English

• Electrocatalytic reduction of CO₂ (CO₂RR), based on the sustainable energy input (solar-wind-geo-hydro, SWGH), has been regarded as a promising pathway to convert the captured CO₂ into value-added chemicals [1-5]. The products of CO₂RR span from simple C1 products (carbon monoxide (CO) and formic acid) to higher-order multi-electron products including ethanol, n-propanol, etc. [6,7]. For CO₂-to-CO, typical working catalysts are precious metals (e.g., Au [8], Ag [9]), however limited by their scarcity and high cost for large scale deployment. In recent years, earth-abundant carbon materials with relevant to proper functionalization turn into more appealing catalysts for CO₂RR. There are three major groups including: (1) metal-free heteroatom (B [10], N [11,12], P [13], S [14], etc.)-doped carbons, most commonly N-doped type (NC); (2) transition metal (M) and nitrogen (N) co-doped carbons, typically M-N-C type [15-17] (3) carbon immobilized molecular catalysts [18].

  Among these, M-N-C type catalysts are a kind of interesting materials, which not only have the advantages of metal catalysts but also have a low cost. For instance, Su et al. reported a novel Ni-N-modified graphene electrocatalysts, Ni-N-Gr that was prepared by the heat treatment of a Ni-N organometallic complex in the presence of graphene oxide [19]. Later on, various transition metals (e.g., Co, Fe, Ni, etc.) and nitrogen co-doped carbon-based catalysts were explored for electrochemical reduction CO₂ to CO. Extensive researches were conducted to unveil the nature of the active sites [20,21], which include their coordination details [22,23], the dispersion [24,25] as well as the surface and volumetric density [26].

  Porous carbons decorated with Ni-N coordination sites have proven to be particularly active for the reduction of CO₂ to CO [27-29]. For instance, Li et al. proposed that Ni-N₄ is the active site proved by EXAFS, which has high selectivity and activity for CO₂ to CO (FE_CO of 99% at −0.81 V and current density of 28.6 mA/cm²) [30]. Similarly, a recent characterization study by Koshy et al. indicated that the Ni-N₄ sites are likely the active centers based on the time-of-flight secondary ion mass spectrometry (TOF-SIMS) [31]. The TOF-SIMS technique provided the fragments information of the Ni-N-C catalysts, however, for samples with a low Ni loading (e.g., < 1 wt%) lacked featured fragment information [32]. The reported CO₂RR activities vary considerably, and some of which are contradictory. In addition, there also lacks determination of the contribution of Ni and N doped C-moieties and pure N-functionalities in Ni-N-C type catalysts.

  Moreover, the low concentration of the dispersed Ni-N_x centers in porous carbon matrix is a challenging problem, which calls for effective methods to enrich their surface/volumetric density, and thus for higher current density towards targeted reduction products. Typical strategies to increase active sites include defect-trapping strategy [33], dangling bond trapping strategy [34], chemical chelation strategy [35], and vapor deposition (CVD) strategy [36]. The direct increase of the metal species in the synthesis often causes the migration and agglomeration of the dispersed metals into CO₂RR-inactive nanoparticles.

  Herein, we report a family of Ni-N-C catalysts with distinct properties on the dispersion and accessibility of atomically dispersed Ni-N₃C active sites for electrocatalytic CO₂RR. The model catalysts were prepared by nanoengineering of oxidized carbon sheets (CSs), which were subject to different functionalization for targeted active sites (Fig. 1). All the model samples were prepared from the same batch of CSs colloids. For instance, the Ni-N-C catalyst was prepared through the complex-assisted method to simultaneously impart the nitrogen and carbon atoms coordinated nickel active sites. In details, we used the tris(ethylenediamine)nickel(Ⅱ) complex (denoted Ni(en)₃) as the Ni dopants, wherein nickel(Ⅱ) serves as Lewis acid readily binding six N atoms with lone pair of electrons. The Ni(en)₃ anchored CSs was thermally treated, and leached with hydrochloric acid subsequently, resulting in the Ni-N-C materials (Ni-N-C_I, Fig. 1a). In parallel, replacing of Ni(en)₃ with en would lead to the formation of N-doped carbons (NC_I). We further coated the Ni-N-C_I materials with a thin layer of nitrogen doped carbon through a known aniline CVD process [37]. By doing so, the model catalysts of NC-t@Ni-N_C (where t indicates the CVD time) were obtained, which featured the shielded coordinated Ni sites (Fig. 1b). Furthermore, we re-imparted the coordinated Ni sites on the NC-120@Ni-N_C samples by using the same Ni(en)₃-assisted method, aiming to double the density of coordinated Ni sites (Ni-N-C_II, Fig. 1c, details can be found in Supporting information). In addition, we measured the thickness of Ni-N-C layer on Ni-N-C_I using atomic force microscopy (AFM) technique. The thickness of the pristine carbon nanosheets is 4.3 ± 0.6 nm (Figs. S1a and b in Supporting information). After CVD treatment, the resultant sample of NC-120@Ni-N-C_I show the thickness of 13.3 ± 2.1 nm (Figs. S1c and d in Supporting information). Accordingly, the thickness of Ni-N-C layer on both sides can be determined to be around 9 nm.

  Figure 1

  

  Figure 1.  Schematic illustration of the designed model catalysts based on nanoengineering towards targeted Ni and N-containing functionalities.

  DownLoad: Full-Size Img PowerPoint

  Structural characterizations were carried out to determine the physicochemical features of the three groups of model catalysts. In the XRD patterns, only a broad peak at approximately 25° was detected, suggesting that the Ni-species were in a highly dispersed state (Fig. S2 in Supporting information). Based on N₂ adsorption isotherms (Fig. S3 in Supporting information), the specific surface area (S_BET) of the Ni-N-C_I catalyst was determined to be 188 m²/g, which is very close to its NC counterpart (189 m²/g). Deposition of a thin NC layer onto Ni-N-C_I catalyst led to a reduced S_BET of 124 and 115 m²/g for NC-10@Ni-N_C and NC-120@Ni-N_C, respectively. This is likely because the deposition thermodynamically prefers at the defect sites, e.g., nano-pore or edge sites, and thus lowers the detectable S_BET (Fig. S3).

  The schematic of Ni-N-C catalysts was shown in Fig. 2a. The dispersion of Ni species was studied by high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM, Fig. 2b) and transmission electron microscopy (TEM, Fig. S4 in Supporting information). The Ni and N species were detected to be evenly dispersed (Fig. 2b, below). By using the HAADF-STEM technique (Figs. 2c–g), the bright dots are observed that are relevant to the Ni atoms in atomically dispersed state (Fig. 2c). In more details, the dispersed Ni species are predominately binding to the edge sites of the randomly stacked graphene layers; in contrast, the density of the Ni atoms is much lower on the basal plane (Fig. 2d). This phenomenon aligns with the findings of previous reports [38]. Upon further magnifying, one can see that the Ni species mainly exist as coordinated Ni centers, 2 Å in Ni-N_x or tiny Ni clusters 4–5 Å (Figs. 2e–g). The single atom or cluster of Ni species were assessed by analysis of the STEM images using Nano Measurer 1.2 software. The average statistical results reveal that 74% of the bright dots are attributed to single Ni coordinated N and C-sites, while 26% are associated with metallic Ni clusters (Table S1 in Supporting information).

  Figure 2

  

  Figure 2.  (a) Typical procedures for the synthesis of Ni-N-C_I catalyst. (b) HAADF-STEM image and the EELS mapping of the Ni-N-C_I. (c–g) Aberration-corrected HAADF-STEM images of Ni-N-C_I under different magnifications.

  DownLoad: Full-Size Img PowerPoint

  Furthermore, the chemical environments of coordinated Ni centers of a typical catalyst were investigated by the X-ray absorption spectroscopy (XAS). The X-ray absorption near-edge structure (XANES) spectra displayed the peaks at 8340 eV of the 1s → 4p_z transitions, which are evident on the Ni-N-C samples than benchmarks of Ni foil and NiO. The position of the rising edge of Ni-N-C was found to fall between that of Ni foil and NiO (Fig. 3a). Hence, the valence of the Ni atoms (Ni^δ+) in the coordinated states should be 0 < δ < 2. We further conducted Electron Paramagnetic Resonance (EPR) measurements at 100 K to track the unpaired electron in Ni atoms (Fig. S5 in Supporting information). However, no signals were obtained. Thus, the existence of unsaturated Ni^δ+ could be excluded. Therefore, we speculated this was caused by the co-existence of two forms of Ni species: (1) coordinated Ni atoms showing +2 valance states, (2) metallic Ni clusters with 0 valance state. Notably, the characteristic peak at 8340 eV observed in our samples is consistent with the nickel phthalocyanine (Ni(Ⅱ)Pc) with square planar coordinated Ni-N₄, as reported in previous studies [39,40]. Similarly, the peaks at 8352 and 8358 eV that are relevant to the 1s → 4p_xp_y transitions and multiple scattering processes, coinciding with the reported results on Ni(Ⅱ)Pc. However, the intensity was reduced, likely due to the distorted geometry and/or the mixed coordinating environment of Ni to neighbored N or C atoms [40].

  Figure 3

  

  Figure 3.  Electronic states of Ni atom and chemical state of surface N-functionalities in this series of catalysts. (a) Ni K-edge XANES profiles of Ni-N-C_I, Ni-N-C_II as well as Ni foil and NiO benchmarks, inset shows the expanded pre-edge region. (b) Corresponding Fourier transforms spectra. (c) Fittings results of Ni-N-C_I and Ni-N-C_II. (d) XPS spectra comparison in N 1s region.

  DownLoad: Full-Size Img PowerPoint

  We also collected the Ni K-edge extended X-ray absorption fine structure (EXAFS) spectra to identify the coordination environment of the Ni atoms in Ni-N-C. The Fourier transformed EXAFS spectra showed that the weak peak at 2.1 Å was ascribed to Ni-Ni interaction (Fig. 3b) [41]. By combining the above XRD and HAADF-STEM results, it can be concluded that the Ni-Ni interaction is associated with the formation of small Ni clusters. The main sharp peak at ca. 1.38 Å was stemmed from the scattering interaction between the Ni atoms and the first shell (Ni-N_x) [42,43]. The fitting results of the EXAFS curves suggest that each Ni site is bonded to 2.4~2.6 N atoms. Based on the XAS characterization, high-resolution XPS spectra as well as the previous findings on the identification of the nickel-coordinated nitrogen centers [25,41,44,45]. We speculate that the active coordinated sites are likely to be Ni-N₃C (Fig. 3c).

  The surface chemical composition and elemental states of NC, Ni-N-C, and NC-t@Ni-N_C were studied by X-ray photoelectron spectroscopy (XPS) (Fig. 3d and Table 1). The surface Ni content for Ni-N-C_I was 0.56 at%. After coating a layer of NC, the surface Ni content reduced to 0.26 at% for the sample of NC-120@Ni-N-C_I. This indicated that ca. 54% of Ni sites were covered by the NC layer, being undetectable by XPS. In contrast, 10 min CVD process barely affected the detectable surface Ni species, i.e., 0.56 vs. 0.55 at% before and after 10-min deposition (Fig. S6 in Supporting information). In contrast, the secondary functionalization with Ni(en)₃ on NC-120@Ni-N-C resulted in a further increase in its Ni contents to 0.72 at%, indicating a 1.3 times higher density of surface Ni sites. Meanwhile, we determined the overall compositions by elemental analysis (Table 1), which showed a similar trend as the XPS results.

  Table 1

  

  Table 1.  Elemental composition in the surface and bulk region of this family of model catalysts.

  DownLoad: CSV

  The fitted high-resolution N 1s spectra for all samples revealed the presence of prominent pyridinic (~398.4 eV) and pyrrolic-N (~400.6 eV), together with the graphitic-N (~401.5 eV), and oxidized N (N-ox ~403.2 eV) species (Fig. 3d and Fig. S7 in Supporting information) [46]. Upon doping with Ni species, the Ni-N-C catalysts evidenced the emerging of Ni-N_x moieties (~399.7 eV). Depositing a thin NC layer onto Ni-N-C catalysts by CVD for 10 min slightly reduced the Ni-N_x proportion, however, a 120-min long deposition rendered the surface state more like pure NC. Another interesting feature is the clear reduction of pyrrolic-N fraction (from 1.61 at% to 0.99 at%) but the increase of the total surface-N contents (from 3.98 at% to 4.17 at%) upon coordinating with Ni species (Tables S2 and S3 in Supporting information). It is reasonable to believe that the re-arrangement of N functionality is induced by the introduction of Ni species, likely forming the thermodynamically more stable Ni-N₃ moieties [47,48]. It also identified that the pyrrolic-N played a crucial role in the above re-arrangement [49].

  Next, we investigated the CO₂ reduction performances of the three model catalysts. Liner sweep voltammetry (LSV) saturated electrolytes than Ar-saturated for all the catalysts (Fig. S8 in Supporting information). Both the recorded current density and the onset potential coincides similar trend, indicating the activity orders: Ni-N-C > NC-10@Ni-N-C > NC-120@Ni-N-C > NC. The challenging issue is that the community lies in how to quantitatively determine the contribution that each type of active site offers [49,50]. We thoroughly investigated the product distribution after stationary electrolysis at each applied potential for 1 h (Figs. 4a and b, Fig. S9 in Supporting information). Over all the catalysts, the detected products are CO and H₂, accounting for a total unit FE. For clarity, we also confirmed the used electrolyte by ¹H NMR, which excluded the formation of any liquid products.

  Figure 4

  

  Figure 4.  CO₂RR evaluation results on NC, Ni-N-C and NC layer coated Ni-N-C counterparts. (a, b) FE_CO at varied applied potentials in the H-cell, E vs. RHE. (c, d) Corresponding CO partial current density (j_CO) for this family of model catalysts.

  DownLoad: Full-Size Img PowerPoint

  We first checked the reliability of the collected data by taking the typical electrocatalyst of Ni-N-C_II as the example. The presented results are the average of three repeated experiments, and the error bar was plotted using the standard deviation for three repeated experiments (Fig. S10 in Supporting information). As shown, it shows quite good reproducibility. In the following, we follow the same measurement procedure and evaluated the CO₂RR performances. Fig. 4a showed the FE_CO for Group 1 catalysts which have direct N coordinated Ni sites. The catalyst Ni-N-C_I reached the highest FE_CO of ca. 80% at −0.7 V vs. RHE. The value is 10%−15% lower than that of the reported single atom catalysts (normally FE_CO > 90%) [51], but precisely reflect its structural features that composed of 74% single Ni coordinated N and C-sites and 26% metallic Ni clusters (Table S1 and Fig. S11 in Supporting information). The 26% Ni clusters are believed to preferably favor the hydrogen evolution reaction (HER), contributing a FE_H2 of 20% (Fig. S12 in Supporting information). To confirm this issue, we further extensively re-leached our Ni-N-C_I sample for prolonged time (5 days), in order to reduce the fraction of Ni clusters of Ni-N-C_III. As expected, its FE_CO indeed reached 90.3% at −0.8 V vs. RHE. We compared the catalytic performance with the atomically dispersed Ni-N-C electrocatalysts in Table S4 (Supporting information), which showed comparable activity with previous observations [32,52]. The Ni-N-C_II with 1.3 times more Ni sites showed similar FE_CO as that of Ni-N-C_I, indicating the identical nature of active sites.

  The current density, governed by the density of active sites, is a key parameter for practical applications [53]. The CO partial current densities (j_CO) of Group 1 catalysts were plotted in Fig. 4c. The current densities reached a plateau when potentials reached in the range of −1.0 ~ −1.1 V vs. RHE, then started to decrease due to the competitive HER reaction. Intriguingly, the Ni-N-C_II with 1.3 times more Ni sites led to a consistent increase of j_CO by 52% to 74% in the kinetic region of −0.7 ~ −1.0 V vs. RHE. We further calculated the electrochemical specific surface areas from the double-layer capacitance (C_dl). Although the specific surface area of Ni-N-C_II is smaller than that of Ni-N-C_I. However, the electrochemical specific surface area (ECSA) of Ni-N-C_I and Ni-N-C_II was very close with the value of 0.50 and 0.55 mF/cm², respectively (Fig. S13 in Supporting information), which is probably due to the presence of higher density of active sites on the Ni-N-C_II. We also compared the TOF of different model carbon materials at different potential. The high-density of coordinated Ni-N₃C centers led to the superior TOF (Fig. S14 in Supporting information). For instance, Ni-N-C_II shows the highest TOF of 674.4 h⁻¹ at −1.0 V vs. RHE.

  The results revealed that transforming the N-doped carbon layer into the second Ni-N₃C layer effectively enhanced the reduction activity of CO₂-to-CO. The layer-by-layer enriching strategy enabled a spatial dispersion of the atomically dispersed Ni-species, and thus avoided its agglomeration to bigger nanoparticles. It could further improve the spatially enriching efficiency if the N-doped layer can be engineered for efficiently anchoring Ni-sites. Alternatively, removal of metallic clusters (Ni-N-C_III) also resulted in the enhancement j_CO by 20%−31% in the applied potential range (Fig. 4c).

  For Group 2 catalysts of pure N-doped carbons, the total surface N content in NC_I is 3.58 at%, while on NC_II prepared by replacing the ethylenediamine by dipyridyl, it is 2.4 times higher. Interestingly, the 2.4 times higher N content led to a marked enhancement on the catalytic performance. Over NC_II, nearly 60–200 times higher FE_CO was observed in the potential range of −0.6 ~ −1.2 V vs. RHE as compared to NC_I (Fig. 4b). The detailed N 1s spectra revealed that the pyridinic-N and pyrrolic-N in NC_II was most increased by 170% and 178%, respectively (Tables S2 and S3). The largest increment of pyridinic-N and pyrrolic-N was likely relevant to the enhancement of the selectivity towards CO, and thus their j_CO (Fig. 4d) [54]. Moreover, the j_CO for NC_I was negligible, while the catalyst of NC_II delivered much higher j_CO, reaching the highest point of 5.5 mA/cm² at −1.0 V vs. RHE. Overall, over N-doped carbon catalysts, we find the obvious disproportions between both the FE_CO and j_CO and the increased density of N functionalities. Thus, we speculate that for N-doped catalysts, a threshold of the density of active sites exists, beyond which their activity is enhanced largely.

  Subsequently, we investigated Group 3 catalysts with indirect N coordinated Ni sites (Fig. 4b). For NC-10@Ni-N-C_I with thinner NC coating, its selectivity trend followed that of the original Ni-N-C_I. For NC-120@Ni-N-C_I with thicker NC coating, at lower potential (−0.6 and −0.7 V vs. RHE), its FE_CO was roughly half of that Ni-N-C_I. While the FE_CO reached the same level at −0.8 V and outperformed Ni-N-C_I when the electrode potential was reached −0.9 V vs. RHE and above. For instance, at potentials of −1.0 and −1.1 V vs. RHE, its FE_CO increase by 10% and 25% as compared to Ni-N-C_I. While the j_CO of Group 3 sits well in between Group 1 and Group 2.

  We conducted the in-situ ATR-SEIRAS measurement to analyze the changes of intermediates for Ni-N-C materials during CO₂RR (Fig. 5). In-situ ATR-SEIRAS was performed in the 0.5 mol/L KHCO₃ at the potential range of −0.1 ~ −1.2 V vs. RHE. The small peaks at 1445 and 2086 cm⁻¹ could be assigned to *COOH and *CO, which are the key intermediates for CO₂ reduction to CO [55]. Besides, as the overpotential increasing, the peak of CO₃²⁻ at 1395 cm⁻¹ was found to be more intense. This result suggested that the hydrogen evolution reaction (HER) of Ni-N-C_I with cluster is more severe compared to Ni-N-C_III, leading to the increased consumption of H⁺. As a result, an excess of OH⁻ is produced and combines with CO₂, which causes the saturation buffer equilibrium of CO₂ towards the formation of CO₃²⁻ [56]. The above results also provide evidences for confirming the Ni-N₃C active sites for CO₂ reduction to CO.

  Figure 5

  

  Figure 5.  In-situ ATR-SEIRAS spectra of (a) Ni-N-C_I and (b) Ni-N-C_III. The range from 2110 cm⁻¹ to 2020 cm⁻¹ is highlighted.

  DownLoad: Full-Size Img PowerPoint

  To get in-depth understanding in the activity trend and N- and Ni-containing functionalities' contribution to the current density, we correlated different active sites with their FE_CO and S_BET normalized j_CO (Fig. 6). In this section, the j_CO refers to the normalized j_CO values normalized by S_BET. The FE_CO in our work reflects the nature of active sites for CO₂-to-CO, while the j_CO is the indicator of the density of active sites contributing to the current for CO₂-to-CO intrinsically. First, we correlated the FE_CO and j_CO of Group 1 and Group 3. Overall, the catalysts of Group 1 with direct Ni-N₃C sites showed higher and constant FE_CO than that of Group 3 with indirect Ni-N₃C sites. At −0.7 V vs. RHE, Group 1 displayed relatively constant FE_CO at ca. 80%. For Group 3, a thinner NC coating was rather dominated by the Ni-N₃C sites, and thus delivered a very close FE_CO to that of direct Ni-N₃C sites. While, the one with thicker NC coatings showed the most indirect Ni-N₃C sites, exhibiting only half of FE_CO. We estimated their contribution to the j_CO by calculating their slopes of the linear fitting. The direct Ni-N₃C sites were observed to contribute 14.2 times higher than the indirect Ni-N₃C sites.

  Figure 6

  

  Figure 6.  The correlation (top) between key structural parameters and core performance indexes on this family of functional carbon-based model catalysts (below).

  DownLoad: Full-Size Img PowerPoint

  At middle potential (−0.9 V vs. RHE), both the direct and indirect Ni-N₃C sites led to a ca. 80% FE_CO, indicating a similar overall activity for Group 3. Surprisingly, the slope of Group 3 was even higher than that of Group 1. We speculated that the NC coating created a confined effect for the covered Ni-N₃C sites that enhanced their activity in catalyzing CO₂-to-CO or suppressing the competitive HER.

  At high potential (−1.1 V vs. RHE), the FE_CO decreased for all the catalysts, while the j_CO continued to increase. The reduction of FE_CO was due to the increasingly competitive HER, which attracted half of the transferred electrons. For the linear fitting of j_CO, the slope of Group 1 is 1.5 times higher than that of Group 3. This revealed, at potential of −1.1 V, the direct Ni-N₃C coordinated sites were more catalytically active for CO₂-to-CO than the indirect type.

  Similarly, we then moved to the correlation between the FE_CO and j_CO with the surface atomic Ni sites of Group 1 and Group 3. Overall, the trend was like the above correlation. However, the slope of j_CO plot correlated to the Ni sites was lower as compared to the correlation with the Ni-N₃C sites. This was because of the existence of metallic Ni clusters that were HER active. For instance, at −1.1 V, the slope of j_CO plot was reduced by 66.2% and 41.3% for Group 3 and Group 1, respectively. Finally, the relationship between the FE_CO and j_CO and the typical N-functionality of pyridinic-N was plotted. On Group 3, we did not observe positive effects of the existence of pyridinic-N in contributing CO₂-to-CO. The higher pyridinic-N content, however, resulted in the lower j_CO. In other word, this revealed the dominated role of the Ni-N₃C coordinated sites in CO₂-to-CO for Ni-N-C type catalysts. We can further quantitatively analyze their contributions by comparing their corresponding slopes. A S1/S2 of 116.67, 34.48 and 80.53 was obtained at −0.7, −0.9 and −1.1 V vs. RHE, respectively. This provided clear clues on the understanding of their activity, and a valuable method for creating high density active sites to enhance the current density for CO₂-to-CO.

  A large set of control experiments combined with aberration-corrected HAADF-STEM images, XAS characterization as well as the in-situ ATR-SEIRAS spectra were applied to identify dispersed Ni-N₃C sites that are likely the active centers for CO₂ reduction to CO. For instance, increasing leaching cycles promoted activity by exposing more coordinated Ni-N₃ sites. This observation was consistent with previous studies on Ni-N-C materials [57]. In addition, quantification of Ni species showed that 74% of the sample consisted of single Ni-coordinated N sites, while 26% were metallic Ni clusters for Ni-N-C_I. The correlation between CO selectivity and metal entities revealed that FE_CO was 10%−15% lower than that of samples dominated by single Ni sites, supporting the proposed N-coordinated Ni active sites [58,59].

  Besides, considering the oxidized carbon sheet as the support to fabricate this series of electrocatalysts, we also carefully analyze the effects of O atom on catalytic performance. The previous literatures reported that the axial O atoms in coordinated Ni-N₄-O structures can induce an axial traction effect, which promote the CO₂ reduction [60]. However, in principle, it is unlikely to form the axial O atoms by introducing the nickel nitrate and ethylenediamine complex into the support in our work. Anther evidence is that the O contents of this family of electrocatalyst are rather close. Only after introducing the metal species, the remarkably increased CO partial current density can be observed. These results suggested that the presence of O atoms should not play evident role in CO₂RR.

  In summary, a family of Ni and N functionalized carbon electrocatalysts were developed through a nickel-diamine complex-assisted nanoengineering method. By correlating the structural features and their CO₂RR performances, we explored the roles of specific functionalities which were responsible for catalyzing CO₂-to-CO conversion. A tiny amount of surface Ni-N₃C active species significantly enhance of the CO partial current density, which was nearly 50 times higher than that of the N-doped sites. Technically important, the increase of j_CO was achieved by a layer-by-layer enriching strategy, in which the density of Ni-N₃C coordinated sites was increased nearly linearly. Our results may inspire the new design and fabrication of advanced nonprecious metal catalysts for other heterogeneous reactions beyond the target in this work.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  Acknowledgments

  This work was financed by National Natural Science Foundation of China (Nos. 22275027, 21975037), the Fundamental Research Funds for the Central Universities (Nos. DUT22LAB607, DUT22QN206). We also thank Prof. Si from Chinese Academy of Sciences for his support in the collection and analysis of the XAS data.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108634.

  
  
• 1. [1]

     M. Rahimi, A. Khurram, T.A. Hatton, et al., Chem. Soc. Rev. 51 (2022) 8676–8695. doi: 10.1039/D2CS00443G

  2. [2]

     J. Liu, C. Chen, K. Zhang, et al., Chin. Chem. Lett. 32 (2021) 649–659. doi: 10.1016/j.cclet.2020.07.040

  3. [3]

     S. Jin, Z. Hao, K. Zhang, et al., Angew. Chem. Int. Ed. 60 (2021) 20627–20648. doi: 10.1002/anie.202101818

  4. [4]

     Z. Gao, J. Li, Z. Zhang, et al., Chin. Chem. Lett. 33 (2022) 2270–2280. doi: 10.1016/j.cclet.2021.09.037

  5. [5]

     T. He, A.R. Puente-Santiago, S. Xia, et al., Adv. Energy Mater. 12 (2022) 2200493. doi: 10.1002/aenm.202200493

  6. [6]

     B. Zhang, Y. Jiang, M. Gao, et al., Nano Energy 80 (2021) 105504. doi: 10.1016/j.nanoen.2020.105504

  7. [7]

     F. Wang, W. Zhang, H. Wan, et al., Chin. Chem. Lett. 33 (2022) 2259–2269. doi: 10.1016/j.cclet.2021.08.074

  8. [8]

     S. Li, A.V. Nagarajan, D.R. Alfonso, et al., Angew. Chem. Int. Ed. 60 (2021) 6351–6356. doi: 10.1002/anie.202016129

  9. [9]

     L. Wei, H. Li, J. Chen, et al., ACS Catal. 10 (2020) 1444–1453. doi: 10.1021/acscatal.9b04633

  10. [10]

      J. Zhao, Z. Chen, J. Zhao, J. Mater. Chem. A 7 (2019) 4026–4035. doi: 10.1039/C8TA11825F

  11. [11]

      P. Yao, Y. Qiu, T. Zhang, et al., ACS Sustain. Chem. Eng. 7 (2019) 5249–5255. doi: 10.1021/acssuschemeng.8b06160

  12. [12]

      C. Li, Y. Wang, N. Xiao, et al., Carbon 151 (2019) 46–52. doi: 10.1016/j.carbon.2019.05.042

  13. [13]

      T. Liu, S. Ali, Z. Lian, et al., J. Mater. Chem. A 6 (2018) 19998–20004. doi: 10.1039/C8TA06649C

  14. [14]

      F. Pan, B. Li, W. Deng, et al., Appl. Catal. B: Environ. 252 (2019) 240–249. doi: 10.1016/j.apcatb.2019.04.025

  15. [15]

      J. Gu, C. Hsu, L. Bai, et al., Science 364 (2019) 1091–1094. doi: 10.1126/science.aaw7515

  16. [16]

      L. Dong, X. Hu, Y. Du, et al., J. Mater. Chem. A 9 (2021) 17821–17829. doi: 10.1039/D1TA05355H

  17. [17]

      W. Ju, A. Bagger, G. Hao, et al., Nat. Commun. 8 (2017) 944. doi: 10.1038/s41467-017-01035-z

  18. [18]

      P. Wang, T. Li, Q. Wu, et al., ACS Nano 16 (2022) 17021–17032. doi: 10.1021/acsnano.2c07138

  19. [19]

      P. Su, K. Iwase, S. Nakanishi, et al., Small 12 (2016) 6083–6089. doi: 10.1002/smll.201602158

  20. [20]

      W. Wang, L. Shang, G. Chang, et al., Adv. Mater. 31 (2019) 1808276. doi: 10.1002/adma.201808276

  21. [21]

      S. Shen, C. Han, B. Wang, et al., Chin. Chem. Lett. 33 (2022) 3721–3725. doi: 10.1016/j.cclet.2021.10.063

  22. [22]

      H. Cheng, X. Wu, X. Li, et al., Chem. Eng. J. 407 (2021) 126842. doi: 10.1016/j.cej.2020.126842

  23. [23]

      B. Lu, Q. Liu, S. Chen, ACS Catal. 10 (2020) 7584–7618. doi: 10.1021/acscatal.0c01950

  24. [24]

      H. Li, K. Gan, R. Li, et al., Adv. Funct. Mater. 33 (2023) 2208622. doi: 10.1002/adfm.202208622

  25. [25]

      Y. Li, N.M. Adli, W. Shan, et al., Energy Environ. Sci. 15 (2022) 2108–2119. doi: 10.1039/D2EE00318J

  26. [26]

      Y. Lu, H. Wang, P. Yu, et al., Nano Energy 77 (2020) 105158. doi: 10.1016/j.nanoen.2020.105158

  27. [27]

      F. Pan, H. Zhang, K. Liu, et al., ACS Catal. 8 (2018) 3116–3122. doi: 10.1021/acscatal.8b00398

  28. [28]

      X. Zhao, Y. Liu, J. Am. Chem. Soc. 142 (2020) 5773–5777. doi: 10.1021/jacs.9b13872

  29. [29]

      X. Fu, P. Zhang, T. Sun, et al., Small 18 (2022) 2107997. doi: 10.1002/smll.202107997

  30. [30]

      X. Li, W. Bi, M. Chen, et al., J. Am. Chem. Soc. 139 (2017) 14889–14892. doi: 10.1021/jacs.7b09074

  31. [31]

      D.M. Koshy, A.T. Landers, D.A. Cullen, et al., Adv. Energy Mater. 10 (2020) 2001836. doi: 10.1002/aenm.202001836

  32. [32]

      Y. Gong, L. Jiao, Y. Qian, et al., Angew. Chem. Int. Ed. 59 (2020) 2705–2709. doi: 10.1002/anie.201914977

  33. [33]

      Z. Liu, S. Li, J. Yang, et al., ACS Nano 14 (2020) 11662–11669. doi: 10.1021/acsnano.0c04210

  34. [34]

      Y. Qu, L. Wang, Z. Li, et al., Adv. Mater. 31 (2019) 1970316. doi: 10.1002/adma.201970316

  35. [35]

      Z. Miao, X. Wang, M. Tsai, et al., Adv. Energy Mater. 8 (2018) 1801226. doi: 10.1002/aenm.201801226

  36. [36]

      S. Liu, M. Wang, X. Yang, et al., Angew. Chem. Int. Ed. 59 (2020) 21698–21705. doi: 10.1002/anie.202009331

  37. [37]

      D. De Fazio, D.G. Purdie, A.K. Ott, et al., ACS Nano 13 (2019) 8926–8935. doi: 10.1021/acsnano.9b02621

  38. [38]

      K. Huang, Z. Wei, J. Liu, et al., Small 18 (2022) 2201139. doi: 10.1002/smll.202201139

  39. [39]

      W. Bi, X. Li, R. You, et al., Adv. Mater. 30 (2018) 1706617. doi: 10.1002/adma.201706617

  40. [40]

      C. Yan, H. Li, Y. Ye, et al., Energy Environ. Sci. 11 (2018) 1204–1210. doi: 10.1039/C8EE00133B

  41. [41]

      Q. Fan, P. Hou, C. Choi, et al., Adv. Energy Mater. 10 (2020) 1903068. doi: 10.1002/aenm.201903068

  42. [42]

      K. Mou, Z. Chen, X. Zhang, et al., Small 15 (2019) 1903668. doi: 10.1002/smll.201903668

  43. [43]

      R. Daiyan, X. Zhu, Z. Tong, et al., Nano Energy 78 (2020) 105213. doi: 10.1016/j.nanoen.2020.105213

  44. [44]

      C. Zhao, X. Dai, T. Yao, et al., J. Am. Chem. Soc. 139 (2017) 8078–8081. doi: 10.1021/jacs.7b02736

  45. [45]

      S. Gong, W. Wang, R. Lu, et al., Appl. Catal. B: Environ. 318 (2022) 121813. doi: 10.1016/j.apcatb.2022.121813

  46. [46]

      R. Ge, L. Dong, X. Hu, et al., Chem. Eng. J. 438 (2022) 135500. doi: 10.1016/j.cej.2022.135500

  47. [47]

      C. Wen, F. Mao, Y. Liu, et al., ACS Catal. 10 (2020) 1086–1093. doi: 10.1021/acscatal.9b02978

  48. [48]

      P. Prslja, N. López, ACS Catal. 11 (2021) 88–94. doi: 10.1021/acscatal.0c01909

  49. [49]

      T. Asset, S.T. Garcia, S. Herrera, et al., ACS Catal. 9 (2019) 7668–7678. doi: 10.1021/acscatal.9b01513

  50. [50]

      R. Daiyan, W.H. Saputera, H. Masood, et al., Adv. Energy Mater. 10 (2020) 1902106. doi: 10.1002/aenm.201902106

  51. [51]

      T. Zheng, K. Jiang, N. Ta, et al., Joule 3 (2019) 265–278. doi: 10.1016/j.joule.2018.10.015

  52. [52]

      C. Lu, J. Yang, S. Wei, et al., Adv. Funct. Mater. 29 (2019) 1806884. doi: 10.1002/adfm.201806884

  53. [53]

      B. Chen, B. Li, Z. Tian, et al., Adv. Energy Mater. 11 (2021) 2102152. doi: 10.1002/aenm.202102152

  54. [54]

      R.M. Yadav, Z. Li, T. Zhang, et al., Adv. Mater. 34 (2022) 2105690. doi: 10.1002/adma.202105690

  55. [55]

      M. Huang, B. Deng, X. Zhao, ACS Nano 16 (2022) 2110–2119. doi: 10.1021/acsnano.1c07746

  56. [56]

      C. Liu, Y. Wu, K. Sun, et al., Chem 7 (2021) 1297–1307. doi: 10.1016/j.chempr.2021.02.001

  57. [57]

      A.S. Varela, N. Ranjbar Sahraie, J. Steinberg, et al., Angew. Chem. Int. Ed. 54 (2015) 10758–10762. doi: 10.1002/anie.201502099

  58. [58]

      D.M. Koshy, S. Chen, D.U. Lee, et al., Angew. Chem. Int. Ed. 59 (2020) 4043–4050. doi: 10.1002/anie.201912857

  59. [59]

      Z. Li, D. He, X. Yan, et al., Angew. Chem. Int. Ed. 59 (2020) 18572–18577. doi: 10.1002/anie.202000318

  60. [60]

      X. Wang, Y. Wang, X. Sang, et al., Angew. Chem. Int. Ed. 60 (2021) 4192–4198. doi: 10.1002/anie.202013427

• 

  Figure 1  Schematic illustration of the designed model catalysts based on nanoengineering towards targeted Ni and N-containing functionalities.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) Typical procedures for the synthesis of Ni-N-C_I catalyst. (b) HAADF-STEM image and the EELS mapping of the Ni-N-C_I. (c–g) Aberration-corrected HAADF-STEM images of Ni-N-C_I under different magnifications.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Electronic states of Ni atom and chemical state of surface N-functionalities in this series of catalysts. (a) Ni K-edge XANES profiles of Ni-N-C_I, Ni-N-C_II as well as Ni foil and NiO benchmarks, inset shows the expanded pre-edge region. (b) Corresponding Fourier transforms spectra. (c) Fittings results of Ni-N-C_I and Ni-N-C_II. (d) XPS spectra comparison in N 1s region.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  CO₂RR evaluation results on NC, Ni-N-C and NC layer coated Ni-N-C counterparts. (a, b) FE_CO at varied applied potentials in the H-cell, E vs. RHE. (c, d) Corresponding CO partial current density (j_CO) for this family of model catalysts.

  下载: 全尺寸图片 幻灯片
  

  Figure 5  In-situ ATR-SEIRAS spectra of (a) Ni-N-C_I and (b) Ni-N-C_III. The range from 2110 cm⁻¹ to 2020 cm⁻¹ is highlighted.

  下载: 全尺寸图片 幻灯片
  

  Figure 6  The correlation (top) between key structural parameters and core performance indexes on this family of functional carbon-based model catalysts (below).

  下载: 全尺寸图片 幻灯片

  Table 1.  Elemental composition in the surface and bulk region of this family of model catalysts.

  下载: 导出CSV
•