Atomically precise Ni clusters inducing active NiN2 sites with uniform-large vacancies towards efficient CO2-to-CO conversion

Xu, Guangyuan; Peng, Xingjie; Wu, Chuanqiang; Xi, Shibo; Xiang, Huixin; Feng, Lei; Liu, Zhendong; Duan, Yi; Gan, Lijin; Chen, Si; Kong, Yuan; Ma, Yanzhe; Nie, Fujing; Zhao, Jie; Hai, Xiao; Wei, Wei; Zhou, Meng; Wang, Tianfu; Yao, Chuanhao; Zhou, Wu; Yan, Huan

doi:10.1038/s41467-025-59079-5

Download PDF

Article
Open access
Published: 22 April 2025

Atomically precise Ni clusters inducing active NiN₂ sites with uniform-large vacancies towards efficient CO₂-to-CO conversion

Nature Communications volume 16, Article number: 3774 (2025) Cite this article

4946 Accesses
5 Citations
Metrics details

Subjects

Abstract

CO₂ electroreduction to CO promises to give an efficient strategy for CO₂ fixation and transformation. However, current reported active sites fail to deliver sufficient activity with high CO Faradic efficiency (FEco) over a wide range of potential. Here, we show a general synthetic protocol to fabricate a batch of highly pure and active NiN₂ catalysts with precise engineering of the uniform-large (UL) vacancy around the active sites, which is accomplished through the ‘pre-deposition + pyrolysis’ of various atomically precise Ni clusters (Ni_n) and in-situ etching of the support by the ‘nano bomb’ (sulfur-ligand in the clusters). The NiN₂ sites with UL vacancies could achieve a high turnover frequency (TOF) of 350000 h⁻¹ with ~100% FEco in a wide potential range of 1500 mV. In-situ infrared spectra and theoretical calculations reveal that a highly pure NiN₂ site with UL vacancy contributes to this remarkable catalytic performance compared to the counterparts. This general synthetic strategy enables us to simultaneously engineer active sites and surrounding vacancies with the employment of atomically precise metal clusters, thereby enhancing catalytic performance for other specific reactions.

Asymmetric dinitrogen-coordinated nickel single-atomic sites for efficient CO₂ electroreduction

Article Open access 24 June 2023

Nickel dual-atom sites for electrochemical carbon dioxide reduction

Article 18 August 2022

The role of promoters on NiO catalysts for methane decomposition and hydrogen production

Article Open access 01 July 2025

Introduction

Electrocatalytic reduction of carbon dioxide is a promising method to achieve carbon neutrality and mitigate the global warming crisis^1,2. Transforming CO₂-to-CO is a cost-effective method for electrocatalytic CO₂ reduction, due to its practical manufacturing advantages in techno-economic assessments³. Researchers have noticed that several metals are capable of converting CO₂ to CO, including gold (Au), silver (Ag), iron (Fe), cobalt (Co), and nickel (Ni), etc^4,5,6,7. Among these catalyst candidates, Ni catalysts promise to provide high activity and FEco while suppressing the side reaction (hydrogen evolution reaction, HER), rendering Ni one of the most valuable and effective candidates for CO₂-to-CO conversion⁸. Meanwhile, for practical application, the work potential range for CO₂-to-CO conversion should be wide enough to match the high-fluctuating renewable electricity and to couple with different anode reactions. Therefore, the promising Ni catalysts should be capable of working in a wide range of potentials with highly intrinsic activity, FEco and stability^9,10,11. Indeed, certain synthetic strategies have been reported to fabricate various Ni catalysts with different species/dispersions, morphologies and coordination environments^{12,13,14,15,16,17,18,19,20}. For instance, Zhang reported that dual Ni atoms could offer a moderate Turnover frequency (TOF) of 77,500 h⁻¹ per site with a potential window from −0.4 to −0.95 V (98–99% FE_CO)¹⁵. Also, Song reported that Ni-N species could achieve a TOF of 274,000 h⁻¹ per site within a narrower potential range from −0.3 to −0.7 V (98–99% FE_CO)²⁰. However, to satisfy the practical conditions, the Ni catalysts with the employment of current synthetic methods still can not offer sufficiently high intrinsic activity and FEco within a wide potential window. Therefore, it is urgent to develop a new synthetic strategy for fabricating a highly active, stable and selective Ni site for CO₂-to-CO conversion, capable of operating within a sufficiently wide range of work potential.

In addition to the species/dispersions, morphologies, and coordination, the micro-environment formed by the vacancy around the active site could also play a key role in a synergistic effect, and even vacancy itself can be an active site in many catalytic processes^{21,22,23,24,25,26,27}. For CO₂-to-CO electrocatalytic conversion, there is barely reported work on simultaneous engineering of the vacancy and construction of highly pure and active sites for boosted CO₂ conversion. Here we report a general strategy to synthesize the efficient Ni catalytic sites with uniform-large (UL) vacancies via the introduction of pre-designed atomically precise Ni clusters with sulfur ligands. Rather than the single role of atomically precise metal clusters reported previously as purely unsupported active sites with ligand-protecting, or precursors for the synthesis of supported ligand-protecting clusters^{13,28,29,30,31,32,33}, in our method, they could serve entirely new dual roles in synthetic process, acting as a Ni precursor for Ni individual atoms and also etching the N (nitrogen)-rich carbon support like an in-situ nano “bomb” to form the uniform-large (UL) vacancy surrounding Ni atoms. With this strategy, highly pure and unsaturated NiN_x species (“x” refers to coordination numbers) with UL vacancy could probably be fabricated via Ni cluster decomposition and in-situ etching of carbon support at high temperatures. As an expectation, this highly unsaturated NiN_x species with UL vacancies could make these catalysts potentially more flexible to the adsorption changes of intermediates and promote this specific catalyst more stable, active and competitive to side reaction (HER). More strikingly, we found that this synthetic strategy demonstrated strong generality, where a series of atomically precise Ni clusters with different Ni atom numbers applied to the fabrication of this highly pure NiN_x site with UL vacancy. As a consequence, with the employment of pre-deposition of atomically precise Ni clusters and in-situ sulfur-ligand-etching, we could fabricate a batch of highly pure, active, and stable NiN_x sites with UL vacancies. Compared to the Ni-N species made by previously reported methods, this highly pure NiN_x site with UL vacancy could dominantly suppress the side reaction (HER) and achieve almost 100% FEco and a high TOF of 350,000 h⁻¹ per site within a wide potential range of about 1500 mV (from −0.12 to −1.6 V).

Results

Synthesis of atomically precise Ni clusters and NiN_x catalysts

We first synthesized the atomically precise Ni clusters with a typical method^34,35,36. Single-crystal X-ray diffraction over these different atomically precise Ni clusters ambiguously demonstrated that all these Ni clusters were ring-like species (inset images in Fig. 1a–f), which could facilitate the pre-deposition of the clusters on the carbon support. The sulfur ligands (including thiol/thiophenol) could behave as a precise in-situ etching agent to form a UL vacancy for the simultaneous Ni deposition. To extend the library of cluster precursors, we fabricated small Ni clusters (Ni₅(SC₂H₄Ph)₁₀ and Ni₆(SC₂H₄Ph)₁₂) and relatively larger clusters (Ni_n(SPhCH₃)_2n, n = 9–12) (Fig. 1a–f). Of note, for simplification, we used Ni₅, Ni₆, Ni₉, Ni₁₀, Ni₁₁, and Ni₁₂ as short forms for their molecular compositions. The as-obtained fresh clusters were determined by the high-resolution electrospray ionization mass spectrometry (ESI-MS) and Ultraviolet-Visible-Near Infrared (UV-Vis-NIR) spectroscopy (Fig. 1a–f and Supplementary Fig. S1). For instance, the dominant peaks at m/z ~ 1666.10, 1999.12, 2745.90, 3050.89, 3355.85, and 1831.93, respectively, could be assigned to the molecular ion peaks of [Ni₅ + H]⁺, [Ni₆ + H]⁺, [Ni₉ + H]⁺, [Ni₁₀ + H]⁺, [Ni₁₁ + H]⁺, [Ni₁₂ + 2H]²⁺ adducts (Fig. 1a–f), as evidenced by the well-matched high-resolution experimental MS spectra compared with the calculated isotopic MS patterns of each cluster (Fig. 1a–f). Subsequently, we selected small clusters (Ni₅, Ni₆) and relatively larger cluster (Ni₁₀) as the deposition precursors to synthesize the NiN_x sites with UL vacancies. The Ni cluster precursors were first grafted onto the nitrogen-rich carbon (NC) support. Then, the catalysts were further purified with ethanol to remove the physisorbed Ni clusters. Next, the as-prepared fresh Ni catalysts (designated as Ni_n/NC, “n” refers to the number of the atomically precise Ni clusters) were treated under high temperature to form UL vacancies by in situ sulfur-ligand-etching, providing a highly stable anchoring environment for the deposition of Ni atoms (Fig. 1g and Supplementary Table 1). We found that highly pure NiN_x sites with UL vacancies were obtained through the deposition and pyrolysis of atomically precise Ni clusters along with in situ sulfur-ligand-etching.

**Fig. 1: High resolution ESI-MS of the Ni clusters and the schematic illustration of NiN_x fabrication from Ni_n clusters.**

Characterization for Ni sites and UL vacancies

To shed light on the specific NiN_x active sites with UL vacancies, we resorted to the atomic resolution aberration-corrected high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) imaging and scanning transmission electron microscopy electron energy loss spectroscopy (STEM-EELS) mapping to identify the local coordination environment around Ni atoms. For the fresh catalysts, Ni_n clusters were highly dispersed on NC support (Supplementary Figs. 2–4). The average particle sizes of Ni₅/NC, Ni₆/NC, and Ni₁₀/NC were about 0.8, 1.0, and 1.3 nm (Supplementary Fig. 5). After pyrolysis at high temperatures, the Ni clusters were decomposed into individual atoms (Fig. 2a, b and Supplementary Figs. 6–10) and deposited on the vacancies formed by precise in situ sulfur-ligand-etching. No any visible clusters or nanoparticles were observed in the NiN₂-Ni_n samples, demonstrating atomic dispersion of Ni species on NC support (Supplementary Figs. 6–10). Based on the quantitative atomic STEM-EELS elemental analysis (Fig. 2a, c, k), in the case of NiN_x site derived from the small cluster (Ni₆), it demonstrated that there were about two N atoms spatially correlated with one Ni atom, which confirmed Ni₁-N₂ coordination configuration. In other randomly selected areas, the atomic STEM-EELS mapping also demonstrated the identical coordination environment around the Ni atoms (Supplementary Fig. 11), suggesting the high purity of NiN₂ sites derived from small Ni clusters (Ni₆). For other small cluster (Ni₅) derived NiN₂-Ni₅ sample, the coordination environment is the same as that of NiN₂-Ni₆ (Supplementary Fig. 12).

**Fig. 2: Microscopic characterizations of NiN₂-Ni_n (n = 6, 10) catalysts.**

Interestingly, the large Ni clusters (Ni₁₀) demonstrated similar features with small clusters. The fresh Ni₁₀ clusters were also highly dispersed on the NC, and the Ni atoms formed from Ni₁₀ decomposition were also individually dispersed without any aggregation (Fig. 2b). As revealed by the quantitative atomic STEM-EELS elemental analysis, one Ni atom was possibly coordinated with two N atoms (Fig. 2d, l and Supplementary Fig. 13), which manifested a similar coordination structure with the catalysts derived from the larger clusters (Ni₁₀) with an adequate homogeneity. It is worth noting that there were no sulfur atoms doped around the Ni atoms (Supplementary Fig. 14). For the coordinated nitrogen in the NiN₂ site derived from both the small clusters (Ni₆) and large cluster (Ni₁₀), the STEM-EELS results (Fig. 2i, j) revealed that the nitrogen atoms were in sp² hybridized form^37,38, suggesting the same coordination environment around these NiN₂ sites synthesized from different atomically precise Ni clusters. Interestingly, the fresh samples were oxidized by ozone at low temperature (denoted Ni_n/NC-O₃), obtaining the mixtures of clusters and isolated atoms (Supplementary Figs. 15–17). The NiN₂ sites derived from Ni_n (Ni₅, Ni₆, and Ni₁₀) were denoted as NiN₂-Ni_n (n = 5, 6, 10). Also, NC with one cycle of Ni ALD (atomic layer deposition), pyrolysis with Ni(NO₃)₂·6H₂O precursor and grafted with NiPc (Nickel Phthalocyanine) were denoted as 1cNi/NC, Ni/NC-T and NiPc/NC respectively for benchmark screening. For the benchmark Ni catalysts, Ni atoms were also atomically dispersed on the NC support (Supplementary Figs. 18, 19). In order to confirm the UL vacancy by sulfur-ligand-etching along with the Ni anchoring, we characterized the samples for more local information. As shown in the Fig. 2e–h and Supplementary Figs. 11e–f and 13e–f, the atomic STEM-EELS carbon mapping indicated that the local vacancy was about 0.6–0.7 nm per Ni atom, which was obviously larger and more uniform than that in normal electrocatalysts^{22,23,24,25,26,39,40,41}. The existence of this UL vacancy aligned with the in situ sulfur-ligand-etching. To further confirm the process of in situ sulfur-ligand etching, we identified the formed gaseous species during this process by mass spectra. We observed some fragments of small sulfide molecules formed during this etching process (Supplementary Figs. 20–22), suggesting that the etching of the NC support indeed occurred in this synthetic process.

In addition to the local information, we also resorted to more general characterizations. Similarly, the surface areas of the NC-related materials were totally different (Fig. 3a, b). The surface area of the NC was 888 m²/g, and the high-temperature annealed NC with small vacancies (designated NC-T) also offered a comparable surface area (712 m²/g). Interestingly, the NiN₂-Ni_n samples derived from Ni₅, Ni₆, and Ni₁₀ atomically precise clusters, demonstrated a doubled surface area (1490–1775 m²/g). The pore size distribution also suggested that 0.7–2.5 nm micropores increased sharply in these NiN₂-Ni_n catalysts (Fig. 3b). The evolution of the pore size distribution confirmed that the UL vacancies were possibly formed along with the Ni deposition on the NC. It is worth noting that the formed pores around 0.7–2.5 nm were unambiguously consistent with the vacancy size revealed by atomic STEM-EELS mapping (Fig. 2e–h and Supplementary Figs. 11 and 13), which suggests the excellent uniformity of the vacancies around the Ni atoms in these NiN₂-Ni_n catalysts. To further confirm the size of the vacancy formed in the synthetic process, we performed Small-angle X-ray Scattering (SAXS). As shown in the Fig. 3c and Supplementary Fig. 23, the pores from 0.7 to 2.5 nm also increased significantly, which aligns with the pore size evolution of BET and atomic STEM-EELS mapping. According to the X-ray diffraction (XRD) over these samples (Supplementary Fig. 24a), compared to the control materials, it should be noted that the NC support lost its certain crystalline morphologies along with the increasing disorder domains as revealed by the Raman spectroscopy (Supplementary Fig. 24b), which was probably ascribed to the in situ sulfur-ligand-etching. These results are also consistent with the atomic STEM-EELS mapping, surface area and trend of the pore size distribution. To further confirm the coordination environment of the as-prepared samples, we performed the X-ray absorption over these NiN₂-Ni_n catalysts and other control samples (Fig. 3d, e) for more local coordination information. From the X-ray absorption near edge structure (XANES), the standard NiPc gave the highest valence state (+2) and the Ni foil showed a classic metallic state (Fig. 3d and Supplementary Fig. 25). The XANES spectra of all as-prepared catalysts located between the NiPc and Ni foil, indicating all catalysts derived from atomically precise Ni clusters were between 0 to +2 valence state. For fresh Ni₅/NC, Ni₆/NC, and Ni₁₀/NC, they exhibited similar absorption features, suggesting that they shared the same local Ni-S coordination in these fresh samples. From the X-ray photoelectron spectroscopy (XPS) results, the fresh Ni_n/NC gave lower binding energy of the Ni 2p_3/2 peak than NiN₂-Ni_n (Supplementary Figs. 26–28), which shows good agreement with XANES (Fig. 3d). Also, the binding energy of the Ni 2p_3/2 peak for Ni_n/NC-O₃ manifested high oxidation states (Supplementary Fig. 29).

**Fig. 3: Characterizations of Ni-based catalysts.**

After high-temperature treatment for these fresh catalysts, the obtained NiN₂-Ni₅, NiN₂-Ni₆, and NiN₂-Ni₁₀ catalysts exhibited almost identical characteristic features and similar curves, suggesting the similar Ni-N species in these samples (Fig. 3d). In addition, the fitted Ni–N peak in the N1s XPS spectra (Supplementary Figs. 30–33) also suggested the existence of Ni–N coordination in these NiN₂-Ni_n samples. The extended X-ray absorption fine structure (EXAFS) was also investigated on all Ni catalysts for the local coordination environments (Fig. 3e). For the standard materials, the peaks located at 1.41 Å (NiPc) and 2.11 Å (Ni foil) were assigned to Ni-N and Ni-Ni coordination shells respectively. For fresh catalysts without high-temperature treatments, the experimental EXAFS curves of all Ni_n/NC catalysts almost demonstrated the same peak around 1.63–1.65 Å, which corresponded to the Ni-S coordination shell. The fitting curves (dashed line in Fig. 3e) with the Ni-S₄ pathway showed excellent agreement with the experimental results. Interestingly, after the fresh catalysts were pretreated with ozone at low temperature (Ni_n/NC-O₃), there were two main peaks between 1.41 and 1.61 Å, indicating there were Ni-S and Ni-O/N coordinations in these samples (Supplementary Figs. 34). After the high-temperature treatment, all NiN₂-Ni_n catalysts gave identical curves with the same main peak around 1.41–1.43 Å, suggesting that all NiN₂-Ni_n catalysts obtained the highly identical Ni-N coordination environment. Also, the fitting curves with Ni₁-N₂ pathway over these NiN₂-Ni_n samples are highly consistent with the experimental curves (dashed line in Fig. 3e, Supplementary Fig. 35 and Table 2), and this NiN₂ coordination from EXAFS also agrees well with the results of atomic STEM-EELS mapping. The Ni catalyst with a potentially small vacancy (1cNi/NC) also showed a main peak around 1.47 Å. The XAFS spectra of Ni/NC-T were also provided for benchmarking (Supplementary Figs. 34, 35).

Taken together, the NiN₂ site with UL vacancy could be fabricated by the employment of Ni₅, Ni₆ and Ni₁₀ atomically precise clusters, providing a general strategy to synthesize highly pure and active NiN₂ sites with UL vacancies.

Evaluation of CO₂-to-CO activity

Electrocatalytic CO₂-to-CO tests over all the catalysts were first performed in the H-type cell (Fig. 4a), and the catalysts (Ni_n/NC) were also screened with treatments at different temperatures (Supplementary Figs. 36–39). The only products were the CO and H₂ in all cluster-related catalysts, no liquid products were detected as revealed by ¹HNMR spectra after the tests (Supplementary Figs. 40–41). As expected, the carbon itself just produced H₂ as the main product. For the 1cNi/NC (small vacancy) and NiPc/NC (NiN₄ site) catalysts, they just offered a 300 mV of potential range with 80% FEco, and ~200 mV of potential range with 60% FEco, respectively. For the catalysts derived from normal Ni precursor (Ni(NO₃)₂·6H₂O) without in situ sulfur-ligand-etching (Ni/NC-T), it only provided a narrow potential range with low FEco. For NiN₂-Ni_n catalysts derived from the Ni_n clusters (Ni₅, Ni₆, and Ni₁₀) could display a 1000 mV of potential range with ~99% FEco, demonstrating a high CO selectivity and a wide potential range. All the tests in the flow-cell mode were also performed as shown in Fig. 4b–h and Supplementary Figs. 42–50. The fresh Ni_n/NC (Ni₅, Ni₆, and Ni₁₀) demonstrated a low FEco (less than 20%), indicating the Ni_n cluster itself was retarded to the CO₂ reduction. Also, we tested the ozone pretreated catalysts, and they were highly active for the side reaction (HER) rather than for CO₂-to-CO reaction (Supplementary Fig. 47). However, the NiN₂-Ni₆ with UL vacancy presented above 99% FEco within almost 1500 mV (from −0.12 to −1.6 V) of potential range (Fig. 4b and Supplementary Fig. 49a). Interestingly, it showed the same scenario in the NiN₂-N₅ and NiN₂-N₁₀, demonstrating about 99% FEco with about 1400 mV of potential range (from −0.2 to −1.6 V). Therefore, these NiN₂-Ni_n catalysts derived from different atomically precise Ni clusters could possibly obtain the highly identical active site (NiN₂), demonstrating similar electrocatalytic performance for CO₂-to-CO process. Compared with the catalysts synthesized by other methods (Fig. 4c and Supplementary Table 3), NiN₂-Ni_n fabricated through the current synthetic protocol demonstrated wide potential window (FEco > 95%). For the cathode energy efficiency (CEE), the fresh and ozone-treated catalysts both gave a low CCE (Supplementary Fig. 49c). In contrast, the NiN₂-Ni_n catalysts could offer an ideal CEE value from −0.12 to −1.6 V, stepping close to the theoretical value of CEE in this system (Supplementary Fig. 49c). The NiN₂-Ni_n could also obtain high (~100%) FEco at high work current (300–1200 mA), suggesting that the NiN₂ sites with UL vacancies derived from different Ni clusters exhibited boosted inhibition towards HER (Fig. 4d, e and Supplementary Fig. 49b). For CO₂ conversion, the NiN₂-Ni_n could demonstrate 7–20 times higher than fresh and/or ozone-treated catalysts (Supplementary Fig. 50). 68% of CO₂ conversion could be achieved with 1 mg NiN₂-Ni_n catalysts at a current of 900 mA for all NiN₂-Ni_n catalysts (Fig. 4f). Compared to other reported Ni catalysts (Fig. 4g), we found that the TOFs in the NiN₂-Ni_n catalysts were definitely the highest, achieving about 350,000 h⁻¹ in these NiN₂-Ni_n samples. The stability tests were also carried out on the all NiN₂-Ni_n catalysts (Fig. 4h and Supplementary Fig. 51), and the catalyst could maintain high FEco (>99%) for more than 50 h without any obvious deactivation or aggregation, and 116 h stability in 0.5 M KHCO₃ at −0.6 V (Supplementary Figs. 52, 53), suggesting that the NiN₂-Ni_n could not only give a high TOF, CO₂ conversion and FEco across a wide potential range but also display excellent stability in the electrocatalytic CO₂-to-CO process.

**Fig. 4: Electrocatalytic CO₂-to-CO test.**

Catalytic mechanism over Ni₂-Ni_n catalysts

In-situ attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) was performed on these NiN₂-Ni_n samples to deeply investigate the reaction intermediates to uncover the electrocatalytic CO₂-to-CO process occurring on the NiN₂ site with UL vacancy. The pathway of electro-catalytic reduction of CO₂-to-CO was reported in previous work, and the adsorbed intermediates are COOH* and CO*^42,43. As reported in previous work^44,45,46,47, the peak at 2100 cm⁻¹ was assigned to CO* vibration on NiN₂-Ni₅, and peaks located at 1671 and 1428 cm⁻¹ were ascribed to the adsorption of COOH* (Fig. 5a). Also, the putative adsorbed H₂O* (1639 cm⁻¹) and HCO₃⁻* (1364 cm⁻¹) species were also observed in the NiN₂-Ni₅ catalyst. As expected, in the NiN₂-Ni₆ and NiN₂-Ni₁₀ catalysts, the adsorbed species were almost identical (Fig. 5b, c), suggesting that same active sites and electrocatalytic reduction mechanism were present in these NiN₂-Ni_n catalysts. These in-situ ATR-FTIR results were consistent with the atomic STEM-EELS mapping and XAFS results. Furthermore, we also performed the quasi in-situ XPS (Supplementary Fig. 54) on NiN₂-Ni_n catalysts to elucidate whether there are the electronic changes of Ni atoms during the CO₂-to-CO conversion. The quasi in-situ XPS spectra demonstrated no obvious electronic changes of Ni atoms in NiN₂-Ni_n catalysts, suggesting that the catalysts reconstruction may not occur in these systems under reaction condition.

**Fig. 5: In situ measurements and theoretical calculations.**

To further elucidate the high performance and the mechanism of the NiN₂-Ni_n catalysts with UL vacancies for electrocatalytic CO₂-to-CO conversion, we performed density functional theory (DFT) calculations. To investigate the vacancy around the NiN₂ site, we constructed three structures, including NiN₂ site with a uniform-large vacancy (designated NiN₂-Ni_n), NiN₂ site with small vacancy (designated NiN₂-SV), and four-coordinated Ni as benchmark (designated NiPc/NC), and the corresponding adsorption intermediates are also provided (Fig. 5d, Supplementary Figs. 55–57). As shown in Fig. 5d, all adsorbed intermediate species on NiN₂ with large vacancy are displayed. The process of CO₂ reduction to CO is generally believed to involve different adsorbed intermediates, including CO₂*, COOH* and CO*. We depicted the changes in Gibbs free energy by calculating the adsorption free energies of these intermediates as shown in Fig. 5e. For benchmark model (NiPc/NC in Fig. 5e), the rate-determining step (RDS) is the formation of the COOH* intermediate. Compared to four-coordinated Ni (NiPc/NC), the enhanced adsorption capability of Ni atoms shifted the RDS from the formation of the intermediate (COOH*) to the desorption of products (CO*) in NiN₂-Ni_n and NiN₂-SV. Based on the RDS in the CO₂-to-CO process (Fig. 5e and Supplementary data 1), the catalytic performance of NiN₂-Ni_n (RDS = 0.92 eV) was significantly better than that of NiPc/NC (RDS = 1.78 eV) and NiN₂-SV (RDS = 1.32 eV). A larger 8×8 graphene supercell was also simulated for NiN₂-Ni_n, which revealed that the free energy of each intermediate varies by ~0.2 eV or less (Supplementary Fig. 58). Furthermore, compared to NiN₂-SV, the UL vacancy around the NiN₂ site helped protect the intermediates from the influence of other atoms in the carbon framework of the catalyst except for the active Ni atoms, ultimately leading to a weakened CO* adsorption strength. Moreover, the “d orbital center” (Supplementary Fig. 59) of NiN₂-Ni_n (−1.12 eV) located between that of NiPc/NC (−1.78 eV) and NiN₂-SV (−0.82 eV), further suggesting that the adsorption capacity of NiN₂-Ni_n was between that of NiPc/NC and NiN₂-SV. This observation was consistent with the trend of COOH* and CO* adsorption. In the electrocatalytic CO₂-to-CO process, the HER is the most competitive side reaction. As shown in Fig. 5f, we illustrated a competitive relationship between HER (Supplementary Figs. 60, 61) and CO₂ reduction at different applied potentials (U). In this context, the ordinate was defined as G_H-G_CO2, where a more positive value indicated a greater tendency for CO₂ adsorption. We observed that the value of G_H-G_CO2 in NiN₂ site with large vacancy (NiN₂-Ni_n) was significantly higher than that in NiN₂-SV, demonstrating higher selectivity for CO₂ reduction in the NiN₂ system with large vacancy within a wide potential. Compared to small vacancy (NiN₂-SV), this trend confirmed that UL vacancy could assist the NiN₂ site in facilitating the reaction towards the CO₂-to-CO pathway across the entire potential range. The changes of reaction energies in NiN₂ with large vacancy (NiN₂-Ni_n) with different bias were also provided in the Supplementary Fig. 62. Taken together, the theoretical results reveal that the large vacancy around NiN₂ could boost the activity, FEco and the stability in wide potential range. This agrees well with the observed intermediates results from in-situ infrared experiments, and the experimental catalytic performance.

Discussion

In this work, with the employment of atomically precise Ni clusters, we developed a general synthetic strategy for the synthesis of highly pure NiN₂ sites with UL vacancy via the “pre-deposition+ pyrolysis”, and in-situ precise sulfur-ligand-etching. The atomic STEM-EELS mapping, N₂ physisorption, SAXS, and XAFS characterizations confirm the NiN₂ species and UL vacancies around the Ni atoms. In the electrocatalytic CO₂-to-CO conversion, this NiN₂ with UL vacancy demonstrates highly electrocatalytic performance, achieving the highest TOF of ~350,000 h⁻¹ and ~100% FEco within a wide potential range (~1500 mV). DFT and in-situ ATR-FTIR further elucidate that the UL vacancies around the NiN₂ species could promote the durability at high potential with high CO selectivity, and facilitate the desorption of the target products (CO). This original synthetic strategy provides a promising method to engineer the vacancy with a highly active site by using atomically precise metal clusters towards highly catalytic performance in other reactions.

Methods

Chemicals and materials

Ni(NO₃)₂·6H₂O (Aladdin, 99.99%), Zn(NO₃)₂·6H₂O (Aladdin, 99.99%), NiCl₂·6H₂O (Aladdin, 99.9%), 4-methylphenthiophenol (Aladdin, 98%), tetrahydrofuran (Aladdin, 99.99%), toluene (Aladdin, 99.8%), C₅H₈N₂ (Aladdin, 99.99%), Methanol (Sinopharm Chemical Reagent Co., Ltd, ≥99.8%), Ni(C₅H₅)₂ (Aladdin, NiCp₂, 98%), anion-exchange membrane (Fuelcellstore, FAA-PK-130), ethanol (Beijing Chemical Works, ≥99.8%), HCl (Sinopharm Chemical Reagent Co., Ltd, 36%-38%), sulfuric acid (H₂SO₄, Aldrich, 98%), Nafion solution (Sigma-Aldrich, 5 wt%), N, N-Dimethylformamide (Sinopharm Chemical Reagent Co., Ltd, ≥99.8%), 2-phenylethanethiol (Aladdin, 97%), NaBH₄ (Sinopharm Chemical Reagent Co., Ltd, 96%), CH₂Cl₂ (Aladdin, HPLC). All gases in the experiments were purchased from Nanjing Special Gas Co., Ltd.

Synthesis of NC support

Solution A: 3.0 g Zn(NO₃)₂·6H₂O dissolved in 50 mL CH₃OH, solution B: 6.5 g dimethylimidazole dissolved in 100 mL CH₃OH. Solution B was added dropwise to A and vigorously stirred for 24 h to obtain a white uniform suspension. ZiF-8 sample was acquired by centrifuge and rinsed three times with methanol and then dried at 60 °C in a vacuum. Then, 0.4 g ZiF-8 was pyrolyzed at 1100 °C for 6 h under the protection of Ar to obtain NC support.

Synthesis of NC-T

To acquire the blank control material, 40 mg NC support was ground evenly with 2.4 g melamine and then pyrolyzed at 950 °C for 1 h in Ar (99.999%, 30 ml/min) atmosphere.

Synthesis of Ni_n clusters

Synthesis of Ni_n (PET)_2n (n = 5, 6; PET, phenylethylthiol, SC₂H₄Ph) clusters

100 mg of nickel(II) chloride hexahydrate (NiCl₂·6H₂O) was dissolved in a 9 ml mixture of tetrahydrofuran (THF) and methanol (MeOH) with a 2:1 ratio. Subsequently, 282 μL of PET was introduced into the solution and stirred for 10 min. 2 ml of an aqueous solution containing 40 mg of sodium borohydride (NaBH₄), pre-chilled to 0 °C, was rapidly added to the above solution. After the gas evolution, the suspension was stirred at 1000 rpm for 3 h. Residual reactants were removed via twice washing with methanol, and the product was subsequently extracted using dichloromethane (DCM) and stored at −8 °C. Ni₅(PET)₁₀ and Ni₆(PET)₁₂ were then isolated and purified through preparative thin-layer chromatography (PTLC) employing a polar silica-gel stationary phase and a mobile phase consisting of mixture of petroleum ether and DCM (volume ratio ~ 2:1).

Synthesis of Ni_n(4MPT)_2n (n = 9,10,11,12; 4MPT, 4-methylphenthiophenol, SPhCH₃) clusters

24 mg of nickel(II) chloride hexahydrate (NiCl₂·6H₂O) were rapidly dissolved in 20 ml ethanol to form a verdant solution. After 15 min of stirring, 15 mg of 4MPT were introduced into the solution, and the mixture was stirred continuously for 30 min, followed by a fast addition of 3 ml of an ethanol solution containing 13.7 mg of sodium borohydride (NaBH₄). The reaction proceeded for 12 h at ambient temperature. After the completion of the reaction, the crude products were harvested via centrifugation and triple ethanol washing. Subsequently, nickel cluster homologs were extracted from the crude cluster product with 2 ml CH₂Cl₂. Thin-layer chromatography (TLC) was employed for the separation of nickel cluster homologs, with a developing solvent comprising a mixture of CH₂Cl₂ and petroleum ether with a 1:3 volume ratio.

Synthesis of Ni_n (n = 5, 6, 10)/NC

5 mg Ni_n (n = 5, 6, 10) cluster were dissolved in 20 ml dichloromethane, then 100 mg NC support was added to the above solution. The mixture was stirred at room temperature (25 °C) for 12 h and then filtered after the addition of 200 ml ethanol to obtain Ni_n/NC precursor. The residual organic solvent was removed by vacuum drying at room temperature.

Synthesis of Ni_n (n = 5, 6, 10)/NC-O₃

The synthesis of Ni_n (n = 5,6,10)/NC-O₃ catalysts were performed in a viscous ALD flow reactor (ACME Beijing Technology Co., Ltd) at a pressure of 0.6 Torr. Typically, 100 mg Ni_n /NC (n = 5, 6,10) powders were placed in the ALD chamber at 150 °C. 50 ml/min of N₂ carrier gas and 150 ml/min of ozone (O₃) were dosed into the chamber. The timing sequence was 300 s and 600 s for O₃ exposure and N₂ purge, respectively.

Synthesis of NiN₂-Ni_n (n = 5, 6, 10)

Ni₆/NC was pyrolyzed at different temperatures, 40 mg of Ni₆/NC and 2.4 g of melamine were ground evenly, and then pyrolyzed at 750 °C, 850 °C, 950 °C and 1050 °C for 1 h under the protection of Ar (99.999% 30 ml/min) to obtain NiN₂-Ni₆. To obtain NiN₂-Ni₅ and NiN₂-Ni_10, 40 mg of Ni₅/NC and Ni₁₀/NC were ground evenly with 2.4 g melamine and then pyrolyzed at 950 °C for 1 h in Ar (99.999% 30 ml/min) atmosphere.

Synthesis of 1cNi/NC

Ni ALD was performed in a viscous ALD flow reactor (ACME Beijing Technology Co., Ltd) at a pressure of 0.6 Torr. Ultrahigh-purity N₂ (99.999%, Nanjing Special Gases) at a flow rate of 150 ml/min was used as the carrier gas. The timing sequence was 100 s, 180 s, 600 s, 300 s for NiCp₂ exposure, N₂ purge, O₂ exposure, and N₂ purge, respectively.

Note: 1cNi/NC, 3 N-coordinated Ni species with small vacancy, was also provided for the comparison with NiN₂-Ni_n (Supplementary Fig. 63).

Synthesis of Ni/NC-T

5 mg Ni(NO₃)₂ and 100 mg NC support were dissolved in 20 ml H₂O, then the mixture was stirred at room temperature for 24 h and then filtered to obtain Ni/NC precursor. 40 mg of Ni/NC and 2.4 g of melamine were ground sufficiently and then pyrolyzed at 950 °C for 1 h under Ar atmosphere to obtain Ni/NC-T.

Synthesis of NiPc/NC

5 mg NiPc were dissolved in 20 ml N, N-Dimethylformamide (DMF), then 100 mg NC were added to the above solution, followed by centrifuging and washing with DMF three times, finally dried at 80 °C to obtain NiPc/NC catalyst.

Characterization

Inductively Coupled Plasma Optical Emission Spectrometer (ICP-OES) analysis was performed on an iCAP 7400 to determine the elemental percentage. X-ray diffraction patterns were collected with a Rigaku TTRΙΙΙ diffractometer with Cu Kα radiation at room temperature. Ex situ X-ray photoelectron spectroscopy (XPS) was operated on a Kratos Axis supra+ spectrometer with an excitation source of monochromatized Al Ka (hν = 1486.6 eV) and a pass energy of 40 eV. Raman results were acquired on LabRAM HR Evolution Laser Confocal Raman Microscope. The XAFS spectroscopy results of Ni K-edge are collected at the X-ray absorption fine structure for the catalysis beamline of the Singapore Synchrotron Light Source, Singapore, and a nickel foil is used to calibrate the energy. AC-HAADF-STEM and EDS mapping were operated on a JEOL-F200 transmission electron microscope operating at 200 kV. Atomic STEM-EELS analysis was performed on the Nion HERMES-100 equipped with a C3/C5 corrector, operated at 60 kV with a convergence angle of 32 mrad. The collection angle is 0–75 for STEM-EELS and 75–210 mrad for HAADF. The probe current for STEM-EELS spectrum image acquisition is about 8–17 pA. Principal component analysis, together with Savitzky-Golay smooth was conducted as the denoising strategy to improve the signal-to-noise ratio for STEM-EELS mapping. It should be noted that all the signals/spetra related to the N hybridization state analysis and the quantitative STEM-EELS analysis were the raw data without processing. The quantification was achieved by Digital Micrograph through the following equation:

$$\frac{{N}_{{Ni}}}{{N}_{N}}=\frac{{I}_{{Ni}-L}}{{I}_{N-K}}\times \frac{{\sigma }_{N-K}}{{\sigma }_{{Ni}-L}}$$

(1)

where N is the areal concentration of the atoms, I is the integrated signal intensity, and σ is the cross-section for ionization of an electron in the associated shell. The Hartree-Slater model was chosen to calculate the theoretical cross-section. The region of 40 eV near the edge was excluded and the following region of 70 eV was taken into account as the integration region. (University of Chinese Academy of Sciences)

In situ attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR) was conducted on a Perkin Elmer Spectrum 3 Spectrometer equipped with a mercury cadmium telluride (MCT) detector. The H-type electrochemical cell consisted of a working electrode with a catalyst loaded onto an ATR crystal (Au-coated Si semi-cylindrical prism), a high-purity graphite counter electrode (separated by Nafion 117 membrane), an Ag/AgCl reference electrode, and CO₂-saturated 0.5 M KHCO₃ solution as the electrolyte, which was installed in the spectrometer. Electrolysis was performed over a potential range from the open circuit potential (OCP) to −1.1 V vs. RHE, with spectra recorded at each potential after stabilization. N₂ physisorption analysis were conducted at −196 °C using a Micromeritics ASAP 2460 surface area and porosity analyzer to ascertain the specific surface area. Preceding the analysis, the samples were degassed at 200 °C for a minimum duration of 12 h. Small-angle X-ray scattering (SAXS) were performed on Xeuss 2.0 (Xenocs, France) with a Cu X-ray source (wave-length of 0.1542 nm) and a detector of Pilatus 300 K (Dectris). Heating experiments of 20 mg Ni₆/NC powder were conducted under He flow (30 mL/min) using a thermogravimetric analyzer (Netzsch STA449F3) coupled to a mass spectrometer (Netzsch QMS 403Q).

Quasi in-situ XPS spectra of all electrodes completing CO₂ electrolysis in CO₂ saturated 0.5 M KHCO₃ solution at different potential, were collected on a Thermo Scientific ESCALAB Xi+ X-ray photoelectron spectrometer (the Vacuum Interconnected Nanotech Workstation of the Suzhou Institute). The transfers of electrodes to XPS vacuum chamber were protected with the inert argon.

Electrochemical measurements

The electrochemical measurements were operated on the CorrTest-CS310X workstation (Wuhan Corrtest Instrument Corp. Ltd). During the H-type electrolytic cell test, the catalyst ink consists of 5 mg catalyst, 950 μl isopropanol, and 50 μl Nafion solution. 0.5 mg catalyst deposited on 2 cm² glassy carbon sheet (0.25 mg/cm²) was employed as the working electrode, which was immersed in 0.5 M KHCO₃ (1 cm² platinum foil as counter electrode and Ag/AgCl as reference electrode). All the electrolytes were prepared and used immediately. For CO₂ CO₂-saturated 0.5 M KHCO₃ solution, the value of pH was 7.20 ± 0.10. The Ag/AgCl reference electrode was calibrated in a three-electrode cell employing dual Pt foil electrodes (as working/counter electrodes) and H₂-saturated electrolyte (1 h purging), with cyclic voltammetry conducted at 1 mV s⁻¹. Nafion117 cation exchange membrane (size: 3 cm × 3 cm, thickness: 178 µm) separates the anode and cathode chambers. Nafion-117 membranes were pretreated by heating in Milli-Q water at 80 °C for 1 h, followed by three repetitions of sequential cycles (For each cycle: 3% H₂O₂ and 1.0 M H₂SO₄ treatments at 80 °C for 1 h with intermediate Milli-Q water rinsing). Finally, the membranes were stored in fresh Milli-Q water for further use.

For experiments in the flow cell, catalyst ink contains 5 mg catalyst, 3 ml ethanol, and 100 μl Nafion (sigma, 5%) solution. A piece of gas diffusion electrode (GDE, 1 × 1 cm²) was loaded with 0.2 mg catalysts, or a 5 cm² GDE was loaded with 1 mg catalysts (1.5 mg for NiN₂-Ni₁₀) as the cathode (Ag/AgCl as reference electrode). A mass flow controller controlled gaseous CO₂ entering the gas chamber and then flowing to the gas chromatograph at a flow rate of 10 ml/min monitored by a mass flowmeter. 1 M KOH at a constant rate of 10 ml/min was circulated through the cathode and anode chamber (nickel foam as counter electrode), which were separated by the FAA-3-PK-130 anion exchange membrane (size: 1 cm × 1 cm or 3 cm × 3 cm, thickness: 130 µm). The FAA-3-PK-130 anion exchange membrane was fully converted to OH⁻ form by soaking in 1.0 M KOH at room temperature for 24 h, then rinsed with ultrapure water to neutral pH (~7). The pretreated membrane was then stored in ultrapure water before further use. All experimental voltages were not iR corrected. For 1 M KOH solution, the value of pH was 13.9 ± 0.10. All potentials reported in this study were converted to the reversible hydrogen electrode (RHE) scale using the following equation:

$${{{\rm{E}}}}({{{\rm{vs}}}}.{{{\rm{RHE}}}})={{{\rm{E}}}}({{{\rm{vs}}}}.{{{\rm{Ag}}}}/{{{\rm{AgCl}}}})+{{{\rm{E}}}}^{{{\rm{\theta }}}}({{{\rm{Ag}}}}/{{{\rm{AgCl}}}})+0.0592\times {{{\rm{pH}}}}$$

(2)

Products analysis

During the electrocatalytic reduction process, the reduction products of CO₂ were analyzed online by gas chromatography (GC9790 Plus, Zhejiang Fuli Analytical Instruments Co., Ltd) and the gaseous products entered into a 1 ml quantitative loop at a flow rate of 10 ml/min. Among them, H₂ and high-concentration CO are quantified by a thermal conductivity detector, and low-concentration carbon monoxide is quantified by a flame ionization detector, the correlation coefficient of H₂ and CO calibration curves is 99.99% (Supplementary Figs. 64, 65). Faradaic efficiency (FE) of H₂ and CO was calculated by the following equation.

$${FE}=\frac{n\times F\times W\times v\times P}{R\times T\times I}$$

(3)

Where, n is the electron transferred during the reaction of carbon dioxide and water to the product (CO and H₂, n = 2). F = 96485 C · mol⁻¹; W is the volume fraction of product confirmed by calibration curves from standard gas determined by gas chromatography; v is the gas flow rate through the quantitative loop of gas chromatography (mL/min); P = 1.01 × 10⁵Pa; R = 8.314 J mol⁻¹ K⁻¹; T = 273.15 K. I refers to the total current of cell.

TOF was calculated by the following equation.

$${TOF}=\frac{{I}_{{CO}}\times {M}_{{Ni}}}{{n\times F\times \beta \times m}_{{cat}}}\times 3600$$

(4)

Where, I_CO (A) is the partial current of CO; M_Ni is the atomic mass of Ni (58.69 g/mol); n is the electrons transferred in reaction; F = 96485 C · mol⁻¹; β (wt%) is the amount of Ni loading in Ni based catalysts from ICP-AES analysis; m_cat (g) is the mass of catalyst on electrode.

The cathode energy efficiency (CEE) of CO is calculated as follows:

$${CEE}=\frac{{FE}\times {E}_{0}}{1.23-V}\times 100\%$$

(5)

Where E₀ is the thermodynamic cell potential of CO₂ reduction to CO at the cathode (E₀ = 1.23 − (−0.106) = 1.336 V); the thermodynamic potential of H₂O oxidation at the anode is 1.23 V and equilibrium electrode potential of CO₂ reduction to CO is −0.106 V. V is the applied potential of a full cell without iR correction.

Single pass CO₂ to CO Conversion (SPCC) of different catalysts were calculated as follows:

$${SPCC}=\frac{{I}_{{CO}}\times 60(s)\times {1000\times V}_{{{{\rm{m}}}}}}{n\times F\times v\times 1(\min )}$$

(6)

I_CO (A) is the partial current of CO; V_m (24.05 L/mol) is the molar volume of gas at 25 °C; n is the electrons transferred in reaction; F = 96485 C · mol⁻¹; v (mL/min) is the gas flow rate through the quantitative loop of gas chromatography.

¹H NMR spectra were collected on AVANCE 400 MH Superconducting Fourier Nuclear Magnetic Resonance Spectrometer. Liquid NMR samples consist of 400 µl of electrolyte collected after electrocatalysis, 100 µl of D₂O as the solvent, and 100 μl DMSO as an internal standard.

Computation details

Density-functional theory (DFT) calculations were performed with the Vienna Ab-initio Simulation Package (VASP) codes 5.4⁴⁸. PAW pseudo-potentials⁴⁹ with projectors up to l = 2 for Ni and l = 1 for C, N, O, and l = 0 for H, and the revised Perdew-Burke-Ernzerhof (RPBE) exchange-correlation functional⁵⁰ were employed in our calculations. A plane wave cutoff of 500 eV was applied in the calculations.The valence electron numbers for Ni, C, N, O, and H are 10, 4, 5, 6, and 1, respectively. The convergences of energy and force were set to 10⁻⁴eV, and 0.02 eV/Å, respectively. Spin-polarized calculations with Gaussian smearing in combination with a width of the smearing 0.05 eV were identified for all the analyzed structures. The spin polarization of Ni in NiN₂-Ni_n were initialized at 0, 1, 2, 3, and 4 unpaired electrons, where all but 0 resulted in a net spin of 1 unpaired electron after optimization, which also yields the lowest energy. The initial magnetic moments of Ni in our calculations were set to one. The calculated structure was set by a 6 × 6 graphene supercell, and we transformed the graphene rhombic unit cell into a rectangular unit cell through coordinate transformation. A vacuum layer of 20 Å and a 3 × 3 × 1 Monkhorst-Pack k-point mesh were used in this work. The model of large vacancy of NiN₂ was set based on the experimental results of pore size (average size, 0.6–0.7 nm). For NiN₂ with a large vacancy (NiN₂-Ni_n), we removed two complete benzene rings with four additional carbon atoms and replaced two adjacent carbon atoms with two nitrogen atoms. Meanwhile, for NiN₂ with a small vacancy (NiN₂-SV), we removed a complete benzene ring and replaced two adjacent carbon atoms with two nitrogen atoms.

The theoretical potentials for NiN₂-Ni_n, NiPc/NC and NiN₂-SV were determined, assuming based on a conventional CO₂RR mechanism. The elementary steps of the conventional pathway, are believed to involve CO₂, COOH, CO adsorbed on the surface (*) according to the following:

$${{{{\rm{CO}}}}}_{2}+* \to {{{{\rm{CO}}}}}_{{2}^{*}}$$

(7)

$${{{{\rm{CO}}}}}_{{2}^{*}}+{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}\to {{{{\rm{COOH}}}}}^{*}$$

(8)

$${{{{\rm{COOH}}}}}^{*}+{{{{\rm{H}}}}}^{+}+{{{{\rm{e}}}}}^{-}\to {{{{\rm{CO}}}}}^{*}+{{{{\rm{H}}}}}_{2}{{{\rm{O}}}}$$

(9)

$${{{{\rm{CO}}}}}^{*}\to {{{\rm{CO}}}}+\ast$$

(10)

For each step, the Gibbs free energy differences of these intermediates include zero-point energy (ZPE) correction, enthalpic temperature correction, and entropy contribution. Thus, the calculated Gibbs free energy is defined as: ∆G = E_DFT + ZPE + ∫CpdT − TdS. The details of ZPE, enthalpic temperature correction, entropy contribution are listed in Supplementary Table S4. The gas values for above three parts correction are sourced from the literature⁵¹. ZPE values for the intermediates are calculated from the relevant frequency calculation of VASP, where ${ZPE}={\sum }_{i}\left(\frac{1}{2}*{{hv}}_{i}\right)$. Enthalpic temperature correction and entropy contribution of the intermediates are calculated by vaspkit⁵². Finite displacement methods are used for computing vibrational modes, with the displacement of each ion set at 0.015 Å. The mean absolute error (MAE) correction of gas-phase molecules for CO₂ and CO were +0.41 and −0.18 eV with the RPBE functional⁵¹. For the solvation effect, we implemented the implicit solvation model code using VASPsol^53,54 after structural optimization, where the corrections of the solvation energy for COOH*, CO*, CO₂*, and H* are −0.24, 0.01, −0.12 and 0 eV.

To deal with the hydrogen electrode, the computational hydrogen electrode proposed by Nørskov in 2004⁵⁵ was used. Where, the free energy of proton−electron pair G (H⁺+ e⁻) = ½ G(H₂).

For these four elementary steps:

$${\Delta {{{\rm{G}}}}}_{1}={{{\rm{G}}}}({{{{\rm{CO}}}}}_{{2}^{*}})-{{{{\rm{E}}}}}^{*}-{{{\rm{G}}}}({{{{\rm{CO}}}}}_{2})$$

(11)

$${\Delta {{{\rm{G}}}}}_{2}={{{\rm{G}}}}({{{{\rm{COOH}}}}}^{*})-{{{\rm{G}}}}({{{{\rm{CO}}}}}_{{2}^{*}})-1/2{{{\rm{G}}}}({{{{\rm{H}}}}}_{2})$$

(12)

$${\Delta {{{\rm{G}}}}}_{3}={{{\rm{G}}}}({{{{\rm{CO}}}}}^{*}){{{\mbox{-}}}}{{{\rm{G}}}}({{{{\rm{COOH}}}}}^{*})-1/2{{{\rm{G}}}}({{{{\rm{H}}}}}_{2})+{{{\rm{G}}}}({{{{\rm{H}}}}}_{2}{{{\rm{O}}}})$$

(13)

$${\Delta {{{\rm{G}}}}}_{4}={{{{\rm{E}}}}}^{*}+{{{\rm{G}}}}({{{\rm{CO}}}}){{{\rm{\hbox{-}}}}}{{{\rm{G}}}}({{{{\rm{CO}}}}}^{*})$$

(14)

Data availability

Data supporting the findings of this study are available in the article and supplementary information. Source data are provided with this paper.

References

Xu, S. & Carter, E. A. Theoretical insights into heterogeneous (photo) electrochemical CO₂ reduction. Chem. Rev. 119, 6631–6669 (2018).
Article PubMed Google Scholar
Sullivan, I. et al. Coupling electrochemical CO₂ conversion with CO₂ capture. Nat. Catal. 4, 952–958 (2021).
Article CAS Google Scholar
Jin, S., Hao, Z., Zhang, K., Yan, Z. & Chen, J. Advances and challenges for the electrochemical reduction of CO₂ to CO: from fundamentals to industrialization. Angew. Chem. 133, 20795–20816 (2021).
Article ADS Google Scholar
Li, S. et al. Boosting CO₂ electrochemical reduction with atomically precise surface modification on gold nanoclusters. Angew. Chem. Int. Ed. 60, 6351–6356 (2021).
Article CAS Google Scholar
Gu, J., Hsu, C.-S., Bai, L., Chen, H. M. & Hu, X. Atomically dispersed Fe³⁺ sites catalyze efficient CO₂ electroreduction to CO. Science 364, 1091–1094 (2019).
Article ADS PubMed CAS Google Scholar
Zhang, N. et al. Silver single‐atom catalyst for efficient electrochemical CO₂ reduction synthesized from thermal transformation and surface reconstruction. Angew. Chem. Int. Ed. 60, 6170–6176 (2021).
Article ADS CAS Google Scholar
Liang, S., Huang, L., Gao, Y., Wang, Q. & Liu, B. Electrochemical reduction of CO₂ to CO over transition metal/N‐doped carbon catalysts: the active sites and reaction mechanism. Adv. Sci. 8, 2102886 (2021).
Article CAS Google Scholar
Xu, C., Vasileff, A., Zheng, Y. & Qiao, S. Z. Recent progress of 3d transition metal single‐atom catalysts for electrochemical CO₂ reduction. Adv. Mater. Interfaces 8, 2001904 (2021).
Article CAS Google Scholar
Chen, Z., Mou, K., Wang, X. & Liu, L. Nitrogen‐doped graphene quantum dots enhance the activity of Bi₂O₃ nanosheets for electrochemical reduction of CO₂ in a wide negative potential region. Angew. Chem. Int. Ed. 57, 12790–12794 (2018).
Article CAS Google Scholar
Cai, C. et al. Heteroatoms induce localization of the electric field and promote a wide potential‐window selectivity towards CO in the CO₂ electroreduction. Angew. Chem. Int. Ed. 61, e202212640 (2022).
Article ADS CAS Google Scholar
Chen, Y. et al. Hierarchical Ni/N/C single-site catalyst achieving industrial‐level current density and ultra‐wide potential plateau of high CO Faradic efficiency for CO₂ electroreduction. Adv. Funct. Mater. 33, 2214658 (2023).
Yang, H. B. et al. Atomically dispersed Ni (i) as the active site for electrochemical CO₂ reduction. Nat. Energy 3, 140–147 (2018).
Article ADS CAS Google Scholar
Ding, T. et al. Atomically precise dinuclear site active toward electrocatalytic CO₂ reduction. J. Am. Chem. Soc. 143, 11317–11324 (2021).
Article PubMed CAS Google Scholar
Jia, C. et al. Sulfur‐dopant‐promoted electroreduction of CO₂ over coordinatively unsaturated Ni‐N₂ moieties. Angew. Chem. Int. Ed. 60, 23342–23348 (2021).
Article CAS Google Scholar
Hao, Q. et al. Nickel dual-atom sites for electrochemical carbon dioxide reduction. Nat. Synth. 1, 719–728 (2022).
Article ADS CAS Google Scholar
Wu, Z.-Y. et al. A general synthesis of single atom catalysts with controllable atomic and mesoporous structures. Nat. Synth. 1, 658–667 (2022).
Article ADS CAS Google Scholar
Yang, X. et al. Sulfur-doped unsaturated Ni-N3 coordination for efficient electroreduction of CO₂. Chem. Eng. J. 450, 137950 (2022).
Article CAS Google Scholar
Huang, J. R. et al. Single‐product faradaic efficiency for electrocatalytic of CO₂ to CO at current density larger than 1.2 A cm^{− 2} in neutral aqueous solution by a single‐atom nanozyme. Angew. Chem. Int. Ed. 61, e202210985 (2022).
Article CAS Google Scholar
Zhao, R. et al. Partially nitrided Ni nanoclusters achieve energy‐efficient electrocatalytic CO₂ reduction to CO at ultralow overpotential. Adv. Mater. 35, 2205262 (2023).
Article CAS Google Scholar
Zhou, Y. et al. Asymmetric dinitrogen-coordinated nickel single-atomic sites for efficient CO₂ electroreduction. Nat. Commun. 14, 3776 (2023).
Article ADS PubMed PubMed Central Google Scholar
Campbell, C. T. & Peden, C. H. Oxygen vacancies and catalysis on ceria surfaces. Science 309, 713–714 (2005).
Article PubMed CAS Google Scholar
Yan, X., Jia, Y. & Yao, X. Defects on carbons for electrocatalytic oxygen reduction. Chem. Soc. Rev. 47, 7628–7658 (2018).
Article PubMed CAS Google Scholar
Wei, S. et al. Self-carbon-thermal-reduction strategy for boosting the Fenton-like activity of single Fe-N4 sites by carbon-defect engineering. Nat. Commun. 14, 7549 (2023).
Article ADS PubMed PubMed Central CAS Google Scholar
Ni, W. et al. Electroreduction of carbon dioxide driven by the intrinsic defects in the carbon plane of a single Fe–N₄ site. Adv. Mater. 33, 2003238 (2021).
Article CAS Google Scholar
Wang, X. et al. Single-atom vacancy defect to trigger high-efficiency hydrogen evolution of MoS₂. J. Am. Chem. Soc. 142, 4298–4308 (2020).
Article PubMed CAS Google Scholar
Rong, X., Wang, H. J., Lu, X. L., Si, R. & Lu, T. B. Controlled synthesis of a vacancy‐defect single‐atom catalyst for boosting CO₂ electroreduction. Angew. Chem. 132, 1977–1981 (2020).
Article ADS Google Scholar
Ling, L. L. et al. Potassium‐assisted fabrication of intrinsic defects in porous carbons for electrocatalytic CO₂ reduction. Adv. Mater. 34, 2205933 (2022).
Article CAS Google Scholar
Li, S. et al. Dissecting critical factors for electrochemical CO2 reduction on atomically precise Au nanoclusters. Angew. Chem. 134, e202211771 (2022).
Article ADS Google Scholar
Seong, H. et al. Transplanting gold active sites into non-precious-metal nanoclusters for efficient CO₂-to-CO electroreduction. J. Am. Chem. Soc. 145, 2152–2160 (2023).
Article PubMed CAS Google Scholar
Wu, Q. J. et al. Atomically precise copper nanoclusters for highly efficient electroreduction of CO₂ towards hydrocarbons via breaking the coordination symmetry of Cu site. Angew. Chem. 135, e202306822 (2023).
Article Google Scholar
Wang, J., Xu, F., Wang, Z. Y., Zang, S. Q. & Mak, T. C. Ligand‐shell engineering of a Au28 nanocluster boosts electrocatalytic CO₂ reduction. Angew. Chem. Int. Ed. 61, e202207492 (2022).
Article ADS CAS Google Scholar
Zhuang, S. et al. Hard‐sphere random close‐packed Au₄₇Cd₂ (TBBT) ₃₁ nanoclusters with a Faradaic efficiency of up to 96% for electrocatalytic CO₂ reduction to CO. Angew. Chem. Int. Ed. 59, 3073–3077 (2020).
Article CAS Google Scholar
Tang, L. et al. Poly‐Hydride [Au_I7 (PPh₃) ₇H₅](SbF₆)₂ cluster complex: structure, transformation, and electrocatalytic CO₂ reduction properties. Angew. Chem. 135, e202300553 (2023).
Article Google Scholar
Yao, C. et al. Atomically-precise dopant-controlled single cluster catalysis for electrochemical nitrogen reduction. Nat. Commun. 11, 4389 (2020).
Article ADS PubMed PubMed Central CAS Google Scholar
Zhou, M. et al. Three-orders-of-magnitude variation of carrier lifetimes with crystal phase of gold nanoclusters. Science 364, 279–282 (2019).
Article ADS PubMed CAS Google Scholar
Li, Y. et al. Double-helical assembly of heterodimeric nanoclusters into supercrystals. Nature 594, 380–384 (2021).
Article ADS PubMed CAS Google Scholar
Bhattacharyya, S., Lübbe, M., Bressler, P., Zahn, D. & Richter, F. Structure of nitrogenated amorphous carbon films from NEXAFS. Diam. Relat. Mater. 11, 8–15 (2002).
Article ADS CAS Google Scholar
Arenal, R. et al. Atomic configuration of nitrogen-doped single-walled carbon nanotubes. Nano Lett. 14, 5509–5516 (2014).
Article ADS PubMed CAS Google Scholar
Jia, C. et al. Nitrogen vacancy induced coordinative reconstruction of single‐atom Ni catalyst for efficient electrochemical CO₂ reduction. Adv. Funct. Mater. 31, 2107072 (2021).
Article CAS Google Scholar
Cho, J. et al. Importance of broken geometric symmetry of single-atom Pt sites for efficient electrocatalysis. Nat. Commun. 14, 3233 (2023).
Article ADS PubMed PubMed Central CAS Google Scholar
Jia, Y. & Yao, X. Defects in carbon-based materials for electrocatalysis: synthesis, recognition, and advances. Acc. Chem. Res. 56, 948–958 (2023).
Article PubMed CAS Google Scholar
Zhang, M.-D., Huang, J.-R., Shi, W., Liao, P.-Q. & Chen, X.-M. Synergistic effect in a metal–organic framework boosting the electrochemical CO₂ overall splitting. J. Am. Chem. Soc. 145, 2439–2447 (2023).
Article PubMed CAS Google Scholar
Li, X. et al. Convergent paired electrosynthesis of dimethyl carbonate from carbon dioxide enabled by designing the superstructure of axial oxygen coordinated nickel single-atom catalysts. Energy Environ. Sci. 16, 502–512 (2023).
Article CAS Google Scholar
Dunwell, M. et al. The central role of bicarbonate in the electrochemical reduction of carbon dioxide on gold. J. Am. Chem. Soc. 139, 3774–3783 (2017).
Article PubMed CAS Google Scholar
Firet, N. J. & Smith, W. A. Probing the reaction mechanism of CO₂ electroreduction over Ag films via operando infrared spectroscopy. ACS Catal. 7, 606–612 (2017).
Article CAS Google Scholar
Katayama, Y. et al. An in situ surface-enhanced infrared absorption spectroscopy study of electrochemical CO₂ reduction: selectivity dependence on surface C-bound and O-bound reaction intermediates. J. Phys. Chem. C. 123, 5951–5963 (2018).
Article Google Scholar
Zhang, L. et al. Atomically dispersed Ni–Cu catalysts for pH‐universal CO₂ electroreduction. Adv. Mater. 35, 2209590 (2023).
Article CAS Google Scholar
Kresse, G. & Furthmüller, J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 11169 (1996).
Article ADS CAS Google Scholar
Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B 50, 17953 (1994).
Article ADS Google Scholar
Hammer, B., Hansen, L. B. & Nørskov, J. K. Improved adsorption energetics within density-functional theory using revised Perdew-Burke-Ernzerhof functionals. Phys. Rev. B 59, 7413 (1999).
Article ADS Google Scholar
Peterson, A. A., Abild-Pedersen, F., Studt, F., Rossmeisl, J. & Nørskov, J. K. How copper catalyzes the electroreduction of carbon dioxide into hydrocarbon fuels. Energy Environ. Sci. 3, 1311–1315 (2010).
Article CAS Google Scholar
Wang, V., Xu, N., Liu, J.-C., Tang, G. & Geng, W.-T. VASPKIT: A user-friendly interface facilitating high-throughput computing and analysis using VASP code. Comput. Phys. Commun. 267, 108033 (2021).
Article CAS Google Scholar
Mathew, K., Sundararaman, R., Letchworth-Weaver, K., Arias, T. & Hennig, R. G. Implicit solvation model for density-functional study of nanocrystal surfaces and reaction pathways. J. Chem. Phys. 140, 084106 (2014).
Mathew, K., Kolluru, V., Mula, S., Steinmann, S. N. & Hennig, R. G. Implicit self-consistent electrolyte model in plane-wave density-functional theory. J. Chem. Phys. 151, 234101 (2019).
Nørskov, J. K. et al. Origin of the overpotential for oxygen reduction at a fuel-cell cathode. J. Phys. Chem. B 108, 17886–17892 (2004).
Article PubMed Google Scholar

Download references

Acknowledgements

H.Y. acknowledges support from National Key Research and Development Program of China (Grant 2023YFA1508200), the fund of USTC start-up grant (KY2060000213 and KY2060000204), and National Natural Science Foundation of China (GG2060007005). C.Y. acknowledges the start-up research funding of Fujian Normal University, the Natural Science Foundation of Fujian Province (2024J01300) and the Natural Science Foundation of Ningbo (202003N4005). W.W. acknowledges the support from the National Natural Science Foundation of China (No. 22302225). This work was also financially supported by the National Natural Science Foundation of China (No. 22103003), Anhui Provincial Natural Science Foundation (No. 2208085Y02), the University Annual Scientific Research Plan of Anhui Province (No. 2022AH010013). The numerical calculations in this paper have been done on the supercomputing system in the supercomputing center, University of Science and Technology of China. This work was partially carried out at the Instruments Center for Physical Science, University of Science and Technology of China. This research also benefited from resources and support from the Electron Microscopy Center at the University of Chinese Academy of Sciences.

Author information

These authors contributed equally: Guangyuan Xu, Xingjie Peng, Chuanqiang Wu, Shibo Xi.

Authors and Affiliations

State Key Laboratory of Precision and Intelligent Chemistry, School of Chemistry and Materials Science, University of Science and Technology of China, Hefei, Anhui, PR China
Guangyuan Xu, Lei Feng, Zhendong Liu, Yi Duan, Lijin Gan, Si Chen, Yuan Kong, Yanzhe Ma, Fujing Nie, Jie Zhao & Huan Yan
School of Physical Sciences and CAS Key Laboratory of Vacuum Physics, University of Chinese Academy of Sciences, Beijing, PR China
Xingjie Peng & Wu Zhou
Information Materials and Intelligent Sensing Laboratory of Anhui Province, Institutes of Physical Science and Information Technology, Anhui University, Hefei, Anhui, PR China
Chuanqiang Wu
Institute of Sustainability for Chemicals, Energy and Environment, A*STAR, Singapore, Singapore
Shibo Xi
Strait Institute of Flexible Electronics (SIFE, Future Technologies), Fujian Key Laboratory of Flexible Electronics, Fujian Normal University and Strait Laboratory of Flexible Electronics (SLoFE), Fuzhou, PR China
Huixin Xiang & Chuanhao Yao
School of Materials Science and Engineering, Peking University, Beijing, PR China
Xiao Hai
Vacuum Interconnected Nanotech Workstation, Suzhou Institute of Nano-Tech and Nano-Bionics, Chinese Academy of Sciences, Suzhou, China
Wei Wei
Hefei National Research Center for Physical Sciences at the Microscale, University of Science and Technology of China, Hefei, Anhui, PR China
Meng Zhou
School of Environmental Science and Engineering, Shanghai Jiao Tong University, Shanghai, PR China
Tianfu Wang

Authors

Guangyuan Xu
View author publications
Search author on:PubMed Google Scholar
Xingjie Peng
View author publications
Search author on:PubMed Google Scholar
Chuanqiang Wu
View author publications
Search author on:PubMed Google Scholar
Shibo Xi
View author publications
Search author on:PubMed Google Scholar
Huixin Xiang
View author publications
Search author on:PubMed Google Scholar
Lei Feng
View author publications
Search author on:PubMed Google Scholar
Zhendong Liu
View author publications
Search author on:PubMed Google Scholar
Yi Duan
View author publications
Search author on:PubMed Google Scholar
Lijin Gan
View author publications
Search author on:PubMed Google Scholar
Si Chen
View author publications
Search author on:PubMed Google Scholar
Yuan Kong
View author publications
Search author on:PubMed Google Scholar
Yanzhe Ma
View author publications
Search author on:PubMed Google Scholar
Fujing Nie
View author publications
Search author on:PubMed Google Scholar
Jie Zhao
View author publications
Search author on:PubMed Google Scholar
Xiao Hai
View author publications
Search author on:PubMed Google Scholar
Wei Wei
View author publications
Search author on:PubMed Google Scholar
Meng Zhou
View author publications
Search author on:PubMed Google Scholar
Tianfu Wang
View author publications
Search author on:PubMed Google Scholar
Chuanhao Yao
View author publications
Search author on:PubMed Google Scholar
Wu Zhou
View author publications
Search author on:PubMed Google Scholar
Huan Yan
View author publications
Search author on:PubMed Google Scholar

Contributions

H.Y. supervised and designed the project. G.X. synthesized the catalysts and performed the activity test. H.X. fabricated and characterized the atomically precise Ni clusters under the supervision of C.Y. X.P. performed the HADDF-STEM, atomic STEM-EELS mapping, and catalyst structure-property correlation study under the supervision of T. W. and W.Z. S.X., and X.H. helped perform the XAFS characterization. W.W. performed the quasi in-situ XPS. Y.K. designed the theoretical model and performed the theoretical calculation. C.W. helped screen the catalysts with STEM characterization. F.N., L.F., Z.L., J.Z., and S.C. helped perform the electric measurements. Y.D., L.G., Y.M., and M.Z. helped perform the activity tests and acquire the spectra. H.Y., G.X., and C.Y. wrote the manuscript. All authors participated in the discussion of the data and polishing of the manuscript.

Corresponding authors

Correspondence to Yuan Kong, Tianfu Wang, Chuanhao Yao or Huan Yan.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Description Of Additional Supplementary File

Supplementary Data 1

Transparent Peer Review file

Source data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions

About this article

Cite this article

Xu, G., Peng, X., Wu, C. et al. Atomically precise Ni clusters inducing active NiN₂ sites with uniform-large vacancies towards efficient CO₂-to-CO conversion. Nat Commun 16, 3774 (2025). https://doi.org/10.1038/s41467-025-59079-5

Download citation

Received: 14 November 2024
Accepted: 10 April 2025
Published: 22 April 2025
Version of record: 22 April 2025
DOI: https://doi.org/10.1038/s41467-025-59079-5