Atomically dispersed Au1 catalyst towards efficient electrochemical synthesis of ammonia

Accepted Manuscript Article Atomically dispersed Au1 catalyst towards efficient electrochemical synthesis of ammonia Xiaoqian Wang, Wenyu Wang, Man Qiao, Geng Wu, Wenxing Chen, Tongwei Yuan, Qian Xu, Min Chen, Yan Zhang, Xin Wang, Jing Wang, Jingjie Ge, Xun Hong, Yafei Li, Yuen Wu, Yadong Li PII: DOI: Reference:

S2095-9273(18)30325-6 https://doi.org/10.1016/j.scib.2018.07.005 SCIB 455

To appear in:

Science Bulletin

Please cite this article as: X. Wang, W. Wang, M. Qiao, G. Wu, W. Chen, T. Yuan, Q. Xu, M. Chen, Y. Zhang, X. Wang, J. Wang, J. Ge, X. Hong, Y. Li, Y. Wu, Y. Li, Atomically dispersed Au1 catalyst towards efficient electrochemical synthesis of ammonia, Science Bulletin (2018), doi: https://doi.org/10.1016/j.scib.2018.07.005

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Atomically dispersed Au1 catalyst towards efficient electrochemical synthesis of ammonia Xiaoqian Wang,1† Wenyu Wang,1† Man Qiao,2 Geng Wu,1 Wenxing Chen,3 Tongwei Yuan,4 Qian Xu,5 Min Chen,1 Yan Zhang,1 Xin Wang,1 Jing Wang,1 Jingjie Ge,1 Xun Hong,1 Yafei Li,2 Yuen Wu1*, Yadong Li3 1

Department of Chemistry, iChEM (Collaborative Innovation Center of Chemistry for Energy Materials), University of Science and Technology of China, Hefei 230026, China 2 Jiangsu Collaborative Innovation Centre of Biomedical Functional Materials, School of Chemistry and Materials Science, Nanjing Normal University, Nanjing 210023, China. 3 Department of Chemistry and Collaborative Innovation Center for Nanomaterial Science and Engineering, Tsinghua University, Beijing 100084, China 4 NEST Lab, Department of Chemistry, College of Science, Shanghai University, Shanghai 200444, China 5 National Synchrotron Radiation Laboratory, University of Science and Technology of China, Hefei 230029, China †

Xiaoqian Wang and Wenyu Wang contributed equally to this work.

*Correspondence: [email protected] Abstract Tremendous efforts have been devoted to explore energy-efficient strategies of ammonia synthesis to replace Haber-Bosch process which accounts for 1.4% of the annual energy consumption. In this study, atomically dispersed Au1 catalyst is synthesized and applied in electrochemical synthesis of ammonia under ambient conditions. A high NH4+ Faradaic efficiency of 11.1% achieved by our Au1 catalyst surpasses most of reported catalysts under comparable conditions. Benefiting from efficient atom utilization, an NH4+ yield rate of 1,305 μg h−1 mgAu−1 has been reached, which is roughly 22.5 times as high as that by supported Au nanoparticles. We also demonstrate that by employing our Au1 catalyst, NH4+ can be electrochemically produced directly from N2 and H2 with an energy utilization rate of 4.02 mmol kJ−1. Our study provides a possibility of replacing the Haber-Bosch process with environmentally benign and energy-efficient electrochemical strategies. Keywords NH3 synthesis, metal single sites, electrocatalysis, Haber–Bosch process, nitrogen reduction 1. Introduction As one of the most essential industrial chemicals, ammonia (NH3) is currently produced on an enormous scale of over 150 megatons per year by the Haber–Bosch process which requires pressures of 200 to 300 atmospheres and temperatures from 300 to 500 oC. To date, this energyand capital-intensive process accounts for 1.4% of the annual energy consumption and around 3% of global CO2 emissions [1-5]. Electrocatalytic approach, especially driven by renewable energy,

is generally regarded as an energy-efficient and sustainable process to synthesis NH3 at ambient conditions [2,6-9]. Despite the tremendous efforts, few attempts achieved comparable energy efficiency to Haber-Bosch process [2,6-11]. The obstacles lie on intolerable working overpotential caused by inertness of N2 [12], and poor Faradaic efficiency induced by unfavorable H2 evolution reaction in aqueous solution [2]. Therefore, to realize energy-efficient electrochemical synthesis of NH3, eligible catalysts are required to possess both activity and selectivity simultaneously. The emerging atomically dispersed metal catalysts, with definite structures as active sites, provide possibility to overcome these issues [13-17]. Benefiting from the minimum metal sizes, atomically dispersed metal catalysts usually exhibit prominent catalytic activities and may therefore convert N2 into NH3 under an acceptable overpotential. Meanwhile, the uniform active sites within atomically dispersed metal catalysts ensure high selectivity and hence a satisfactory Faradaic efficiency [18]. In this study, we prepare atomically dispersed Au on carbon nitride (Au1/C3N4) and investigate its catalytic performance towards N2 reduction to ammonium ions (NH4+) in sulfuric acid aqueous solution. Compared with Au nanoparticles supported on C3N4 (Au NPs/C3N4), Au1/C3N4 exhibited outstanding NH4+ formation Faradaic efficiency which achieved 11.1% at −0.10 V vs. RHE (reversible hydrogen electrode), outperforming most of the reported catalysts. We then assembled a full electrolytic cell with Au1/C3N4 as cathode to reduce N2 into NH4+ and a platinum foil as anode to catalyze hydrogen oxidation reaction. Our experimental results demonstrate that such an electrolytic cell allows us to synthesize NH3 directly from N2 and H2 with an energy utilization rate of 4.01 mol kJ−1. 2. Materials and method 2.1 Chemicals and materials Analytical grade chloroauric acid tetrahydrate (HAuCl4·4H2O), ethanol, salicylic acid, sodium nitroprusside dihydrate (Na2[Fe(CN)5NO]·2H2O), sodium hydroxide (NaOH), trisodium citrate dihydrate (C6H5Na3O7·2H2O), ammonium sulfate ((NH4)2SO4)), sulfuric acid (H2SO4), urea, sodium borohydride (NaBH4), sodium hypochlorite (NaClO) were obtained from Shanghai Chemical Reagents. Nafion was acquired from Sigma-Aldrich. Carbon Paper (Toray) was from Alfa Aesar and ultrasonically cleaned in ethanol. Deionized (DI) water from Milli-Q System (Millipore, Billerica, MA) was used in all experiments. All chemicals used in this experiment were from commercial and used without further purification. 2.2 Preparation procedure 2.2.1 Synthesis of polymeric g-C3N4. The g-C3N4 was prepared by a facile and efficient method. Typically, 10 g of urea powder was placed in a porcelain boat with a cover, put into a tube furnace and then heated to 550 oC for 4 h at a heating rate of 10 oC min−1 under air. The obtained yellow colored products were ground in a mortar to obtain g-C3N4 powder [19-22]. 2.2.2 Synthesis of Au1/C3N4. In detail, 20 mg of g-C3N4 powder was dispersed in 10 mL deionized water and then ultrasonicated for 2h to get homogeneous solution. After that, HAuCl4·4H2O (2 mg mL−1, 60 μL)

aqueous solution was added dropwise into the g-C3N4 aqueous dispersion under ultrasound. Next, the mixed aqueous dispersion was stirred for 3 h at room temperature. The resulting product was washed with deionized water and ethanol for several times. Then the sample was redispersed in 10 mL deionized water for later use. The as-prepared samples were stirred for 3 h under room temperature under 1 atm H2 atmosphere (1 atm= 101,325 Pa), following by stirred at 40 oC overnight under 1 atm H2 atmosphere. Then the final products were centrifuged and dried in vacuum at 60 oC overnight. The actual loading of Au1/C3N4 Au NPs/C3N4 was 0.15%, which was decided by inductively coupled plasma optical emission spectroscopy (ICP-OES). 2.2.3 Synthesis of Au NPs/C3N4. Typically, 30 mg of g-C3N4 powder was dispersed in 10 mL deionized water and then ultrasonicated for 1h to get homogeneous solution. After that, HAuCl4·4H2O (20 mg mL−1, 100 μL) aqueous solution was added dropwise into the g-C3N4 aqueous dispersion under ultrasound. Next, the mixture was stirred for 3h at room temperature. Then NaBH4 (0.05 mol L−1, 1 mL) aqueous solution was added dropwise into the mixture, following by stirred overnight under room temperature. As-obtained deep purple products were collected by centrifugation at 11,000 r min−1 for 7 min, washed three times with water and ethanol, and finally dried in vacuum at 60 oC overnight. The actual loading of Au NPs/C3N4 was 3.4%, which was decided by ICP-OES. The obtained Au NPs possess a diameter of ca. 8 nm. 2.3 Characterization. Powder X-ray diffraction ( XRD) patterns of samples were recorded on a Rigaku Miniflex-600 with Cu Kα radiation (λ=0.15406 nm, 40 kV and 15 mA). The morphologies are characterized by transmission electron microscopy (TEM, Hitachi-7700, 100 kV). The high-resolution TEM, high-angle annular dark-field scanning transmission electron microscope (HAADF-STEM) images and the corresponding electron energy-loss spectroscopy were recorded by a FEI Tecnai G2 F20 S-Twin high-resolution transmission electron microscope working at 200 kV and on a JEOL JEM-ARM200F TEM/STEM with a spherical aberration corrector working at 300 kV. X-ray photoemission spectroscopy (XPS) experiments and near edge X-ray absorption fine structure (NEXAFS) were carried out at the Catalysis and Surface Science Endstation at the BL11U beamline and Photoemission endstation at the BL10B beamline in the National Synchrotron Radiation Laboratory (NSRL) in Hefei, China. The elemental content of Au in the solid samples was monitored by an Optima 7300 DV inductively coupled plasma atomic emission spectrometer (ICP-AES). The Fourier-transform infrared (FT-IR) spectrum was record on the Nicolet 8700 FT-IR spectrometer. The concentration of NH3 and N2H4·H2O were determined by the UV-3600 UV-VIS-NIR spectrophotometer. The Au L3-edge X-ray absorption spectra were collected at room temperature in fluorescence mode at beamline 1W1B of Beijing Synchrotron Radiation Facility (BSRF) using a Si (111) double crystal monochromator. The storage rings of BSRF were operated at 2.5 GeV with a maximum current of 250 mA in decay mode. The energy was calibrated using Au foil. The electrocatalytic N2 reduction reaction (NRR) tests were carried out on the CHI760E electrochemical workstation. 2.4 Electrochemical measurements.

2.4.1 NRR measurements Before NRR tests, the Nafion membrane was pretreated by heating in H2O2 5% aqueous solution at 80 °C for 1 h and ultrapure water at 80 °C for another 1 h, respectively. The experiments were performed in a gas-tight cell with two-compartments separated by a cation exchange membrane (Nafion® 115, Dupont). Each compartment contained 30 mL electrolyte (5 mmol L−1 H2SO4 aqueous solution). Pt plate was used as a counter electrode with an Ag/AgCl (saturated KCl) electrode as reference electrode. In a typical prepared procedure of the working electrode, 60 μL of the homogeneous ink, which was prepared by dispersing 8 mg sample and 80 μL Nafion solution (5 wt%) in 1 mL of water-ethanol solution with a volume ratio of 1:1, was loaded onto the two sides of a carbon fiber paper electrode with 1.0 cm× 0.5 cm. Cyclic voltammetry (CV) measurements were performed with a scan rate of 10 mV s−1 between −0.2 V to −0.9 V vs. Ag/AgCl in N2-saturated 5 mmol L−1 H2SO4. The potentials in the study were reported versus RHE with the conversion E (vs. RHE) = E (vs. Ag/AgCl) + 0.1989 V + 0.0592 × pH. The current density was obtained by normalized with the carbon fiber paper geometric surface area. 2.4.2 Determination of ammonia. Concentration of generated ammonia was spectrophotometrically determined by the indophenol blue method [23]. In detail, 10 mL of the solution was drew from the electrochemical reaction vessel. Then, 0.5 mL of the solution with 0.625 mol L−1 NaOH, 0.36 mol L−1 salicylic acid and 0.17 mol L−1 sodium citrate were added, followed by adding 100 μL of an aqueous solution of sodium nitroferricyanide (10 mg mL−1) and 100 μL of 1.5 mol L−1 NaClO to it. Finally, the solution was shaked several minutes to make it distributed uniformly. After 2 h standing at room temperature, the absorption spectrum was measured using an ultraviolet-visible spectrophotometer. The formation of indophenol blue was determined by using the absorbance at the wavelength of 695 nm. The concentration-absorbance curves were calibrated using standard ammonium sulfate solution, which contained the same concentration of H2SO4 as used in the electrolysis experiments. 2.4.3 Determination of hydrazine. The hydrazine present in the electrolyte was detected by the method of Watt and Chrisp [24]. A mixture of para-(dimethylamino) benzaldehyde (4.0 g), HCI (concentrated, 24 mL) and ethanol (200 mL) were used as a color reagent. The hydrazine hydrochloride solution (10 g mL−1) with 5 mmol L−1 H2SO4 was used as the standard solution. Calibration solutions were prepared as follows: First, adding suitable volumes of the standard solution in colorimetric tubes; Second, making up to 10 mL with 5 m mol L−1 H2SO4 solution; Third, adding 5 mL above prepared color reagent followed by shaked to make it mix well, then stand for 20 min at room temperature; Fourth, the absorbance of the resulting solution was measured at 457.5 nm. The yields of hydrazine were measured by drawing 10 mL electrolyte from the electrochemical reaction vessel and adding 5 mL color reagent to it and then comparing the absorbance of it with the calibration curves to obtain the concentration of the hydrazine produced in the NRR process. 2.4.4 Faradaic efficiency. The Faradaic efficiency for NRR was defined as the quantity of electric charge used for synthesizing ammonia divided the total charge passed through the electrodes during electrolysis.

The total amount of NH3 produced was measured using colorimetric methods, assuming that three electrons were consumed to produce one NH3 molecule, the Faradaic efficiency can be calculated as follows:

where F is the Faraday constant,

is the NH3 concentration, V is the volume of electrolyte and

Q is the total amount of charge transferred during catalysis. The rate of ammonia production was calculated using the following equation: (2) where

is molecular weight of NH3, mAu is the mass of Au in used catalyst and t is the

testing time. 2.5 Computational Method. Density functional theory (DFT) computations were performed using the plane-wave technique implemented in Vienna ab initio simulation package (VASP) [25,26]. The ion-electron interaction was described using the projector-augmented plane wave (PAW) approach [27]. The generalized gradient approximation (GGA) expressed by PBE functional [28] and a 420 eV cutoff for the plane-wave basis set were adopted in all the computations. The convergence threshold was set as 10−4 eV in energy and 0.05 eV Å−1 in force. The Brillouin zones was sampled with a 441  centered k points grid. The PBE-D3 method was adopted to describe the van der Waals interactions. The solvent effect on adsorbates was simulated using the Poissson-Boltzmann implicit solavtion model with a dielectric constant of 80 [29]. In this work, we built a 2×2 supercell of C3N4 monolayer consist of 24 C and 32 N atoms. The optimized lattice constants of C3N4 are a = b = 7.09 Å, which match well with experimental measurements. One C atom was replaced to by an Au atom to construct Au1/C3N4. The optimized Au−N bond length are 2.05, 2.29 and 1.96 Å, which further prove the three coordination environment of Au atom. We used a slab model of Au (211) plane with three atomic layers to represent Au nanoparticle. The free energy (G) of each species is estimated at T=298 K according to: G = EDFT + EZPE  TS, (3) where EDFT, EZPE and S are electronic energy, zero point energy, and entropy, respectively. For adsorbed intermediates, EZPE and S were determined by vibration frequencies calculations, where all 3N degrees of freedom are treated as harmonic vibration motions with neglecting contributions from the slab, while for molecules those were taken from the NIST database [30]. The contribution of zero point energy and entropy corrections to G is provided in Table S1 (online).

3. Results and discussion 3.1 Characterization The structure was confirmed by TEM (Fig. S1 online), FT-IR (Fig. S2 online) [31] and XRD (Fig. S3 online) [32]. To synthesize Au1/C3N4, the as-prepared C3N4 was first soaked in HAuCl4 aqueous solution. After sufficient adsorption, the dispersion was reduced by H2 under 40 oC

overnight. The light yellow product was collected by centrifugation and washing for several times (Fig. S4 online). For better comparison, we also prepared Au NPs/C3N4 by a similar method except that NaBH4 was used as reducing agent (Figs. S1 and S4 online). TEM studies indicate that Au1/C3N4 retains the free-standing-nanosheet structure from C3N4 (Fig. 1a). As shown in Fig. 1b, the energy dispersive X-ray (EDX) spectroscopy elemental mappings revealed the distributions of Au, C and N, indicating that Au atoms are homogenous over the entire nanosheets. The atomic dispersion of Au atoms were directly observed by aberration corrected HAADF-STEM measurements with sub-angstrom resolution, in which the bright spots marked with red cycles represent Au atoms due to a higher Z-contrast (Figs. 1c and d). The ring-like pattern collected in selected area electron diffraction (SAED) suggests poor crystallinity of Au1/C3N4, implying the atomic distribution of Au (inset in Fig. 1a). Similar result can be reached by XRD measurements. As shown in Fig. 2a, only one characteristic peak assigned to C3N4 can be observed for Au1/C3N4, indicating the Au-crystal-free feature. In contrary, a set of peaks for Au crystal confirmed the crystallization of Au NPs/C3N4. To further verify the distribution form of Au species, EXAFS was performed to investigate the coordination environment around Au atoms (Fig. 2b). The absence of Au−Au bonding can be affirmed due to the lack of corresponding signal at 2.86 Å. Instead, one can identify an obvious signal assigned to Au−N bond with a bond length of ca. 1.65 Å. Our XAS fitting results further indicate the Au−N coordination number is nearly to 3 (Fig. S5 and Table S2 online). Based on the above experimental results, the atomic dispersion of Au in Au1/C3N4 was strongly proved. The valence state of Au in Au1/C3N4 was explored by XANES spectra, in which the Au valence state can be derived from the absorption threshold and white line peak intensity of Au L3-edge (Fig. 2c) [33]. One may notice that, not like HAuCl4 with strong white line peak intensity or Au foil with illegible white line peak, Au1/C3N4 and AuCl exhibit similar white line peaks with comparable intensities, demonstrating that the Au valence state in Au1/C3N4 is around +1. Likewise, the close peak positions in the first derivative curves from Au1/C3N4 and AuCl XANES spectra also support above conclusion (Fig. S6 online). Moreover, XPS of Au 4f in Au1/C3N4 reaches the same result, in which Au1/C3N4 shows almost the same signal from Au (I) [34,35]. Besides, benefiting from the nanosheets structure, Au1/C3N4 possesses a surface area of 58 m2 g−1 as revealed by Brunauer-Emmett-Teller (BET) adsorption-desorption isotherms (Fig. S7 online). 3.2 Evaluation of Activity for Electrochemical Reduction of N2 The electrochemical synthesis of ammonia was achieved in a custom-made three-electrode set-up with 5 mmol L−1 H2SO4 solution as electrolyte under ambient conditions (Fig. 3a). As working electrode, carbon paper coated by catalysts converts supplied N2 and H+ into NH4+ under suitable potentials. All potentials were measured against an Ag/AgCl (saturated KCl) reference electrode and were converted to those against a RHE. Pt plate was used as a counter electrode on which water would be oxidized to oxygen (O2), leaving H+. Altogether, NH4+ and O2 would be obtained from N2, H+ and water during such a catalysis process. CV mearsuremens was performed to show the total current under different applied potentials (Fig. S8 online). After long-term operation under a certain potential, the concentration of generated NH4+ and N2H4 in electrolyte can be quantitatively analyzed by reported colorimetric methods (Figs. S9−S11 online) [23,24,36]. Compared with Au NPs/C3N4, Au1/C3N4 exhibited higher NH4+ Faradaic efficiencies at the applied potential from −0.1 to −0.5 V (Fig. 3b). Notably, Au1/C3N4 achieved an NH4+ Faradaic efficiency

of 11.1% at −0.1 V which is, to the best of our knowledge, the highest reported value under comparable conditions (Table S3 online) [37-41]. The extremely poor NH4+ Faradaic efficiencies obtained by pure C3N4 underline the crucial role of the Au species in NH4+ synthesis. Besides, NH4+ can be hardly obtained when N2 feed changed into argon (Ar), which indicates that the N atoms in produced NH4+ are from N2 feed rather than from C3N4 (Figs. 3b and S11 (online)). No N2H4 can be detected at all tested potential for Au1/C3N4, Au NPs/C3N4 or pure C3N4. The NH4+ yield rates normalized by Au mass in catalysts are shown in Fig. 3c. At −0.1 V, Au1/C3N4 reached an NH4+ yield rate of 1,305 μg h−1 mgAu−1 which is roughly 22.5 times as high as that reached by Au NPs/C3N4, mainly due to high NH4+ Faradaic efficiency and optimal Au atom utilization of Au1/C3N4. Also, as listed in Table S3 (online), such a performance exceeds most reported results. A higher yield rate normalized by metal mass, especially precious metal, signifies a more satisfactory atom economy and thus a lower catalyst cost, which is of significance in industrial applications. At potentials from −0.1 to −0.3 V, negligible current density decay in long-term operation testing indicates the catalytic durability of Au1/C3N4. The TEM and HAADF-STEM images of recycled Au1/C3N4 after long-term operation verifies that the nanosheets morphology and the atomic dispersion of Au remain almost unchanged, confirming its robustness during the catalysis process (Fig. S12 online). 3.3 Computational Studies and Further Research DFT calculations were further carried out to gain some fundamental insights into the superior catalytic activity and selectivity of Au1/C3N4 towards the electrochemical reduction of N2 to NH3. Generally, the NRR can proceed two N atoms being hydrogenated simultaneously to produce two NH3 molecules at the same time (alternating mechanism), or one N atom being hydrogenated firstly to produce a NH3 molecule then the left N atom can be further hydrogenated to yield the second NH3 molecule (distal mechanism). For Au1/C3N4 and Au NPs both distal and alternating mechanisms were considered, and the adsorption free energies of NRR intermediates (e.g., *NNH, *NHNH, *NHNH2, *NH2, etc.) and the Gibbs free energies (ΔG) of all elementary steps were computed by employing the computational hydrogen electrode methodology (Table S1 online) [42]. A slab model of Au (211) plane was adopted to represent the Au NPs. We first investigated the NRR on Au1/C3N4 following the alternating mechanism. The corresponding atomic configurations at various states along the reaction path are displayed in Fig. 4a, and the free energy profiles are summarized in Fig. 4b. The rate-determining step (RDS) is the reduction of N2 to *NNH, which has a positive ΔG of 1.33 eV. Similarly, the RDS for NRR on Au (211) through the alternating mechanism is also the reduction of N2 to *NNH with a higher ΔG of 2.01 eV. The formation of N2H4 is energetically unfavorable for neither Au1/C3N4 nor Au (211) compared with the formation of NH3, which is coincident with our experimental results. The distal mechanism for both Au1/C3N4 and Au (211) should be overwhelmed due to the much higher ΔG of RDS (Fig. S13 online). Therefore, our computations revealed Au1/C3N4 favors the alternating mechanism to catalyze the reduction of N2 to NH3, and it has better catalytic performance than Au (211) due to the lower ΔG of RDS. As the better NRR catalytic activity of Au1/C3N4 is due to the stronger binding interaction with *NNH, we also plotted the electron density difference of Au1/C3N4 caused by anchoring of Au atom to get some deep insight. As vividly shown in Fig. 3c, there is pronounced charge transfer from the Au atom to the g-C3N4, resulting in a 0.56 |e| positive charge for the Au atom according

to the Bader charge population analysis. The electron depletion of Au atom may shift its d-orbital position toward the Fermi level, which could enhance its interaction with intermediates (e.g., NNH*) can subsequently lead to a better NRR performance. Moreover, we are aware of that the hydrogen evolution reaction (HER) is competitive to NRR, thus we also computed the relative thermodynamic limiting potentials between NRR and HER (denoted as UL(NRR)  UL(HER) for Au1/C3N4 and Au (211) to access the selectivity. Previous studies suggested that a more positive value of UL(NRR)  UL(HER) means a higher NRR selectivity. According to our computations, Au1/C3N4 and Au (211) has a UL(NRR)  UL(HER) value of 0.88 and 1.67 V, respectively (Fig. 4d), indicating that the Au1/C3N4 shows a better selectivity for reduction of N2 to NH3, which achieves a good agreement with experimental results. It is noteworthy that the by-product O2 leads to tremendous unnecessary energy input during NH3 synthesis due to the high thermodynamic equilibrium potential of O2/H2O (1.23 V vs. RHE) [43]. To improve the energy utilization rate and atom economy, O2 evolution reaction should be avoided. Thus, we refitted the electrolyzer as Fig. 4e. In such a set-up, H2 gas is continuously supplied to Pt planet anode and thus H2 oxidation reaction is able to occur, which directly provides H+ for NH4+ synthesis on cathode. Altogether, NH4+ can be produced by only N2 and H2 without any by-product, which is similar to Haber–Bosch process. Besides, the actual input potential between cathode and anode can be controlled when Pt plate is set as both counter electrode and reference electrode. Thus, the input electric power during electro-catalysis process can be given as p=E×I, (3) where E is the potential applied between cathode and anode, and I is the current. Further, after the quantitative analysis of produced NH4+, the energy utilization rate can be calculated as ,

(4)

where Q is the total amount of transferred charge between cathode and anode, and

is the

+

amount of produced NH4 . Based on above inference, the energy utilization at different applied potential were investigated using Au1/C3N4 as the cathode catalyst (Fig. 4f). Our catalyzer can produce NH4+ with an promising energy utilization rate of 4.02 mmol kJ−1 (Fig. S14 online) [2]. Despite the achieved high energy utilization, there are still obstacles required to be overcome before our system possesses the ability to challenge the Haber-Bosch process. The major one is that the slow synthetic rate cannot meet the demand of mankind, which is mainly restricted by low N2 solubility in aqueous solution. Another one is low N2 and H2 conversion rate caused by air-blowing method. Thus, our future research will focus on enhancing electrolytic rate and improving our equipment with N2 and H2 circulating system. 4. Conclusion In summary, using atomically dispersed Au1/C3N4 catalyst, electrochemical reduction of N2 into NH4+ has been achieved under ambient conditions with a high Faradaic efficiency of 11.1%. Benefiting from efficient atom utilization and high Faradaic efficiency, an NH4+ yield rate of 1,305 μg h−1 mgAu−1 has been reached, which is roughly 22.5 times as high as that by supported Au nanoparticles. Our DTF calculation results reveal that an alternating hydrogenation mechanism is carried out on both Au1/C3N4 and Au NPs/C3N4, and that lower required energy in the first hydrogenation step explains the enhanced NH4+ formation Faradaic efficiency on Au1/C3N4. Furthermore, a by-product-free synthesis of NH3 from N2 and H2 has been realized in a

two-electrode electrolyzer with an energy utilization rate of 4.02 mmol kJ−1. Although the Haber–Bosch process will likely dominate the industrial NH3 synthesis for a long time, this study provides a possibility of a more energy-efficient NH3 electrochemical synthesis strategy.

Fig. 1. (Color online) TEM observation of Au1/C3N4. (a) TEM images of Au1/C3N4 and corresponding SEAD pattern (inset). (b) Examination of the corresponding EDX mapping reveals the homogeneous distribution of Au, C and N on the nanosheets. (c), (d) Magnified HAADF-STEM images of Au1/C3N4 directly show the atomic dispersion of Au atoms.

Fig. 2. (Color online) Characterizations of Au1/C3N4. (a) XRD patterns imply poor crystallinity of Au1/C3N4 but crystalline features of Au NPs/C3N4. (b) EXAFS spectra confirm the atomic dispersion of Au1/C3N4. (c) The normalized Au L3-edge XANES and (d) XPS measurements reveals the Au valence state in Au1/C3N4 is about +1.

Fig. 3. (Color online) Performance of N2 electroreduction. (a) Schematic electrolyzer for N2 electroreduction test. (b) NH4+ formation Faradaic efficiencies for Au1/C3N4, Au NPs/C3N4, pure C3N4 and Au1/C3N4 with Ar feed instead of N2. (c) NH4+ yield rates normalized by Au mass. (d) Catalytic durability test for Au1/C3N4 at different potentials.

Fig. 4. (Color online) DFT studies and NH4+ electrochemical synthesis from N2 and H2. (a) Optimized geometric structures of various states (*NNH, *NHNH, *NHNH, *NH 2NH2, *NH2) of NRR proceeded on Au1/C3N4 following the alternating mechanism. (b) Free energy profile of NRR on Au1/C3N4 and Au (211) at zero potential following the alternating mechanism. (c) Electron density different of Au1/C3N4 caused by anchoring of Au atom. Cyan and pink represents electron accumulation and depletion, respectively. (d) Difference in limiting potentials for NRR and HER. (e) Schematic electrolyzer for NH4+ synthesis directly from N2 and H2. (f) Energy utilization rates under different applied potential between cathode and anode.

Conflict of interest The authors declare that they have no conflict of interest. Acknowledgments This work was supported by the National Key R&D Program of China (2017YFA 0208300), and the National Natural Science Foundation of China (21522107, 21671180, 21521091, 21390393, U1463202, and 21522305). We thank the photoemission end-stations BL1W1B in Beijing Synchrotron Radiation Facility (BSRF), BL14W1 in Shanghai Synchrotron Radiation Facility (SSRF), BL10B and BL11U in National Synchrotron Radiation Laboratory (NSRL) for the help in characterizations.

References 1.

Service R F. New recipe produces ammonia from air, water, and sunlight. Science 2014; 345: 610-610.

2.

van der Ham C J, Koper M T, Hetterscheid D G. Challenges in reduction of dinitrogen by proton and electron transfer. Chem Soc Rev 2014; 43: 5183-5191.

3.

Rosca V, Duca M, de Groot M T, et al. Nitrogen cycle electrocatalysis. Chem Rev 2009; 109: 2209-2244.

4.

Appl M. Ammonia principles and industrial practice. Future 1999; 245: 12.

5.

Burford R J, Fryzuk M D. Examining the relationship between coordination mode and reactivity of dinitrogen. Nat Rev Chem 2017; 1: 0026.

6.

Guo C, Ran J, Vasileff A, et al. Rational design of electrocatalysts and photo (electro) catalysts for nitrogen reduction to ammonia (NH3) under ambient conditions. Energy Environ Sci 2018; 11: 45-56.

7.

Kyriakou V, Garagounis I, Vasileiou E, et al. Progress in the electrochemical synthesis of ammonia. Catal Today 2017; 286: 2-13.

8.

Zhao J, Chen Z. Single Mo atom supported on defective boron nitride monolayer as an efficient electrocatalyst for nitrogen fixation: a computational study. J Am Chem Soc 2017; 139: 12480-12487.

9.

Seh Z W, Kibsgaard J, Dickens C F, et al. Combining theory and experiment in electrocatalysis: insights into materials design. Science 2017; 355: eaad4998.

10. Lan R, Irvine J T, Tao S. Synthesis of ammonia directly from air and water at ambient temperature and pressure. Sci Rep 2013; 3: 1145. 11. Kordali V, Kyriacou G, Lambrou C. Electrochemical synthesis of ammonia at atmospheric pressure and low temperature in a solid polymer electrolyte cell. Chem Commun 2000; (17):1673-1674. 12. Haynes W M. CRC handbook of chemistry and physics. Boca Raton: CRC press, 2014 13. Yang XF, Wang A, Qiao B, et al. Single-atom catalysts: a new frontier in heterogeneous catalysis. Acc Chem Res 2013; 46: 1740-1748. 14. Liu P, Zhao Y, Qin R, et al. Photochemical route for synthesizing atomically dispersed palladium catalysts. Science 2016; 352: 797-800. 15. Yin P, Yao T, Wu Y, et al. Single cobalt atoms with precise N-coordination as superior oxygen

reduction reaction catalysts. Angew Chem Int Ed 2016; 55: 10800-10805. 16. Ge J, He D, Chen W, et al. Atomically dispersed Ru on ultrathin Pd nanoribbons. J Am Chem Soc 2016; 138: 13850-13853. 17. Liu P, Zhao Y, Qin R, et al. A vicinal effect for promoting catalysis of Pd1/TiO 2: supports of atomically dispersed catalysts play more roles than simply serving as ligands. Sci Bull 2018; 63: 675-682 18. Wang X, Chen Z, Zhao X, et al. Regulation of coordination number over single Co sites: triggering the efficient electroreduction of CO2. Angew Chem 2018; 130: 1962-1966. 19. Dong F, Wu L, Sun Y, et al. Efficient synthesis of polymeric gC 3N4 layered materials as novel efficient visible light driven photocatalysts. J Mater Chem 2011; 21: 15171-15174. 20. Wang H, He W, Dong X a, et al. In situ FT-IR investigation on the reaction mechanism of visible light photocatalytic NO oxidation with defective g-C3N4. Sci Bull 2018 63(2): 117-125 21. Jin H, Guo C, Liu X, et al. Emerging two-dimensional nanomaterials for electrocatalysis. Chem Rev 2018; doi: 10.1021/acs.chemrev.7b00689 22. Vasileff A, Xu C, Jiao Y, et al. Surface and Interface engineering in copper-based bimetallic materials for selective CO2 electroreduction. Chem 2018; doi: 10.1016/j.chempr.2018.05.001 23. Zhu D, Zhang L, Ruther R E, et al. Photo-illuminated diamond as a solid-state source of solvated electrons in water for nitrogen reduction. Nat Mater 2013; 12: 836. 24. Watt G W, Chrisp J D. Spectrophotometric method for determination of hydrazine. Anal Chem 1952; 24: 2006-2008. 25. Kresse G, Furthmüller J. Efficient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys Rev B 1996; 54: 11169. 26. Kresse G, Furthmüller J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Compu Mater Sci 1996; 6: 15-50. 27. Kresse G, Joubert D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys Rev B 1999; 59: 1758. 28. Perdew J P, Burke K, Ernzerhof M. Generalized gradient approximation made simple. Phys Rev Lett 1996; 77: 3865. 29. Mathew K, Sundararaman R, Letchworth-Weaver K, et al. Implicit solvation model for density-functional study of nanocrystal surfaces and reaction pathways. J Chem Phys 2014; 140: 084106. 30. Johnson R D. NIST computational chemistry comparison and benchmark database. http://srdata. nist. gov/cccbdb 2006 31. Niu P, Zhang L, Liu G, et al. Graphene-like carbon nitride nanosheets for improved photocatalytic activities. Adv Funct Mater 2012; 22: 4763-4770. 32. Xu S, Zhou P, Zhang Z, et al. Selective oxidation of 5-hydroxymethylfurfural to 2,5-furandicarboxylic acid using O2 and a Photocatalyst of Co-thioporphyrazine Bonded to g-C3N4. J Am Chem Soc 2017; 139: 14775-14782. 33. Lytle F, Wei P, Greegor R, et al. Effect of chemical environment on magnitude of X-ray absorption resonance at LIII edges. Studies on metallic elements, compounds, and catalysts. J Chem Phys 1979; 70: 4849-4855. 34. Thomas T D, Weightman P. Valence electronic structure of AuZn and AuMg alloys derived from a new way of analyzing Auger-parameter shifts. Phys Rev B 1986; 33: 5406. 35. Kitagawa H, Kojima N, Nakajima T. Studies of mixed-valence states in three-dimensional

halogen-bridged gold compounds, Cs2AuIAuIIIX6 (X= Cl, Br or I). Part 2. X-Ray photoelectron spectroscopic study. J Chem Soc, Dalton Trans. 1991; (11):3121-3125. 36. Boltz D F. Book reviews: colorimetric determination of non-metals. Science, 1958, 128: 709-710. 37. Bao D, Zhang Q, Meng F L, et al. Electrochemical reduction of N 2 under ambient conditions for artificial N2 fixation and renewable energy storage using N2/NH3 cycle. Adv Mater 2017; 29: 1604799 38. Shi M M, Bao D, Wulan B R, et al. Au Sub-nanoclusters on TiO2 toward highly efficient and selective electrocatalyst for N2 conversion to NH3 at ambient conditions. Adv Mater 2017; 29: 1606550 39. Li S J, Bao D, Shi M M, et al. Amorphizing of Au Nanoparticles by CeOx-RGO hybrid support towards highly efficient electrocatalyst for N2 reduction under ambient conditions. Adv Mater. 2017; 29: 1700001 40. Chen S, Perathoner S, Ampelli C, et al. Electrocatalytic synthesis of ammonia at room temperature and atmospheric pressure from water and nitrogen on a carbon-nanotube-based electrocatalyst. Angew Chem. 2017; 129: 2743-2747. 41. Chen GF, Cao X, Wu S, et al. Ammonia electrosynthesis with high selectivity under ambient conditions via a Li+ incorporation strategy. J Am Chem Soc 2017; 139: 9771-9774. 42. Nørskov J K, Rossmeisl J, Logadottir A, et al. Origin of the overpotential for oxygen reduction at a fuel-cell cathode. J Phys Chem B 2004; 108: 17886-17892. 43. Mistry H, Varela A S, Kühl S, et al. Nanostructured electrocatalysts with tunable activity and selectivity. Nat Rev Mater 2016; 1: 16009.

Xiaoqian Wang received his B.Sc. in materials physics in Jiaxi Lu Talent Program from University of Science and Technology of China in 2016. He is now pursing his Ph.D. degree under supervision of Prof. Yuen Wu at iChEM (Collaborative Innovation Center of Chemistry for Energy Materials), University of Science and Technology of China. His current research interests focus on the application of nanomaterials in energy conversion and storage, such as carbon dioxide electro-catalytic reduction and electro-catalysis in fuel cells.

Yuen Wu received his B.Sc. and Ph.D. degrees from the Department of Chemistry, Tsinghua University in 2009 and 2014, respectively. He is currently a professor in the Department of Chemistry, University of Science and Technology of China. His research interests are focused on the synthesis, assembly, characterization and application of functional nanomaterials.

Graphical abstract The electrochemical reduction of N2 offers a compelling strategy for environmentally friendly and energy-efficient synthesis of NH3, which suffers from low Faradaic efficiency. Atomically dispersed catalysts are regarded to perform high Faradaic efficiency due to uniform catalytic active sites. In this study, an atomically dispersed Au 1 catalyst achieves a high NH4+ formation Faradaic efficiency as well as a promising energy efficiency.