High-Performance N2-to-NH3 Conversion Electrocatalyzed by Mo2C Nanorod.

The synthesis of NH3 is mainly dominated by the traditional energy-consuming Haber-Bosch process with a mass of CO2 emission. Electrochemical conversion of N2 to NH3 emerges as a carbon-free process for the sustainable artificial N2 reduction reaction (NRR), but requires an efficient and stable electrocatalyst. Here, we report that the Mo2C nanorod serves as an excellent NRR electrocatalyst for artificial N2 fixation to NH3 with strong durability and acceptable selectivity under ambient conditions. Such a catalyst shows a high Faradaic efficiency of 8.13% and NH3 yield of 95.1 μg h-1 mg-1 cat at -0.3 V in 0.1 M HCl, surpassing the majority of reported electrochemical conversion NRR catalysts. Density functional theory calculation was carried out to gain further insight into the catalytic mechanism involved.

As a necessary industrial chemical, NH3 has been employed in medication, fertilizer, fuel, and explosives, etc.[1−5] Today, ever-increasing NH3 consumption stimulates intensive research on artificial N2 fixation technology.[6−10] However, industrial-scale NH3 production mainly depends on the Haber–Bosch process, which is performed under rigorous conditions (350–550 °C and 150–350 atm) with rather high energy consumption and CO2 emission.[11,12] Therefore, it is urgently desired to develop an energy-saving and environmentally benign technological process for NH3 production.

Electrochemical reduction has emerged as a promising method for artificial N2 fixation under ambient conditions.[13] However, the N2 reduction reaction (NRR) process needs to break a rather inert molecular structure of N2 with extremely high bond energy of about 941 kJ mol–1.[6] Thus, electrocatalysts with high activity for the NRR are a prerequisite.[13−15] In nature, N2 fixation can be catalyzed under ambient conditions by Mo-dependent nitrogenases, via multiple proton and electron transfer steps.[16−18] Mo has also emerged as an interesting metal for homogeneous N2 functionalization reactions, and some Mo-based molecular complexes have been designed[19,20] and synthesized for artificial N2 fixation.[21−24] However, other than stability of these catalysts, it is also challenging to effectively graft such catalysts onto electrodes for electrochemical tests and applications. Therefore, it is highly pressing to develop Mo-based heterogeneous electrocatalysts to solve these problems. Recently, (110)-oriented Mo nanofilm was reported for N2 reduction electrocatalysis with only a Faradaic efficiency (FE) of 0.72%.[25] Our recent studies suggest that MoS2,[26] MoO3,[27] MoN,[28] and Mo2N[29] are effective for the NRR process with Faradaic efficiencies of 1.17%, 1.9%, 1.15%, and 4.5%, respectively. As such, to develop new Mo-based electrocatalysts for the NRR with improved activity is highly desired.

Here, we present our recent study in developing the Mo2C nanorod as a superb NRR catalyst for artificial N2 fixation to NH3 with strong electrochemical durability and acceptable selectivity under ambient conditions. Such Mo2C achieves an FE as high as 8.13% with NH3 yield of 95.1 μg h–1 mg–1cat at −0.3 V versus reversible hydrogen electrode (RHE) in 0.1 M HCl. The possible NRR catalytic mechanism is also studied by density functional theory (DFT) calculations.

The Mo2C nanorod was derived from a Mo-based nanorod precursor (Figure S1) under Ar atmosphere (see the Supporting Information for preparation details). X-ray diffraction (XRD) analysis suggests a hexagonal phase structure of the resulting Mo2C product (JCPDS 79-0744) in Figure a. The scanning electron microscopy (SEM) image confirms formation of the Mo2C nanorod (Figure b), and the transmission electron microscopy (TEM) image indicates the nanoporous nature (Figure c and Figure S2). The high-resolution TEM (HRTEM) image of a nanorod (inset of Figure c) suggests an interplanar distance of 2.3 Å, in accordance with Mo2C (121). Energy-dispersive X-ray (EDX) analysis displays a homogeneous distribution of Mo and C in Mo2C (Figure d). The X-ray photoelectron spectroscopy (XPS) survey spectrum (Figure S3) suggests the presence of C and Mo elements. In the Mo 3d region (Figure e), the binding energies (BEs) at 228.5 and 231.5 eV are matched well with Mo2+ in Mo2C.[30,31] The BEs at 229.5, 232.6, and 235.4 eV are assigned to Mo 3d5/2, Mo 3d3/2, and Mo6+ regions, respectively, indicating that the surface of the catalyst is oxidized, and the BE at 230.3 eV is ascribed to Mo4+.[31−33] In Figure f, the BEs of the C 1s region at 284.8 and 283.5 eV are consistent with carbidic carbon, in accordance with the reported Mo2C.[31] The BEs at 286.3 and 288.4 eV are ascribed to C—O and C=O bonds.[34]

Figure 1

(a) XRD spectrum and (b) SEM image of the Mo2C nanorod. (c) TEM and (inset) HRTEM images of the Mo2C nanorod. (d) TEM image and corresponding EDX elemental mapping images of the Mo2C nanorod. XPS spectra in the (e) Mo 3d and (f) C 1s regions of Mo2C.

Mo2C catalyst was dropped on a glassy carbon electrode, and the resulting Mo2C/GCE was tested as an NRR working electrode under ambient conditions in an electrolytic cell separated by a Nafion membrane (Figure a). During N2 electrolysis, N2 was supplied to the cathode, and the H+ in the electrolytic cell can transfer through the HCl electrolyte solution to react with N2 for NH3 production. The potential was reported on an RHE scale. Linear sweep voltammetry (LSV) curves of Mo2C/GCE are provided in Figure S4 under N2- and Ar-saturated electrolytes (0.1 M HCl). The black line of the LSV curve is gained by N2 bubbling, which exhibits a slightly higher catalytic current, indicating that N2 is reduced to NH3. The NH3 is detected by indophenol blue method,[35] and the byproduct of N2H4 is also determined by means of Watt and Chrisp.[36] The calibration curves are given in Figures S5 and S6. Figure b shows UV–vis absorption spectra, which are colored with indophenol indicator of the 3 h NRR process at some constant potentials. Mo2C/GCE exhibits excellent selectivity without N2H4 production (Figure S7). The NRR process of Mo2C/GCE is operated from −0.2 to −0.6 V. In Figure c, the FEs and average NH3 yields are provided. The most effective NRR rate is located at −0.3 V versus RHE, and FE and NH3 yield are calculated as 8.13% and 95.1 μg h–1 mg–1cat, respectively. Our Mo2C/GCE compares favorably to the NRR behaviors of many aqueous-based NRR electrocatalysts under ambient conditions like Au/TiO2 (21.4 μg h–1 mg–1cat, 8.11%),[37] Au nanorods (6.042 μg h–1 mg–1cat, 4%),[38] and γ-Fe2O3 (0.212 μg h–1 mg–1cat, 1.9%),[39] etc. Table S1 gives a detailed comparison of NRR electrocatalysts at ambient conditions. Of note, Mo2C/GCE is even superior to some reported catalysts under severe reaction conditions (Table S2). When applied potential surpasses −0.3 V, the NRR properties (NH3 yield and FE) reduce obviously, because of competitive adsorption of nitrogen and H+/H2 on Mo2C/GCE.[9] Moreover, in Figure S8, corresponding H2 yield and the relevant FEs of the hydrogen evolution reaction (HER, a competitive process) are provided. Upon combination of the data of the HER and NRR, the remaining nearly 50% Faraday loss at low potential (−0.2 V versus RHE) could be attributed to the following three aspects: (1) capacitance consumption of catalyst, (2) dynamic hydrogen adsorption and absorption on the catalyst, and (3) uncontrollable experimental error.[40−43]Figure d compares the amount of NH3 generated using Mo2C/GCE and blank GCE as the working electrodes at −0.3 V after 3 h of electrolysis. Clearly, the bare GCE exhibits poor electrocatalytic NRR activity. It is also noted that there is almost no NH3 detected in the electrolytes after similar cathodic tests in comparative experiments (Figure S9). All of these results strongly support that NH3 originates from electrochemical N2 reduction effectively catalyzed by the Mo2C nanorod.

Figure 2

(a) Schematic graph to illustrate the electrocatalytic setup for the NRR test. (b) UV–vis spectra of obtained electrolyte solutions colored with indophenol indicator after 3 h electrolysis reactions. (c) NH3 yield rates and corresponding FEs at each given potential. (d) Amount of NH3 generated using Mo2C/GCE and a blank GCE at potential of −0.3 V after 3 h of electrolysis under ambient conditions.

Durability is a critical parameter to assess the properties of NRR catalysts.[44,45] In Figure a, Mo2C/GCE exhibits excellent durability at each given potential. The catalytic experiments were performed six times using the same catalyst. Because of the strong durability of the catalyst (Figure b), the FE and NH3 yield are rather stable at ambient conditions for the NRR (at potential of −0.3 V versus RHE in Figure c). The values of NH3 produced are listed as an average of six catalytic runs. Meanwhile, 25 h long-term electrolysis at −0.3 V demonstrates nearly unchanged current density (Figure b). After electrolysis, such Mo2C/GCE exhibits an excellent NRR performance (Figure S10) before its real longevity of 58 h (Figure S11). After the durability test, the pH of the electrolyte is changed from 1 to 1.2. The negligible change is due to the usage of a Nafion membrane (Figure a) in the two-compartment cell, where H+ is supplied from the anode to cathode continuously during H+ consumption when the NRR occurs. Meanwhile, we also carried out an NRR electrocatalytic experiment in 0.05 M H2SO4 electrolyte. The results indicate that there are no variations in both conditions (Figure S12). XRD analysis (Figure S13) confirms the Mo2C nature, and XPS analysis (Figure S14) indicates that there are no changes of valence states of Mo after the NRR process. The SEM image and corresponding element images (Figure S15) further suggest that such a catalyst can maintain initial morphology and composition after NRR electrolysis. Furthermore, TEM and HRTEM images (Figure S16) indicate that the Mo2C porous nanorod is well-maintained after the long-term stability test. A N2 flow rate test was also carried out in this electrocatalysis process. As shown in Figure d, the approximate unchanged NH3 yields and FEs certify that N2 diffusion is non-rate-determined because of its independent reaction interface. Large-scale reactions have been performed by this catalyst for the NRR, demonstrating excellent catalytic performance (Figure S17).

Figure 3

(a) Time–current density curves of Mo2C/GCE for the NRR process in 0.1 M HCl solution. (b) Time–current density curve at a potential of −0.3 V using Mo2C/GCE for 25 h of electrolysis. (c) NH3 yields and corresponding FEs during six recycling tests. (d) NH3 yields and corresponding FEs of Mo2C/GCE under different flow rates of N2. (e) Curve of ammonia production vs reaction time at −0.3 V vs RHE. (f) 1H NMR spectra for the 15NH4+ standard sample and electrolytes after electrolysis using Ar and 15N2 as the feeding gas.

Figure S18 shows the NRR performance of the catalyst electrode by alternating 3 h cycles of N2-/Ar-saturated electrolyte solutions, respectively. The NH3 can only be detected in the electrocatalysis process of the N2-saturated HCl solution, while Ar-saturated electrolytes exhibit blank results, further ensuring the electrocatalytic NRR process on Mo2C/GCE. Moreover, the time-dependent experiment further confirms the results (Figure e).[46,47] For verification that the ammonia is derived from the N2 gas supplied, a 15N isotopic labeling experiment is carried out. In Figure f of the 1H nuclear magnetic resonance (1H NMR) spectra, the standard sample displays a doublet coupling for 15NH4+. 15N2 is used as the feeding gas with 15NH4+ formation, and Ar is used as the feeding gas without any NH4+ formation. These results provide vital evidence to ensure that the produced NH3 is electrocatalytically reduced by Mo2C.

For an understanding of the mechanism of the NRR process occurring on the catalyst Mo2C (121) surface, extensive first-principles calculations were performed to investigate the adsorption of NH species, the free energy profile, as well as the electronic structures of the critical steps. The results are presented in Figure . It is known that the N2 molecule adsorption and the first hydrogenation reaction (*N2 → *NNH) to open the N≡N bond are the critical steps to determine the NRR performance. Upon full relaxation, our results show that the N2 molecule is tilted on the exposed Mo atom of the Mo2C (121) surface with the elongated N—N bond length of 1.13 Å, yielding a large adsorption energy of 1.39 eV and short N—Mo distance (2.08 Å). The strong chemical adsorption of N2 implies that Mo2C would be quite active to catalyze the dissociation of the N2 molecule. The atom-projected density of state (DOS) (Figure b) shows different features of the two N atoms in the *N2 configuration. The N2 packet below the Fermi level is entirely covered by the Mo atom within −1.0 and 0 eV, verifying the Mo—N2 chemical bonding. However, the curve of the N1 atom below the Fermi level is negligible. The charge density difference of the *N2 state visualized in Figure d shows that the electron depletion occurs at the N1—N2 region and accumulates at the Mo—N2 bond, inducing the dissociation of the N2 molecule. The chemical bonding of the N2 atom to the outmost Mo atom leads to a significant decrease of the free energy of *N2 down by 0.91 eV under the zero electrical potential. Realizing the first hydrogenation step is of great importance for the NRR process. As shown in Figure a, the moderate free energy change of 0.74 eV suggests that it is feasible for this step on the Mo2C (121) surface to take place without the electrical potential. The DOS of the *NNH configuration in Figure c makes evident the chemical bonding interactions of H—N1—N2—Mo at the energy range from −1.0 to 0 eV. The amplified DOS of the N2 atom suggests the enhanced bonding interactions with the anchored Mo atom. The results are consistent with the charge density difference analysis in Figure e.

Figure 4

(a) Free energy profile for the NRR process under different electrical potentials. An asterisk (*) denotes as the adsorption site. The competitive processes are represented as the light lines. The inserted images are the reactions with uphill free energy changes. (b, c) Atom-projected DOS of the *N2 and *NNH configurations; (d, e) charge density difference of the *N2 and *NNH configurations, respectively. The isosurface level is 0.0013 e Å–3.

Sequentially, there are different pathways for other hydrogen atoms to go to the next hydrogenation steps. In principle, the second hydrogen might go either on the distal pathway to bond with the lower Mo-anchored N atom to form *NHNH, or on the alternating path to form *NNH2 with the upper H-bonded N atom. The free energy diagram shows that *NNH2 formation is an exothermic reaction and thus more energetically preferable, while it is uphill in free energy with 0.49 eV energy change to form *NHNH. Sequential addition of the third hydrogen to form *NNH3 is also favorable since it is steeply downhill in the free energy. Because of the relatively strong binding interactions with the exposed Mo atoms, NH3 needs an external energy of 0.76 eV to escape from the surface, which is the potential-determining step. Hydrogenation of the retained *N is feasibly realized to form *NH because the uphill free energy change is only 0.08 eV. Of note, the subsequent processes for *NH2 and *NH3 formation are downhill. The release of the second NH3 needs a high free energy change of 1.15 eV. The competitive reaction steps represented by the light black line might also take place under the experimental conditions because it is easy to overcome the 0.49 eV uphill free energy change for the *NNH → *NHNH step, and it is downhill in free energy from *NHNH to *NH2NH3. After the release of the first NH3 molecule, there is no difference in the rest of the steps in both pathways. As for the influence of the applied electric potential on the NRR process, the energy profiles under U = −0.1, −0.3, and −0.5 V are shown in Figure a. The results indicate that application of the electrical potential is beneficial for the decrease of all the uphill energy changes; e.g., the value of 0.74 eV without external electric potential for addition of the first H atom is reduced to 0.24 eV under the electric potential of U = −0.5 V. The uphill energy change of 0.49 eV on the alternating pathway (light black line) is gradually reduced to 0.39, 0.19, and 0 eV with the applied electrical potential U = −0.1, −0.3, and −0.5 V, respectively. Therefore, the external electric potential promotes the NRR process.

In summary, the Mo2C nanorod is proven experimentally as an efficient and durable catalyst for electrochemical conversion of N2 to NH3 with good selectivity under ambient conditions. Such Mo2C/GCE achieves a high FE of 8.13% and NH3 yield of 95.1 μg h–1 mg–1cat at potential of −0.3 V, with strong electrochemical durability in 0.1 M HCl electrolyte. This research not only supplies an efficient earth-abundant nanocatalyst toward NH3 electrosynthesis, but also opens up a new path to the rational design and development of Mo-based nanomaterials as effective electrocatalysts for artificial N2 fixation.[48]