Molybdenum-Modified Titanium Dioxide Nanotube Arrays as an Efficient Electrode for the Electroreduction of Nitrate to Ammonia

Electrochemical nitrate reduction (NO3−RR) has been recognized as a promising strategy for sustainable ammonia (NH3) production due to its environmental friendliness and economical nature. However, the NO3−RR reaction involves an eight-electron coupled proton transfer process with many by-products and low Faraday efficiency. In this work, a molybdenum oxide (MoOx)-decorated titanium dioxide nanotube on Ti foil (Mo/TiO2) was prepared by means of an electrodeposition and calcination process. The structure of MoOx can be controlled by regulating the concentration of molybdate during the electrodeposition process, which can further influence the electron transfer from Ti to Mo atoms, and enhance the binding energy of intermediate species in NO3−RR. The optimized Mo/TiO2-M with more Mo(IV) sites exhibited a better activity for NO3−RR. The Mo/TiO2-M electrode delivered a NH3 yield of 5.18 mg h−1 cm−2 at −1.7 V vs. Ag/AgCl, and exhibited a Faraday efficiency of 88.05% at −1.4 V vs. Ag/AgCl. In addition, the cycling test demonstrated that the Mo/TiO2-M electrode possessed a good stability. This work not only provides an attractive electrode material, but also offers new insights into the rational design of catalysts for NO3−RR.


Introduction
Ammonia (NH 3 ) is essential in various industrial sectors, such as chemical production, refrigeration, and pharmaceuticals [1,2].It also serves as a carbon-free fuel with a high hydrogen density (the NH 3 molecule has a H content of 17.75% by weight), making it easily storable and transportable through existing infrastructures [3].The Haber-Bosch process for NH 3 production stands as the most significant inventions in chemical engineering.It relies on high-temperature (400-600 • C) and high-pressure (200-350 atm) reactions between N 2 and H 2 with a suitable catalyst [4][5][6].Apart from the energy required for heating and pumping, the high-purity H 2 utilized in this process is predominantly produced from natural gas, resulting in substantial greenhouse gas emissions [7,8].Hence, there is considerable importance in developing a sustainable and environmentally friendly strategy for NH 3 production.Electrocatalytic approaches are recognized for their energy-saving and environmentally benign nature, distinguished by simple equipment, mild state, high efficiency, and immense potential for application on a large scale [9,10].
Among these approaches, NH 3 production from N 2 and H 2 O has garnered significant attention [11,12].However, the N≡N bond energy is notably high at 941 kJ mol −1 , and N 2 only dissolves sparingly in water [13].Consequently, the energy utilization of nitrogen reduction reactions in the aqueous environment is limited, with NH 3 production yields up to three orders of magnitude lower than those of the Haber-Bosch process [14].Considering this, researchers have explored other nitrogen-containing compounds as NH 3 synthesis sources and identified NO 3 − as a promising candidate.NO 3 − exhibits good solubility in water, and the N=O bond energy is relatively low at 204 kJ mol −1 [15,16].In addition, NO 3 − is also a significant pollutant due to human activities, including excessive nitrogen fertilizer usage, fossil fuel combustion, and wastewater discharge, contributing to increased nitrate levels in water [17].The World Health Organization (WHO) recommends that the level of nitrates in drinking water not exceed 50 mg L −1 [18].Once nitrates are ingested by the human body, they undergo metabolism in the digestive system where they are converted into nitrites, thereby presenting a potential carcinogenic hazard [19].Therefore, utilizing NO 3 − as a nitrogen source for NH 3 synthesis not only enhances energy utilization efficiency and reduces greenhouse gas emissions, but also addresses the issue of NO 3 − pollution in the environment.
Recently, the electrocatalytic nitrate synthesis of ammonia (NO 3 − RR) has been investigated using metals such as Cu, Fe, Pt, Ag, and Mo [20][21][22][23][24].Among them, Mo has attracted much attention due to its low price and good catalytic performance.Moreover, the nitrate reduction reaction process in nature is facilitated by an enzymatic cascade, where the Mo(IV) cofactor of nitrate reductase catalyzes the conversion of NO 3 − to NO 2 − [25].This step plays a crucial role in determining the overall reaction rate [26].Wang  − and contributes to the conversion of *NO to *NOH [29].Thus, MoO x with a Mo(IV) site can accelerate the conversion of electrochemical NO 3 − to NH 3 and limit NO 2 − generation, resulting in the efficient production of NH 3 .
Generally, the nanostructure is favorable to enlarge the interface between electrode and electrolyte, and the construction of a nanocomposite can endow the catalyst with more active sites [30][31][32].Xiong et al. prepared platinum nanoparticles embedded on nickel oxide nanosheets, which can serve as a electrocatalyst for boosting NO 3 − transfer [33].Wang et al. constructed ultra-small iron oxide nanoparticles on carbon nanotubes, which can enlarge the active surface area and accelerate ion transfer [34].All these studies have demonstrated that nanomaterials have the effect of promoting full solid-liquid phase contact.As a nontoxic and abundant material, TiO 2 is one of the hot spots in photochemistry and electrochemistry.TiO 2 nanotube arrays (TNTAs) can be easily prepared and serve as the substrate with a large surface area to support active materials [35][36][37].Inspired by this, we anticipated that decorating MoO x onto the surface of TNTAs to form a nanocomposite structure can enlarge the active area and enhance the ion transfer from electrolyte to electrode.
In this work, MoO x -loaded TNTAs was prepared on Ti foil (Mo/TiO 2 ) by means of an electrodeposition and calcination process.This can directly serve as the electrode for NO 3 − RR.The TNTAs as the substrate can offer a large surface area to support MoO x .The Mo/TiO 2 nanocomposite structure was confirmed by SEM and TEM measurement.Notably, XPS measurement further revealed the electron transfer behavior between Ti and Mo atoms.The electron transfer can be modulated by regulating the concentration of molybdate during the electrodeposition process.By optimizing the electrodeposition conditions, the obtained Mo/TiO 2 -M delivered a high NH 3 yield of 5.18 mg h −1 cm −2 at −1.7 V vs. Ag/AgCl and a Faraday efficiency of 88.05% at −1.4 V vs. Ag/AgCl in 0.1 M NaNO 3 solution.It also maintained a Faraday efficiency of over 80% under five consecutive cycle tests.This work not only presents a highly promising electrode material, but also offers new insights into the rational design of Mo-based nanocomposites for NO 3 − RR.

Morphological and Structural Analysis of Catalysts
Figure 1 shows the two-step process to fabricate the Mo/TiO 2 electrode.Firstly, TiO 2 nanotube arrays were formed on Ti foil (TNTAs) by using anodization process (Figure 1a) [38], and then Mo was further loaded onto the TNTAs using the electrodeposition method and finally annealed under 3% H 2 /Ar atmosphere to prepare Mo/TiO 2 (Figure 1b).The Mo/TiO 2 can directly serve as the electrode for NO 3 − RR.

Morphological and Structural Analysis of Catalysts
Figure 1 shows the two−step process to fabricate the Mo/TiO2 electrode.Firstly, T nanotube arrays were formed on Ti foil (TNTAs) by using anodization process (Figure [38], and then Mo was further loaded onto the TNTAs using the electrodeposition meth and finally annealed under 3% H2/Ar atmosphere to prepare Mo/TiO2 (Figure 1b).T Mo/TiO2 can directly serve as the electrode for NO3 − RR.As shown in Figure 2a, the nanotube structure of the TNTAs was observed, a TNTAs with a tube diameter of 100 nm uniformly covered the Ti foil.The EDS mappi images (Figure 2b,c) show that Ti and O elements are distributed on TNTAs.TNTAs w a nanotube array structure can be a promising substrate to support active materials, a can offer a further pathway for ion transfer from electrolyte to electrode [35].It was o served that the particle size of molybdenum oxide particles grown on the surface of nanotubes gradually decreased with the increase in molybdate concentration in the el trodeposition solution.The SEM image of Mo/TiO2−L (Figure 2d) displays some nanop ticles on TNTAs.With increased molybdate concentration, relatively small particles c be found on the surface of TNTAs, as shown in Figure 2e,f, which is ascribed to the kinet of electrochemical deposition.In general, the concentration of the electrolyte is prop tional to the uniformity of the electrodeposition.A high concentration of molybdate lows a sufficient amount of Mo species to be adsorbed onto the electrode surface, th resulting in the uniform growth of the Mo layer.In the electrolyte with low molybd concentration, the nucleation of Mo is controlled by the diffusion of Mo species.This because there are insufficient Mo species near the electrode region, causing the Mo spec to preferentially adsorb on the nucleated particles with a larger radius of curvature.The fore, the particles with a larger size can be formed for Mo/TiO2−L (Figure 2d).To furth confirm the nanocomposite structure, TEM measurement was carried out.The Mo/T active material was scraped off and dispersed in ethanol for TEM measurement.As show in Figure 2g, the nanotube structure can be clearly observed with a wall thickness of nm.Notably, some nanoparticles can be found inside and outside the nanotube with size of 25-30 nm.The high−magnification TEM image (Figure 2h inset) shows a cl As shown in Figure 2a, the nanotube structure of the TNTAs was observed, and TNTAs with a tube diameter of 100 nm uniformly covered the Ti foil.The EDS mapping images (Figure 2b,c) show that Ti and O elements are distributed on TNTAs.TNTAs with a nanotube array structure can be a promising substrate to support active materials, and can offer a further pathway for ion transfer from electrolyte to electrode [35].It was observed that the particle size of molybdenum oxide particles grown on the surface of the nanotubes gradually decreased with the increase in molybdate concentration in the electrodeposition solution.The SEM image of Mo/TiO 2 -L (Figure 2d) displays some nanoparticles on TNTAs.With increased molybdate concentration, relatively small particles can be found on the surface of TNTAs, as shown in Figure 2e,f, which is ascribed to the kinetics of electrochemical deposition.In general, the concentration of the electrolyte is proportional to the uniformity of the electrodeposition.A high concentration of molybdate allows a sufficient amount of Mo species to be adsorbed onto the electrode surface, thus resulting in the uniform growth of the Mo layer.In the electrolyte with low molybdate concentration, the nucleation of Mo is controlled by the diffusion of Mo species.This is because there are insufficient Mo species near the electrode region, causing the Mo species to preferentially adsorb on the nucleated particles with a larger radius of curvature.Therefore, the particles with a larger size can be formed for Mo/TiO 2 -L (Figure 2d).To further confirm the nanocomposite structure, TEM measurement was carried out.The Mo/TiO 2 active material was scraped off and dispersed in ethanol for TEM measurement.As shown in Figure 2g, the nanotube structure can be clearly observed with a wall thickness of 15 nm.Notably, some nanoparticles can be found inside and outside the nanotube with a size of 25-30 nm.The high-magnification TEM image (Figure 2h inset) shows a clear boundary between nanoparticles and nanotubes.A highly ordered fringe with an interplanar distance of 0.176 nm can be observed in the nanotube region (Figure 2h inset), which corresponds to the (111) plane of the TiO 2 phase [39].Moreover, the high-magnification TEM image of the nanoparticles (Figure 2i) displays a lattice fringe with a d-space of 0.324 nm, suggesting the (−313) plane of MoO 2 .The above results indicate the successful formation of a Mo/TiO 2 nanocomposite structure by means of the electrochemical deposition process [40].
boundary between nanoparticles and nanotubes.A highly ordered fringe with an i planar distance of 0.176 nm can be observed in the nanotube region (Figure 2h in which corresponds to the (111) plane of the TiO2 phase [39].Moreover, the high-mag cation TEM image of the nanoparticles (Figure 2i) displays a lattice fringe with a d−s of 0.324 nm, suggesting the (−313) plane of MoO2.The above results indicate the succe formation of a Mo/TiO2 nanocomposite structure by means of the electrochemical dep tion process [40].200) plane of MoO3, respectively.The peaks of TiO2 can sti observed after Mo loading, indicating that the electrodeposition process cannot influ the crystal structure of TiO2.In addition, the intensity of new peaks correspondin MoOx decreased with the increase in molybdate concentration.This could be due to smaller grain size of MoOx, which is ascribed to the SEM results.In addition, the g size of samples was calculated using the Scherrer equation, D = Kγ/(Bcosθ), where Scherrer's constant (0.89), γ is the wavelength of the X-rays (1.54056 Å), B is the half−p height width, and θ is the Bragg angle.Therefore, the grain size of Mo/TiO Mo/TiO2−M, and Mo/TiO2−H was calculated to be 30.8,26.8, and 22.6 nm, respectivel In addition, the intensity of new peaks corresponding to MoO x decreased with the increase in molybdate concentration.This could be due to the smaller grain size of MoO x , which is ascribed to the SEM results.In addition, the grain size of samples was calculated using the Scherrer equation, D = Kγ/(Bcosθ), where K is Scherrer's constant (0.89), γ is the wavelength of the X-rays (1.54056 Å), B is the half-peak height width, and θ is the Bragg angle.Therefore, the grain size of Mo/TiO 2 -L, Mo/TiO 2 -M, and Mo/TiO 2 -H was calculated to be 30.8,26.8, and 22.6 nm, respectively.

Electrocatalytic Performance of Electrodes for Mo/TiO2
CV, LSV, and EIS measurements were conducted in an H−type three−electrode system in 0.05 M Na2SO4 electrolyte to compare the electrochemical performance of different samples.The CV tests were firstly conducted at different scan rates (Figure S2) to determine the electric double−layer capacitances (Cdl).Notably, the electrochemical active surface area (ECSA) was positively correlated with Cdl [50,51].The Cdl for TiO2, Mo/TiO2−L, Mo/TiO2−M, and Mo/TiO2−H were 0.93, 9.65, 16.43, and 11.01 mF cm −2 , respectively (Figure 4a), demonstrating that Mo/TiO2−M has a significantly greater ECSA than other concentrations.The greater ECSA for Mo/TiO2−M can be ascribed to its appropriate particle size.As shown in Figure 2d-f, the high concentration of molybdate results in the formation of a dense MoOx layer on the surface of TNTAs, thus leading to the reduction of the surface area for the nanoarray electrode.Conversely, Mo/TiO2−L, prepared with a low molybdate concentration, exhibits a larger size of particles, which decreases the solid−liquid contact surface.Therefore, the result indicates that Mo/TiO2−M could behave better for NO3 − RR.This result can also be proven by the EIS measurement.Figure 4b shows the Nyquist plots of TiO2, Mo/TiO2−L, Mo/TiO2−M, and Mo/TiO2−H.In the high frequency region, the charge transfer resistance (Rct) and the electrolyte contact resistance (Re) are reflected by the intercepts of the radius of the high frequency arc on the real axis and the Nyquist plots, respectively [52].The Rct of the Mo/TiO2−M electrode is much smaller than that of the other electrode, indicating faster charge transition [53].Moreover, since TiO2 is The surface chemistry of Mo/TiO 2 -M and Mo/TiO 2 -H was further investigated using XPS.The survey scan XPS spectrum shows the photoelectron lines with binding energies (BEs) at three peaks of 532, 460, and 233 eV corresponding to the O 1s, Ti 2p, and Mo 3d signals (Figure S1).As shown in Figure 3b, the peaks of O 1s with BEs at 530.38 and 532.85 eV are attributed to lattice oxygen and physically adsorbed oxygen, respectively [41,42].In addition, the peak with BE at 531.74 eV corresponds to chemically adsorbed oxygen.This could be due to the fact that the oxygen defects on the surface after H 2 treatment, and the positively charged defection sites can adsorb O 2 to become reactive oxygen species [34,43].As displayed in Figure 3c, the peaks with BEs at 460.18 and 464.90 eV correspond to the spin-orbit splitting peak of Ti 2p 1/2 and Ti 2p 3/2 , respectively, proving the existence of Ti 4+ [44].Notably, compared with the Mo/TiO 2 -H sample, the Ti 2p 1/2 and Ti 2p 3/2 peaks display positive shifts of 0.18 and 0.27 eV for Mo/TiO 2 -M, indicating that the structure of MoO x could influence the local chemical states of Ti 4+ .According to a previous report, oxygen defect could be formed after H 2 treatment, which can reduce Ti 4+ into Ti 3+ [45].However, almost no Ti 3+ was detected in the two samples, possibly due to the low amount of Ti 3+ .The positive shift could be explained by the grain size of MoO x , in which more interaction could occur between Mo and Ti atoms.Figure 3d demonstrates the Mo 3d XPS spectra of Mo/TiO 2 -M and Mo/TiO 2 -H.The Mo 3d in the samples are consist of three spin-orbit splitting components.The two peaks at BEs of 229.64 and 233.37 eV are attributed to the Mo 2d 5/2 and Mo 2d 3/2 of Mo 2+ [46].Additionally, the peaks at BEs of 230.91, 234.26, 232.54, and 235.74 eV are attributed to the Mo 2d 5/2 and Mo 2d 3/2 of Mo 4+ and the Mo 2d 5/2 and Mo 2d 3/2 of Mo 6+ , respectively [46,47].Notably, negative shifts in the peak position of Mo/TiO 2 -M can be found compared with Mo/TiO 2 -H.This indicates the electron transfer to Mo atoms.Therefore, the existence of electron transfer from Ti to Mo atoms is speculated.According to the d-band center theory proposed by Norskov et al. [48], the active site of Mo in Mo/TiO 2 -M could exhibit a rising d-band center compared with Mo/TiO 2 -H.This suggested that less electrons would fill the antibonding orbitals, thus increasing the adsorption energy between intermediate species and active site [49].Hence, the binding energy of intermediate species in NO 3 − RR could be adjusted by rationally regulating the grain size of MoO x .Furthermore, the relative amounts of Mo in each valence state were calculated through the peak areas.As displayed in Figure 3d, it was found that Mo(IV) is more abundant in Mo/TiO 2 -M.This implies that the structure of MoO x (Figure 2d-f and 3a) and the valence of Mo (Figure 3d) can be influenced by rationally designing the concentration of molybdate during electrodeposition.

Electrocatalytic Performance of Electrodes for Mo/TiO 2
CV, LSV, and EIS measurements were conducted in an H-type three-electrode system in 0.05 M Na 2 SO 4 electrolyte to compare the electrochemical performance of different samples.The CV tests were firstly conducted at different scan rates (Figure S2) to determine the electric double-layer capacitances (C dl ).Notably, the electrochemical active surface area (ECSA) was positively correlated with C dl [50,51].The C dl for TiO 2 , Mo/TiO 2 -L, Mo/TiO 2 -M, and Mo/TiO 2 -H were 0.93, 9.65, 16.43, and 11.01 mF cm −2 , respectively (Figure 4a), demonstrating that Mo/TiO 2 -M has a significantly greater ECSA than other concentrations.The greater ECSA for Mo/TiO 2 -M can be ascribed to its appropriate particle size.As shown in Figure 2d-f, the high concentration of molybdate results in the formation of a dense MoO x layer on the surface of TNTAs, thus leading to the reduction of the surface area for the nanoarray electrode.Conversely, Mo/TiO 2 -L, prepared with a low molybdate concentration, exhibits a larger size of particles, which decreases the solid-liquid contact surface.Therefore, the result indicates that Mo/TiO 2 -M could behave better for NO 3 − RR.This result can also be proven by the EIS measurement.Figure 4b shows the Nyquist plots of TiO 2 , Mo/TiO 2 -L, Mo/TiO 2 -M, and Mo/TiO 2 -H.In the high frequency region, the charge transfer resistance (R ct ) and the electrolyte contact resistance (R e ) are reflected by the intercepts of the radius of the high frequency arc on the real axis and the Nyquist plots, respectively [52].The R ct of the Mo/TiO 2 -M electrode is much smaller than that of the other electrode, indicating faster charge transition [53].Moreover, since TiO 2 is a semiconductor, its conductivity is the worst, resulting in the smallest R ct .At low frequency, the perpendicularity of the lines of Mo/TiO 2 -L and Mo/TiO 2 -H are as similar as TiO 2 , indicating that their ion diffusion is close.
The LSV curves of the Mo/TiO 2 -M catalyst were tested in the electrolyte with and without nitrate-N at a scan rate of 5 mV s −1 to characterize whether it has NO 3 − RR catalytic properties.As shown in Figure 5a, it is evident that the current density of the LSV curve with NO 3 − in 0.05 M Na 2 SO 4 is larger than that of the other one, ranging from −1.1 V vs. Ag/AgCl to −1.4 V vs. Ag/AgCl, which proves that Mo/TiO 2 -M has NO 3 − RR properties.Moreover, it is widely known that nitrate reduction is an eight-electron transfer process.In addition, the NH 3 yield and Faraday efficiency of samples are essential factors in evaluating NO 3 − RR electrocatalytic properties.For this reason, UV spectroscopy was used to measure the concentration of NH 3 + and NO 2 − .As shown in Figure S3, the linear fitting results correspond to the absorbance versus concentration curves of NH 3 + and NO 2 − .The concentrations of the corresponding ions can be obtained from the measured absorbance and the standard curve.a semiconductor, its conductivity is the worst, resulting in the smallest Rct.At low quency, the perpendicularity of the lines of Mo/TiO2−L and Mo/TiO2−H are as simila TiO2, indicating that their ion diffusion is close.The LSV curves of the Mo/TiO2−M catalyst were tested in the electrolyte with a without nitrate−N at a scan rate of 5 mV s −1 to characterize whether it has NO3 − RR catal properties.As shown in Figure 5a, it is evident that the current density of the LSV cu with NO3 − in 0.05 M Na2SO4 is larger than that of the other one, ranging from −1.1 V Ag/AgCl to −1.4 V vs. Ag/AgCl, which proves that Mo/TiO2−M has NO3 − RR propert Moreover, it is widely known that nitrate reduction is an eight−electron transfer proc In addition, the NH3 yield and Faraday efficiency of samples are essential factors in ev uating NO3 − RR electrocatalytic properties.For this reason, UV spectroscopy was used measure the concentration of NH3 + and NO2 − .As shown in Figure S3, the linear fitt results correspond to the absorbance versus concentration curves of NH3 + and NO2 − .T concentrations of the corresponding ions can be obtained from the measured absorba and the standard curve.
Figure 5b shows that the most preferred NH3 production of TiO2, Mo/TiO2 Mo/TiO2−M, and Mo/TiO2−H was reached at −1.6 V vs. Ag/AgCl, as the NH3 yield did change much as the voltage continued to increase.Additionally, the highest NH3 prod tions of samples are reached at −1.7 V vs. Ag/AgCl.Among them, Mo/TiO2−M a Mo/TiO2−H exhibit the highest NH3 yields, around 5.18 mg h −1 cm −2 and 5.20 mg h −1 cm respectively.Moreover, as shown in Figure 5c, the highest FE was 88.05%, correspond to Mo/TiO2−M at −1.4 V vs. Ag/AgCl.Meanwhile, the highest FEs of 65.50%, 85.98%, a 63.91% were achieved for TiO2, Mo/TiO2−L, and Mo/TiO2−H, respectively.It can be s that the FEs of Mo/TiO2−M remain at a high value at different voltages.This indicates t Mo/TiO2−M has superior NO3 − RR performance, which may be due to the appropr grain size of MoOx in Mo/TiO2−M.Moreover, Table S1 compares the NO3 − RR performa of Mo/TiO2−M with other previously reported electrodes.The FE and NH3 yields of Mo/TiO2−M electrode are comparable to most of the previous cathodes, further indicat the good activity of the as−prepared Mo/TiO2−M electrode [27,[54][55][56][57][58][59].
Furthermore, the generation properties of the byproduct NO2 − at each potential w also evaluated, as shown in Figure 5d.It was found that all MoOx−loaded samples inh ited NO2 − generation compared to TiO2, with the strongest inhibition achieved Mo/TiO2−M at voltages of −1.5-1.7 V vs. Ag/AgCl.Additionally, the change in the amo of NO3 − in the electrolyte (Figure S5a) was measured, and then the amount of N2 produ during the NO3 − RR reaction was calculated.As shown in Figure S5b, the quantity of decreases with the increase in voltage, and no N2 is produced at −1.7 V vs. Ag/AgCl Based on the above test and analysis, Mo/TiO2−M was selected to operate a cycling test at −1.4 V vs. Ag/AgCl.As shown in Figure 6a, NH3 production exceeded 3 mg h −1 cm −2 and that Faraday efficiency stabilized over 80% in all groups.Cycling evaluation further highlighted Mo/TiO2−M's outstanding and steady NO3 − RR performance at −1.4 V vs. Ag/AgCl.Furthermore, a leaching test was conducted to determine possible Mo species in the electrolyte [60].After the NO3 − RR process, the concentration of Mo elements in the electrolyte was measured at only 0.0148 mg/L, suggesting almost no dissolution of Mo. Figure 5b shows that the most preferred NH 3 production of TiO 2 , Mo/TiO 2 -L, Mo/TiO 2 -M, and Mo/TiO 2 -H was reached at −1.6 V vs. Ag/AgCl, as the NH 3 yield did not change much as the voltage continued to increase.Additionally, the highest NH 3 productions of samples 2024, 29, 2782 8 of 14 are reached at −1.7 V vs. Ag/AgCl.Among them, Mo/TiO 2 -M and Mo/TiO 2 -H exhibit the highest NH 3 yields, around 5.18 mg h −1 cm −2 and 5.20 mg h −1 cm −2 , respectively .Moreover, as shown in Figure 5c, the highest FE was 88.05%, corresponding to Mo/TiO 2 -M at −1.4 V vs. Ag/AgCl.Meanwhile, the highest FEs of 65.50%, 85.98%, and 63.91% were achieved for TiO 2 , Mo/TiO 2 -L, and Mo/TiO 2 -H, respectively.It can be seen that the FEs of Mo/TiO 2 -M remain at a high value at different voltages.This indicates that Mo/TiO 2 -M has superior NO 3 − RR performance, which may be due to the appropriate grain size of MoO x in Mo/TiO 2 -M.Moreover, Table S1 compares the NO 3 − RR performance of Mo/TiO 2 -M with other previously reported electrodes.The FE and NH 3 yields of the Mo/TiO 2 -M electrode are comparable to most of the previous cathodes, further indicating the good activity of the as-prepared Mo/TiO 2 -M electrode [27,[54][55][56][57][58][59].
Furthermore, the generation properties of the byproduct NO 2 − at each potential were also evaluated, as shown in Figure 5d.It was found that all MoO x -loaded samples inhibited NO 2 − generation compared to TiO 2 , with the strongest inhibition achieved by Mo/TiO 2 -M at voltages of −1.5-1.7 V vs. Ag/AgCl.Additionally, the change in the amount of NO 3 − in the electrolyte (Figure S5a) was measured, and then the amount N 2 produced during the NO 3 − RR reaction was calculated.As shown in Figure S5b, the quantity of N 2 decreases with the increase in voltage, and no N 2 is produced at −1.7 V vs. Ag/AgCl for any of the electrodes.
Based on the above test and analysis, Mo/TiO 2 -M was selected to operate a cycling test at −1.4 V vs. Ag/AgCl.As shown in Figure 6a, NH 3 production exceeded 3 mg h −1 cm −2 and that Faraday efficiency stabilized over 80% in all groups.Cycling evaluation further highlighted Mo/TiO 2 -M's outstanding and steady NO 3 − RR performance at −1.4 V vs. Ag/AgCl.Furthermore, a leaching test was conducted to determine possible Mo species in the electrolyte [60].After the NO 3 − RR process, the concentration of Mo elements in the electrolyte was measured at only 0.0148 mg/L, suggesting almost no dissolution of Mo. Figure S6

Pre-Treatment of Ti
The pre-cut Ti foil with 2 cm × 1 cm × 0.03 cm dimensions was sanded with 800and 1200-mesh metallographic sandpaper to produce a silvery luster on the surface.Subsequently, the processed Ti sheets were ultrasonically washed with ethanol for 20 min and deionized water for 20 min [61].Then, the Ti foil was rinsed with deionized water to effectively remove the residual organic impurities on the surface during ultrasonic cleaning.

Preparation of TNTAs
Uniformly aligned TNTAs were grown on the surface of Ti foil using an anodic oxidation strategy.Dissolve ammonium fluoride (NH 4 F) (0.5 g) in deionized water (2 mL) and ethylene glycol (C 2 H 6 O 2 ) (98 mL) to prepare electrolyte solution successively.Then, the Ti foil and platinum (Pt) electrode, which were cleaned as described above, were used as anode and cathode, respectively.The distance between the anode and the cathode was adjusted to be 3 cm approximately and reacted by a constant voltage device at 30 V for two hours.After that, the Ti foil was washed with anhydrous ethanol and placed in a tube furnace to be calcined for 2 h at 450 • C in an air environment.Finally, TNTAs with anatase phase were successfully produced.

Preparation of Mo/TiO 2 Electrode
A certain mass of ammonium molybdate tetrahydrate solid ((NH 4 ) 6 Mo 7 O 24 •4H 2 O) and 2.2075 g of sodium citrate (C 6 O 7 H 5 Na 3 •2H 2 O) were added into deionized water (50 mL)  to prepare four groups of samples (0 M, 0.05 M, 0.1 M, and 0.2 M molybdate).NH 3 •H 2 O was added to the mixture after stirring to bring the pH up to 9. The prepared TNTA was used as the working electrode, a platinum (Pt) sheet as the auxiliary electrode, and Ag/AgCl as the reference electrode.Meanwhile, the current density was −20 mA cm −2 , and the electrodeposition time was set to 20 min.After the process was completed, the samples were cleaned with deionized water and then put in a tube furnace to be calcined in 3% H 2 /Ar at the rate of 50 mL min −1 for 2 h.Finally, the electrodes were successfully produced, and the materials with different concentration of molybdate (0, 0.05 M, 0.1 M, and 0.2 M) were named as TiO 2 , Mo/TiO 2 -L, Mo/TiO 2 -M, and Mo/TiO 2 -H, respectively.

Characterization
The Ag/AgCl potential was converted to a reversible hydrogen electrode (RHE) using the Nernst equation: E RHE = E Ag/AgCl + 0.059 × pH + 0.197.All data were collected on a CHI660E electrochemical workstation (Shanghai CH Instruments, Shanghai, China).The crystal structure of the processed samples was characterized through X-ray diffraction (XRD) using a MiniFlex600 (Rigaku, Tokyo, Japan) with Cu-Kα radiation (λ = 0.154056 nm) at an ambient temperature (25 • C) and 2θ values ranging from 10 to 80 • .Energy dispersive spectroscopy (EDS) spectra and scanning electron microscopy (SEM) graphics were obtained using an FEI Quanta 250 (Regulus 8230U, Hitachi, Japan).A K-alpha spectrometer (Thermo Scientific, Waltham, MA, USA) equipped with a monochromatic Al Kα X-ray source (1486.6 eV photons) was used to perform X-ray photoelectron spectroscopy (XPS).Images from a transmission electron microscope (TEM) were captured at 200 kV using a Libra 200FE (Zeiss, Oberkochen, Germany).The concentration of ions in electrolyte was determined using an optima 7000DV inductively coupled plasma optical emission spectrometer (ICP-OES, Thermo Scientific, Waltham, MA, USA)

Electrochemical Measurement
The above-made electrode was used as the working electrode (1 cm × 1 cm), a Pt sheet as the counter electrode (1 cm × 1 cm), and Ag/AgCl as the reference electrode, placed in an H-type electrolyzer.The electrolyte was 0.1 M NO 3 − -N solution containing 0.05 M Na 2 SO 4 .The nitrate solution was then tested by i-t for one hour, testing the NH 3 production at constant voltage to obtain the optimum operating voltage.The reacted cathode solution was collected for subsequent measurements.
Cyclic voltammetry (CV) was performed at −0.1 to 0 V against Ag/AgCl with a sampling rate of 20 to 100 mV s −1 with an interval of 20 mV s −1 in order to estimate the double layer capacitance (Cdl) of samples.Electrochemical impedance spectroscopy (EIS) was performed in an aqueous solution comprising 0.05 M Na 2 SO 4 .Additionally, the EIS measurements were performed in a frequency range from 0.01 to 100,000 Hz with an amplitude of sinusoidal AC voltage of 5 mV and 2 points per decade.Then, linear scanning voltammetry (LSV) was performed in the voltage range of −1.4-0 V (vs.RHE) for the tests.

Determination of Ion Concentration
UV spectroscopy was used to measure the concentrations of ammonium and nitrite ions.The method is as follows:

Nitrite-N Detection
Five groups of sodium nitrite solutions (0, 0.25, 0.5, 1, 2, and 3 µg/mL) were prepared separately, 1 mL of the Griess reagent was added, and then the solution was left to develop color for 10 min.The absorbance was measured at 540 nm using a UV spectrophotometer, and the standard concentration curve of nitrite was plotted.Diluting the cathode solution after the reaction to a measurable concentration range, 1 mL of the Griess reagent was added into it, and then the absorbance was measured.The corresponding nitrite concentration was calculated according to its standard concentration graph.

NH 3 -N Detection
Five groups of NH 4 + solution (0, 1, 2, 3, and 4 µg/mL, respectively) were prepared.A total of 2 mL of colorant, 1 mL of oxidant, and 200 µL of catalyst were added sequentially, and then the solution was left to develop color by avoiding light for one hour.The absorbance was measured at 660 nm using a UV spectrophotometer, and the standard concentration curve of NH 4 + was plotted.Diluting the NH 4 + concentration after the reaction to a measurable concentration range, the three solutions were added in turn like the steps mentioned above, and then the absorbance was measured.The corresponding NH 4 + concentration was calculated according to its standard concentration graph.

Product Calculation (Yield and Faraday Efficiency)
The NH 3 yield was calculated using the following equation: where C (NH 3 ) is the measured concentration of NH 3 −N (aq), V (50 mL) is the volume of the cathode cell electrolyte, t (3600 s) is the electrochemical reaction time, and S (1 cm × 1 cm) is the surface area of the working electrode.

Figure 1 .
Figure 1.Schematic illustration of the two-step synthesis process of the Mo/TiO2 electrodes: (a) odization and then calcination under a N2 atmosphere at 450 °C for 2 h to prepare TNTAs; (b) e trodeposition and then calcination in 3% H2/Ar at 500 °C for 2 h.

Figure 1 .
Figure 1.Schematic illustration of the two-step synthesis process of the Mo/TiO 2 electrodes: (a) anodization and then calcination under a N 2 atmosphere at 450 • C for 2 h to prepare TNTAs; (b) electrodeposition and then calcination in 3% H 2 /Ar at 500 • C for 2 h.

Figure 6 .
Figure 6.(a) NH3 yields and FE of Mo/TiO2−M at −1.4 V (vs.RHE) for five cycling tests.(b) patterns of Mo/TiO2−M before and after cycling tests.

Figure 6 .
Figure 6.(a) NH 3 yields and FE of Mo/TiO 2 -M at −1.4 V (vs.RHE) for five cycling tests.(b) XRD patterns of Mo/TiO 2 -M before and after cycling tests.