Semi-metal 1T 0 phase MoS 2 nanosheets for promoted electrocatalytic nitrogen reduction

Herein,

), as one of the most important chemicals of industrial production, is not only widely used in the synthesis of fertilizers, fibers, and explosives, and so forth, 1-3 but is considered as a hopeful carbon-free energy carrier, so it can be used as a possible substitute to hydrogen. [4][5][6][7] Till now, the industrial-scale ammonia synthesis is dominated by the Haber-Bosch process 8 with high temperature and pressure. 9,10 Hence, develop and design a high-performance, cost-effective, and environmentally friendly alternative technology for the production of NH 3 is necessary.
Up till now, several strategies have been developed toward ammonia synthesis under ambient condition, including biological nitrogen fixation, 11,12 photochemical nitrogen reduction, [13][14][15] and electrochemical nitrogen reduction (NRR). 16,17 Among them, electrochemical NRR, as a novel ammonia synthesis route, has a higher ammonia yield than biological nitrogen fixation. Compared with photochemical nitrogen reduction reaction, it also has higher Faradaic efficiency (FE) and NH 3 yield. This is because, for the photochemical NRR, not all of the photos can be utilized efficiently due to multiple wavelengths and fast charge carrier recombination. 16 In addition, compared with the Haber-Bosch method, the electrochemical NRR has many advantages, such as using renewable solar, wind, or hydro energy-driven electricity, reducing the energy input by about 20%, decreasing the CO 2 emission, simplifying the reactor design, and easing the complicacy of ammonia production. 16 Despite many advantages of electrochemical NRR, achieving efficient NRR has been proved to be extremely challenging in practice. Among them, the grand challenge is that the catalyst surface that catalyzes the reduction of N 2 in the aqueous electrolyte also has a strong activity for the reduction of H 2 O to H 2 . More importantly, most of the protons and electrons in the electrochemical reduction system tend to shift toward the HER rather than the NRR, causing lower NH 3 yields and Faradaic efficiencies (FEs). 18 Therefore, developing a suitable catalyst for achieving efficient electrochemical NRR is needed.
Since 2016, noble metals (Au, Ag, Ru, Rh, etc.), transition metals and their derivatives (Ti, Nb, Bi, etc.), metalfree materials (organic conducting polymers and carbonaceous catalysts), single-atom catalysts, and their hybrids, have been developed as potential catalysts for electrochemical NRR. 19 Among them, noble metals have outstanding catalytic activity for electrocatalytic NRR. El-Sayed et al. reported that the Au nanocages can achieve a FE of 30.2% in 0.5 M LiClO 4 aqueous solution, 20 but their low abundance and high cost limit the possibility of its large-scale application. In contrast, low-cost and earth-abundant transition metals and their derivatives are studied as NRR electrocatalysts. 19 Considering the fact that in nature, the nitrogenase enzymes with FeMo cofactor as catalytic active sites can achieve electron efficiency of 70% under room temperature and pressure, 21 therefore Mobased nanomaterials are promising NRR electrocatalysts.
Molybdenum sulfide (MoS 2 ), as a typical layered twodimensional transition metal sulfide, has been widely used in industry as a hydrodesulphurization catalyst. With the rise of the field of electrocatalytic NRR, its performance in the direction of NRR began to be explored. Sun et al. demonstrated firstly that 2H-MoS 2 was active for NRR under ambient condition but presented a lower Faradaic efficiency (1.17%). The density functional theory (DFT) calculations display that the edge of MoS 2 is the electrocatalytic active site for NRR, similar to the case of HER. However, the basal plane of MoS 2 is inert for N 2 molecules. 22 Subsequently, defect-rich MoS 2 nanoflowers with excellent selectivity were developed to boost electrocatalytic N 2 reduction to NH 3 , which can attain a FE of 8.34% and a NH 3 yield of 29.28 μg h À1 mg À1 cat. in 0.1 M Na 2 SO 4 aqueous electrolyte. 23 In addition, Liu and coworkers introduced orbital hybridization by introducing S vacancies at the basal plane of MoS 2 , which activated the basal plane's catalytic activity and promoted the dissociation of N N bonds. Furthermore, the substitution of Mo atoms by Co doping reduces the Gibbs free energy of N 2 adsorption step and hydrogenation intermediate, thereby achieving a FE exceeding 10%. 24 Considering that the phase structure of the catalyst will affect its catalytic performance, our group previously loaded 1T-MoS 2 nanodots onto g-C 3 N 4 nanosheets, achieving a FE of 20.48% and a NH 3 yield of 29.97 μg h À1 mg À1 cat. 25 1T 0 -MoS 2 , as a derivation of 1T-MoS 2 , not only retains the physical and chemical properties of 1T-MoS 2 , but also has better stability. However, its performance on NRR has not been studied.
TiO 2 , as a common semiconductor functional material, has huge potential in the field of catalysis, 26,27 electrochemical energy storage, 28,29 and solar cell, 30 35 and more transition metal oxides as NRR electrocatalysts were developed, and the performance of TiO 2 in NRR began to be explored. Sun's group reported that the TiO 2 nanosheets array is efficient for electrochemical NRR with a FE of 2.50% at ambient condition. 36 Then, TiO 2 nanoparticles were loaded on the rGO with high conductivity, which presented a FE of 3.3% with the corresponding NH 3 yield of 15.13 μg h À1 mg À1 cat. in 0.1 M Na 2 SO 4 . 37 In addition, Xu and coworker prepared the TiO 2 nanoparticles on Ti 3 C 2 T x nanosheets for NRR, achieving a FE of 16.07% and the NH 3 yield of 32.17 μg h À1 mg À1 cat. 38 Thus, TiO 2 based NRR electrocatalysts that simultaneously achieve a large NH 3 yield and a high FE are still worthy of our research.
Herein, we report that 1T 0 -MoS 2 nanosheets loaded on urchin-like TiO 2 hollow nanospheres (HNSs) can achieve effectively the electrochemical conversion of N 2 to NH 3 . In 0.1 M Na 2 SO 4 , the catalyst exhibits excellent NRR electrocatalytic performance with the highest NH 3 yield of 29.62 μg h À1 mg À1 cat. at À0.75 V versus RHE. Besides, the highest FE of 24.9% is achieved at À0.65 V versus RHE for 1T 0 -MoS 2 /TiO 2 HNSs composites. Notably, 1T 0 -MoS 2 /TiO 2 HNSs electrocatalyst also displays outstanding selectivity and stability to NH 3 , without the formation of by-products such as N 2 H 4 . 15 N isotopic labeling test is investigated to confirm the nitrogen source of produced NH 3 .

| RESULTS AND DISCUSSION
The preparation of 1T 0 -MoS 2 /TiO 2 HNSs composites is described in Scheme 1. Firstly, TiO 2 SNSs were synthesized by direct hydrolysis of titanium isopropoxide in an ethanol-acetonitrile mixed solution containing trace of H 2 O and NH 3 (Scheme 1A), where NH 3 acts as an aggregation driver and morphological controller to conduct titanium isopropoxide to assemble spherical agglomerates. 39,40 Subsequently, the prepared TiO 2 SNSs were fluorinated through ligand exchange between surface hydroxyl groups and F À , after the F À are evenly distributed throughout the spheres, PVP was introduced into the suspension of TiO 2 SNSs and coated TiO 2 SNSs. 39 Then, the F-TiO 2 /PVP was targeted etched by hydrothermal reaction, assembling an urchin-like hollow nanosphere structure (Scheme 1B). Finally, 1T 0 -MoS 2 nanosheets were loaded on the nanothorns by the hydrothermal reaction on TiO 2 HNSs as the substrate, forming 1T 0 -MoS 2 /TiO 2 HNSs composites (Scheme 1C).
The phase composites and crystal structure of TiO 2 SNSs, urchin-like TiO 2 HNSs and 1T 0 -MoS 2 /TiO 2 HNSs composites with various amount of 1T 0 -MoS 2 (1, 3, 5, 7, and 10 wt%) were studied by XRD ( Figure 1A). The XRD pattern of TiO 2 SNSs obtained by hydrolysis using TTIP S C H E M E 1 Schematic illustration of the preparation of 1T 0 -MoS 2 /TiO 2 HNSs composites as a precursor does not show any diffraction peaks, indicating that the TiO 2 SNSs structure is amorphous. After hydrothermal targeted etching and calcination, the obtained urchin-like TiO 2 HNSs show excellent crystallinity, and the strongest diffraction peak at~25.32 can be attributed to the (101) plane of TiO 2 . 29 26 After loading the 1T 0 -MoS 2 nanosheets by hydrothermal reaction, the diffraction pattern did not change significantly. In addition, all of the 1T 0 -MoS 2 /TiO 2 HNSs composites do not present obvious signals that can be assigned to MoS 2 , which may be caused by a small amount and highly dispersed of 1T 0 -MoS 2 . 26 As shown in Figure 41 Regarding the O 1 s spectrum, the two peaks at~530.08 and~531.65 eV can be assigned to the Ti O band and the OH adsorbed on the surface of TiO 2 ( Figure 1D). 42 The Mo 3d spectrum can be divided into five peaks ( Figure 1E). The two main peaks located at~232.2 and~228.9 eV can be ascribed to the Mo 3d 3/2 and Mo 3d 5/2 of 1T 0 -MoS 2 , indicating the formation of 1T 0 phase MoS 2 . 43 Furthermore, the two small peaks at~233.2 and~230.3 eV correspond to Mo 3d 3/2 and Mo 3d 5/2 of 2H-MoS 2 , respectively. 44 The extra peak at~235.3 eV corresponds to Mo 6+ of MoO 3 . 33 Similarly, the two main peaks located at~162.7 and~161.3 eV are attributed to S 2p 1/2 and S 2p 3/2 of 1T 0 -MoS 2 ( Figure 1F), respectively. While the two weak peaks at~163.8 and~162.3 eV are assigned to S 2p 1/2 and S 2p 3/2 of 2H-MoS 2 , respectively. 45 The proportion of 1T 0 phase in the MoS 2 component is calculated~63.6% on the basis of the deconvolution of the XPS spectra, indicating the 1T 0 phase in the composite is a major phase.
The microstructures and morphologies of TiO 2 SNSs, urchin-like TiO 2 HNSs and 1T 0 -MoS 2 /TiO 2 HNSs composites are attained using SEM and TEM (Figures 2 and  3). The prepared TiO 2 SNSs exhibit the uniform sphere and the size is around 500 nm. The surface of TiO 2 SNSs is rough and granular protrusions can be clearly seen, indicating that TiO 2 SNSs are composed of smaller size nanoparticles (Figure 2A,B). As shown in Figure 2C,D, after TiO 2 SNSs are targeted etched by hydrothermal reaction, the microstructure of TiO 2 SNSs is transformed into a hollow structure. In addition, the surface of the TiO 2 HNSs exhibits an urchin-like morphology due to the spontaneous reorganization of the structure during the reaction, which is conducive to the MoS 2 loading and the contact of N 2 molecules. After loading 1T 0 -MoS 2 with urchin-like TiO 2 HNSs as the substrate, the hollow structure of TiO 2 HNSs did not collapse, but the nanothorns on the surface changed to granular or sheets shape ( Figure 2E,F). In addition, the EDS mapping images of 1T 0 -MoS 2 /TiO 2 HNSs composites show the MoS 2 are homogeneously distributed on urchin-like TiO 2 HNSs ( Figure S1). Figure 3A,B show that the TiO 2 HNSs exhibits typical hollow structure characteristics, and a number of nanothorns grow on the surface. The whole displays a good urchin-like hollow structure, which can improve significantly TiO 2 specific surface area. The TEM images of 1T 0 -MoS 2 /TiO 2 HNSs composites are displayed in Figure 3C, D. As shown in Figure 3C that TiO 2 HNSs still maintain a hollow structure after loading 1T 0 -MoS 2 , but the urchin-like morphologies changes to a sheet shape. A closer scrutiny of the surface morphologies image clearly shows the size of MoS 2 nanosheets is 20~30 nm ( Figure 3D). The HRTEM images of 1T 0 -MoS 2 /TiO 2 HNSs composites are shown in Figure 3E,F. The internal composition of nanosheets is TiO 2 ( Figure 3E) and the lattice spacing of 0.35 and 0.28 nm can assign to the (101) and (111) planes of TiO 2 , respectively. 39 In addition, the typical fold-like morphology attributed to MoS 2 can be observed at the edge position. 46 The typical triangle arrangement of Mo atoms indicates the presence of 1T 0 -MoS 2 ( Figure 3F). 25 The electrocatalytic nitrogen fixation activity of 1T 0 -MoS 2 /TiO 2 HNSs composites was measured in an H-type cell. The quantities of the produced NH 3 and possible byproduct N 2 H 4 were measured by the indophenols blue and Watt/Chrisp method ( Figure S2). In order to preliminary test the electrocatalytic NRR performance of 1T 0 -MoS 2 /TiO 2 HNSs composites, the LSV curves are recorded in N 2 -and Ar-saturated 0.1 M Na 2 SO 4 electrolytes ( Figure S3a). There is a difference in current density between the LSV curves obtained under the two test conditions in the potential range of À0.65 to À0.95 V versus RHE, which indicates that 1T 0 -MoS 2 /TiO 2 HNSs exhibit NRR catalytic activity in this potential range. Similar to previous studies, 25,47 two LSV curves recorded in Ar-and N 2 -saturated electrolyte gradually coincide owing to the drastic HER dominant the cathode reaction. Figure 4A shows the UV-Vis absorption spectra of electrolytes attained at various potentials, indicating the nitrogen fixation effect of 1T 0 -MoS 2 /TiO 2 HNSs composites reaches the best at À0.75 V. The specific results are displayed in Figure 4B. The highest NH 3 yield of 29.62 μg h À1 mg À1 cat. and corresponding FE of 12.4% is obtained at À0.75 V for 7.0 wt% 1T 0 -MoS 2 /TiO 2 HNSs composites.
Besides, the NH 3 yield of 12.50 μg h À1 mg À1 cat. and corresponding highest FE of 24.9% is obtained at À0.65 V for 7.0 wt% 1T 0 -MoS 2 /TiO 2 HNSs composites. With the applied potential becomes more negative, both NH 3 yield and FE decrease sharply to 6.94 μg h À1 mg À1 cat. and 0.49% (À0.90 V), which can also be attributed to the intense HER at lower applied potential, consisting with the results of LSV curves. 48 In order to prevent the cathode electrolyte from being affected by NH 3 escape and atmosphere during the reaction, an in-line acid trap is installed outside of the cathode chamber and tests the NH 3 concentration. The results show that the presence of NH 3 could not be detected in-line acid trap ( Figure S4a). Furthermore, considering that there may be slight NH 3 in the supplied gas, we also test the NH 3 content in the electrolyte before NRR. The corresponding UV-Vis absorption spectrum showed that NH 3 is not generated in the electrolyte after supplying N 2 for 30 min, indicating that the experimental results were hardly interfered by the supplied gas ( Figure S4b). Furthermore, to verify that the NH 3 detected in the cathode chamber is produced by electrocatalytic NRR, the reaction is performed in Ar-saturated electrolyte at À0.75 V and in N 2 -saturated electrolyte at open circuit potential, respectively. Figure S4c and d compare the UV-Vis absorption spectra of the electrolyte solutions acquired under the above two test conditions. The results show that the presence of NH 3 is not detected under the above two conditions, indicating that the previously detected NH 3 originated from electrocatalytic nitrogen reduction of 1T 0 -MoS 2 /TiO 2 HNSs composites. In view of the fact that electrocatalytic nitrogen fixation is a multi-electron reaction, it is possible to produce other products except for NH 3 , therefore, we also examine the possible by-products N 2 H 4 , and the results indicate that 1T 0 -MoS 2 /TiO 2 HNSs composites have the excellent selectivity ( Figure S5).
Stability is another important parameter to evaluate the property of catalysts, which was investigated through recycling and chronoamperometric measurements. Timedependent current density curves at different potentials in N 2 -saturated 0.1 M Na 2 SO 4 are presented in Figure S6. Remarkably, the current densities keep steady under higher potential (À0.65 to À0.80 V), demonstrating 1T 0 -MoS 2 /TiO 2 HNSs composites have the excellent chemical stability. The NH 3 yield and corresponding FEs at À0.75 V during the recycling test is shown in Figure 5A, the results show the NH 3 yield and corresponding FEs do not decline obviously after eight recycling tests, which also indicates the prepared 1T 0 -MoS 2 /TiO 2 HNSs composites has admirable chemical stability. In addition, no obvious current density loss can be found during the long-term electrochemical nitrogen reduction for 24 h at À0.75 V ( Figure 5B), which further displays the good stability of 1T 0 -MoS 2 /TiO 2 HNSs composites. The NH 3 yield increases proportionally with time ( Figure S7), demonstrating the efficient NRR activity of 1T 0 -MoS 2 /TiO 2 HNSs composites. In addition, XRD analysis confirms that 1T 0 -MoS 2 /TiO 2 HNSs composites present no F I G U R E 4 (A) UV-Vis absorption spectra of 0.1 M Na 2 SO 4 electrolyte at different potentials for 2 h; (B) NH 3 yields and FEs of 7.0 wt% 1T 0 -MoS 2 /TiO 2 HNSs composites at a series of potentials observable crystalline phase change ( Figure S8) and SEM image shows the structure of 1T 0 -MoS 2 /TiO 2 HNSs composites remain stable after NRR electrolysis ( Figure S9).
The effect of 1T 0 -MoS 2 content on the property of electrocatalytic nitrogen fixation was further explored. As shown in Figure 6A, the NH 3 yield and FE of pure TiO 2 HNSs are only 9.26 μg h À1 mg À1 cat. and 5.76%, significantly smaller than 1T 0 -MoS 2 /TiO 2 HNSs composites, indicating that 1T 0 -MoS 2 plays an important role in the NRR. As the loading of 1T 0 -MoS 2 increases from 0 to 10.0 wt%, the electrocatalytic NRR activity of 1T 0 -MoS 2 / TiO 2 HNSs composites is expressively improved, and the highest NH 3 yield and FE are simultaneously attained at a loading of 7.0 wt% 1T 0 -MoS 2 . In addition, we also explored the influence of N 2 flow rate on the NRR performance ( Figure 6B). The NH 3 yield and corresponding FE are stable with no change in different N 2 flow rates, which shows that N 2 diffusion is not the rate-determining step in the NRR. To confirm NH 3 production originating from N 2 , 15 N isotopic labeling experiment is measured ( Figure S10). 1  DFT calculation is tested to explore the origin of the enhanced activities of 1T 0 -MoS 2 /TiO 2 toward NRR. The most stable configurations of reaction intermediates on the 1T 0 -MoS 2 /TiO 2 surface are shown in Figure 7A, and the corresponding structures on the pristine TiO 2 surface are included in Figure S11. The NRR activity can be evaluated through the Gibbs free energy diagram, where the most favorable reaction pathway is the alternating pathway for both structural models. The distal pathway is also investigated and shown in Figure S12 for comparison. To comply with the experimental conditions, free energies were calculated at pH = 7 and  Figure 7B, the reduction of *N 2 to *NNH is the potential-determining step (PDS) for both structural models. However, the energy barrier for the reduction of *N 2 on the 1T 0 -MoS 2 /TiO 2 surface (0.84 eV) is lower than that on the pristine TiO 2 surface (1.48 eV), which indicates that the 1T 0 -MoS 2 on the 1T 0 -MoS 2 /TiO 2 surface makes the activation and further reduction of *N 2 more thermodynamically favorable than pristine TiO 2 , contributing to better NRR catalytic activities.
The excellent electrocatalytic NRR performances of 1T 0 -MoS 2 /TiO 2 HNSs composites are mainly attributed to the following points: Firstly, 1T 0 -MoS 2 nanosheets retain the high electrical conductivity of 1T-MoS 2 , which is favorable to the electron transfer in the process of nitrogen fixation. In addition, there are abundant active sites at the edge position and plane, which can be used to activate nitrogen molecules. Besides, the urchin-like structure of TiO 2 HNSs provides a large specific surface area, which is helpful in the adsorption of nitrogen molecules during the reaction. Finally, the semiconducting properties of TiO 2 can effectively inhibit the HER during nitrogen fixation (Figure 8).
F I G U R E 8 Schematic illustration of the NRR mechanism of 1T 0 -MoS 2 /TiO 2 HNSs composites F I G U R E 7 (A) Most stable configurations of reaction intermediates on the 1T 0 -MoS 2 surface along the NRR alternating pathway (color notation: purple-Mo, gray-Ti, yellow-S, red-O, blue-N, white-H). (B) Gibbs free energy diagram of the NRR alternating pathway on TiO 2 and 1T 0 -MoS 2 /TiO 2 surfaces at pH = 7 and U RHE = À0.65 V

| CONCLUSION
In summary, we constructed urchin-like TiO 2 HNSs by hydrothermal targeted etching, and used it as a substrate to loading 1T 0 -MoS 2 nanosheets as an effective NRR electrocatalyst. The 1T 0 -MoS 2 /TiO 2 HNSs composite attains the highest NH 3 yield of 29.62 μg h À1 mg À1 cat. and corresponding FE of 12.4% at À0.75 V. Besides, the NH 3 yield of 12.50 μg h À1 mg À1 cat. and corresponding highest FE of 24.9% is achieved at À0.65 V for 1T 0 -MoS 2 /TiO 2 HNSs composites. Meanwhile, the 1T 0 -MoS 2 /TiO 2 HNSs composites also display excellent selectivity and stability. DFT calculation reveals that the 1T 0 -MoS 2 on the surface of 1T 0 -MoS 2 /TiO 2 makes the activation and further reduction of *N 2 more thermodynamically favorable than pristine TiO 2 , contributing to better NRR catalytic activities. 15 N isotopic labeling experiment indicates that NH 3 is yield from NRR by 1T 0 -MoS 2 /TiO 2 .

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of this article.