Boosting Efficient Ammonia Synthesis over Atomically Dispersed Co-Based Catalyst via the Modulation of Geometric and Electronic Structures

Abstract

Open AccessCCS ChemistryRESEARCH ARTICLE1 May 2022Boosting Efficient Ammonia Synthesis over Atomically Dispersed Co-Based Catalyst via the Modulation of Geometric and Electronic Structures Yanliang Zhou†, Congying Wang†, Xuanbei Peng†, Tianhua Zhang, Xiuyun Wang, Yafei Jiang, Haifeng Qi, Lirong Zheng, Jianxin Lin and Lilong Jiang Yanliang Zhou† National Engineering Research Center of Chemical Fertilizer Catalyst, Fuzhou University, Fuzhou, Fujian 350002 †Y. Zhou, C. Wang, and X. Peng contributed equally to this work.Google Scholar More articles by this author , Congying Wang† National Engineering Research Center of Chemical Fertilizer Catalyst, Fuzhou University, Fuzhou, Fujian 350002 †Y. Zhou, C. Wang, and X. Peng contributed equally to this work.Google Scholar More articles by this author , Xuanbei Peng† National Engineering Research Center of Chemical Fertilizer Catalyst, Fuzhou University, Fuzhou, Fujian 350002 †Y. Zhou, C. Wang, and X. Peng contributed equally to this work.Google Scholar More articles by this author , Tianhua Zhang National Engineering Research Center of Chemical Fertilizer Catalyst, Fuzhou University, Fuzhou, Fujian 350002 Google Scholar More articles by this author , Xiuyun Wang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] National Engineering Research Center of Chemical Fertilizer Catalyst, Fuzhou University, Fuzhou, Fujian 350002 Google Scholar More articles by this author , Yafei Jiang Department of Chemistry, Southern University of Science and Technology, Shenzhen 518055 Google Scholar More articles by this author , Haifeng Qi Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023 Google Scholar More articles by this author , Lirong Zheng Institute of High Energy Physics, Chinese Academy of Sciences, Beijing 100049 Google Scholar More articles by this author , Jianxin Lin National Engineering Research Center of Chemical Fertilizer Catalyst, Fuzhou University, Fuzhou, Fujian 350002 Google Scholar More articles by this author and Lilong Jiang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] National Engineering Research Center of Chemical Fertilizer Catalyst, Fuzhou University, Fuzhou, Fujian 350002 Google Scholar More articles by this author https://doi.org/10.31635/ccschem.021.202100912 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail Ammonia (NH3) synthesis at mild conditions is of great significance, while the significant bottleneck of this process is the activation of N2 to realize the desired NH3 synthesis performance, which requires deep insight and rational design of active sites at the atomic level. Here, were synthesized atomically dispersed Co-based catalysts with different Co-N coordination numbers (CNs) to explore the coordination-sensitive NH3 synthesis reaction for the first time. Our studies showed that Co-based catalysts increased the NH3 synthesis rate gradually with a decrease in CN. The Co-N2 catalyst exhibited the highest NH3 synthesis rate of 85.3 mmol gCo−1 h−1 at 300 °C and 1 MPa, which outperformed most of the previously reported Co-based catalysts. Various characterizations and theoretical calculations demonstrated that atomically dispersed Co catalyst with low CN could generate more unoccupied Co 3d charges and tetrahedral cobalt(II) sites. The unoccupied Co 3d charge, in turn, promoted the electron donation from the Co active center to the antibonding π-orbital (π*) of N2 and expedites N2 hydrogenation. Furthermore, the Co-N2 catalyst with more tetrahedral cobalt(II) sites could effectively facilitate the desorption of N-containing intermediate species (such as *NH3 and *N2H4) to obtain a high NH3 synthesis rate. Download figure Download PowerPoint Introduction Ammonia (NH3) is not only an essential raw material for the synthesis of nitric acid, fertilizer, ammonium salt, and so on but also a promising renewable energy carrier.1,2 The state-of-the-art manufacture of NH3 comes from the Haber–Bosch (HB) process using iron-based catalysts that require harsh reaction conditions (400–500 °C, 10–30 MPa), thereby suffering from high energy consumption and environmental pollution problems.3–5 Therefore, it is urgent to develop effective routes that are both energy-saving and environment-friendly. Most recently, electric power generated from renewable energy sources is widely used for hydrogen (H2) production via water electrolysis. It becomes economically acceptable to NH3 synthesis from renewable H2 and N2 via an electrolysis-driven HB (eHB) process.6 The eHB route is attractive for realizing efficient utilization of renewable energy and a “carbon-free” society. However, the major challenge of NH3 synthesis lies in the activation of N2 (945 kJ mol−1) at mild conditions to realize desired NH3 synthesis rate,7 which would require a deep insight into the reaction mechanism; therefore, the development of advanced, efficient catalysts that can produce NH3 at mild conditions is beneficial. Currently, industrial applied Fe- or Ru-based catalysts that exist as nanoparticles usually obey the dissociative mechanism,8,9 requiring harsh conditions to dissociate the stable N≡N triple bond. In contrast to the N2 dissociative mechanism, most of the enzyme-catalyzed N2 fixation processes followed the associative mechanism, wherein the N≡N triple bond breaks stepwise after partial hydrogenation of the N2 molecule.10,11 Notably, the dissociation energy of the N–N bond (297 kJ/mol) from *N2H4 intermediates is less than a third of that from the N≡N triple bond. Consequently, the exploitation of advanced catalysts that obey the N2 hydrogenation process instead of direct N2 dissociation is promising for realizing the desired NH3 synthesis performance at mild conditions. Further, NH3 synthesis is a typical structure sensitivity reaction, where the pathway of N2 activation is closely related to the geometric and electronic structures of active sites.8,12 For instance, theoretical studies have demonstrated that N2 could preferentially undergo hydrogenation via an associative pathway other than direct dissociation on Ru(0001) surface,13 RhCo3 clusters,12 and Co3Mo3N catalysts.14 In addition, very recently, a limited number of researches have provided us with experimental evidence that the N2 hydrogenation to N2Hx species is preferred over atomically dispersed Ru and Co, as well as Li-promoted Ru catalysts.6,15,16 Among them, atomically dispersed catalysts present well, having the potential for NH3 synthesis at mild conditions, as the direct N2 dissociation is difficult while the N2 hydrogenation is feasible on single-atom sites.15 Meanwhile, atomically dispersed catalysts with definite structures offer the possibility to investigate the geometric and electronic effects on NH3 synthesis at the atomic level. Differing from nanoparticle catalysts, in which the particle size and crystal face usually play an important part in the reactivity, catalytic activity of atomically dispersed catalysts is mainly determined by the coordination environment of single-atom sites.17–19 However, up to now, the modulation of the coordination structure of single-atom active sites to optimize the catalytic activity of NH3 synthesis has not yet been researched. Moreover, the effects of geometric and electronic structures in these single-atom sites on the dynamic transformation of intermediates remain unclear. Herein, we synthesized a series of atomically dispersed Co-Nx (x = 2∼4) catalysts through the pyrolysis of nitrogen-anchored Co-based precursors. The coordination numbers (CNs) of Co-N over the as-prepared catalysts were modulated by varying the calcination temperature. The coordination-sensitive NH3 synthesis reaction in the case of Co-based catalyst is disclosed through a combination of a suite of elaborated experimental characterizations and theoretical calculations. The geometric and electronic structures of Co active sites could be simultaneously tuned via the modulation of Co-N CNs, which are closely related to the N2 activation and desorption of N-containing intermediate species (such as *N2H4 and *NH3), thus resulting in different catalytic behaviors in the NH3 synthesis. This finding is beneficial to rationally design advanced Co-based catalysts by regulating the geometric and electronic structures to realize efficient NH3 synthesis at mild conditions. Experimental Methods Catalyst preparation Synthesis of hollow N-doped porous carbon spheres support The synthetic method of hollow N-doped porous carbon spheres (HNPCSs) support was referenced in previous work.20 Briefly, NH3 solution (10 mL) and ethanol (240 mL) were dissolved in deionized water (80 mL) and stirred evenly at room temperature (RT). Then, tetraethyl orthosilicate (11.2 mL) was added to this solution and stirred further for 1 h. Subsequently, poloxamer, resorcinol, and formaldehyde (2.24 mL) were added to the above solution and continually stirred for 0.5 h. After the addition of melamine (1.26 g) and formaldehyde (1.68 mL), the mixture was treated at 100 °C for 24 h. After filtration and washing, the sample obtained was calcined under an Ar atmosphere at 700 °C for 2 h. Finally, the sample was etched by hydrofluoric acid (HF) solution to remove the SiO2 substrate. After filtration, washing, and drying, the resulting substance was herein referred to as HNPCSs support. Synthesis of Co-Nx and Co-NPs catalysts In detail, HNPCSs (120 mg) and cobalt phthalocyanine (CoPc; 80 mg) was dispersed in dimethylformamide (DMF; 120 mL), respectively, and the mixtures were stirred for 2 h. Then the CoPc and DMF mixed solution was added slowly to the HNPCSs suspension. After stirring for 24 h, the sample was dried in a vacuum dryer at 60 °C for 12 h. Finally, the samples were pyrolyzed under Ar atmosphere at different temperatures of 450, 550, and 650 °C for 2 h, denoted as Co-N4, Co-N3, and Co-N2, respectively. The synthetic procedure used for the Co-NPs catalyst was similar to that of Co-N2, except that the calcination was performed at 700 °C for 10 h. Synthesis of catalysts as reference The Cs-Co/MgO catalyst was synthesized by an impregnation method. In detail, a certain amount of cobalt nitrate hexahydrate and cesium nitrate was dissolved in deionized water. MgO support was added to this solution, and the mixture was stirred evenly. After impregnation at RT for 24 h, the mixture was dried at 80 °C for 24 h. The acquired catalyst was denoted as Cs-Co/MgO, in which the content of Co was the same as that over Co-N2, and the molar ratio of Cs to Co was ∼1.0. A Fe-based catalyst was purchased commercially from Fujian SJ. FD. Co. Ltd. (Fuzhou, Fujian, China) for comparative studies. NH3 synthesis performance test The NH3 synthesis activities over the as-synthesized catalysts were tested in a pressurized fixed bed reactor. In detail, the catalyst (0.2 g, 20–30 mesh) was loaded into the stainless steel reactor (inner diameter 10 mm). Before measurement, the catalysts were treated in a 25%N2–75%H2 feed gas at 300 °C for 2 h. Subsequently, the catalyst performance was evaluated in 300 °C at a weight hourly space velocity (WHSV) of 60,000 mL g−1 h−1 and 1 MPa. The outlet NH3 was trapped by diluted H2SO4 solution and then detected by ion chromatography (DIONEX, ICS-600; Thermo Scientific, Waltham, MA, USA). The turnover frequency (TOFCo) was calculated based on the total molar of Co metal using the following equation: TOF Co = r n M (1)where, r represents the NH3 synthesis rate (mol g−1 s−1), nM is the molar amounts of Co obtained from inductively coupled plasma (ICP) analysis. Catalyst characterization X-ray diffraction (XRD) measurement was recorded on a PANalytical X’Pert Pro diffractometer (PANalytical, Almelo, The Netherlands) using a Cu Kα radiation source. N2 adsorption/desorption isotherms were collected on an ASAP 2020 apparatus (MicroMetric, Lincolnshire, United Kingdom). Elemental analysis (EA) was conducted with a Vario EL-Cube instrument (Elementar, Jiangsu, China) to measure the N contents of the samples. The Co contents of the catalysts were measured by ICP atomic emission spectroscopy (ICP-AES) analysis equipped with an Ultima 2 spectrometer (Perkin-Elmer, Waltham, MA, USA). Catalyst morphologies were detected on scanning electron microscopy (SEM; a Hitachi Model S-4800; Hitachi, Tokyo, Japan). Transmission electron microscopy (TEM) images of the catalysts were obtained on a JEM-2010 instrument (JEOL Ltd., Tokyo, Japan). Aberration-corrected high-angle annular dark-field scanning TEM (AC-HAADF-STEM) analysis was performed on a JEOL JEM-ARM 200 F instrument (JEOL Ltd., Tokyo, Japan). Raman spectroscopy was measured on a multichannel modular triple Raman system (Renishaw Co., London, United Kingdom). Ultraviolet photoelectron spectroscopy (UPS) was measured using a helium resonance lamp offering He I (hν = 21.2 eV) photons. X-ray photoelectron spectroscopy (XPS) was performed on an ESCALAB 250Xi photoelectron spectrometer (Thermo Fisher Scientific, Shanghai, China) with a monochromatic Al Kα source (hν = 1486.6 eV) as the X-ray source. The XPS binding energy was referenced to the C 1s peak at 284.6 eV of adventitious carbon. Electron-paramagnetic resonance (EPR) test was performed on an E500 spectrometer (Bruker-BioSpin, Shanghai, China) with a 100 KHz magnetic field modulation at RT. UV–vis diffuse reflection spectrum (DRS) data were obtained from a Perkin Elmer Lambda 750s UV–vis spectrometer (Perkin-Elmer, Waltham, MA, USA). The productions of NH3 synthesis were trapped at the aqueous solution containing sulfuric acid and para-(dimethylamino) benzaldehyde for the UV–vis DRS measurement. X-ray absorption spectroscopy (XAS), containing X-ray absorption near-edge structure (XANES) and extended X-ray absorption fine structure (EXAFS) analyses were performed at the 1W1B beamline of the Beijing Synchrotron Radiation Facility (BSRF), China. Spectra were collected at the Co K-edge in transmission mode with a Si (111) double crystal monochromator. The Co foil and CoPc samples were used as references. Density functional theory (DFT) calculations First-principle calculations were performed using the generalized gradient approximation (GGA) of Perdew–Burke–Ernzerhof (PBE),21 as implemented in the Vienna ab initio Simulation Package (VASP 5.4.4).22 The valence electronic states of all atoms were expanded in a plane-wave basis set with a cutoff energy of 400 eV, and gamma points were used for Brillouin zone integration. All atoms were allowed to relax until the forces fell below 0.02 eV Å−1. The energy convergence criterion was set to 10−5 eV. Grimme D3 method with the zero-damping function was used to describe van der Waals interactions.23 The graphene supercell with a surface periodicity of 10 × 10, including 200 atoms as a basis, was employed to construct the Co-Nx (x = 4, 3, and 2) moieties. A vacuum space of 20 Å was used in the normal direction of the graphene plane to ensure negligible interaction between the mirror images. Results and Discussion Structure characterization The process for the synthesis of Co-Nx catalysts with different Co-N CNs is depicted in Figure 1a. The atomically dispersed Co-Nx catalysts were prepared by CoPc pyrolysis supported on HNPCSs precursor. The CN of Co-Nx catalysts was modulated by varying the pyrolysis temperature (as shown in the “Catalyst preparation” section). To highlight the coordination structure of atomically dispersed Co center, Co nanoparticle loaded on the same support (Co-NPs) was prepared for comparison. EXAFS analyses of the Co K-edge were conducted to demonstrate the coordination structure of Co-Nx catalysts. The Co K-edge EXAFS spectra of as-synthesized catalysts and reference Co foil and CoPc, are illustrated in Figure 1b. An apparent peak located at 1.5 Å for the as-synthesized Co-Nx samples was attributed to Co-N coordination, compared with CoPc reference.24 Meanwhile, no typical peaks for Co-Co coordination at 2.2 Å were visible, indicating the atomically dispersed Co atoms on these Co-Nx catalysts. Besides, the peak intensity of Co-N coordination decreased with increased pyrolysis temperature, suggesting a decrease in Co-N CN. Based on fitting results, the average CNs of Co-N of the three samples were 3.9, 3.0, and 2.1 (Table 1), and thus, the samples were denoted as Co-N4, Co-N3, and Co-N2, respectively. Additionally, the fitting result of Co K-edge EXAFS spectra of Co-N2 based on our hypothetical model was consistent with the experimental result (Figure 1c), demonstrating the validity of the depicted models. In the case of Co-NPs, the Co-Co coordination existed predominant beside the weak Co-N coordination, indicating that Co entities were distributed mainly in the form of nanoparticles. Figure 1 | (a) Schematic diagram of the preparation process of Co-Nx with different CNs. (b) Co K-edge EXAFS spectra of as-prepared samples. (c) EXAFS spectra and the corresponding curve-fitting results of Co-N2 catalyst. Download figure Download PowerPoint Table 1 | EXAFS Data Fitting Results of the As-Synthesized Catalysts Samples Shell CN R (Å)a σ2 × 102 (Å2)b ΔE0 (eV)c R-Factor (%) Co foil Co-Co 12.0 2.49 0.7 6.3 0.9 Co-N4 Co-N 3.9 1.92 0.4 8.1 0.07 Co-N3 Co-N 3.0 1.91 0.3 7.2 0.09 Co-N2 Co-N 2.1 1.89 0.3 −1.1 0.8 Co-NPs Co-N 2.0 2.03 0.7 −1.8 0.7 Co-Co 2.1 2.49 0.7 −1.8 0.7 Note: The accuracy of parameters: CN, ±20%; R, ±1%; σ2, ±20%; ΔE0, ±10%. The fitting range in k- and R-space is 3.0–12.3 Å−1 and 1.0–2.7 Å, respectively. aR, bonding distance. bσ2, the Debye–Waller factor. cΔE0, inner potential shift. SEM ( Supporting Information Figure S1) and TEM images (Figures 2a and 2b) simultaneously displayed uniform microspheres with the hollow structure of Co-Nx and Co-NPs catalysts. The average diameter of the microspheres was ∼150 nm of these samples. The high-resolution TEM (HR-TEM) images (Figure 2c and Supporting Information Figure S2) did not show any Co nanoparticles sights, implying a high dispersion of Co species as clusters or single atoms over Co-Nx catalysts. We further used the AC-HAADF-STEM characterization to inspect the Co species over Co-Nx catalysts. We found that plenty of highly dispersed Co single atoms were visible over Co-N2 (Figure 2d), Co-N3 (Figure 2e), and Co-N4 (Figure 2f) catalysts. In comparison with Co-Nx catalysts, distinct nanoparticles were observed over the Co-NPs catalyst ( Supporting Information Figure S3). Combined with the EXAFS results, we deduced that atomically dispersed Co-Nx catalysts with different Co-N CNs were synthesized successfully. Figure 2 | (a and b) TEM and (c) HR-TEM images of Co-N2 catalyst. AC-HAADF-STEM images of (d) Co-N2, (e) Co-N3, and (f) Co-N4 catalysts. Download figure Download PowerPoint The XRD patterns ( Supporting Information Figure S4) showed that all Co-Nx displayed a broad diffraction peak at ∼26.4°, belonging to the (002) crystal plane of graphitic carbon (JCPDS card no. 00-041-1487). The highly graphitic feature was inspected by Raman measurement. The Raman spectra ( Supporting Information Figure S5) displayed that Co-NP samples presented peaks at 1333 and 1596 cm−1, corresponding to the D and G bands of defective or sp3 carbon and sp2-bonded graphitic carbon, respectively.25 Furthermore, the intensity ratios of the D to G band of these samples were ∼0.95 ( Supporting Information Table S1), confirming that the coordination structure has a small impact on defective carbon configuration. Notably, the diffraction peaks on the Co-N4 sample are attributed to the CoPc phase (JCPDS card no. 00-014-0948), suggesting an incomplete decomposition of CoPc species and the existence of Co-N4 planar structure, which coincided with the fitting results of EXAFS. No XRD diffraction peaks were related to metallic Co or CoPc species in the Co-N3 and Co-N2 samples. Referring to the above EXAFS and AC-STEM data, our findings undoubtedly, demonstrated that the pyrolysis of CoPc would form atomically dispersed Co sites over Co-Nx catalysts without agglomeration. Nevertheless, after the pyrolysis of the precursor at 700 °C for 10 h, the emergence of diffraction peaks at 44.4° and 51.6° in Co-NPs catalyst were attributable to the (111) and (200) crystal faces of metallic Co (JCPDS card no. 00-001-1255), respectively. The average size of Co nanoparticles over Co-NPs was ∼20 nm, according to the calculation of (111) crystal plane via Scherrer equation, which coincided with the particle sizes observed in the HR-TEM ( Supporting Information Figure S3). ICP-AES detection showed that the Co contents in these samples were distributed in the range of 3.0–3.7 wt % ( Supporting Information Table S1). According to the EA results, the N contents of the synthesized Co catalysts ( Supporting Information Table S1) slightly decreased with increased pyrolysis temperature. N2 physical sorption ( Supporting Information Table S1) results showed that the specific surface areas of Co-Nx catalysts slightly increased from 191 to 230 m2 g−1 with increased pyrolysis temperature. However, at such a high pyrolysis temperature, the BET surface area of Co-NPs catalyst decreased to 138 m2 g−1. Identification of the geometric and electronic structure XANES experiment was performed to gain in-depth insight into the effect of different CNs on the geometric and electronic structure of Co sites. Referring to the magnified XANES results (Figure 3a), the preedge peak intensity in the range of 7706–7711 eV was indicative of the geometric configuration of Co sites, that is, tetrahedral or octahedral symmetry.26 The p → d transitions from the noncentrosymmetric tetrahedral symmetry of Co sites were supposed to contribute to this preedge peak.27 The intensity of this preedge peak over Co-N2 was higher than that of Co-N3 or Co-N4, indicating that more Co ions (II) were localized at the tetrahedral sites in the Co-N2 catalyst.26 Besides, the position of the absorption edge in the magnified images (Figure 3b) could be used to evaluate the oxidation states of Co species. A negative shift of absorption edge position from Co-N4 to Co-N2 was found, indicating that the valence state of Co species decreased along with decreased CN of the Co-Nx catalysts. Figure 3 | Normalized Co K-edge XANES spectra and the magnified images in the range of (a) 7706–7711 eV and (b) 7717–7721 eV. Download figure Download PowerPoint Further, XPS investigations were performed to obtain the valence state of Co and N species over the as-synthesized catalysts. As shown in Figure 4a, the Co 2p3/2 spectra were fitting into the characteristic peaks with binding energies of 779.0, 780.8, 782.6, and 787.0 eV, ascribed to Co0, Co2+, Co3+, and satellite peak, respectively.28 No characteristic peaks of surface metallic Co were observed on these catalysts, except for Co-NPs, which coincided with the results of XRD and EXAFS. The atomic ratio of Co2+/Co3+ over the Co-N4, Co-N3, and Co-N2 catalysts determined from XPS results were 1.86, 2.23, and 2.70, respectively (Table 2). This finding was supported by the lower oxidation state of Co entities over Co-N2 as shown in the XANES (Figure 3) results, revealing that the decreased CN might have been promoted to generate more unoccupied Co 3d charge over the Co-N2 catalyst.28 These results were further verified by EPR measurements. The EPR spectra ( Supporting Information Figure S6) showed that the g values of 2.540 and 2.005 over these catalysts were attributed to an unpaired electron in the 3dx2-y2 orbital of Co2+ and carbon radical signal, respectively.29 We found that the signal intensity of 2.540 increased with the lowering of CN, indicating that more unpaired electrons on the Co-N2 catalyst. The N 1s spectra (Figure 4b) revealed three types of N species, viz, pyridinic N, pyrrolic N, and graphitic N over Co-Nx catalysts.24,27 Through the comparison of the content of different N species from the respective characteristic peak (Table 2), we observed that the pyridinic N content increased from 24% to 50%, while the pyrrolic N content decreased from 60% to 30%, along with lowering of CN over Co-Nx catalysts. This result suggested that the atomically dispersed Co coordinated mainly with the pyrrolic N. Collectively, these observations demonstrated that the geometric and electronic structures of the atomically dispersed Co catalyst were modulated through different CNs of the Co-N, which, in turn, influenced the catalytic activity during the NH3 synthesis. Figure 4 | (a) Co 2p and (b) N 1s XPS spectra of the as-synthesized catalysts. Download figure Download PowerPoint Table 2 | The XPS Spectra Results of the As-Synthesized Catalysts Samples Pyridinic N (%) Pyrrolic N (%) Graphitic N (%) Co2+ (%) Co3+ (%) Co0 (%) Co-N4 24 60 16 0.65 0.35 0 Co-N3 41 47 12 0.69 0.31 0 Co-N2 50 30 20 0.73 0.27 0 Co NPs 40 22 38 0.39 0.34 0.27 Catalyst activity and stability Figure 5a shows the NH3 synthesis rate over the as-synthesized catalysts with a 25%N2–75%H2 gas and a WHSV of 60000 mL gcat−1 h−1. The NH3 synthesis rate in the case of Co-N2 was 85.3 mmolNH3 gCo−1 h−1 at mild conditions of 300 °C and 1 MPa, evidently higher than those over Co-N3 (68.8 mmolNH3 gCo−1 h−1) and Co-N4 (38.8 mmolNH3 gCo−1 h−1). An NH3 synthesis rate of 21.7 mmolNH3 gCo−1 h−1 was acquired over Co-NPs, which is more inferior to those over the atomically dispersed Co-Nx catalysts. Meanwhile, the NH3 synthesis rate over the Co-free nitrogen-doped carbon (denoted as N–C) support was 0.05 mmol gcat−1 h−1 at 300 °C and 1 MPa, as shown in Supporting Information Table S2, suggesting that the contribution of support to the catalytic activity of Co-Nx could be negligible. These results indicated that NH3 synthesis is a coordination-sensitive reaction, and the Co-N2 catalyst with the lowest CN presented the highest NH3 synthesis rate at mild conditions. It has been reported that the addition of a suitable electronic promoter to Co-based catalysts could improve NH3 synthesis performance, referred to as an electronic promoting effect.30–33 Herein, we researched the impact of Ba promoter to the highly active Co-N2 catalyst on NH3 synthesis activity. It showed that the NH3 synthesis rate over Ba/Co-N2 was ∼1.3 times that of the nonpromoted one (Figure 5a). The typical Cs-Co/MgO and the commercial Fe-based catalysts reactions were also performed and evaluated for comparison ( Supporting Information Table S2). At 300 °C and 1 MPa, the NH3 synthesis rate over Co-N2 (2.7 mmol gCo−1 h−1) was much higher than those of Cs-Co/MgO (0.48 mmol gcat−1 h−1) and commercial Fe-based (0.72 mmol gcat−1 h−1) catalysts, indicating the superior performance of NH3 synthesis activity over Co-N2 catalyst. For intrinsic comparison, TOFCo values, based on the total number of Co active sites on these catalysts, were calculated and depicted in Figure 5b. Notably, the TOFCo values of Co-N2 and Ba/Co-N2 were 1.40*10−3 s−1 and 1.80*10−3 s−1, respectively, which outperformed most of the previously reported Co-based catalysts presented in Supporting Information Table S2. Figure 5 | (a) NH3 synthesis rate based on Co content over as-synthesized catalysts at 1 MPa and 300 °C. (b) TOF was calculated on the total number of Co atoms basis over the catalysts. (c) N2 and (d) H2 reaction orders over Co-N2 and Co-NPs catalysts. Download figure Download PowerPoint A stability test of the Co-N2 catalyst was conducted at 300 °C and 1 MPa. As shown in Supporting Information Figure S7, the NH3 synthesis rate sustained a stable value of 2.7 mmol gCo−1 h−1 during a 50-h time-on stream. The used catalyst was further tested using HR-TEM and EXAFS techniques. As shown in the Supporting Information Figure S8, no particles or agglomeration at different selected areas were observed over the used Co-N2 catalyst. Moreover, the Co K-edge EXAFS spectrum of the used Co-N2 catalyst ( Supporting Information Figure S9) showed that there still existed the Co-N coordination, while no Co-Co coordination was observable, indicating the maintenance of a highly stable Co-N2 active site in the process of NH3 synthesis. Thus, the outcomes of these tests demonstrated the superior stability of Co-N2 catalyst in NH3 synthesis at mild conditions. Kinetic studies Kinetic studies were conducted to demonstrate the differences of N2 and H2 activation over the Co-Nx catalysts. The reaction orders of N2 (α) and H2 (β) usually serve as indicators related to the rate-limiting step and the degree of hydrogen poisoning, respectively.34 For most conventional catalysts, the reaction order