra. XANES calculations were conducted based on the FeS1N2-FeN3 moieties with different Fe-S or C-S positions (Supplementary Fig. 21). It appears that the FeS1N2-FeN3 structural motif with Fe-Fe direct bonding, one Fe bonded with one S atom and two N atoms, the other Fe connected by three N atoms in Supplementary Fig. 21a meet an excellent agreement with the experimental spectra. In addition, the C, N, and S types in the carbon matrix were confirmed by soft XAS and XPS spectra (Supplementary Figs. 22–26). It appears that the π* (C = C ring), π* (C–N–Fe or S–C–Fe), and σ* (C–C ring) exist in the support for A-Fe2S1N5/SNC, Fe2N6/NC, and Fe1N4/NC. Furthermore, three obvious peaks in the N K-edge soft XAS and N 1s XPS spectra for A-Fe2S1N5/SNC, Fe2N6/NC, Fe1N4/NC, and SNC represent the coexist of pyridinic N, pyrrolic N, and graphitic N sites on the carbon matrix. Additionally, the C-S-N (C) species suggest the successful doping of the S atom into the NC fragment for A-Fe2S1N5/SNC.25
Electrocatalytic OER performances of A-Fe2S1N5/SNC. To evaluate the alkaline OER electrocatalytic performance of A-Fe2S1N5/SNC, we conducted electrochemical measurements in 1.0 M KOH solution using a standard three-electrode configuration. The obtained A-Fe2S1N5/SNC displayed more obvious advantages (Fig. 3a) in terms of overpotential (193 mV at 10 mA cm− 2) than those of the Fe2N6/NC (281 mV), Fe1N4/NC (369 mV), SNC (409 mV), and even RuO2 (255 mV). The A-Fe2S1N5/SNC delivered a superior Tafel slope than Fe2N6/NC, Fe1N4/NC, SNC, and RuO2, suggesting faster OER kinetics (Supplementary Fig. 27).26 It’s noted that A-Fe2S1N5/SNC was the most active OER electrocatalyst among most reported SACs/DACs catalysts (Supplementary Fig. 28 and Supplementary Table 2), which might give the credit to the asymmetric iron dual-atom sites. Electrochemically active surface area (ECSA) represented the actual number of active sites of the electrocatalytic catalysts, which was evaluated through the double-layer capacitance (Cdl) of the catalysts.27 As derivate from cyclic voltammograms at various sweep rates (Supplementary Fig. 29), the ECSA of the A-Fe2S1N5/SNC was significantly larger than Fe2N6/NC, Fe1N4/NC, and SNC (Supplementary Fig. 30), confirming that the asymmetric configuration in A-Fe2S1N5/SNC possessed more exposed reactive sites and hence enhanced the OER activity.28,29 Importantly, from the ECSA-normalized LSV curves (Supplementary Fig. 31a), the A-Fe2S1N5/SNC catalyst still exhibits the highest activity, suggesting intrinsically improved OER activity on the A-Fe2S1N5/SNC catalyst due to the asymmetric iron dual-atom sites.30 Electrochemical impedance spectroscopy (EIS) was a common indicator to probe interface resistance change between the catalyst electrode and the electrolyte.31 As illustrated in Supplementary Fig. 31b, the A-Fe2S1N5/SNC had the minimum impedance fitting semicircles (Rct) among the contrast electrodes, suggesting its enhanced charge migration ability across the catalyst electrode and electrolyte interface.32 Subsequently, the intrinsic catalytic activity of the catalysts was further assessed by the calculated turnover frequency (TOF) in Fig. 3b and Supplementary Fig. 32a. Notably, the A-Fe2S1N5/SNC catalyst displayed extremely high TOF at the whole overpotentials windows. Extremely, the TOF values for A-Fe2S1N5/SNC were high up to 2880.0 O2 h− 1 at the overpotential of 300 mV, 11 times and 183 times relative to those of Fe2N6/NC (273.0 O2 h− 1) and Fe1N4/NC (15.7 O2 h− 1). In addition, the mass activities of the A-Fe2S1N5/SNC were also higher than those of Fe2N6/NC and Fe1N4/NC (Supplementary Fig. 32b), demonstrating the high intrinsic catalytic activity of A-Fe2S1N5/SNC.33
The in-situ XAFS analysis was performed to elucidate the structural evolution and the real-time change in the Fe chemical state during OER.34,35 A homemade cell was applied to perform the in-situ XAS tests (Supplementary Fig. 33), through the commonly used Lytle detector (in a fluorescence model) to collect all the spectra. As exhibited in Fig. 3c, the Fe K-edge absorption edge was gradually shifted to the higher energy accompanied by an increase in the applied potential from ex-situ condition to 1.42 V vs. RHE, manifesting the average valance state of the Fe sites increased as the OER proceeds.36,37 The formation of higher valence Fe could be further supported by the first derivative XANES curves in Supplementary Fig. 34. To further probe the evolution of the coordination configuration for Fe atoms in A-Fe2S1N5/SNC, FT-EXAFS spectra were recorded at Fe K-edges (Fig. 3d). Apparently, the Fe-N, Fe-S and Fe-Fe shell displayed an obvious dynamic change during OER. The coordination peaks shift to the low-R region, which might be derived from the bridging adsorption of the oxygen-containing intermediates (i.e., O*, OH*, OOH*) on dual Fe sites.38 By considering Fe-N, Fe-S, Fe-Fe, and Fe-O scattering paths, the first coordinated shell of EXAFS curve-fitting analysis for A-Fe2S1N5/SNC was displayed in Supplementary Figs. 35–36 and Supplementary Table 3. Accordingly, slight shrinkage for the Fe-N bond length could fix the dual Fe sites during OER, thus avoiding possible dissolution and improving stability.39
Satisfactorily, the A-Fe2S1N5/SNC electrode could hold excellent electrochemical durability after 2000 h of continuous operation under constant current test conditions (10 mA cm− 2) with > 97% retention (Fig. 3e), which proved its feasibility in industrial applications.40 Meanwhile, the good stability of A-Fe2S1N5/SNC was also validated by LSV curves before and after 50000 CV cycles (merely no loss) (Supplementary Fig. 37). The comparison of the long-term OER durability for A-Fe2S1N5/SNC and other reported electrocatalysts was summarized in Supplementary Fig. 38 and Supplementary Table 2. Undoubtedly, A-Fe2S1N5/SNC was one of the most stable reported OER catalysts. The morphology and structural characterizations for the A-Fe2S1N5/SNC after the durability tests were examined using TEM, XRD, Raman, and XPS analyses (Supplementary Figs. 39–40), which show negligible change.
To determine the universality of the synthesis method, some other 3d transition-metal (for example, M = Co [6.43 wt%], Cu [7.15 wt%], Ni [6.98 wt%], and Mn [7.01 wt%]) diatomic A-M2S1N5/SNC were produced (Supplementary Fig. 41). As demonstrated in Supplementary Figs. 42–45, A-M2S1N5/SNC possess ultra-thin and curly nanosheet morphology with no obvious metal particles. In the AC HAADF-STEM image, plentiful dense dimers are observed, indicating the presence of metal dual-atom pairs in A-M2S1N5/SNC. The unsymmetrically arranged MS1N2-MN3 coordination configuration was further confirmed by XAS spectra and the quantitative EXAFS fittings (Supplementary Figs. 46–52 and Supplementary Table 4). The corresponding OER performance of A-M2S1N5/SNC is also shown in Supplementary Fig. 53. The successful preparation of A-M2S1N5/SNC demonstrates the feasibility and scalability of the strategy proposed in this paper for preparing asymmetric homonuclear diatomic catalysts, providing potential for the industrialization of diatomic catalysts.
Theoretical study of A-Fe2S1N5/SNC on OER. To deeply unearth the activity origin of oxygen electrochemical reaction on the asymmetrical Fe2S1N5/SNC catalyst in comparison with the Fe2N6/NC catalyst, the first-principles method was adopted to study their proton-coupling electron-transfer (PCET) processes.41 Supplementary Fig. 54 shows the optimized configurations of Fe2N6/NC and A-Fe2S1N5/SNC catalysts, two Fe atoms were named Fe-1 and Fe-2 site for ease of distinction, which Fe-2 site connects with one S atom and two N atoms, and the Fe-1 site coordinates with only N atoms in unsymmetrical A-Fe2S1N5/SNC model. Initially, the adsorption of intermediates in OER/ORR was evaluated on the said models, corresponding adsorbed configurations and adsorption-free energies were shown in Supplementary Figs. 55–56, and Supplementary Tables 5–7, respectively. The O and OH species possess strong binding energy with −0.11 and −0.61 eV on the bare A-Fe2S1N5/SNC surface, −0.10 and −0.38 eV on the bare Fe2N6/NC surface, respectively. It meant that the O or OH intermediates were prone to occupy the Fe sites in the electrochemical surroundings.42 As known, the Pourbaix diagram can clearly show the stable states of the catalyst surface under the electrochemical conditions, which are relative to the pH and electrode potential.43 Fig. 4a and Supplementary Fig. 57 depicts the free energies of a series of coverage surfaces sampled for A-Fe2S1N5/SNC and Fe2N6/NC model following electrode potential U vs. SHE at pH = 14, and their optimal models and adsorbed free energy were displayed in Supplementary Figs. 58–61. At the real electrochemical OER/ORR conditions, such as 1.50 V vs RHE (0.67 V vs SHE) for OER, 0.90 V vs RHE (0.07 V vs SHE) for ORR, their surfaces were all covered by 1.5 O coverage (three O atom by two Fe sites) on the above-mentioned models, named as 1.5 O-Fe2S1N5/SNC and 1.5 O-Fe2N6/NC, respectively. Based on their practical models, the adsorptions of OER/ORR intermediates were verified on the Fe-1 and Fe-2 sites on the 1.5 O- Fe2S1N5/SNC configuration, corresponding adsorption energies were shown in Supplementary Table 8, compared to that of 1.5 O-Fe2N6/NC model in Supplementary Table 9. For the unsymmetric 1.5 O-Fe2S1N5/SNC model, the adsorption energy of intermediate species on the Fe-2 site possesses lower energy than the Fe-1 site, while Fe-1 and Fe-2 sites have similar energy for 1.5O-Fe2N6/NC catalyst. As shown in Fig. 4b, compared to 1.5 O-Fe2N6/NC, *OH and *OOH intermediates have smaller Gibbs free energies on 1.5 O-Fe2S1N5/SNC (on Fe-2 site), indicating a lower OER energy barrier.44,45
Hence, Figs. 4c and 4d display their free energy changes of cascade elementary steps for OER on the Fe-2 site, which reveals that 1.5 O-Fe2S1N5/SNC has a lower overpotential (0.34 V) than that of 1.5O-Fe2N6/NC catalyst (0.43 V), consistent with the experimental results. Meanwhile, the asymmetric A-Fe2S1N5/SNC catalyst also owns a larger onset potential (0.98 V) for ORR than the symmetric Fe2N6/NC model (0.93 V) in Supplementary Fig. 62, indicating better ORR activity of asymmetric configuration under the involving S source.46 Similarly, the OER/ORR performance of the Fe-1 site on the said models was shown in Supplementary Figs. 63–64, asymmetric 1.5O-Fe2S1N5/SNC model also owns superior OER activity (overpotential = 0.32 V)/ORR (onset potential = 0.94 V), compared to 1.5O-Fe2N6/NC catalyst (overpotential = 0.44 V)/ORR (onset potential = 0.92 V). Interestingly, for 1.5 O-Fe2S1N5/SNC, the Fe-1 site features better OER performance than the Fe-2 site. Further, differential charge density and Bader charge analysis results demonstrated that the Fe-1 site reduces much more electrons to surrounding coordinate atoms than the Fe-2 site on the bare- and 1.5 O-Fe2S1N5/SNC models, inversely, Fe-1 and Fe-2 sites on symmetric model Fe2N6/NC have similar charge transfer,47,48 as shown in Supplementary Figs. 65–66. It meant that the asymmetric model shows spontaneous electric polarization directing from the Fe-1 site to the Fe-2 site, which results in a disparate activity of OER/ORR.
AEMWE device performances of A-Fe2S1N5/SNC. We further assembled an alkaline AEMWE (anion exchange membrane water electrolyzer) catalyzed by A-Fe2S1N5/SNC as shown in Fig. 5a. The A-Fe2S1N5/SNC-based cell displayed a better performance: a current density of 500 mA cm–2 was reached at 1.94 V, while an even higher current density of 1.0 A cm–2 was reached at 2.18 V, in contrast, the current density of the RuO2||Pt/C cell at 2.0 V was only 200 mA cm–2 (Fig. 5b). Consequently, compared to those catalyzed by precious metal electrocatalysts, the performance of the AEM electrolyzer catalyzed by A-Fe2S1N5/SNC was improved. To give a comprehensive assessment of the alkaline electrolyzer enabled by A-Fe2S1N5/SNC, as shown in Fig. 5c, the performance of the AEM electrolyzer using the A-Fe2S1N5/SNC electrocatalyst exceeded most of the reported state-of-the-art catalysts. To characterize the stability of the electrolyzer, when assembled with A-Fe2S1N5/SNC, the AEM electrolyzer could operate at 100 mA cm− 2 for at least 200 h with a negligible increase in voltage (Fig. 5d). Furthermore, the catalytic performance of ORR (Supplementary Figs. 67–71) and the zinc–air battery (ZAB) were tested to validate the potential application of bifunctional A-Fe2S1N5/SNC catalyst. As depicted in Supplementary Fig. 72, the A-Fe2S1N5/SNC-based battery exhibits superior performance than that of the Pt/C + RuO2), which indicates the potential in commercial applications for A-Fe2S1N5/SNC-based devices.