catalysts Characteristics of High Surface Area Molybdenum Nitride and Its Activity for the Catalytic Decomposition of Ammonia

: High surface area (>170 m 2 g − 1 ) molybdenum nitride was prepared by the temperature-programmed nitridation of α -MoO 3 with pure ammonia. The process was optimized by adjusting the experimental variables: the reaction temperature, heating rate, and molar ﬂow rate of ammonia. The physicochemical properties of the as-formed molybdenum nitride were characterized by X-ray diffraction, N 2 sorption, transmission electron microscopy, temperature-programmed oxida-tion/reduction, and X-ray photoelectron spectroscopy. Of the experimental variables, the nitridation temperature was found to be the most critical parameter determining the surface area of the molybdenum nitride. When the prepared molybdenum nitride was exposed to air, the speciﬁc surface area rapidly decreased because of the partial oxidation of molybdenum nitride to molybdenum oxynitride. However, the surface area recovered to 90% the initial value after H 2 treatment. The catalyst with the highest degree of nitridation showed the best catalytic activity, superior to that of unmodiﬁed α -MoO 3 , for the decomposition of ammonia because of its high surface area.


Introduction
Transition metal oxides, carbides, nitrides, and phosphides are widely used as catalysts or catalyst supports [1]. Well-known oxide materials include single component oxides, such as SiO 2 , Al 2 O 3 , TiO 2 , V 2 O 5 , MoO 3 , and WO 3 , and their combinations have been used as catalysts and catalyst supports for reactions such as metathesis, isomerization, hydrogenation, and partial oxidation [2]. In addition, transition metal nitrides, such as Si 3 N 4 , AlN, TiN, VN, Mo 2 N, and WN/W 2 N, and carbides, such as SiC, Al 4 C 3 , TiC, VC, Mo 2 C, and WC/W 2 C, have been used in cutting tools and refractory materials owing to their high thermal and chemical stabilities, as well as mechanical strengths [3]. Thus, because of their remarkable thermal and chemical stabilities, these materials are potential supports or catalysts for use in heterogeneous catalysis. However, the application of transition metal nitrides and carbides has been limited because of their low specific surface areas, which are a result of sintering during their high-temperature synthesis [4,5]. To use these transition metal nitrides and carbides as catalysts or catalyst supports, it is necessary to maintain a high surface area to increase the reactant-catalyst contact as much as possible. Volpe and Boudart reported the preparation of high surface area Mo 2 N and W 2 N Figure 1a shows the plots of NH 3 decomposition and H 2 and H 2 O production during temperature-programmed nitridation, as well as the reduction of α-MoO 3 and MoO 2 , as a function of reaction temperature. H 2 O, which is generated by the reduction of oxides by NH 3 , was detected from ca. 320 • C, and the evolution of H 2 via the decomposition of NH 3 was observed at ca. 400 • C. The evolution of H 2 over both oxides started from the peak temperature in the H 2 O evolution curves, indicating that the decomposition of NH 3 proceeds over nitrided α-MoO 3 (γ-Mo 2 N). Dewangan et al. reported that the formation of γ-Mo 2 N by the nitridation of α-MoO 3 with NH 3 proceeds via an intermediate γ-Mo 2 O x N 1-x or MoO 2 phase [32]. Thus, the hydrogen produced by the initial dissociation of NH 3 was not detected and was mostly consumed during the reduction of both oxides. In addition, the decomposition of NH 3 and generation of H 2 over α-MoO 3 require higher temperatures than MoO 2 because of the formation of an intermediate γ-

Results and Discussion
The reaction parameters affecting the Brunauer-Emmett-Teller (BET) surface area, total pore volume, and average pore diameter during the nitridation of α-MoO 3 were investigated by changing the nitridation temperature, heating rate, and MHSV (Figure 1b-d, respectively). As shown in Figure 1b, the BET surface area of α-MoO 3 increased proportionally with increase in nitridation temperature, reaching its maximum value of 150 m 2 g −1 at 630 • C; this is approximately 50 times higher than that (3 m 2 g −1 ) of pristine α-MoO 3 . In line with the increase in the BET surface area, the total pore volume of α-MoO 3 increased proportionally to the nitridation temperature up to 770 • C, subsequently decreasing rapidly because of the phase transformation of γ-Mo 2 N to the other phase, such as β-Mo 2 N. Wei et al. reported the transformation of γ-Mo 2 N to β-Mo 2 N at 600 • C and β-Mo 2 N to metallic Mo at around 880 • C based on the observation of N 2 release in their temperature-programmed reduction (TPR) experiments under flowing H 2 [33]. Thus, the reduction of the specific surface area at nitridation temperatures higher than 630 • C can be attributed to the formation of the β-Mo 2 N phase. However, a continuous increase in the total pore volume up to 770 • C was observed, probably because of the formation of mesopores during the partial transformation of γ-Mo 2 N to β-Mo 2 N [34]. The differentiation of γ-Mo 2 N and β-Mo 2 N by X-ray diffraction is quite difficult because of the heavy overlap of the characteristic peaks; however, the BET surface areas of γ-Mo 2 N and β-Mo 2 N fall into the range of 40-145 and 2-17 m 2 g −1 , respectively [25,35,36]. Thus, it can be reasonably speculated that the phase change of γ-Mo 2 N to β-Mo 2 N occurred at the temperatures higher than 630 • C. The rapid increase in the average pore diameter with associated reduction in the surface area and pore volume over the samples nitrided at temperatures higher than 800 • C is a result of the formation of metallic Mo particles. This can be further supported by the observation of X-ray diffraction peaks corresponding to metallic Mo at the temperatures higher than 800 • C, as shown in Table S1. In fact, the heating rate during high-temperature treatment of oxide materials is an important factor determining the surface area and particle size [2]. Figure 1c shows the textural properties (i.e., BET surface area, total pore volume, and average pore diameter) of α-MoO 3 nitrided in flowing NH 3 at 630 • C and a MHSV of 6.2 h −1 at different heating rates (30-240 • C h −1 ). For the nitridation of α-MoO 3 in flowing NH 3 , a low heating rate at temperatures below 450 • C is advantageous for the generation of γ-Mo 2 O x N 1-x and H x MoO 3 rather than MoO 2 [32]. Under these conditions, the H 2 formed by the dissociation of NH 3 diffuses into the α-MoO 3 lattice and produces the H x MoO 3 phase by a topotactic reaction in which the loosely held double-thick layers of [MoO 6 ] octahedra in α-MoO 3 are broken by H 2 insertion [37]. On the other hand, the fast heating of α-MoO 3 below 450 • C disturbs the diffusion of H 2 into the α-MoO 3 lattice, and, as a result, the decomposition of α-MoO 3 to MoO 2 dominates. As shown in Figure 1c, the surface area of nitrided α-MoO 3 was maintained higher than 150 m 2 g −1 when a heating rate of less than 120 • C h −1 was used but steadily reduced at higher heating rates. This implies that the diffusion of H 2 is limited at heating rates higher than 120 • C h −1 , resulting in the decomposition of α-MoO 3 to MoO 2 . This result is also well correlated with the rapid increase in the average pore diameter of this material at heating rates higher than 120 • C h −1 . Unlike the changes in surface area and average pore diameter, the total pore volume of nitrided α-MoO 3 hardly changed over the entire range of heating rates. Figure 1d shows the changes in the BET surface area, total pore volume, and average pore diameter of α-MoO 3 nitrided with NH 3 at 630 • C as a function of MHSV. The α-MoO 3 nitrided with NH 3 at a low MHSV of 10 h −1 has a relatively low surface area and total pore volume, probably because of the thermal reduction of α-MoO 3 to MoO 2 at this reaction temperature. Spevack and McIntyre also observed the thermal reduction of α-MoO 3 to MoO 2 at 600 • C using Raman spectroscopy [38]. The surface area of nitrided α-MoO 3 gradually increased with increasing MHSV and leveled off at MHSV values higher than 40 h −1 . In addition, the total pore volume of the nitrided sample rapidly increased until the MHSV reached 20 h −1 and was slightly reduced at higher MHSV.
The theoretical mass reduction in the transformation of α-MoO 3 to γ-Mo 2 N is 28.5 wt.%, and the nitrogen content in γ-Mo 2 N is 6.8 wt.%. As shown in Table S1, the mass reduction occurred below the theoretical value for all nitride samples obtained after nitridation, and, in most cases, the MoO 2 phase was detected. Thus, to prevent the rapid surface oxidation of the nitrided sample on exposure to air, the sample was passivated with 1% O 2 /Ar at room temperature, and the weight difference between the theoretical and practical values of the sample can be explained by oxygen adsorption and diffusion into the nitride samples. Above a nitridation temperature of 700 • C, the MoO 2 phase was no longer detected, and only the γ-Mo 2 N phase was observed by X-ray diffraction (XRD) analysis ( Figure S1 and Table S1). The lattice parameter of the sample nitrided at 700 • C was similar (0.4166 nm) to that of stoichiometric γ-Mo 2 N (a = 0.4165 nm), which has a face-centered cubic (fcc) structure [31,32]. Above a nitridation temperature of 800 • C, metallic Mo was detected by XRD analysis, and this was likely formed by the partial decomposition of γ-Mo 2 N Catalysts 2021, 11,192 4 of 15 (Table S1). This formation of metallic Mo caused a rapid decrease in the surface area and total pore volume, and the porous nature of the sample was lost. The reaction parameters affecting the Brunauer-Emmett-Teller (BET) surface area, total pore volume, and average pore diameter during the nitridation of α-MoO3 were investigated by changing the nitridation temperature, heating rate, and MHSV (Figure 1bd, respectively). As shown in Figure 1b, the BET surface area of α-MoO3 increased proportionally with increase in nitridation temperature, reaching its maximum value of 150 m 2 g −1 at 630 °C; this is approximately 50 times higher than that (3 m 2 g −1 ) of pristine α-MoO3. In line with the increase in the BET surface area, the total pore volume of α-MoO3 increased proportionally to the nitridation temperature up to 770 °C, subsequently decreasing rapidly because of the phase transformation of γ-Mo2N to the other phase, such as β-Mo2N. Wei et al. reported the transformation of γ-Mo2N to β-Mo2N at 600 °C and β-Mo2N to metallic Mo at around 880 °C based on the observation of N2 release in their temperature-programmed reduction (TPR) experiments under flowing H2 [33]. Thus, the reduction of the specific surface area at nitridation temperatures higher than 630 °C can be attributed to the formation of the β-Mo2N phase. However, a continuous increase in the total pore volume up to 770 °C was observed, probably because of the formation of mesopores during the partial transformation of γ-Mo2N to β-Mo2N [34]. The differentiation of γ-Mo2N and β-Mo2N by X-ray diffraction is quite difficult because of the heavy overlap of the characteristic peaks; however, the BET surface areas of γ-Mo2N and β-Mo2N fall into the range of 40-145 and 2-17 m 2 g −1 , respectively [25,35,36]. Thus, it can be reasonably speculated that the phase change of γ-Mo2N to β-Mo2N occurred at the temperatures higher than 630 °C. The rapid increase in the average pore diameter with associated reduction in the surface area and pore volume over the samples nitrided at temperatures higher than 800 °C is a result of the formation of metallic Mo particles. This can be further supported by the observation of X-ray diffraction peaks corresponding to metallic Mo at the temperatures higher than 800 °C, as shown in Table S1. In fact, the heating rate during  Figure 2 shows the N 2 adsorption-desorption isotherms of α-MoO 3 and the analogous samples nitrided at different temperatures for 7 h. The α-MoO 3 sample shows a typical type II isotherm, which is characteristic of non-porous materials [39]. This is further supported by its very low BET surface area (2.8 m 2 g −1 ) and total pore volume (0.007 cm 3 g −1 ). However, the shape of the N 2 adsorption-desorption isotherm of nitrided α-MoO 3 gradually changed to type IV, characteristic of a mesoporous material, as the nitridation temperature increased from 450 to 700 • C, and the hysteresis loop related to the narrow slit-type pores was observed over all nitrided α-MoO 3 . The maximum BET surface area (145 m 2 g −1 ) and total pore volume (0.106 cm 3 g −1 ) were observed for α-MoO 3 nitrided at 650 • C. The maximum Barrett-Joyner-Halenda (BJH) pore diameter calculated from the N 2 adsorption branch also gradually increased from 1.2 to 2.2 nm with the increase in nitridation temperature. Figure 3 shows the results of simultaneous thermogravimetric (TG)/differential thermal analysis (DTA)-mass spectrometry (MS) analyses of the nitrided α-MoO 3 (i.e., γ-Mo 2 N) under inert, oxidizing, and reducing atmospheres (He, O 2 , and 20% H 2 /He, respectively). The TG-DTA traces shown in Figure 3a,b contain three steps of weight loss and changes in the heat flow in flowing O 2 : 0-250, 250-700, and 700-900 • C, corresponding to the exothermic oxidation of γ-Mo 2 N to γ-Mo 2 O x N 1-x or MoO 2 (14% mass increase), total oxidation to α-MoO 3 (5% mass increase), and melting and evaporation of α-MoO 3 , respectively. The sharp endothermic peak observed in the DTA curve around 800 • C represents the phase transformation of solid α-MoO 3 (melting point = 795 • C) to the liquid phase. As shown in Figure 3c, massive consumption of O 2 (m/z = 32) and evolution of N 2 (m/z = 28) at ca. 450 • C in an oxidizing atmosphere were observed by quadrupole mass spectrometry (QMS), indicating that the total oxidation of γ-Mo 2 O x N 1-x to α-MoO 3 occurred at this temperature. This is further confirmed by the fact that only peaks corresponding to α-MoO 3 were observed in the XRD pattern ( Figure S2c) of γ-Mo 2 N treated at 700 • C in flowing 5% O 2 /He (100 cm 3 min −1 ). The theoretical mass increases for the oxidation of γ-Mo 2 N to MoO 2 and α-MoO 3 are 24.3% and 39.8%, respectively. The actual mass increase of less than 20% under an O 2 atmosphere is probably due to the partial oxidation of the prepared γ-Mo 2 N to γ-Mo 2 O x N 1-x or MoO 2 , especially in the surface region during the passivation process. The mass losses observed up to 400 • C under H 2 and He atmospheres mainly originate from the reaction adsorbed ammonia on the partially oxidized α-MoO 3 (MoO 3 ·NH 3 → γ-MoO x N 1-x ), and the mass reduction at temperatures higher than 600 • C is probably due to the phase transformation of γ-Mo 2 N to β-Mo 2 N, which is well correlated with the decrease in the surface area of this material, as shown in Figure 1b. However, as shown in Figure S2b, a significant amount of the MoO 2 phase was observed after treatment with γ-Mo 2 N in a flow of pure He (100 cm 3 min −1 ), probably caused by the decomposition of surface γ-Mo 2 O x N 1-x formed during passivation in 1% O 2 /He to MoO 2 . The generation of nitrogen-containing compounds (i.e., NO, NO 2 , and N 2 O) and H 2 below 300 • C in the He atmosphere supports the presence of reaction with adsorbed ammonia and lattice oxygen ( Figure 3d). In addition, the increase in N 2 release at temperatures higher than 600 • C confirms the phase transformation to the nitrogen-deficient β-Mo 2 N phase, and this change was more significant under the reducing conditions (20% H 2 /He), resulting in the higher weight loss shown in Figure 3a.  Figure 3 shows the results of simultaneous thermogravimetric (TG)/differential thermal analysis (DTA)-mass spectrometry (MS) analyses of the nitrided α-MoO3 (i.e., γ-Mo2N) under inert, oxidizing, and reducing atmospheres (He, O2, and 20% H2/He, respectively). The TG-DTA traces shown in Figure 3a,b contain three steps of weight loss and changes in the heat flow in flowing O2: 0-250, 250-700, and 700-900 °C, corresponding to the exothermic oxidation of γ-Mo2N to γ-Mo2OxN1-x or MoO2 (14% mass increase), total oxidation to α-MoO3 (5% mass increase), and melting and evaporation of α-MoO3, respectively. The sharp endothermic peak observed in the DTA curve around 800 °C represents the phase transformation of solid α-MoO3 (melting point = 795 °C) to the liquid phase. As shown in Figure 3c, massive consumption of O2 (m/z = 32) and evolution of N2 (m/z = 28) at ca. 450 °C in an oxidizing atmosphere were observed by quadrupole mass spectrometry containing compounds (i.e., NO, NO2, and N2O) and H2 below 300 °C in the He atm phere supports the presence of reaction with adsorbed ammonia and lattice oxygen (F ure 3d). In addition, the increase in N2 release at temperatures higher than 600 °C confir the phase transformation to the nitrogen-deficient β-Mo2N phase, and this change w more significant under the reducing conditions (20% H2/He), resulting in the high weight loss shown in Figure 3a.  Figure 4 shows the transmission electron microscopy (TEM) images and selected a electron diffraction (SAED) patterns of α-MoO3 and its nitrided analogs prepared at d ferent temperatures (550-700 °C). α-MoO3 has plate-like morphology with an orthorho bic crystal lattice (a = 3.96, b = 13.86, and c = 3.70 Å) [40]. Because the nitridation procee at high temperatures, the generation of new pores observed in the samples is probab due to the extraction of lattice oxygen from the crystalline oxynitride phase. Volpe a Boudart [6] and Jaggers [5] reported that α-MoO3 is transformed to γ-Mo2N under an N atmosphere via a reaction pathway in which γ-Mo2OxN1-x or MoO2 are reaction interm diates. The reaction pathway for α-MoO3 nitridation is mainly dependent on the heati conditions (i.e., temperature and heating rate) [32,41]. In addition, γ-Mo2OxN1-x is know to be transformed to γ-Mo2N at a relatively low nitridation temperature (ca. 500 ° whereas MoO2 starts to react with NH3 only at temperatures higher than 657 °C and completely converted to γ-Mo2N at 785 °C [6,31]. This is further supported by the obs vation of characteristic peaks corresponding to pure γ-Mo2N at 700 °C, as shown in Figu S1.  Figure 4 shows the transmission electron microscopy (TEM) images and selected area electron diffraction (SAED) patterns of α-MoO 3 and its nitrided analogs prepared at different temperatures (550-700 • C). α-MoO 3 has plate-like morphology with an orthorhombic crystal lattice (a = 3.96, b = 13.86, and c = 3.70 Å) [40]. Because the nitridation proceeds at high temperatures, the generation of new pores observed in the samples is probably due to the extraction of lattice oxygen from the crystalline oxynitride phase. Volpe and Boudart [6] and Jaggers [5] reported that α-MoO 3 is transformed to γ-Mo 2 N under an NH 3 atmosphere via a reaction pathway in which γ-Mo 2 O x N 1-x or MoO 2 are reaction intermediates. The reaction pathway for α-MoO 3 nitridation is mainly dependent on the heating conditions (i.e., temperature and heating rate) [32,41]. In addition, γ-Mo 2 O x N 1-x is known to be transformed to γ-Mo 2 N at a relatively low nitridation temperature (ca. 500 • C), whereas MoO 2 starts to react with NH 3 only at temperatures higher than 657 • C and is completely converted to γ-Mo 2 N at 785 • C [6,31]. This is further supported by the observation of characteristic peaks corresponding to pure γ-Mo 2 N at 700 • C, as shown in Figure S1.
As shown in Figure 4a, the SAED image of α-MoO 3 clearly shows reflections corresponding to the (100), (101), and (001) facets along the [010] direction [5]. After the transformation of α-MoO 3 to the fcc-structured γ-Mo 2 O x N 1-x by nitridation at high temperatures, reflections corresponding to the (002), (020), and (022) facets were also observed along the [100] direction [11,32]. Energy dispersive X-ray spectroscopy (EDS) mapping (Figure 4f) shows the uniform distribution of nitrogen on the γ-Mo 2 N sample prepared at 700 • C by nitridation and also indicates the presence of molybdenum oxynitrides, which are generated during the passivation process. However, the large amount of oxygen observed in the EDS maps of γ-Mo 2 N prepared at 700 • C suggests the presence of XRD-invisible MoO 2 clusters in the sample. As shown in Figure 4a, the SAED image of α-MoO3 clearly shows reflections corresponding to the (100), (101), and (001) facets along the [010] direction [5]. After the transformation of α-MoO3 to the fcc-structured γ-Mo2OxN1-x by nitridation at high temperatures, reflections corresponding to the (002), (020), and (022) facets were also observed along the [100] direction [11,32]. Energy dispersive X-ray spectroscopy (EDS) mapping (Figure 4f) shows the uniform distribution of nitrogen on the γ-Mo2N sample prepared at 700 °C by nitridation and also indicates the presence of molybdenum oxynitrides, which are generated during the passivation process. However, the large amount of oxygen observed in the EDS maps of γ-Mo2N prepared at 700 °C suggests the presence of XRDinvisible MoO2 clusters in the sample.
To examine the stability of γ-Mo2N under oxidizing conditions, a sample was exposed to air for up to 60 days, and the textural properties are compared in Figure 5. Compared to the initial surface area (120 m 2 g −1 ) of γ-Mo2N prepared at 700 °C for 7 h, those of the samples exposed to air gradually decreased over time, as did their total pore volumes. After 60 days, we measured the N2 adsorption-desorption isotherm of the γ-Mo2N sample, which revealed a change to a type II nonporous material, as well as reductions in the surface area and total pore volume of 88% and 81%, respectively, compared to the initial values. This is a result of the diffusion of oxygen into the nitride crystal lattice, which results in the formation of molybdenum oxynitride and the observed reduction in pore volume and specific surface area. After exposure to air for seven days, the initial γ-Mo2N phase did not change significantly, as shown by the XRD analysis, which contained no peaks corresponding to molybdenum oxide phases ( Figure S3). However, after air exposure for seven days, the reflection corresponding to the (111) facet shifted from 37.4° to 37.2° in 2θ. This small shift (Δ = 0.2°) indicates the expansion of the fcc unit cell from 0.4163 to 0.4180 nm. This volume expansion induced by oxygen diffusion into bulk γ-Mo2N causes pore blocking and, finally, the transformation to γ-Mo2OxN1-x, which results in a To examine the stability of γ-Mo 2 N under oxidizing conditions, a sample was exposed to air for up to 60 days, and the textural properties are compared in Figure 5. Compared to the initial surface area (120 m 2 g −1 ) of γ-Mo 2 N prepared at 700 • C for 7 h, those of the samples exposed to air gradually decreased over time, as did their total pore volumes. After 60 days, we measured the N 2 adsorption-desorption isotherm of the γ-Mo 2 N sample, which revealed a change to a type II nonporous material, as well as reductions in the surface area and total pore volume of 88% and 81%, respectively, compared to the initial values. This is a result of the diffusion of oxygen into the nitride crystal lattice, which results in the formation of molybdenum oxynitride and the observed reduction in pore volume and specific surface area. After exposure to air for seven days, the initial γ-Mo 2 N phase did not change significantly, as shown by the XRD analysis, which contained no peaks corresponding to molybdenum oxide phases ( Figure S3). However, after air exposure for seven days, the reflection corresponding to the (111) facet shifted from 37.4 • to 37.2 • in 2θ. This small shift (∆ = 0.2 • ) indicates the expansion of the fcc unit cell from 0.4163 to 0.4180 nm. This volume expansion induced by oxygen diffusion into bulk γ-Mo 2 N causes pore blocking and, finally, the transformation to γ-Mo 2 O x N 1-x , which results in a decrease in the surface area and total pore volume. However, the oxynitride sample exposed to air recovered the textural properties of the initial nitride (by more than 90% the initial values) after H 2 treatment at 500 • C for 2 h ( Figure S4).
As shown in Figure 6, the distribution of Mo, N, and O in the γ-Mo 2 N sample exposed to air for three days was characterized by X-ray photoelectron spectroscopy (XPS) depth profiling analysis. As decrease in the surface area and total pore volume. However, the oxynitride sample exposed to air recovered the textural properties of the initial nitride (by more than 90% the initial values) after H2 treatment at 500 °C for 2 h ( Figure S4). As shown in Figure 6, the distribution of Mo, N, and O in the γ-Mo2N sample exposed to air for three days was characterized by X-ray photoelectron spectroscopy (XPS) depth profiling analysis. As    As shown in Figure 6, the distribution of Mo, N, and O in the γ-Mo2N sample exposed to air for three days was characterized by X-ray photoelectron spectroscopy (XPS) depth profiling analysis. As  The XPS results are shown in Figure 7 and Table 1. XPS measurements were performed to investigate the degree of nitridation, reduction of MoO3, and the oxidation state of Mo with respect to the increase in nitridation temperature. The N 1s peak in the spectra was deconvoluted into three peaks: O-N, NH3, and NH4 + species centered at 397.4-397.7, 399.0-399.4, and 401.2-401.3 eV, respectively, and an overlapped Mo 3p3/2 peak (395.2-395.6 eV) was also observed [42,43]. The binding energy of Mo 3p3/2 (398.7 eV) in the unmodified α-MoO3 shifted to lower energy after high-temperature nitridation because of The XPS results are shown in Figure 7 and Table 1. XPS measurements were performed to investigate the degree of nitridation, reduction of MoO 3 , and the oxidation state of Mo with respect to the increase in nitridation temperature. The N 1s peak in the spectra was deconvoluted into three peaks: O-N, NH 3 , and NH 4 + species centered at 397.4-397.7, 399.0-399.4, and 401.2-401.3 eV, respectively, and an overlapped Mo 3p 3/2 peak (395.2-395.6 eV) was also observed [42,43]. The binding energy of Mo 3p 3/2 (398.7 eV) in the unmodified α-MoO 3 shifted to lower energy after high-temperature nitridation because of the changes in the oxidation state (Mo 6+ → Mo 6+ , Mo 5+ , and Mo 4+ ) and electroneg-ativity (3.44 → 3.04, Pauling scale) of the bonding elements. In addition, the intensities of the peaks corresponding to Mo 3p 3/2 and O-N increased as high-temperature nitridation proceeded because of the increased proportion of Mo in Mo 2 N compared to that in α-MoO 3 and the presence of γ-Mo 2 O x N 1-x . On the other hand, the peak intensities of adsorbed species, such as NH 3 and NH 4 + , were reduced at higher nitridation temperatures because desorption was easier. The O 1s peaks in the spectra were deconvoluted into three peaks corresponding to O 2 − (530.6-530.7 eV), OH -(531.8-531.9 eV), and O-N (532.9-533.1 eV) species originating from α-MoO 3 , adsorbed H 2 O, and γ-Mo 2 O x N 1-x , respectively [44]. As the nitridation temperature increased, the atomic concentration of O 2 − and OHdecreased, whereas that of O-N increased. However, of the three components, the peak corresponding to O 2 − was dominant, suggesting the formation of oxide layers after passivation with dilute O 2 and the presence of XRD-invisible MoO 2 clusters. The Mo 3d peak was deconvoluted into five peaks, representing Mo 6+ , Mo 5+ , Mo 4+ , Mo-N, and Mo-OH centered at 232.5-232.8, 230.7-230.8, 229.7, 228.9-229.0, and 233.6-233.8 eV, respectively [45,46]. As the nitridation temperature was increased, the atomic concentration of Mo 6+ decreased, whereas that of Mo-N increased. Moreover, the peak of Mo-OH was drastically reduced after nitridation at high temperatures.  [45,46]. As the nitridation temperature was increased, the atomic concentration of Mo 6+ decreased, whereas that of Mo-N increased. Moreover, the peak of Mo-OH was drastically reduced after nitridation at high temperatures.    The conversions of NH 3 during the nitridation of α-MoO 3 at various temperatures were shown in Figure 8. It is obvious that the nitridation of α-MoO 3 and catalytic decomposition of NH 3 over nitrided α-MoO 3 proceed with the concurrent manner, and their performances are increases with the elevation of temperature. However, a sharp increase of NH 3 conversion was observed between 550 and 600 • C, probably due to the facilitated nitridation of α-MoO 3 . As shown in Figure 1b, the BET surface areas of nitrided α-MoO 3 was almost proportional to the temperatures up to 600 • C, while no specific correlation between the surface area and the conversion of NH 3 was observed. However, the atomic content of Mo-N species in Mo 3d XPS analysis (Table 1) was 4.7 times increased at this temperature range (550-600 • C) indicating the facilitated structural change from α-MoO 3 or Mo 2 O x N 1-x to Mo 2 N. In addition, the induction period in the conversion of NH 3 was observed during the decomposition of NH 3 at all temperatures except 700 • C indicating that the transition of α-MoO 3 or Mo 2 O x N 1-x to Mo 2 N still proceeds at the reaction condition.

PEER REVIEW 10 of 15
The conversions of NH3 during the nitridation of α-MoO3 at various temperatures were shown in Figure 8. It is obvious that the nitridation of α-MoO3 and catalytic decomposition of NH3 over nitrided α-MoO3 proceed with the concurrent manner, and their performances are increases with the elevation of temperature. However, a sharp increase of NH3 conversion was observed between 550 and 600 °C, probably due to the facilitated nitridation of α-MoO3. As shown in Figure 1b, the BET surface areas of nitrided α-MoO3 was almost proportional to the temperatures up to 600 °C, while no specific correlation between the surface area and the conversion of NH3 was observed. However, the atomic content of Mo-N species in Mo 3d XPS analysis (Table 1) was 4.7 times increased at this temperature range (550-600 °C) indicating the facilitated structural change from α-MoO3 or Mo2OxN1-x to Mo2N. In addition, the induction period in the conversion of NH3 was observed during the decomposition of NH3 at all temperatures except 700 °C indicating that the transition of α-MoO3 or Mo2OxN1-x to Mo2N still proceeds at the reaction condition.    Figure 9a shows the XRD patterns of molybdenum nitride samples prepared at 650 • C for 7 and 24 h, respectively. It is obvious that the remaining portion of MoO 2 on the molybdenum nitride sample reduced at the prolonged reactions (nitridation and NH 3 decomposition). In addition, this further nitridation of sample during the reaction conditions results in the increase of both BET surface area and total pore volume (Figure 9b), which are favorable for the catalytic performance of molybdenum nitride (Figure 9c). The apparent activation energies (E a ) of ammonia decomposition between 400 and 525 • C over two molybdenum nitride samples prepared at 650 • C for nitridation times from 7 and 24 h were calculated using the following relationship: −r NH3 = k o exp(E a /RT)P NH3 α . Here, r NH3 is the number of moles of NH 3 consumed per second per gram of base MoO 3 . As shown in Figure 8a, the NH 3 conversion was higher over the sample prepared for a longer nitridation time, indicating that the nitridation period is another decisive parameter determining the catalytic activity of molybdenum nitrides. Figure 8b shows the Arrhenius plots for NH 3 consumption over the two molybdenum nitride catalysts. The apparent activation energies of ammonia decomposition were calculated to be 28.4 and 29.4 kcal/mol over the samples nitrided for 7 and 24 h, respectively. The activation energies of the molybdenum nitride catalysts were slightly higher than those of molybdenum carbide (21.3 kcal/mol) and supported Ru catalysts (17.9-20.3 kcal/mol) reported in the literature [47]. Overall, the results of this study reveal the potential of high surface area molybdenum nitrides as catalysts for H 2 production via ammonia decomposition.
R PEER REVIEW 11 of 15 are favorable for the catalytic performance of molybdenum nitride (Figure 9c). The apparent activation energies (Ea) of ammonia decomposition between 400 and 525 °C over two molybdenum nitride samples prepared at 650 °C for nitridation times from 7 and 24 h were calculated using the following relationship: −rNH3 = ko exp(Ea/RT)PNH3 α . Here, rNH3 is the number of moles of NH3 consumed per second per gram of base MoO3. As shown in Figure 8a, the NH3 conversion was higher over the sample prepared for a longer nitridation time, indicating that the nitridation period is another decisive parameter determining the catalytic activity of molybdenum nitrides. Figure 8b shows the Arrhenius plots for NH3 consumption over the two molybdenum nitride catalysts. The apparent activation energies of ammonia decomposition were calculated to be 28.4 and 29.4 kcal/mol over the samples nitrided for 7 and 24 h, respectively. The activation energies of the molybdenum nitride catalysts were slightly higher than those of molybdenum carbide (21.3 kcal/mol) and supported Ru catalysts (17.9-20.3 kcal/mol) reported in the literature [47]. Overall, the results of this study reveal the potential of high surface area molybdenum nitrides as catalysts for H2 production via ammonia decomposition.

Catalyst Preparation
Molybdenum nitride having a high specific surface area was synthesized by the temperature-programmed nitridation of α-MoO 3 (Sigma-Aldrich, St. Louis, MO, USA) in a flow of pure ammonia. The optimal synthetic parameters were investigated by changing the pure ammonia flow rate (2-11.5 L h −1 ), heating rate (30-240 • C h −1 ), and nitridation temperature (425-846 • C). The molar hourly space velocity (MHSV, h −1 ) is defined as the molar flow rate of pure ammonia per mole of α-MoO 3 as the molybdenum nitride precursor. Typically, pure molybdenum nitride phase was obtained by the nitridation of α-MoO 3 at a reaction temperature of 700 • C for 7 h at a heating rate of 200 • C h −1 to 200 • C and 60 • C h −1 to 700 • C in flowing ammonia (10 L h −1 for 1 g of α-MoO 3 , i.e., MHSV = 64.2 h −1 ). After cooling, the samples to room temperature in flowing N 2 (100 cm 3 g −1 ), passivation was performed using 1% O 2 /N 2 (100 cm 3 min −1 ) for a minimum of 1 h to avoid the rapid oxidation of the external nitride surface.

Catalyst Characterization
XRD analyses were performed to examine the crystal phases of the prepared catalysts using a D8 Discover with a GADDS detector (Bruker AXS, Billericay, MA, USA). The generator current and voltage were 30 mA and 50 kV, respectively, and Cu K α X-rays (0.15418 nm) were used. The analyses were carried out between 2θ values of 10 • and 90 • at a scanning rate of 0.4 • min −1 . To measure the specific surface area, pore size distribution, and total pore volume of the catalyst used, nitrogen adsorption-desorption measurements were performed at liquid nitrogen temperature (−196 • C) using an ASAP 2020 (Micromeritics, Norcross, GA, USA). The specific surface area was measured in the relative pressure (P/P 0 ) range of 0.05-0.2 using the BET formula. The total pore volume was calculated from the amount of nitrogen adsorbed at P/P 0 = 0.995. The pore size distribution was calculated from the adsorption isotherm using the BJH equation. XPS measurements were performed on a PHI Quantera II analyzer (Al K α =1486.6 Ev, Ulvac-PHI Inc., Chigasaki, Kanagawa, Japan), and the results were used to investigate the electronic states of the surface elements in the molybdenum nitride samples. The binding energy was corrected based on the C 1s peak at 284.6 eV. Changes in the morphology of the catalysts were observed through TEM, and elemental analyses of the catalyst were performed via EDS analysis. SAED measurements were performed to confirm the crystalline phases. An FEI/Talos F200X (Thermo Fisher Scientific, Waltham, MA, USA) was used for TEM measurements, and an Oxford X-Max 50 (Oxford Instruments, Abingdon-on-Thames, UK) was used for EDS analyses; Mn K α radiation (hυ ≤ 129 eV) produced at an acceleration voltage of 200 kV was used. Samples used for TEM and EDS analyses were dispersed in toluene (>99.8%, SAMCHUN, Pyungtaek, Korea) for 5 min using a small fraction of the prepared catalyst. Temperature-programmed decomposition (TPD) experiments were performed to investigate the decomposition of molybdenum nitride in pure He, O 2 , and 20% H 2 /He at flow rates of 100 cm 3 min −1 . The TPD experiments were carried out at room temperature to 700 • C at a heating rate of 10 • C min −1 , and the decomposed gases were analyzed by QMS 200 (Balzers, Nashua, NH, USA). Simultaneous TG-DTA (TA Instruments, New Castle, DE, USA) analyses of molybdenum nitride was performed in flowing helium, pure oxygen, and 20% H 2 /He at flow rates of 50 cm 3 min −1 .

Catalytic Activity Tests
The catalytic activity of molybdenum nitride prepared by the temperature-programmed nitridation of α-MoO 3 for ammonia decomposition was tested at atmospheric pressure. The conversion of ammonia is defined using Equation (1).
Conversion of ammonia (%) = 2 3 F H 2 produced F NH 3 fed × 100 Here, F H2 and F NH3 are the molar flow rates of produced hydrogen and reactant ammonia, respectively. The reactants and products (ammonia, hydrogen, and water) were analyzed using nitrogen as carrier gas and a gas chromatograph (Varian 450GC, Varian Inc., Palo Alto, CA, USA) equipped with a Carbosphere-packed column (1/8" × 1.8 m) and a thermal conductivity detector (TCD).

Conclusions
Temperature-programmed nitridation of α-MoO 3 was performed to prepare molybdenum nitride with a high surface area (>170 m 2 g −1 ). MoO 2 was observed as the intermediate phase. A mixture of molybdenum metal and molybdenum nitride was observed by XRD analyses after the nitridation of α-MoO 3 at temperatures over 800 • C, and this phase transformation caused a rapid decrease in the specific surface area. When the prepared molybdenum nitride was exposed to air, the diffusion of oxygen into the bulk nitride resulted in a volume expansion, and a corresponding gradual reduction in the surface area and total pore volume arising from pore blocking was also observed. The easy diffusion of oxygen into the molybdenum nitride resulted in the formation of oxynitrides, but up to 90% of the initial specific surface area was recovered after H 2 treatment. The catalytic activity of molybdenum nitride for ammonia decomposition was dependent on the degree of nitridation, and, of the molybdenum nitride samples, that with the highest degree of nitridation showed the highest catalytic activity. The molybdenum nitride obtained by the nitridation at 650 • C for 24 h exhibits the conversion of ammonia higher than 94% at the same temperature without significant catalyst deactivation.  Figure S2: (a) XRD pattern of nitride sample and those after temperature-programmed decomposition of the nitride sample at (b) 700 • C at a heating rate of 10 • C min −1 in flow of pure He (100 cm 3 min −1 ) and at (c) 700 • C at a heating rate of 10 • C min −1 in flow of 5% O 2 /He (100 cm 3 min −1 ). The nitride sample in (a) was obtained by the nitridation of MoO 3 under the following reaction conditions: reaction temperature of 700 • C for 7 h at a heating rate of 60 • C h −1 in pure ammonia at a flow rate of 10 L h −1 for 1 g of MoO 3 (MHSV = 64.2 h −1 ). Figure S3: XRD patterns of molybdenum oxynitride samples exposed to air as a function of exposure time: (a) passivated samples after nitridation at 700 • C for (b) 3, (c) 5, and (d) 7 days; Figure S4: N 2 adsorption-desorption isotherms and pore size distribution of molybdenum nitride exposed to air for 60 days and (b) treated at 500 • C for 2 h in pure H 2 (30 cm 3 min −1 ); Table S1: Mass variation, chemical compositions, and XRD phases of nitride samples obtained after the nitridation of MoO 3 as a function of reaction temperature, heating rate, and MHSV. Nitridation was carried out at the final temperature for 7 h.

Conflicts of Interest:
The authors declare no conflict of interest.