A universal and scalable transformation of bulk metals into single-atom catalysts in ionic liquids

Significance

Atomically dispersed metal catalysts, including monometallic and multimetallic single-atom catalysts (SACs), still face certain challenges, i.e., general and scalable synthesis process, compatibility with various support materials, etc. This work proposed a general electrodissolution–electrodeposition strategy in alternative ionic liquids media to convert bulk metals into dispersed atoms. Twenty different monometallic SACs and one multimetallic SAC with five distinct elements have been synthesized on different support materials, the synthesized Pt-SAC exhibits excellent catalytic performance for alkaline hydrogen evolution reaction and oxygen evolution reaction. This work provides an insight into the accurate control of the single-atom electrodeposition kinetics, which has excellent promise to serve as a general strategy for the electrosynthesis of SACs for various applications.

Abstract

Single-atom catalysts (SACs) with maximized metal atom utilization and intriguing properties are of utmost importance for energy conversion and catalysis science. However, the lack of a straightforward and scalable synthesis strategy of SACs on diverse support materials remains the bottleneck for their large-scale industrial applications. Herein, we report a general approach to directly transform bulk metals into single atoms through the precise control of the electrodissolution–electrodeposition kinetics in ionic liquids and demonstrate the successful applicability of up to twenty different monometallic SACs and one multimetallic SAC with five distinct elements. As a case study, the atomically dispersed Pt was electrodeposited onto Ni₃N/Ni-Co-graphene oxide heterostructures in varied scales (up to 5 cm × 5 cm) as bifunctional catalysts with the electronic metal–support interaction, which exhibits low overpotentials at 10 mA cm⁻² for hydrogen evolution reaction (HER, 30 mV) and oxygen evolution reaction (OER, 263 mV) with a relatively low Pt loading (0.98 wt%). This work provides a simple and practical route for large-scale synthesis of various SACs with favorable catalytic properties on diversified supports using alternative ionic liquids and inspires the methodology on precise synthesis of multimetallic single-atom materials with tunable compositions.

Exploiting sustainable energy resources and developing low-carbon technologies are critical for resolving the emerging energy crisis and are the pathway toward a society with carbon neutrality (1, 2). Single-atom catalysts (SACs) with isolated metallic atoms anchored on supports have received soaring scientific interest global-wide in recent years, exhibiting tremendous catalytic capabilities for various catalysis reactions, such as hydrogen evolution reaction (HER), oxygen evolution reaction (OER), nitrogen reduction reaction (NRR), CO₂ reduction reaction (CO₂RR), etc. (3–5). Numerous strategies have been proposed to produce versatile SACs, such as wet chemistry, atomic layer deposition, chemical vapor deposition, etc. (6–8) (SI Appendix, Fig. S1). In particular, some pioneering efforts have recently focused on exploring a universal and straightforward strategy to achieve the mass production of SACs and even multimetallic SACs. For instance, a gas-migration strategy was put forward to directly transform bulk metals into single atoms (SAs) on nitrogen-doped carbon (N-C) supports and demonstrated with three different metals (M-SAs/N-C, M = Cu, Ni, Co) (7). After that, an extensive library of monometallic SACs with 37 different metallic elements and multimetallic SACs, such as 2-, 8- and 12-metal SACs, were realized using a dissolution-and-carbonization synthesis method (9). In 2023, Zou et al. reported a laser-planting strategy to synthesize several SACs including a multiple metallic (e.g., PtIrCuCoNi) (10). However, the practical applications of SACs remain hindered due to the strict requirements and tedious synthesis steps (e.g., high-temperature treatment) in the synthesis process, as well as the substantial gap between lab-scale demonstrations and industrial-scale utilization with high affordability. More importantly, the precise control of the interactions between SAs and supports remains the bottleneck for the large-scale synthesis of monometallic SACs as well as constructing SACs with multimetallic centers for their diversified applications.

Electrochemical methods show advantages of simple operation and promising expansibility for industrial applications and have recently been applied to prepare various SACs (SI Appendix, Table S1), including Ru₁/D-NiFe LDH (11), Pt/np-Co_0.85Se (12), C-Ir₁/Co_0.8Fe_0.2Se₂ (13), Pt_SA-Co(OH)₂@Ag NW (14), etc. Nevertheless, most of these works generally utilize salt-precursors [e.g., H₂PtCl₆ (15)] and share the same issues in manipulating the electrodeposition kinetics due to the instantaneous concentration variation of metal ions and the rapid nucleation of metal nuclei (16). Meanwhile, the prevailing electrochemical synthesis of SACs with aqueous solutions as electrolytes was limited by the intrinsic properties (e.g., narrow electrochemical windows), which excludes the applicability for versatile metals (e.g., W, Mg, Al, etc.). Additionally, the electrochemical process in aqueous electrolytes is commonly accompanied by hydrogen evolution, resulting in inevitable cathode metal embrittlement (17, 18) and metal agglomerations triggered by the surface blocking effect (19, 20). Therefore, a universal and scalable electrochemical synthesis strategy for SACs should allow the accurate control of the single-atom deposition kinetics, whereby electrolytes with moderate electrochemical reaction rates and wide electrochemical windows are required to achieve generality and homogeneousness.

In this work, we report a simple electrochemical strategy that enables direct transformation of various bulk metals into SACs in ionic liquid electrolytes [e.g., the choline chloride–ethylene glycol (ethaline)] via an in situ electrodissolution–electrodeposition route, whereby the wide electrochemical window, excellent stability, and adjustable viscosity of ionic liquids (21) offer an exciting access to the controllable generation of diverse SAs. As case studies, 20 bulk metals (Mg, Al, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ru, In, Sn, W, Ir, Pd, Ag, Pt, and Au) have been successfully transformed into SAs, and a multimetallic SAC with 5 different elements, i.e., Co, Ni, Ir, Pt, and Au (5-metal SAC) has also been realized. Among them, single Pt atoms (0.98 wt% loading) have been incorporated into Ni₃N/Ni-Co-graphene oxide (GO) heterostructure composite supported by carbon cloth (Pt_SA-Ni₃N/Ni-Co-GO@CC). The resultant catalyst with the electronic metal–support interaction modulation exhibits prominent bifunctional electrocatalysis activity in alkaline media, with ultralow overpotentials of 30 mV for HER and 263 mV for OER to achieve a current density of 10 mA cm⁻². A two-electrode electrolyzer was directly assembled using the self-supported Pt_SA-Ni₃N/Ni-Co-GO@CC electrodes for overall water splitting, exhibiting excellent stability over 200 h. Moreover, the scaling capability was demonstrated by preparing a 5 cm × 5 cm catalytic electrode. Intriguingly, the proposed strategy can also be adapted to diverse supports and ionic liquids systems, which suggests that it can be extended to prepare various monometallic and multimetallic SACs with favorable catalytic properties toward multifunctional applications.

Results

Synthesis Strategy and Construction Mechanism of SACs.

Generally, electrodeposition of SACs in liquid electrolytes requires an equilibrium electrochemical process to achieve the atomic-scale control of SAs formation, suggesting that an in situ electrodissolution–electrodeposition process may offer access for accurately fabricating SACs. However, it still remains challenges to produce SACs directly from bulk metals in aqueous electrolytes since their limited in situ electrodissolution reactions. Ionic liquids generally hold a wide potential window (2.0 to 5.8 V) (22, 23) and viscosity range (20 to 40,000 cP) (24), providing possibilities to control metal electrodisslution, ion diffusion, and metal electrodeposition kinetics during the electrochemical process (25, 26), and thus facilitating the subtly regulation of single-atom materials formation (Fig. 1A). Besides, ionic liquids can provide electrostatic protection against aggregation, which have been proven to be effective in improving the stability of metal nanoparticles and homogeneous/single-site catalysts on the supports (27). Here, a halogen-containing deep eutectic solvent (DES, ethaline) with low cost, proper viscosity (52 cP for ethaline vs. 1 cP for water at 20 °C) (28) and wide electrochemical window (~2.0 V vs. ~1.23 V for water) (29) was selected as the electrolyte for SACs production. The halide ions (e.g., Cl⁻) can promote the electrochemical dissolution of bulk metal, as evidenced in our recent work (30), while the moderate viscosity can also guarantee a suitable metal ion transportation velocity for precisely tuning preferred kinetics. These key features highlight that the designed equilibrium electrodissolution–electrodeposition process in ionic liquids holds excellent promise to serve as a general approach for the electrochemical synthesis of SACs (Fig. 1B).

Fig. 1.

Schematic representation of the strategy for synthesizing SACs and their mechanisms. (A) Radar plots of the properties of ionic liquid electrolytes (deep solvent eutectic and other ionic liquid electrolytes) and aqueous electrolytes. The values range from 0 (the center) to 5, where 5 means the highest value of a property, based on the reported works (29, 31, 32). (B) Schematic of the designed equilibrium relationship between the electrodissolution of bulk metals and the electrodeposition of SAs. (C) The illustration of the fabrication route for the self-supported Pt_SA-Ni₃N/Ni-Co-GO on CC substrate. (D) The illustration of the electrolytic synthesis mechanism of SACs as a catalyst for overall water splitting and the digital photo of the synthesized Pt_SA-Ni₃N/Ni-Co-GO (5 cm × 5 cm).

The electrodeposition of Pt SAs on a Ni₃N/Ni-Co-GO@CC support was initially selected to verify the applicability of the as-proposed strategy. Specifically, Ni-Co with embedded GO sheets was grown on CC substrate by electrodeposition in ethaline, followed by nitridation to form a Ni₃N/Ni-Co-GO@CC support for Pt SAs anchoring (Fig. 1C and SI Appendix, Figs. S2 and S6). The in situ electrodissolution–electrodeposition was conducted by cyclic voltammetry (CV) technology, with the prepared Ni₃N/Ni-Co-GO@CC electrode as the working electrode, a Pt foil counter electrode used as the bulk Pt metal precursor, and a nonaqueous Ag wire electrode as the reference electrode. During the electrochemical process, ethaline provides sufficient Cl⁻ surroundings to promote the anodic dissolution of bulk Pt counter electrode into the electrolyte, forming PtCl₄²⁻ and PtCl₆²⁻ (SI Appendix, Figs. S7 and S8). Ni/Co atoms are also slightly dissolved into ethaline, creating specific Ni/Co vacancies (SI Appendix, Fig. S9). Simultaneously, the Pt complex ions tardily electromigrate to the cathode surface region where they were electrodeposited and fixed by vacancy sites, and eventually produce desired Pt SACs (Pt_SA-Ni₃N/Ni-Co- GO@CC) with tunable loadings (0.18 to 1.12 wt%) controlled by regulating CV cycles (Fig. 1D and SI Appendix, Figs. S10 and S11). A scaled-up process was also demonstrated by increasing electrode sizes from 1 cm × 1 cm up to 5 cm × 5 cm, as shown in Fig. 1D. It is experimentally indicated that the in situ electrodissolution–electrodeposition process in ethaline can ensure the continuous Pt dissolution and stable redeposition simultaneously, which implies that the electrochemical process can reach optimum equilibrium electrodeposition for the synthesis of SACs. By simply changing the counter electrode, i.e., using different bulk metals as precursors, the proposed strategy can be facilely extended to produce various monometallic and multimetallic single-atom materials. In this work, Mg, Al, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ru, In, Sn, W, Ir, Pd, Ag, and Au single-atom materials, as well as a multimetallic SAC, i.e., 5-metal SAC (Co, Ni, Ir, Pt, and Au) were achieved.

To explore the electrochemical synthesis process of SACs in ionic liquids, the electrodeposition of Pt on the same support material (Ni₃N/Ni-Co-GO@CC) was also carried out in several typical aqueous electrolytes [i.e., 1.0 M PBS (phosphate buffer solution), 1.0 M KOH, 1.0 and 5.0 M NaCl solutions] for comparison (SI Appendix, Figs. S12 and S14 and Movies S1–S5). The electrodeposition process of Pt in aqueous electrolytes was always accompanied by hydrogen evolution, whereby the insulating hydrogen bubbles would reduce the efficient active surface for anchoring metal atoms (17, 18) (SI Appendix, Fig. S15), and thus may easily cause agglomeration of SAs. Compared with PBS, KOH, and NaCl (including 1.0 and 5.0 M) electrolytes, the ethaline electrolyte also shows obvious superiority for in situ Pt electrodissolution (SI Appendix, Fig. S16). Additionally, uneven microstructures and formation of Pt clusters commonly occurred at the cathode in aqueous electrolytes, which is relatively hard to achieve precise control (SI Appendix, Figs. S17 and S18). To further understand the difference in the electrochemical processes, the electrochemical reduction behaviors of Pt ions in aqueous and ethaline electrolytes were further studied by CV. For the electrolytes with the same concentration of Pt species (K₂PtCl₄, 0.048 M), the reduction of Pt(II) in ethaline is diffusion-controlled with a diffusion coefficient of 1.44 × 10⁻⁶ cm² s⁻¹ (SI Appendix, Supporting Text 1), which is two to three orders of magnitude lower than that in conventional solvents (i.e., aqueous, organic solvents) (26, 33). While in the aqueous electrolyte, this process is controlled by electron transfer rates (SI Appendix, Fig. S19), implying that Pt ions have relatively slower transport kinetic processes in ethaline (26), and accordingly the electrochemical reaction speed in ionic liquids is commonly slower than that in aqueous electrolytes (25), which may contribute to avoiding the atoms aggregation caused by the fast migration of metal ions and the rapid nucleation as well as the growth of nuclei. Furthermore, the electrostatic double layer generated by ionic liquids can maintain the stability of the electrocatalysts, which also cannot be achieved in aqueous solutions (27, 34). These results suggest that the in situ electrodissolution–electrodeposition process in ionic liquids can serve as an alternative strategy for synthesizing SACs.

Atomic Structure Characterization.

The atomic-resolution high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) and energy-dispersive spectroscopy (EDS) images (Fig. 2 and SI Appendix, Figs. S20 and S40) confirm that a family of monometallic SAs (including Mg, Al, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ru, In, Sn, W, Ir, Pd, Ag, Pt, and Au), as well as multimetallic SAs containing five metal elements (Co, Ni, Ir, Pt, and Au with total metals loading of ~0.46 wt%), were atomically dispersed on different supports, including the Ni₃N/Ni-Co-GO, N-doped graphene (NG), NiCo layered double hydroxide, and Ti₃C₂T_x MXene supports. The results of X-ray diffraction (XRD), X-ray photoelectron spectroscopy (XPS), X-ray absorption near-edge structure (XANES) spectroscopy, Fourier-transformed extended X-ray absorption fine structure (FT-EXAFS) and corresponding fittings also demonstrate that the successful construction of SACs (SI Appendix, Figs. S20, S38, S40, and S41 and Table S2). These results indicate that this method is universal and flexible for synthesizing various monometallic (SI Appendix, Fig. S42) and multimetallic SACs by selecting appropriate bulk metal electrodes, ionic liquids systems (SI Appendix, Fig. S39) as well as support materials (SI Appendix, Fig. S40).

Fig. 2.

Structural characterization of the SAs. HAADF-STEM images of Mg (A), Al (B), V (C), Cr (D), Mn (E), Fe (F), Co (G), Ni (H), Cu (I), Zn (J), Mo (K), Ru (L), In (M), Sn (N), W (O), and Ir (P) SAs supported on NG, and Pd (Q), Ag (R), Pt (S), Au (T) SAs supported on Ni₃N/Ni-Co-GO obtained in ethaline. (U) HAADF-STEM image of 5-metal SAC (Co, Ni, Ir, Pt, and Au) supported on NG. The Inset is the close-up view of the boxed area, and the white circles mark the isolated metal atoms. (V) Line-scanning intensity profile along X-Y in the region highlighted with white rectangles in the Inset of (U). (W) HAADF-STEM image of 5-metal SAC and its corresponding EDS elemental maps.

Systematical characterizations were performed on the typically produced Pt_SA-Ni₃N/Ni-Co-GO to analyze the structural information of the as-obtained SACs. The X-ray diffraction (XRD) patterns of the synthesized Pt_SA-Ni₃N/Ni-Co-GO and Pt_SA-Ni₃N-Co-GO show no prominent characteristic peaks of Pt metal except for light shifts (SI Appendix, Fig. S43), elucidating that atomic-level Pt is doped into the phases (Ni₃N/Ni-Co and Ni₃N-Co) and nicely incorporated into the original crystal lattice (12, 35). Additionally, a phase interface of the Pt_SA-Ni₃N/Ni-Co-GO can be further determined through transition electron microscopy (TEM) and high-resolution TEM (HRTEM) images (SI Appendix, Figs. S44 and S45), and the salient interplanar distances of 0.157 and 0.203 nm on both sides of the interface can be ascribed to the (112) and (111) planes of the hexagonal Ni₃N as well as the cubic Ni phase, respectively (36, 37), which confirms the formation of Ni₃N/Ni heterostructure with Co incorporating. Isolated Pt atoms (circled) were manifested by the HAADF-STEM analysis (Fig. 3 A and D), whereby the heavy constituent atoms (bright spots) can be found to be commendably dispersed in the crystal lattice of Ni₃N/Ni-Co and primarily concentrated at the phase interface, implying that Pt SAs are immobilized at Ni/Co positions (38–40). Besides, the different line intensity profiles of two typical areas show the loosely interatomic distances between bright spots, which also confirms the isolated Pt atoms (Fig. 3B). As a comparison, Pt SAs were also fixed in Ni₃N-Co-GO catalyst without heterostructure by the same method (SI Appendix, Fig. S46A). The elemental mapping images further demonstrate the homogeneous distribution of Pt atoms throughout the synthesized Pt_SA-Ni₃N/Ni-Co-GO and Pt_SA-Ni₃N-Co-GO catalysts (Fig. 3D and SI Appendix, Fig. S46B).

Fig. 3.

Surface electronic state and atomic structure characterizations. (A) HAADF-STEM image of Pt_SA-Ni₃N/Ni-Co-GO and the isolated Pt atoms are marked by the white circles. (B) Line-scanning intensity profiles along X-Y in regions 1 and 2 highlighted with white rectangles in (A). (C) HAADF-STEM images of Pt_SA-Ni₃N/Ni-Co-GO with heterostructure and its corresponding interface structure constructed by DFT. (D) HAADF-STEM image and its corresponding EDS elemental mapping of Pt_SA-Ni₃N/Ni-Co-GO. (E) The normalized XANES spectra at Pt L₃-edge of Pt_SA-Ni₃N/Ni-Co-GO, Pt foil, and PtO₂. The Inset shows the corresponding local enlargement. (F) FT-EXAFS spectra corresponding to (E). (G) Wavelet transform contour spectra of PtO₂, Pt_SA-Ni₃N/Ni-Co-GO, and Pt foil. (H) XPS spectra of the Pt_SA-Ni₃N/Ni-Co-GO and commercial Pt/C in Pt 4f region. (I) In situ FTIR spectra of CO adsorption over the Pt_SA-Ni₃N/Ni-Co-GO.

The XANES spectra display that the white-line intensity of Pt_SA-Ni₃N/Ni-Co-GO is lower than that of PtO₂ but higher than that of Pt foil (Fig. 3E), affirming that Pt atoms in Pt_SA-Ni₃N/Ni- Co-GO are more positive (Pt^δ+) in oxidation state relative to Pt foil (Pt⁰), which may be attributed to the electron transfer from Pt to the Ni₃N/Ni-Co-GO support. Moreover, the FT-EXAFS oscillations for the Pt_SA-Ni₃N/Ni-Co-GO shown in Fig. 3F, one conspicuous peak at ~1.7 Å can be ascribed to Pt-N coordination. There is no Pt-Pt contribution of Pt foil at ~2.6 Å (metallic bonding), suggesting no Pt particles or clusters formation. Additionally, the first-shell fitting results from FT-EXAFS features give the coordination number (CN) of approximately 2.0 for the Pt-N shell, 0.6 for Pt-Ni shell or Pt-Co shell in Pt_SA-Ni₃N/Ni-Co-GO (SI Appendix, Figs. S46 B and D and Table S2). It is suggested that the Pt atoms coordinate with N and Ni/Co atoms, occupying the interfacial and superficial Ni or Co positions in the lattice of Ni₃N/Ni-Co. The Ni and Co K-edge XANES spectra of Ni₃N/Ni-Co-GO and Pt_SA-Ni₃N/Ni-Co-GO show similar but slightly different adsorption features (SI Appendix, Fig. S48 A and C). Compared with Ni₃N/Ni-Co-GO, the higher intensity white-line peak of Pt_SA-Ni₃N/Ni-Co-GO verifies the local atomic arrangement of Ni or Co atoms originating from the Pt SAs anchoring. For the FT-EXAFS spectra (SI Appendix, Fig. S48 B and D), the significant peaks of Pt_SA-Ni₃N/Ni-Co-GO shift to a higher value of 0.06 and 0.04 Å compared to that of Ni₃N/Ni-Co-GO, respectively, which might be attributed to the structural lattice distortion caused by the substitution doping of Pt SAs. As displayed in Fig. 3G, the wavelet transform (WT) contour maps in Pt L₃-edge illuminate that the location of the WT intensity maxima (~6.0 Å⁻¹, ~1.6 Å) for the Pt_SA-Ni₃N/Ni-Co-GO is different with that of PtO₂ (~5.1 Å⁻¹, ~1.5 Å) and Pt foil (~11.4 Å⁻¹, ~2.6 Å) in k- and R-spaces, respectively, verifying the existence of isolated Pt sites.

The XPS spectra (Fig. 3H and SI Appendix, Figs. S49 and S50) display that two characteristic signals (Pt 4f_5/2 and 4f_7/2) of the Pt_SA-Ni₃N/Ni-Co-GO exhibit a slight positive shift of 0.5 eV with respect to Pt/C indicated Pt⁰, proving the presence of partially positive charged Pt atoms (Pt^δ+), due to the strong interaction with charge-transfer between the single Pt atoms and the Ni₃N/Ni-Co-GO support (41). To dissect the dispersion of Pt, the CO adsorption over the fabricated Pt_SA-Ni₃N/Ni-Co-GO has also been investigated by in situ Fourier-transform infrared (FTIR) spectroscopy measurement (Fig. 3I). A narrow and quasi-symmetrical band appears at ~2,083 cm⁻¹ during the CO desorption process, which can be rationally assigned to linearly absorbed CO on Pt^δ+ single-atom sites, as commonly observed in Pt single-atom catalysts (42–45). These analyses prove that the separated Pt atoms are atomically decentralized in the Ni₃N/Ni-Co-GO. After anchoring isolated Pt atoms, the Pt_SA-Ni₃N/Ni-Co-GO catalyst exhibits enhanced absorption energy and conductivity (SI Appendix, Fig. S51), thus potentially promoting the electrocatalytic processes (46).

Electrocatalytic Performance Evaluation.

The HER activities of the Pt_SA-Ni₃N/Ni-Co-GO catalysts obtained under different CV cycles, Pt_SA-Ni₃N-Co-GO, Ni₃N/Ni-Co-GO, Ni-Co-GO, 20 wt% Pt/C, and other 19 metal SACs were evaluated by linear sweep voltammetry (LSV) in 1.0 M KOH solution (Fig. 4 A and C and SI Appendix, Figs. S52 and S54 and Movie S6). Notably, the Pt_SA-Ni₃N/Ni-Co-GO catalyst obtained within 4000 CV cycles (only 0.98 wt% Pt SAs) displays the most remarkable catalytic performance, exhibiting overpotentials of 30 (η₁₀) and 61 mV (η₁₀₀) at the current densities of 10 and 100 mA cm⁻², respectively. Its overpotential is lower than those of 20 wt% Pt/C (η₁₀ = 36 mV), Ni₃N/Ni-Co-GO (η₁₀ = 78 mV), Pt_SA-Ni₃N-Co-GO (η₁₀ = 87 mV), and Ni-Co-GO (η₁₀ = 94 mV), demonstrating that the enhanced HER activity of Pt_SA-Ni₃N/Ni-Co-GO is mainly attributed to the efficient interaction between Pt SAs and the noncarbon support. Meanwhile, a Tafel slope (b) of 27.4 mV dec⁻¹ for the Pt_SA-Ni₃N/Ni-Co-GO catalyst was also observed, which is significantly lower than those of Pt/C (28.5 mV dec⁻¹), wNi₃N/Ni-Co-GO (69.8 mV dec⁻¹), Pt_SA-Ni₃N-Co-GO (93.0 mV dec⁻¹), and Ni-Co-GO (78.8 mV dec⁻¹) (Fig. 4B and SI Appendix, Fig. S52D), unambiguously suggesting the rapid HER kinetics originated from the incorporation of isolated Pt atoms. Based on the b of 27.4 mV dec⁻¹, which is close to the b (30 mV dec⁻¹) of Tafel reaction (M-H + M-H → H₂ + 2M) (47), the H₂ generated by the Pt_SA-Ni₃N/Ni-Co-GO catalysis follows a Volmer-Tafel mechanism with the Tafel step (chemical desorption) as the rate-limiting step. Note that the unusual mass activity of this catalyst at an overpotential of 100 mV reaches 6.33 A mg⁻¹, nearly 58 times greater than that of Pt/C (0.11 A mg⁻¹) under the same conditions (Fig. 4D and SI Appendix, Supporting Text 2). Moreover, the turnover frequency (TOF) of Pt_SA-Ni₃N/Ni-Co-GO at −80 mV was calculated to be 5.65 H₂ s⁻¹, approximately 71 times higher than that of Pt/C (0.08 H₂ s⁻¹), indicating the highly efficient utilization of Pt SAs (Fig. 4E and SI Appendix, Supporting Text 3).

Fig. 4.

Alkaline electrocatalytic performance and comparison. (A) LSV curves, (B) Tafel plots, and (C) corresponding overpotentials at 10 and 100 mA cm⁻² for the Pt_SA-Ni₃N/Ni-Co-GO, Ni₃N/Ni-Co-GO, Ni-Co-GO, and Pt/C. (D) The mass activity of the Pt_SA-Ni₃N/Ni-Co-GO and Pt/C. (E) TOFs plots of the Pt_SA-Ni₃N/Ni-Co-GO and Pt/C. (F) Comparison of Tafel slope and overpotentials (at 10 mA cm⁻²) with references all measured in 1.0 M KOH. (G) Comparison of HER performance, Pt contents, and electrodeposition conditions for Pt_SA-Ni₃N/Ni-Co-GO with reported SACs obtained from aqueous electrolytes. (H) Comparison of the overpotentials at 10 mA cm⁻² for Pt_SA-Ni₃N/Ni-Co-GO with previously reported values. Inset is the OER polarization curves of Pt_SA-Ni₃N/Ni-Co-GO and Ru₂O. (I) Polarization curves of Pt_SA-Ni₃N/Ni-Co-GO and metric catalysts for overall water splitting. Inset is the digital photo during the process. (J) Comparison of the cell voltages at 10 mA cm⁻² for Pt_SA-Ni₃N/Ni-Co-GO with reported electrocatalysts. (K) The water electrolysis stability test for Pt_SA-Ni₃N/Ni-Co-GO at 10 mA cm⁻² for 200 h in 1.0 M KOH. Error bars in (C) and (D) illustrate the SD of 10 independent samples.

Apparently, the Pt_SA-Ni₃N/Ni-Co-GO catalyst exhibits ultralow overpotential and Tafel slope (η₁₀ = 30 mV, b = 27.4 mV dec⁻¹), surpassing most of the reported state-of-the-art electrocatalysts obtained by diverse methods (Fig. 4F and SI Appendix, Tables S3 and S4), such as Pt@PCM (48) (η₁₀ = 139 mV, b = 73.6 mV dec⁻¹), Pt₁/N-C (42) (η₁₀ = 46 mV, b = 36.8 mV dec⁻¹), Pt SA/MXene (49) (η₁₀ = 33 mV, b = 43.9 mV dec⁻¹). Its excellent HER catalytic activity is mainly attributed to the coordination species (N) with lone electron pairs, which can capture isolated Pt metal centers and regulate their electronic structures (50). Compared with the previously reported various SACs obtained by electrochemical methods in aqueous solutions (Fig. 4G and SI Appendix, Table S1), the Pt_SA-Ni₃N/Ni-Co-GO synthesized in ethaline exhibits comparable HER activity with relatively low Pt loading, which is also superior to the catalysts prepared in PBS, KOH, and NaCl aqueous solutions in this work (SI Appendix, Figs. S55–S57). Additionally, the maximum current density of the electrochemical synthesis process in ethaline for SACs is approximately 28 mA cm⁻² (SI Appendix, Fig. S12A), much smaller than those of the reported synthesis processes in aqueous solutions (Fig. 4G, i.e., reaching up to 80 to 1,000 mA cm⁻²) (12, 14, 51, 52), which is mainly created by the electrodeposition of Pt and the accompanied hydrogen evolution reaction. These results confirm that the reaction speed in ethaline for forming SAs has been slowed down, suggesting the electrochemical synthesis of SACs with tunable Pt loadings (SI Appendix, Figs. S52B and S56) can be facilely controlled in ionic liquids with slower kinetics. Notably, the Pt_SA-Ni₃N/Ni-Co-GO synthesized in ethaline has a mass activity of 6.33 A mg⁻¹ with 0.98 wt% Pt, outperforming other reported Pt SACs obtained in aqueous solutions [e.g., the mass activity of 1.32 A mg⁻¹ with 1.03 wt% Pt in Pt/np-Co_0.85Se (12)], demonstrating that ionic liquids have certain advantages for the synthesis of efficient SACs.

In addition, the Pt_SA-Ni₃N/Ni-Co-GO gives a larger double-layer capacitance (C_dl) of 33.9 mF cm⁻² compared with the Ni₃N/Ni-Co-GO (17.8 mF cm⁻²), declaring that more active sites were exposed on the catalyst with Pt atoms immobilization (SI Appendix, Fig. S58). Besides, only 3.36 Ω of charge transfer resistance (R_ct2) between the electrocatalyst and electrolyte was measured for the Pt_SA-Ni₃N/Ni-Co-GO catalyst, which is significantly lower than that of Ni₃N/Ni-Co-GO (12.73 Ω), Ni-Co-GO (21.17 Ω) and CC (520.85 Ω) (SI Appendix, Fig. S59), suggesting the enhanced HER charge transfer kinetics and the fast Faradaic reaction process. This observation may be attributed to the incorporation of monatomic Pt altering the electronic structure of SAC and the construction of a 3D porous Ni₃N/Ni-Co-GO heterostructure. As reflected in HER catalyzing durability from long-term electrolysis studies (SI Appendix, Figs. S60 and S61), only weeny potential changes after 3,000 cycles or 21 h were observed for Pt_SA-Ni₃N/Ni-Co-GO, indicating the admirable cycling and long-term stability. All these results confirmed the outstanding stability of Pt_SA-Ni₃N/Ni-Co-GO, highlighting the advance by coupling with single Pt atoms in Ni₃N/Ni-Co-GO. Notably, large area electrodes (3 cm × 3 cm, 5 cm × 5 cm) and the 5-metal (Co, Ni, Ir, Pt, and Au) SAC (η₁₀ = 13 mV, b = 19.5 mV dec⁻¹) also exhibit excellent HER activities (SI Appendix, Figs. S62 and S63), demonstrating the proposed strategy has good universality and extensibility for the electrochemical synthesis of SACs as well as multimetallic SACs with outstanding catalysis performances.

The OER activity of the Pt_SA-Ni₃N/Ni-Co-GO was also measured by LSV. The Pt_SA-Ni₃N/Ni-Co-GO only delivers an overpotential of 263 mV at 10 mA cm⁻² and a low Tafel slope of 61.4 mV dec⁻¹, which is lower than RuO₂ (η₁₀ = 295 mV, b = 103.7 mV dec⁻¹) and most of the recently reported electrocatalysts (Fig. 4H and SI Appendix, Fig. S64 and Table S5). After HER and OER tests, there was no change in the phase of the sample (SI Appendix, Fig. S65). Inspired by the excellent activities of the as-synthesized Pt_SA-Ni₃N/Ni-Co-GO toward HER and OER, a two-electrode cell was assembled for overall water splitting in 1.0 M KOH by using Pt_SA-Ni₃N/Ni-Co-GO as both cathode and anode (Fig. 4I). Compared with Pt/C || RuO₂ (1.591 V), the Pt_SA-Ni₃N/Ni-Co-GO electrolyzer only requires a cell voltage of 1.536 V to drive current density of 10 mA cm⁻². The outstanding electrochemical activity also overmatches the most reported bifunctional electrocatalysts (Fig. 4J and SI Appendix, Table S6). As expected, the Pt_SA-Ni₃N/Ni-Co-GO electrolyzer exhibits significant stability over 200 h test at low current density but less ideal stability under high current density (Fig. 4K and SI Appendix, Fig. S66), and the enlarged electrode (5 cm × 5 cm) also possesses comparable catalytic performance for overall water splitting (SI Appendix, Fig. S60 and Movie S7).

Theoretical Calculation Analysis.

To further investigate the electronic structure and catalysis mechanism of the Ni₃N/Ni-Co-GO tailored Pt SAs with insights into the influences of interface engineering and atom doping on the enhanced catalytic activity, the density functional theory (DFT) calculations were studied. Constructing the optimized Pt_SA interface-Ni₃N/Ni-Co and Pt_SA surface-Ni₃N/Ni-Co models by single Pt atom replacing certain Ni atom (interface and surface) in Ni₃N/Ni-Co heterostructure, the Ni₃N/Ni-Co and Pt_SA-Ni₃N-Co models were also built for comparison (Fig. 5 A and C and SI Appendix, Figs. S67 and S69). The negative formation energies of Pt_SA surface-Ni₃N/Ni-Co and Pt_SA interface-Ni₃N/Ni-Co indicate that it is theoretically possible for Pt single atom doping on the surface and interface of the Ni₃N/Ni-Co heterostructure (SI Appendix, Fig. S70 and Supporting Text 4). After a Pt atom is immobilized in Ni₃N/Ni-Co heterostructure or Ni₃N-Co, the charge accumulation is primarily distributed around the Pt site to form a local electron-rich region by metal–support coordination (Fig. 5 B and D and SI Appendix, Fig. S71). The electronic interaction involving the electron transfer between Pt SAs and Ni₃N/Ni-Co support, results in charge distribution changes, which is conducive to enhancing the HER kinetics (53, 54). The projected density of states (PDOS) of Pt_SA-Ni₃N-Co, Pt_SA surface-Ni₃N/Ni-Co, and Pt_SA interface-Ni₃N/Ni-Co reveal a higher occupation at the Pt site near the Fermi level compared to Ni₃N/Ni-Co (Fig. 5E), indicating their increased intrinsic electrical conductivity and enhanced electron mobility (55). This result demonstrates that the introduction of Pt atom and Ni₃N/Ni-Co heterostructure can effectively improve the total d-electron domination of the catalyst, further boosting the activation of H₂O and energetically catalytic performance, which is consistent with the experimental observations (Fig. 4). The d-band features of Pt atom in Ni₃N-Co and Ni₃N/Ni-Co coordinated configurations were also investigated. The Ni₃N/Ni-Co loaded Pt atoms (Pt_SA interface-Ni₃N/Ni-Co and Pt_SA surface-Ni₃N/Ni-Co) have a wider 5d band and higher density near the Fermi level than that of Pt_SA-Ni₃N-Co (Fig. 5F), suggesting that Ni₃N/Ni-Co-coupled Pt molecules induce more free electrons near the Pt site than Pt_SA-Ni₃N-Co, which is more beneficial to the adsorption and transfer of H reactants.

Fig. 5.

Theoretical calculation investigations. (A) Side view of the optimized structure and (B) calculated electron density difference for Pt_SA interface-Ni₃N/Ni-Co. (C) Top view of the optimized structure and (D) calculated electron density difference for Pt_SA surface-Ni₃N/Ni-Co. Yellow color represents charge density accumulation, and the cyan areas show charge density reduction (the isosurface value is 0.003 e Å⁻¹). (E) Calculated PDOS of Ni₃N/Ni-Co, Pt_SA-Ni₃N-Co, Pt_SA surface-Ni₃N/Ni-Co, and Pt_SA interface-Ni₃N/Ni-Co. (F) Calculated Pt 5d band of Pt_SA-Ni₃N-Co, Pt_SA surface-Ni₃N/Ni-Co, and Pt_SA interface-Ni₃N/Ni-Co. (G) Calculated H₂O adsorption free energies on Ni₃N/Ni-Co, Pt_SA-Ni₃N-Co, Pt_SA surface-Ni₃N/Ni-Co, and Pt_SA interface-Ni₃N/Ni-Co. (H) Optimized configuration of the HER process on the Pt_SA interface-Ni₃N/Ni-Co. (I) Energy barriers of water dissociation and (J) calculated H^* adsorption free energies diagrams toward HER on Ni₃N/Ni-Co, Pt_SA-Ni₃N-Co, Pt_SA surface-Ni₃N/Ni-Co, and Pt_SA interface-Ni₃N/Ni-Co, respectively.

For the HER and OER processes in alkaline solutions, it can be seen that the adsorption energy of H₂O ( $∆G_{{{\mathrm{H}}_2\mathrm{O}}^{\ast }}$   ) plays an important role. As shown in Fig. 5G, the $∆G_{{{\mathrm{H}}_2\mathrm{O}}^{\ast }}$ for Ni₃N/Ni-Co, Pt_SA-Ni₃N-Co, Pt_SA surface-Ni₃N/Ni-Co, and Pt_SA interface-Ni₃N/Ni-Co are −0.02, −0.46, −0.47, and −0.52 eV, respectively. The Pt_SA surface-Ni₃N/Ni-Co and Pt_SA interface-Ni₃N/Ni-Co display the more negative binding energies than Ni₃N/Ni-Co and Pt_SA-Ni₃N-Co, suggesting that the Ni₃N/Ni-Co structure with atomic-level Pt dopant is favorable for the adsorption of H₂O in the Volmer step, which would promote subsequent reactions. Furthermore, the activation energies of water dissociation on Ni₃N/Ni-Co, Pt_SA-Ni₃N-Co, Pt_SA surface-Ni₃N/Ni-Co, and Pt_SA interface-Ni₃N/Ni-Co are 0.88, 0.67, 0.65, and 0.48 eV, respectively (Fig. 5 H and I and SI Appendix, Fig. S72). These results demonstrate that the Pt_SA interface-Ni₃N/Ni-Co heterostructure has the lowest energy barrier for water dissociation into *OH and H*, thereby facilitating HER and OER processes. Moreover, the H atom was anchored in the hollow site of the Ni₃N/Ni-Co surface, and the charge density equally distributed on three Ni sites, forming a stable H* adsorption configuration. However, after anchoring Pt atoms, the charge density redistributes on the surface of Pt_SA-Ni₃N-Co, Pt_SA surface-Ni₃N/Ni-Co, and Pt_SA interface-Ni₃N/Ni-Co, and is mainly concentrated near the Pt atoms (SI Appendix, Fig. S73), weakening the H* adsorption strength and further improving the HER performance. The free energies of hydrogen adsorption (∆G_H*) for the Pt_SA interface-Ni₃N/Ni-Co (−0.11 eV) and Pt_SA surface-Ni₃N/Ni-Co (−0.13 eV) are closer to 0 eV than that of Ni₃N/Ni-Co (−0.38 eV) and Pt_SA-Ni₃N-Co (−0.26 eV), indicating the Pt_SA-Ni₃N/Ni-Co has weaker hydrogen adsorption capacity and easier to desorb hydrogen, and thus expediting HER dynamics (Fig. 5J).

For the OER process, the transition of *OOH from *O is the rate-determining step (Rds) among four steps for Ni₃N/Ni-Co and Pt_SA surface-Ni₃N/Ni-Co (SI Appendix, Fig. S74). Compared with the Ni₃N/Ni-Co (η = 0.76 V), the Pt_SA surface-Ni₃N/Ni-Co is thermodynamically more suitable for OER with a lower overpotential of 0.57 V. The electron-deficient and positively charged Ni/Co species in Pt_SA-Ni₃N/Ni-Co-GO possibly facilitate the generation of surface oxide species as discussed before, which act as catalytically active sites for the OER. The addition of Pt SAs is beneficial to the OER of the Pt_SA-Ni₃N/Ni-Co-GO catalyst, mainly because Pt SAs adjust the electronic structure of the Ni₃N/Ni-Co-GO support to obtain electron-deficient and positively charged Ni/Co species. These results confirm that the Pt_SA-Ni₃N/Ni-Co obtained in an ionic liquid has promising HER and OER catalytic performances.

Discussion

In summary, we proposed and validated a universal strategy to engineer atomic-scale materials by electrochemically transforming bulk metals into single metal atoms in ionic liquids. Twenty monometallic single-atom materials (Mg, Al, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ru, In, Sn, W, Ir, Pd, Ag, Pt, and Au) and a five-metal (Co, Ni, Ir, Pt, and Au) SAC were effectively prepared through this strategy. When 0.98 wt% Pt atoms are anchored onto the self-standing Ni₃N/Ni-Co-GO heterostructure, the resulting Pt_SA-Ni₃N/Ni-Co-GO exhibits exceptional HER and OER performances with a pretty low overpotential of 30 and 263 mV at 10 mA cm⁻², a smaller Tafel slope of 27.4 and 61.4 mV dec⁻¹, respectively. Moreover, the Pt_SA-Ni₃N/Ni-Co-GO@CC coupled in a water electrolyzer only delivers a low cell voltage of 1.536 V to reach 10 mA cm⁻² for overall water splitting and exhibits significant stability. The successful preparation of bifunctionally amplified electrodes with efficient catalytic activity suggests that the method is promising for industrial applications. Theoretical calculations further imply that the substitution of Pt atoms for Ni atoms on the surface and interface of this catalyst can modulate its internal electronic structure, contributing to boosting the adsorption and dissociation of H₂O, the desorption of H*, as well as the O–O coupling, which intuitively reflects the superior HER and OER activities. This general strategy enables the synthesis of various single-atom materials, including multimetallic SACs, by designing ionic liquids and support materials. It broadens the pathways for electrochemical synthesis of SACs with the potential for industrialization, and may also expand a direction for the applications of ionic liquids.

Materials and Methods

Raw Materials.

All the raw materials, including choline chloride [HOC₂H₄N(CH₃)₃Cl], ethylene glycol [(CH₂OH)₂], nickel chloride hexahydrate (NiCl₂·6H₂O), cobalt chloride hexahydrate (CoCl₂·6H₂O), K₂PtCl₄, graphene oxide (GO), 1-ethyl-3-methylimidazolium tetrafluoroborate, 1-butyl-3-methylimidazolium tetrafluoroborate, urea, potassium hydroxide (KOH), Pt/C (20 wt % Pt), ruthenium dioxide (RuO₂), Nafion solution, Mg, Al, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, In, Sn and W plate, Ru bulk, Ir wire, and alcohol were purchased from commercial sources and directly used without further purification.

Synthesis of the Pt_SA-Ni₃N/Ni-Co-GO@CC.

The Pt_SA-Ni₃N/Ni-Co-GO@CC was fabricated by the CV method on an electrochemical station (Biologic-803). The three-electrode system consists of the prepared Ni₃N/Ni-Co-GO@CC as a working electrode, Pt plate as a counter electrode, and nonaqueous Ag wire electrode as a reference electrode. Three ionic liquids (ethaline, 1-ethyl-3-methylimidazolium tetrafluoroborate, or 1-butyl-3-methylimidazolium tetrafluoroborate) were used as the electrolytes. The electrochemical synthesis in ethaline electrolyte was carried out from −0.6 to −1.2 V (vs. Ag) at 60 °C with a scan rate of 100 mV s⁻¹ for different cycles (500, 1,000, 2,000, 3,000, 4,000, 5,000, 6,000, 8,000, 10,000, and 12,000 cycles). The process was repeated in 1-ethyl-3-methylimidazolium tetrafluoroborate or 1-butyl-3-methylimidazolium tetrafluoroborate electrolyte for 4,000 cycles. As comparisons, the same fabricated procedures were performed in 1.0 M KOH and 1.0 M PBS, respectively, by the CV method at 0 ~ −0.6 V (vs. RHE) for different cycles (6,000, 8,000, 10,000, and 12,000 cycles). Similarly, the fabricated procedures were also conducted in 1.0 and 5.0 M NaCl solutions at 0 ~ −0.6 V (vs. RHE) for 4000 cycles at room temperature, where the Hg/HgO electrode was used as the reference electrode. The synthesis processes of the Ni-Co-GO@CC, Ni₃N/Ni-Co-GO@CC, Ni₃N-Co-GO@CC, and enlarged Pt_SA-Ni₃N/Ni-Co-GO@CC (3 cm × 3 cm and 5 cm × 5 cm) are detailed in SI Appendix, Supporting Text 5.

Synthesis of the Pt SAs on Different Supports and M_SA (M = Ag, Au, and Pd)-Ni₃N/Ni-Co-GO@CC.

The detailed synthesis processes are provided in SI Appendix, Supporting Text 5.

Synthesis of the M_SA (M = Mg, Al, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, Ru, In, Sn, W, and Ir)-NG and 5-metal SACs.

The nitrogen-doped graphene (NG) powders were fabricated as supported materials (see details in SI Appendix, Supporting Text 5). After mixing the NG powders (40 mg) with 0.2 mL Nafion solution (5 wt%), 1.4 mL alcohol, and 0.4 mL deionized water and sonicating, we took 0.6 mL of this liquid and inked it on the CC and then dried it for use as a substrate. The process of preparing SAs is the same as the above method for fabricating Pt_SA-Ni₃N/Ni-Co-GO@CC, where Mg, Al, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, Mo, In, Sn or W plate, Ru bulk or Ir wire was replaced to the Pt plate at different conditions (see details in SI Appendix, Supporting Text 5). We note that SAs materials with tunable contents can be facilely obtained through adjusting the electrochemical parameters.

Characterizations.

X-ray diffraction (XRD) patterns were conducted on a Bruker-AXS D8 Advance with a Cu Kα radiation at a scan rate of 3 ° min^–1. The morphology and structure of the as-prepared samples were performed by scanning electron microscopy (SEM, FEI Nova Nano SEM 450) equipped with X-ray energy dispersive spectrometer (EDS, Oxford INCA EDS system) and transmission electron microscope (TEM, FEI TF20) equipped with Super-X EDS. The HAADF-STEM images were obtained on a JEM-ARM300F and Thermo Fisher Scientific Titan Themis Z. The surface elemental analysis was detected by X-ray photoelectron spectra (XPS, Thermo Scientific K-Alpha⁺). The UPS spectroscopy and the Fourier transform infrared (FTIR) measurements of CO adsorption for catalysts were characterized by PHI5000 Versa Probe III and Tensor 27, respectively. The UV-visible spectra of several solutions were detected by UV-3600. The contents of Pt in the catalysts and solutions were collected by inductively coupled plasma mass spectrometry (ICP-MS, Agilent 7700). The X-ray absorption spectroscopy (XAS) measurements, including X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) of the samples, were performed at the Beamline of TPS44A1 in the National Synchrotron Radiation Research Center (NSRRC), Taiwan, China.

Electrochemical Measurements.

Electrochemical tests were examined on a CHI 660e or PARSTAT 4000A electrochemical workstation with a three-electrode system, using Hg/HgO electrode, carbon rod, and the fabricated samples as the reference, counter, and working electrodes, respectively. For comparison, commercial Pt/C (20 wt%) and RuO₂ catalysts inked on CC substrate were also prepared (SI Appendix, Supporting Text 5). Except for large-sized electrocatalysts, the geometric surface area of other as-prepared electrocatalysts is 2 cm². LSV measurements were performed at a scan rate of 2 mV s⁻¹ in 1.0 M KOH solution at room temperature with 95% iR-compensations at −0.6 ~ −1.5 V (vs. Hg/HgO) for HER and 0.9 ~ 0.1 V (vs. Hg/HgO) for OER, respectively. In 1.0 M KOH, E_RHE = E_Hg/HgO + 0.0591 × pH + 0.098 V. The electrochemically active surface areas (ECSAs) were estimated by cyclic voltammetry (CV) with different scan rates (20, 40, 60, 80, 100, and 120 mV s⁻¹) at a potential range of −0.4 ~ −0.5 V (vs. Hg/HgO). AC impedance measurements were evaluated from 5 × 10⁵ Hz to 10⁻² Hz at the overpotential of 100 mV with 5 mV AC voltage. The stability tests and overall water splitting performance tests are detailed in SI Appendix, Supporting Text 5.

Theoretical Computational Methods.

Spin-polarized DFT calculations were conducted to optimize structures by the Vienna ab initio simulation package (VASP) (56). The projector augmented wave (PAW) potentials with a plane-wave cutoff energy of 450 eV were implemented to compute the interaction between the ionic cores and valence electrons (57, 58). The generalized gradient approximation (GGA) functional of Perdew–Burke–Ernzerhof (PBE) functional was applied as the exchange-correlation functional (59). The Grimme’s semiempirical DFT-D3 dispersion correction was utilized to describe the van der Waals (vdW) interactions (60). The convergence criteria of electronic energies and atomic forces for all calculations were 10⁻⁵ eV and 0.03 eV Å⁻¹.

We construct a 4-layered (111) slab structure of Ni₃N from the experimental results as a catalyst surface. A vacuum layer of 15 Å was added to prevent the effects of two adjacent layers. A Ni₃N-Co model was constructed based on the optimized Ni₃N model, and one Ni atom was replaced by Co atom. While the Ni₃N-Co-Pt model was built by replacing two Ni atoms with Co and Pt, respectively. In order to facilitate the establishment of the heterostructure model, the Ni₃N-Ni interface model was constructed by assembling Ni₃N (001) layers with a 2 × 2 supercell onto a 3-layer-thick 4 × 4 Ni (111) slab. During the DFT calculation, the bottom two metal layers were fixed. We applied 4 × 4 × 1 Gamma-centered k-points for the Brillouin zone. To study the HER reaction pathways, the H₂O* and H* absorbed intermediates were optimized by DFT calculation to obtain the energy of each structure. The free energy for H and H₂O adsorption was defined as follows:

$$\mathrm{\Delta }{G}_{{\mathrm{H}}^{*}}=\Delta E_{{\mathrm{H}}^{*}}+\Delta \text{ZPE}-\mathrm{T}\mathrm{\Delta }\mathrm{S},$$ [1]

$$∆G_{{{\mathrm{H}}_2\mathrm{O}}^{\ast }}=\mathrm{\Delta }{E}_{{{\mathrm{H}}_2\mathrm{O}}^{\ast }}+\mathrm{\Delta }\text{ZPE}-\mathrm{T}\mathrm{\Delta }\mathrm{S},$$ [2]

where $∆$E_H* and $∆E_{{{\mathrm{H}}_2\mathrm{O}}^{\ast }}$ are calculated adsorption energy of H and H₂O. The zero-point energy (ZPE) and entropy corrections were performed through frequency calculations. We use the computational hydrogen electrode (CHE) model proposed by Nørskov et al. (61) to calculate the Gibbs free energy of the intermediates. The dissociation barrier was computed using the climbing image nudged elastic band (cl-NEB) method (62).

Formation Energy, Quantum Chemistry, and OER Calculations.

The calculation details were described in SI Appendix, Supporting Texts 4 and 5, respectively.

Supplementary Material

Appendix 01 (PDF)

Movie S1.

CV process for depositing Pt on the Ni3N/Ni-Co-GO support in ethaline electrolyte.

Movie S2.

CV process for depositing Pt on the Ni3N/Ni-Co-GO support in 1.0 M KOH electrolyte.

Movie S3.

CV process for depositing Pt on the Ni₃N/Ni-Co-GO support in 1.0 M PBS electrolyte.

Movie S4.

CV process for depositing Pt on the Ni₃N/Ni-Co-GO support in 1.0 M NaCl electrolyte.

Movie S5.

CV process for depositing Pt on the Ni₃N/Ni-Co-GO support in 5.0 M NaCl electrolyte.

Movie S6.

The Pt_SA-Ni₃N/Ni-Co-GO electrode (1 cm × 1 cm) for HER during the LSV process in 1.0 M KOH.

Movie S7.

Large Pt_SA-Ni3N/Ni-Co-GO electrodes (5 cm × 5 cm) for overall water splitting at 1.8 V in 1.0 M KOH.

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.