Atomically Dispersed Zinc Active Sites Efficiently Promote the Electrochemical Conversion of N2 to NH3

At present, the research on highly active and stable nitrogen reduction reaction catalysts is still challenging work for the electrosynthesis of ammonia (NH3). Herein, we synthesized atomically dispersed zinc active sites supported on N‐doped carbon nanosheets (Zn/NC NSs) as an efficient nitrogen reduction reaction catalyst, which achieves a high ammonia yield of 46.62 μg h−1 mg−1cat. at −0.85 V (vs RHE) and Faradaic efficiency of 95.8% at −0.70 V (vs RHE). In addition, Zn/NC NSs present great stability and selectivity, and there is no significant change in NH3 rate and Faradaic efficiencies after multiple cycles. The structural characterization shows that the active center in the nitrogen reduction reaction process is the Zn–N4 sites in the catalyst. DFT calculation confirms that Zn/NC with Zn–N4 configuration has a lower energy barrier for the formation of *NNH intermediate compared with pure N‐doped carbon nanosheets (N‐C NSs), thus promoting the hydrogenation kinetics in the whole nitrogen reduction reaction process.


Introduction
[3] At present, the industrial production of ammonia is usually carried out by the Haber-Bosch method, but the reaction conditions of the Haber-Bosch method are extremely harsh, [4][5][6] requiring hightemperature and high-pressure environment (350-550 °C, 150-350 bar), and produce huge energy consumption and CO 2 emissions. [7,8]herefore, an ammonia synthesis strategy under mild conditions is urgently needed to meet the needs of sustainable development and green production. [9]Electrochemical nitrogen reduction reaction (eNRR), as one of the important methods to synthesize NH 3 under environmental conditions, has received extensive attention in recent years. [10,11]Driven by renewable electricity, NH 3 can be directly synthesized through eNRR using N 2 and water (N 2 + 3H 2 O → 2NH 3 + 3/ 2 O 2 ) at normal temperature and pressure.The method has low energy consumption and does not generate greenhouse gases, so it has a wide development prospect. [12]However, the largescale application of eNRR has some challenges due to the high breaking energy of the N≡N triple bond (941 kJ mol −1 ). [13]Besides, there will be a competitive hydrogen evolution reaction (HER) in the reaction process, which will lower the selectivity of NRR. [14,15]Therefore, it is crucial to design new nitrogen reduction electrocatalysts with great catalytic activity and selectivity.
So far, noble metal-based catalysts (such as Ru, [16] Au, [8] etc.) have been extensively studied, the results show their excellent catalytic activity for NRR.Nevertheless, the high cost and scarcity of precious metals limit their widespread application. [17,18]herefore, current research is devoted to exploring low-cost and highefficiency nonnoble metal catalysts to promote the conversion of N 2 to NH 3 .21][22] The synthesis mechanism of single-atom catalysts (SACs) is due to the interaction between metal and carrier, the metal particles and adjacent atoms will form chemical bonds so that the particles can be anchored on the matrix, and there is charge transfer and strong electronic coupling between the metal sites and surrounding elements. [23,24]Therefore, the coordination configuration and the number of the central metal have a great influence on the catalytic activity of the catalyst. [25]he SACs have excellent catalytic activity, which is because almost all metal particles in SACs can be exposed to the surface to participate in the reaction, and the atom utilization rate is significantly improved compared with metal clusters. [26,27]In addition, the single atom usually has a positive charge, and the adsorption of reactants on SACs is relatively mild, [16] that is, the energy barrier required for the activation of N 2 adsorbed on SACs catalyst is lower, which makes NRR easier to occur. [28]Some studies have pointed out that SACs can effectively inhibit HER due to the ensemble effect, the selectivity for NRR is significantly higher than that of bulk metal surfaces. [29]Moreover, the "top" binding of nitrogen on single metal sites can increase the rate of ammonia synthesis. [30]Furthermore, in the presence of monometallic Lewis acid ions, nitrogen species in the catalyst material can promote N 2 dissociation. [31]In general, SACs have excellent NRR performance due to their unique structure.
Herein, we synthesized atomically dispersed zinc active sites supported on N-doped carbon nanosheets (Zn/NC NSs) as nitrogen DOI: 10.1002/eem2.12630 At present, the research on highly active and stable nitrogen reduction reaction catalysts is still challenging work for the electrosynthesis of ammonia (NH 3 ).Herein, we synthesized atomically dispersed zinc active sites supported on N-doped carbon nanosheets (Zn/NC NSs) as an efficient nitrogen reduction reaction catalyst, which achieves a high ammonia yield of 46.62 μg h −1 mg −1 cat. at −0.85 V (vs RHE) and Faradaic efficiency of 95.8% at −0.70 V (vs RHE).In addition, Zn/NC NSs present great stability and selectivity, and there is no significant change in NH 3 rate and Faradaic efficiencies after multiple cycles.The structural characterization shows that the active center in the nitrogen reduction reaction process is the Zn-N 4 sites in the catalyst.DFT calculation confirms that Zn/NC with Zn-N 4 configuration has a lower energy barrier for the formation of *NNH intermediate compared with pure N-doped carbon nanosheets (N-C NSs), thus promoting the hydrogenation kinetics in the whole nitrogen reduction reaction process.
reduction catalysts by molten salt-assisted solid-phase method.This method uses metal-organic frameworks (MOFs) as the precursor for pyrolysis.At high temperatures, it can form a carbon matrix rich in defects and make metal atoms diffuse out.By enhancing the charge transfer between individual atoms and defect sites, it can capture and stabilize isolated metal atoms.[34] In 0.1 M Na 2 SO 4 electrolyte, Zn/NC NSs exhibit good NRR activity with an average ammonia yield of 46.62 μg h −1 mg −1 cat. at −0.85 V (vs RHE) and Faradaic efficiency (FE) of 95.8% at −0.70 V (vs RHE).At the same time, it should be noted that Zn/NC NSs also showed good selectivity and stability.As revealed by the DFT calculation, the excellent NRR performance can be owing to the reduction of the energy barrier of *N 2 activation.

Catalyst Synthesis, Structure, and Morphology
As shown in Figure 1, the single-atom NRR catalysts (Zn/NC NSs) composed of nitrogen-doped carbon nanosheets and isolated Zn atoms are synthesized by molten salt-assisted solid-phase synthesis.NaCl, 2meylimidazole, and ZnO were weighed and ground in a mortar to form a uniform mixture, and then seal them in a Teflonlined reactor.39] After washing with deionized water to remove excess NaCl, the final product was got, denoted Zn/NC NSs.Meanwhile, NaCl was not added to synthesize atomically dispersed Zn loaded on bulk N-doped carbon matrix (bulk Zn/NC), and N-doped carbon nanosheets (N-C NSs) were synthesized without ZnO as the reference samples.Obviously, compared with bulk Zn/NC, Zn/NC NSs have a larger electrochemical surface area, so Zn/NC NSs may have more active sites to obtain higher ammonia yield.
To optimize the annealing temperature, we synthesized the catalysts at different annealing temperatures, which are recorded as Zn/NC NSs-T (T is the annealing temperature).Field emission scanning electron microscopy (FE-SEM) reveals the typical lamellar morphology of the Zn/NC NSs catalyst (Figure 2a).The wrinkled nanosheet morphology is also determined by transmission electron microscopy (TEM) images (Figure S1, Supporting Information).The FE-SEM images of bulk Zn/NC, N-C NSs, and Zn/NC NSs-T are presented in Figure S2, Supporting Information.For bulk Zn/NC (Figure S2a, Supporting Information), it presents the bulk structure.As shown in Figure 2a and Figure S2b-d, Supporting Information, the morphologies of N-C NSs and Zn/NC NSs annealed at different annealing temperatures all present the lamellar structures.The energy dispersive spectroscopy (EDS) mapping images of Zn/ NC NSs show a uniform distribution of Zn, N, and C elements throughout the nanosheet structure (Figure 2b), further demonstrating the successful preparation of the samples.The X-ray diffraction (XRD)  patterns of the Zn/NC NSs (Figure 2c) show the features of partial graphitization and prove the absence of Zn clusters. [40]The XRD patterns of Zn/NC NSs and bulk Zn/NC slightly shift compared with the N-C NSs, which may be due to the partial lattice distortion caused by the incorporation of Zn atoms.The total Zn content in the Zn/NC NSs is determined to be 4.87 wt.% by inductively coupled plasma optical emission spectroscopy (ICP-OES).As shown in Figure 2c and Figure S3, Supporting Information, Zn/NC NSs annealed at different annealing temperatures all have the characteristics of partial graphitization.Moreover, in the aberration-corrected scanning transmission electron microscope (AC-STEM) image (Figure 2d), the bright spots with an atomic diameter of ~0.2 nm reveal that the zinc species are in the form of atomically dispersed zinc atoms rather than in clusters. [41,42]o further reveal the chemical structure of Zn species in Zn/NC NSs, we performed X-ray photoelectron spectroscopy (XPS) analysis and Xray absorption spectroscopy (XAS) analysis.The survey XPS spectrum of Zn/NC NSs (Figure 3a) confirms the presence of Zn, C, and N elements.For the N 1s spectrum (Figure 3b), these peaks are ascribed to oxidized N (403.3eV), graphitic N (N1: 401.8 eV), pyrrolic N (N2: 401.0 eV), Zn-N (N3: 399.9 eV), and pyridinic N (N4: 398.5 eV), respectively. [42]It can be seen from Figure 3b that although the nitrogen-doped carbon matrix decomposes to form a certain amount of nitrogen-containing compounds during the synthesis process, the relative number of Zn-N bonds is still high (~22%).For the C 1s spectrum (Figure S4, Supporting Information), five distinct peaks at ~284.8, ~285.9, ~286.5, ~288.11, and ~290.8 eV are attributed to C=C, C-N, C-O, C=O, and graphite C bonds, respectively. [42,43]Zn 2p XPS spectrum further confirms the existence of Zn atoms in Zn/NC NSs.As shown in Figure 3c, the two main peaks located at ~1021.9 and ~1045.0eV are assigned to Zn 2p 3/2 and Zn 2p 1/2 of Zn atoms, respectively.According to the XPS electronic binding energy comparison table, the valence state of Zn element can be determined as +2. [44]igure 3d-f shows the X-ray absorption spectra (XAS) of Zn/NC NSs, with Zn foil, ZnO, and zinc phthalocyanine (ZnPc) as the test standards.From the normalized X-ray absorption close edge structure (XANES) (Figure 3d), the front edge of Zn/NC NSs is characterized between the ZnO and ZnPc, which indicates that the valence of Zn species in Zn/NC NSs is +2.It is consistent with the above XPS results.The Fourier transform (FT) k 2 -weighted extended X-ray absorption fine structure (FT-EXAFS) of Zn/NC NSs is shown in Figure 3e.Only one peak at 1.7 Å appears, which is attributed to the Zn-N coordination.Compared with Zn foil, Zn/NC NSs have no Zn-Zn coordination peak at 2.3 Å, further confirming the formation of the single Zn-N sites, [45] which is consistent with the spherical aberration-corrected HAADF-STEM results (Figure 2d).Furthermore, the chemical coordination environment of Zn/NC NSs is obtained using EXAFS fitting (Figure 3f).Through the fitting results in R space, it can be determined that the Zn-N coordination number of Zn/NC NSs is 4.2 (Table S1, Supporting Information).It is shown that the central Zn atom directly coordinates with the four N atoms and the average Zn-N bond length is 1.99 Å. [46][47][48] Additionally, to determine the coordination environment of Zn/NC NSs more accurately, we performed wavelet transform (WT) analysis on K-space (Figure 3g).The results show that there is only a Zn-N bond signal near 5 Å−1 in Zn/NC NSs, but no χ (k) signal is detected at ~7.3 Å−1 , which fully proves that there is no Zn-Zn bond in Zn/NC NSs.

Electrochemical NRR Performance Tests
All electrochemical experiments are performed at room temperature in an H-type three-electrode cell system using a CHI660E electrochemical workstation.The cathode chamber and anode chamber are separated by a proton exchange membrane (Nafion 117).As shown in the previous studies, [49] pretreated Nafion may cause ammonia contamination.To exclude this effect, we carry out a blank control experiment to exclude other sources of N, thus proving that N in NH þ 4 generated by subsequent electrolysis is from N 2 .In this experiment, the glassy carbon electrode, Ag/AgCl electrode, and graphite electrode are used as working, reference, and counter electrodes, respectively.All reported potentials in this work are converted to RHE.The electrolyte used in this experiment is 0.1 M Na 2 SO 4 .To reduce the error of the experiment, it is necessary to check the air tightness of the electrolytic cell before the experiment and purge the electrolyte with N 2 for 30 min to reduce the influence of the external environment.And an acid trap is installed at the other outlet of the electrolytic cell to prevent ammonia from escaping.The NH 3 produced by electrolysis is detected by the indoxyl blue method, and the produced by-product N 2 H 4 is measured by the method of Watt and Chrisp. [50]Before starting the test, the electrocatalyst is activated by cyclic voltammetry for 50 cycles (0 to −1 V) with a scan rate of 100 mV −1 .
To determine the optimal potential for electrocatalyst materials, we perform linear sweep voltammetry (LSV) experiments.As shown in Figure 4a, LSV tests are performed on the Zn/NC NSs catalyst materials in N 2 -saturated and Ar-saturated 0.1 M Na 2 SO 4 solutions, respectively.The sweep range is 0 to −1.5 V and the sweep rate is 5 mV −1 .It can be seen that the Zn/NC NSs catalyst presents higher current densities under N 2 -saturated conditions compared with Ar-saturated atmospheres, which proves the NRR activity of the catalyst over such a wide potential range.In addition, to discover the intrinsic active sites of Zn/ NC NSs, we implant thiocyanate ions (SCN − ) into N 2 -saturated 0.1 M Na 2 SO 4 electrolyte during LSV.The SCN − can act as an inhibitor that blocks metal-N sites, so it can block the Zn-N sites of Zn/NC NSs and make them ineffective. [51]As can be seen in Figure 4a, the current density of Zn/NC NSs dropped significantly after adding SCN − .It confirms that Zn-N sites are NRR active centers on Zn/NC NSs.
Next, we systematically investigate the electrocatalytic performance of Zn/NC NSs in the potential range of −0.75 to −0.95 V.The corresponding calibration curves of NH þ 4 and N 2 H 4 are made by the absorbance of the gradient solution (Figure S5, Supporting Information), which is used to correspond to the NH þ 4 and N 2 H 4 concentrations obtained after electrolysis.Figure 4b shows the absorbance curves after electrolysis at different potentials for 2 h, and the average ammonia yield and Faradaic efficiencies (FEs) within 2 h are calculated (Figure 4c).For atomically dispersed Zn/NC NSs, the ammonia yield shows a volcano-like trend in the range of −0.7 to −0.95 V, reaching a maximum value of 46.62 μg h −1 mg −1 cat. at −0.85 V.The FEs continue to decrease with the increase of the potential.The FE at −0.70 V is up to 95.8%, and it still has a high FE of 20.7% at −0.85 V. Figure S6, Supporting Information proves that the Zn/NC NSs synthesized by annealing at 900 °C present the best NRR performance.Compared with bulk Zn/NC and N-C NSs, the ammonia production rate of atomically dispersed Zn/NC NSs is significantly improved (Figure S7, Supporting Information).The comparison of NRR performance between our prepared Zn/NC-NSs catalyst and other single-atom catalysts is shown in Table 1.The obtained results show that our prepared atomically dispersed Zn/NC NSs are one of the best NRR electrocatalysts (Table 1).
By measuring the absorbance of hydrazine in the electrolyte before and after NRR (Figure S8, Supporting Information), we find that no hydrazine is produced during the electrolysis process, which indicates that the Zn/NC NSs composite has excellent selectivity.Furthermore, considering that there may be a small amount of NO x or NH 3 in the air, we measure the NH 3 content in 0.1 M Na 2 SO 4 after 30 min of N 2 bubbling before the NRR test (Figure S9, Supporting Information).There is no NH 3 found, suggesting that the NRR results are not interfered with by possible impurities.Moreover, to demonstrate that NH 3 in the cathodic compartment is all produced by NRR, we perform electrolysis in Ar-saturated electrolyte at −0.85 V and in N 2 -saturated electrolyte at open-circuit potential for 2 h, respectively.From the UV-Vis spectra under these two test conditions (Figure S10a,b, Supporting Information), it can be seen that NH 3 is not produced during the experiment, which indicates that all the NH 3 produced in the previous experiments came from the NRR reaction.
To further prove the source of NH 3 , we used 14 N 2 and 15 N 2 as the feed gas, respectively, and conducted a 1 H NMR test on the product after 2 h of eNRR.As shown in Figure 4d, the product 14 NH þ 4 using 14 N 2 as feed gas shows triple coupling, while 15 NH þ 4 shows double coupling.It fully proves that the generated NH 3 comes from the N 2 introduced during the experiment.
Furthermore, we also study the stability of the catalysts by cycling experiments and chronoamperometry, which is an important indicator for evaluating the property of catalysts.Since atomically dispersed Zn/NC NSs have the highest ammonia yield at −0.85 V, the research on catalyst durability is carried out at this potential.In the cycling experiment, the electrolyte is replaced every 2 h without replacing the electrode and Nafion membrane, and the concentration of NH 3 in the electrolyte is detected.After nine cycles, neither the ammonia yield nor the FE of the catalyst changes significantly (Figure 4e), indicating the good electrochemical stability of the 13.27 [52]   Co-SAs/NC-L 0.005 M H 2 SO 4 16.9 μg h −1 mg −1 cat.

Figure 1 .
Figure 1.Schematic diagram of the preparation process for Zn/NC NSs.

Figure 2 .
Figure 2. a) FE-SEM images of Zn/NC NSs; b) HAADF-TEM and EDS mapping images of Zn/NC NSs; c) XRD patterns of bulk Zn/NC, N-C NSs and Zn/NC NSs; d) Aberration-corrected HAADF-STEM images of Zn/NC NSs (The atom-level dispersed Zn species are highlighted in yellow circles).

Figure 3 .
Figure 3. a) Survey, b) N 1s, and c) Zn 2p XPS spectra of Zn/NC NSs; d) XANES spectra of Zn K-edge for Zn/NC NSs with Zn foil, ZnO, and ZnPc; e) The k 2 -weighted χ (k) function of the EXAFS spectra for Zn/NC NSs with Zn foil, ZnO, and ZnPc; f) Fourier transform-EXAFS fitting results of Zn/NC NSs, inset: the atomic structure model of the Zn-N 4 site in Zn/NC NSs; and g) Wavelet transform-EXAFS spectra of Zn/NC NSs, ZnO, Zn foil, and ZnPc.

Figure 4 .
Figure 4. a) Linear sweep voltammetry (LSV) curves of Zn/NC NSs in N 2 -saturated and Ar-saturated 0.1 M Na 2 SO 4 ; b) UV-Vis absorption spectra of catholyte saturated with N 2 after 2 h of electrolysis at a series of potentials; c) NH 3 yield rates and FE of Zn/NC NSs at different potentials; d) 1 H NMR spectrum after eNRR using 15 N 2 and 14 N 2 as feed gas; e) NH 3 yield rate and FE of Zn/NC NSs under nine cycles test at −0.85 V; f) Time-current density curve after 12 h of eNRR at −0.85 V.

Figure 5 .
Figure 5.The schematic diagrams of a) alternate and b) distal pathways of NRR (Blue-N; White-H).c) Gibbs free energy diagram of the NRR alternating pathway on the surface of Zn/NC NSs and N-C NSs.d) Gibbs free energy diagram of the alternate and distal pathways for Zn/NC NSs.

Table 1 .
The NRR performance comparison of Zn/NC NSs and other singleatom catalysts.