Key Role of Precursor Nature in Phase Composition of Supported Molybdenum Carbides and Nitrides

In this work, we studied the effect of molybdenum precursors and the synthesis conditions on the final phase composition of bulk and supported molybdenum carbides and nitrides. Ammonium heptamolybdate, its mixture with hexamethylenetetramine, and their complex were used as the precursors at different temperatures. It was investigated that the synthesis of the target molybdenum nitrides strongly depended on the structure of the precursor and temperature conditions, while the synthesis of carbide samples always led to the target phase composition. Unlike the carbide samples, where the α-Mo2C phase was predominant, the mixture of β-Mo2N, MoO2 with a small amount of metal molybdenum was generally formed during the nitridation. All supported samples showed a very good dispersion of the carbide or nitride phases.


Introduction
There are three types of bonding between the transition metal and carbon or nitrogen atoms: metal bonding (metal-metal), covalent bonding (metal and non-metal), and ionic bonding (charge between metal and non-metal) [1]. The special crystal structure of transition metal carbides and nitrides is created by inserting carbon or nitrogen into the metal-metal bond is what makes its distance longer than the original. This special bond has exclusive electronic properties, which provide catalytic activity similar to the platinum group metals (Pt, Pd, Ru, etc.) in various reactions [2]. Molybdenum carbides exist in three basic forms: a face-centered cubic (fcc, α-MoC 1-x ), a hexagonal closed packed (hcp, β-Mo 2 C), and a simple hexagonal (hex, MoC) structure, while molybdenum nitrides mainly have a cubic structure (fcc, γ-Mo 2 N) [3].
Molybdenum carbides are widely used as catalysts due to their activity in many reactions, particularly in the water gas shift reaction, deoxygenation, denitrification, desulfurization, oxidation, partial oxidation, hydrotreating (HDS, HDO, HDN), dehydrogenation, isomerization, hydrogenolysis, hydrodemetallization, and methane reforming [4]. Molybdenum nitrides possess a series of unique and superior catalytic properties for HDS, HDO, HDN [4,5], and electrochemical catalysis [6,7]. Possible difficulties in the application of these types of catalysts that may appear are most often related to obtaining those materials with high specific surface areas (usually less than 10 m 2 /g) or high porosity. However, these parameters may be varied by changing of synthesis conditions. High surface area Mo 2 N and Mo 2 C are synthesized by various methods [8,9]: (a) Direct reaction between metal and non-metal; (b) reaction of metal oxide in the presence of solid carbon; (c) reaction of metal or compounds with gas phase reagent; (d) temperature-programmed methods; (e) reaction between metal oxide vapor and solid carbon under vacuum; (f) pyrolysis of an organometallic complex have already presented two basic preparation methods. In the first, authors followed the method reported by Afanasiev [43] and obtained Mo 2 C, while in the second, they used mechanically mixed AHM and HMT with a molar ratio of 1:4 with obtaining metallic Mo and MoO 2 . The researchers also described an impact of different parameters in hydrogen thermal treatment preparation of HMT and AHM, namely, AHM:HMT molar ratios, temperature, or heating rates [44]. They reported that pure Mo 2 C was obtained when the molar ratio of precursors AHM and HMT reached 1:4, and nitride, carbide, and carbonitride composite materials were obtained when the molar ratio was 1:2. They showed that Mo 2 N became a major product phase at a lower temperature (500 • C), and for Mo 2 C a higher temperature (650 • C) was needed.

Preparation of Precursors
Based on the method reported by Afanasiev [43], the initial HMT-AHM precursor complex was synthesized using 50 g of AHM and 86 g of HMT dissolved in 300 mL and 400 mL of distilled water, respectively. The solutions were mixed out together and left for 48 h at 3 • C. The sedimented crystals were separated using a paper filter, rinsed with demineralized water, and dried at room temperature for 3 days. The resulted material was named HMT-AHM 7.5×. The samples of HMT-AHM 5× and HMT-AHM 10× with lesser and greater amounts of HMT, correspondingly, were synthesized with the same method.
Precursors for the supported samples were prepared by incipient wetness impregnation of the HMT-AHM 7.5× ammonia solution on the supports Al 2 O 3 , TiO 2 , ZrO 2 , SBA, BEA, and AZF. Al 2 O 3, impregnated with HMT+AHM-S (2:1), was signed as Al 2 O 3 # . To obtain a high content of molybdenum nitride or carbide phase, the impregnation was performed with a saturated solution. The impregnated support from the same batch was used for the carbide and nitride synthesis. In the case of the AZF support after drying at 120 • C for 6 h, the impregnation was repeated once more.

Synthesis of Molybdenum Carbides and Nitrides
The synthesis of the final nitrides and carbides was carried out in a vertical quartz tubular reactor (UniCRE, Litvínov, Czech Republic) with an internal diameter of 27 mm and length of 1 m (Figure 1), heated to the working temperature by a triple-zone electric oven (CLASIC CZ, spol. s.r.o.,Řevnice, Czech Republic) that was regulated by a PID (proportional-integral-derivative) controller. Each precursor or the impregnated support was placed in a fritted quartz cuvette and placed in the center of the reactor. Further processing of the precursors was done in several steps:

Characterisation
The chemical composition of the supported samples was determined by X-ray fluorescence analysis (XRF) of powder materials using S8 Tiger (Bruker AXS GmbH, Karlsruhe, Germany) with the Rh cathode. The results were interpreted using the Spectra plus software. The non-supported samples were analyzed by the ICP method using ICP-EOS Agilent 725 (Agilent Technologies Inc., Santa Clara, CA, United States). The carbon and nitrogen content was determined by the elemental analysis of the catalyst powder using Flash2000 Elemental Analyzer (Thermo Fisher Scientific S.p.A., Milan, Italy).
The crystallography of all synthesized material catalysts in powder form was analyzed by X-ray diffraction (XRD) analysis using D8 Advance ECO (Bruker AXS GmbH, Karlsruhe, Germany), applying CuKα radiation (λ = 1.5406 Å) with a resolution of 0.02 • and a period of 0.5 s. The patterns were collected in the 2 theta range of 5-70 • and evaluated by using the DIFFRAC.EVA software (Bruker AXS GmbH, Karlsruhe, Germany) with the Powder Diffraction File database (PDF 4+ 2018, International Centre for Diffraction Data).
The textural properties of the samples were determined by N 2 physisorption and mercury porosimetry. The specific surface area (BET) was measured by N 2 adsorption/desorption at 196 • C by using Autosorb iQ (Quantachrome Instruments, Boynton Beach, FL, United States). All the samples were dried under vacuum before the analysis in a glass-cell at 200 • C for 16 h.
The visual appearance was studied by scanning electron microscope (SEM) JSM-7500F (JEOL Ltd., Tokyo, Japan) with a cold cathode-field emission SEM (parameters of measurements: 1 kV, GB high mode). Additional images were obtained by using an optical microscope Jenavert (Carl Zeiss Microscopy GmbH, Jena, Germany) equipped with a Canon EOS 1200D camera (CMOS chip 18 Mpx, Canon, Taiwan). Images with different focusing were folded by the QuickPHOTO CAMERA software (PROMICRA, Prague, Czech Republic).
Physical properties and stability of the precursors and samples were studied by thermogravimetric analysis (TGA) using TGA Discovery series (TA Instruments, New Castle, DE, United States) operating in the temperature range of 40-900 • C (heating 10 • C/min) in the nitrogen flow (20 mL/min, Linde 5.0). A Quadrupole mass detector OmniStar GSD320 (Pfeiffer Vacuum GmbH, Wien, Austria) was used for detection of fragments in SCAN mode with 1450 V voltage of the electron multiplier. Thermal behavior of the samples was analyzed by differential scanning calorimetry (DSC) using Q2000 (TA Instruments, New Castle, DE, United States). Approximately 5-10 mg of a sample was placed into Tzero pierced aluminum pans. An initial temperature was equilibrated at 0 • C, then the samples were cooled down to −50 • C at a rate of 10 • C/min and held for 1 min. After this, the samples were heated up to 450 • C (10 • C/min).

Precursors
Initially, before the synthesis, all of the precursors were characterized by several analytical methods. As seen from Figure 2, X-ray patterns of the precursor structures differ. In the sample HMT+AHM-M (8:1) produced by mechanical mixing of HMT and AHM, the presence of diffraction lines appropriated to AHM and HMT was observed. However, the line corresponding to the formation of the HMT-AHM complex was also present in the sample. The precursor complexes, synthesized using HMT+AHM-S, were similar, but not completely the same to those reported by Afanasiev. The MoO 3 sample had typical diffraction patterns corresponding to molybdenum trioxide (molybdite).
The influence of the HMT excess on the HMT-AHM complex prepared by the Afanasiev method was investigated by varying the molar ratio of HMT:AHM in the sequence of 15:1 (sample HMT-AHM 7.5×), 10:1, and 20:1 (samples HMT-AHM 5× and 10×). The TGA results ( Figure 3) show that the forming complexes were absolutely identical. Their similarity was confirmed by the same decomposition in terms of mass depletion of the final residues and decomposition rates at the corresponding temperatures. Decomposition of other precursors differed from the HMT-AHM 7.5× sample. Thus, in the case of HMT+AHM-M (2:1) and HMT+AHM-M (8:1), a conspicuous signal at 200 • C inherent to HMT decomposition was observed. The crystallization of the HMT and AHM ammonia solution with a molar ratio of 1:1 in the sample HMT+AHM-S (1:1) led to the formation of large transparent crystals that decomposed differently than when only tiny crystals were formed in the sample HMT+AHM-S (2:1). In spite of the fact that the TGA curves of the latter and HMT-AHM 7.5× went through the same trajectory and were very similar, a slight difference in their structures was noticeable.
TGA decompositions of AHM and HMT-AHM were determined using a mass spectrometer. Using the literature data [47,48], it was possible to determine the individual decomposition steps. It was found that thermal destruction at temperatures from 100 to 190 • C resulted in a simultaneous release of H 2 O and NH 3 , giving transition polymolybdate phases. At the same time, fragments corresponding to NO, N 2 O, O 2 , and N 2 were also observed, which may have been caused by the presence of air in the thermogravimetric furnace, which was not hermetically sealed. The HMT-AHM complex differed by the signal of CO 2 detected in the range of 190-275 • C, which further gradually released up to 700 • C.
DSC results displayed in Figure 4 show that the HMT-AHM complexes prepared with different HMT excesses (HMT-AHM 5×, 7.5× and 10×), as in TGA, had identical profiles. In comparison with the samples HMT+AHM-M (2:1) and HMT+AHM-M (8:1), some differences to the HMT-AHM complex in the DSC signals were recorded. HMT+AHM-M (2:1) had a sharp endothermic peak, which occurred in the temperature range of 110-130 • C, and the subsequent endothermic peaks at 250-350 • C correspond to AHM. The change in the molar ratio of HMT:AHM in HMT+AHM-M (8:1) sample also changed the DSC curve. Only two distinct peaks were observable on the record: the first endothermic peak at 170.51 • C corresponded to AHM decomposition and the second one at 248.73 • C to HMT sublimation.
According to the DSC records, the complexes obtained by crystallization of the HMT and AHM ammonia solutions differed among themselves and also in comparison with the HMT-AHM complex prepared by the Afanasiev method. Both HMT+AHM-S (1:1) and HMT+AHM-S (2:1), as well as HMT+AHM-M (2:1) complexes, were characterized by the endothermic peak around 120 • C, which was probably related to the sudden formation of another type of a complex. Despite the fact that the HMT-AHM complex had the same peak, but not quite at intensely, we can conclude that the complex is very similar, but not exactly identical to HMT+AHM-S (2:1). This difference also affected the final structure of molybdenum nitrides and carbides.
SEM results ( Figure 5) show that the surface morphology of the HMT+AHM-S (2:1) sample was more crystallized than the HMT-AHM complex. The MoO 3 sample has small molybdite crystals produced by thermal decomposition of AHM.

Non-Supported MoCx
Non-supported (bulk) molybdenum carbides were synthesised using 20 vol% CH 4 in H 2 with a flowrate of 75 cm 3 /min. All the prepared samples showed a high carbon content (Table 1), indicating their complete conversion to carbides. Only the sample where the AHM precursor was used and the reaction temperature was 600 • C contained much less carbon. This was confirmed by X-ray diffraction, where the majority phase was monoclinic MoO 2 and at the same time a minor orthorhombic modification of α-Mo 2 C occurred. The fractions of the carbide phases and their crystallite sizes are shown in Table 1. As seen from the diffraction patterns of the samples synthesised from the AHM and HMT-AHM precursors at different temperatures ( Figure 6), both of them provided a mixture of αand β-Mo 2 C in the ratio of about 2:1 at 700 • C and only α-Mo 2 C at 800 • C. There was no complete conversion to any carbide phase at low temperatures for AHM (Figure 6a), while HMT-AHM (Figure 6b) gave hexagonal β-Mo 2 C already at 600 • C. Other precursors provided a pure α-Mo 2 C phase only at 700 • C (Figure 7).  The prepared carbides showed a relatively wide range of specific surface area values ( Table 1). The samples synthesized from HMT+AHM-S (2:1) had an area two times higher than when HMT-AHM was used. The specific surface areas of the other samples were very small or even equal to zero, though they all contained the carbide phase.
The crystallites sizes (Table 1) were relatively equal. These values were mostly in the range from 14.3 to 21.3 nm for the α-Mo 2 C phase. The exception was AHM-800, the crystallite size of wich was 40.2 nm. The size of the β-Mo 2 C crystalline phase was possible to determine only for the sample HMT-AHM-600, where this phase appears separately. The size of β-Mo 2 C crystallites in this sample was 4.9 nm. In the case of AHM-700 and HMT-AHM-700, crystallite sizes could not be measured due to the overlap of corresponding diffraction lines. The crystallites sizes were calculated from the reflection of 39.5 • 2 theta for α-Mo 2 C and 37.1 • 2 theta for β-Mo 2 C.
The structure of the prepared non-supported molybdenum carbides was analysed using scanning electron and optical microscopes. The samples synthesised from the HMT-AHM and HMT+AHM-S (2:1) complexes showed a well-crystallized structure, while the microcrystalline structure was peculiar to the samples obtained from AHM (Figure 8). Evaluation of the structure using SEM revealed a distinctive spongy (foam) structure of the prepared samples ( Figure 9).

Non-Supported MoNx
All the non-supported (bulk) molybdenum nitrides were synthesised using the same conditions as for carbides. A mixture of 20 vol% H 2 in N 2 was used as a working gas. The specific surface area of the prepared materials and the nitrogen content were very dependent on the content and type of the nitride phase ( Table 2). The maximum was 29 m 2 /g for the sample prepared from HMT+AHM-M (8:1), followed by HMT-AHM-700. Nitrogen content ( Table 2) clearly showed that the nitride phase was produced only under certain conditions, while carbides were formed in the case of each precursor and temperature. A distinctive feature in nitride samples was the presence of carbon, indicating the presence of the carbide phase. The highest carbon and nitrogen were typical for the samples produced from HMT-AHM (700, 800 • C) and HMT+AHM-S (2:1) precursors. The presence of carbon can be explained by the incomplete decomposition of HMT, which was present in the structure of the used precursor complex. The minor phase was confirmed as orthorhombic carbide phase α-Mo 2 C, based on the XRD data. The use of the AHM precursor ( Figure 10a) did not lead to the formation of molybdenum nitride; only traces of tetragonal β-Mo 2 N modification, the cubic phase of metallic molybdenum, and monoclinic MoO 2 started to occur at 700 • C. Higher temperatures resulted in metal molybdenum and MoO 2 . Nitridation of HMT-AHM (Figure 10b) gave a cubic modification of γ-Mo 2 N accompanied by a small α-Mo 2 C fraction at 700 • C. Formation of β-Mo 2 N was observed at 800 • C and only metallic molybdenum was formed at 900 • C. The sample prepared from the HMT+AHM-S (2:1) precursor at 700 • C was cubic Mo 3 N 2 with a small amount of α-Mo 2 C. The other precursors examined by XRD ( Figure 11) produced a mixture of β-Mo 2 N with MoO 2 , in the case of the precursor HMT+AHM-S (1:1) (the mixture with a small proportion of metal molybdenum). An exception was for the previously mentioned HMT+AHM-S (2:1), where cubic Mo 3 N 2 was formed with a small proportion of α-Mo 2 C. The MoO 3 precursor was only partially reduced to MoO 2 and produced a very small proportion of metallic molybdenum.
The crystallite sizes of the nitride phases are shown in Table 2. The biggest crystallite size was related to the β-Mo 2 N phase and determined at 37.4 • 2 theta. The value was in two ranges of about 18 and 27 nm. Crystallite sizes of other presented phases γ-Mo 2 N (13.8 nm) and β-Mo 2 N (11.7 nm) were calculated at 37.5 • and 37.7 • 2 theta, respectively.
Optical and electron micrographs did not show any noticeable difference between molybdenum carbide and nitride samples (Figures 12 and 13). Nitridation of HMT-AHM and HMT+AHM-S (2:1) precursor complexes also resulted in a very crystalline product, while HMT+AHM-M produced a microcrystalline product (Figure 12). The sponge structure was also typical for these materials (Figure 13), but was more distinct and produced labyrinths in comparison to carbides, where relatively small isolated circular pores were observed. When the HMT+AHM-S (1:1) precursor was used, a mixture of two different phases was observed (Figure 14a). A crystalline phase was composed of tugarinovite (MoO 2 ) and sponge phases consisted of β-Mo 2 N phases. The same crystalline phase was observed when using the MoO 3 precursor (Figure 14b).     The obtained data show that the investigated precursors had a high influence on the physico-chemical properties of the prepared materials. It was found that the carbide phase was easier to obtain than the pure nitride phase, which was observed only when using the HMT-AHM precursor at 800 • C, while the carbide phase was present at a temperature above 600 • C for all used precursors except AHM, giving mainly MoO 2 . The summarized scheme of the carburization/nitridation process of HMT-AHM is presented on Figure 15.
Another fact, confirming that the carbides were produced more easily (compare to nitrides), was evidenced by the presence of α-Mo 2 C formed from HMT-AHM and HMT+AHM-S (2:1) at 700 • C as a concomitant phase. Basing on these results, we can conclude that HMT-AHM and HMT+AHM-S (2:1) were the most effective precursors of molybdenum carbides and nitrides synthesis. Even though they were very similar, they were not identical materials.

Supported Samples
Basing on the synthesis of the non-supported molybdenum carbides and nitrides, the HMT-AHM precursor was chosen to prepare supported samples. HMT+AHM-S (2:1) on Al 2 O 3 # was used for the comparative study. The preparation temperature was 700 • C as a compromise between the formation of the desired phase and the avoidance of structural changes in the supports. The samples exhibited high molybdenum contents ranging from 22.1 to 38.4 wt% ( Table 3). As a consequence, the high content of the carbide (MoC x ) or nitride (MoN x ) phases decreased the initial specific surface area of the SBA-15 and BEA supports. When using the AZF support, the main purpose was to achieve a higher possible molybdenum loading to get a composite material with large cavities (up to tens of micrometres) filled with molybdenum carbide or nitride crystal particles. This was reasoned by the fact that the larger particles of carbide and nitride phases have better stability during the reactions and increased resistance to complete oxidation to crystalline metal oxides [49,50]. These materials were characterized by having the lowest specific surface area (16-28 m 2 /g), determined mainly by the surface of the bulk nitride or carbide phases located in macroporous cavities of the AZF support. The significant decrease of the area from the original was due to clogging micropores, as is the case of the BEA samples. However, there was no noticeable difference in comparing nitride and carbide samples in S BET reduction. Both groups showed very similar surface area values without any prevalence in blocking pores ( Table 3).
The carbides except supported on TiO 2 (0.16% of N) contained no nitrogen, or below the detection limit of the device. The carbon content varied between 0.49 and 2.57% for all samples. A small amount of carbon (0.08-0.45%) was inherent for each nitride sample and nitrogen was in similar values (0.21-2.38%) as carbon in carbides ( Table 3). The low nitrogen and carbon contents could be caused by low amounts of impregnating complexes on ZrO 2 and TiO 2 , which in turn was due to their low pore volumes.
XRD results (Figure 16) showed that when Al 2 O 3 was used as the support, well-dispersed carbides and nitrides phases were formed. In carbides samples, apart from α-Mo 2 C and Al 2 O 3 , the main phase was detected as amorphous. The nitrides, besides the main amorphous phase, also contained β-Mo 2 N accompanied by cubic Mo 3 N 2 and Al 2 O 3 . In addition, the formation of a certain proportion of a cubic AlN phase cannot be excluded. The phase composition of MoC x on Al 2 O 3 # and Al 2 O 3 was identical, but in the case of MoN x , the samples differed by predominant monoclinic MoO 2 on Al 2 O 3 # . The synthesis of Al 2 O 3 # supported materials is more often used than the method with precursor complexes due to its simplicity [51][52][53]. However, in the case of nitrides, there was no complete conversion to the desired phase in these conditions. TiO 2 , in the case of carbide preparation, comprised the crystalline phase of anatase accompanied by the well-dispersed carbide phase containing α-Mo 2 C and β-Mo 2 C in the ratio of 40:60. The possible formation of a small fraction of the cubic delta phase cannot be excluded. The nitride pathway led the β-Mo 2 N phase accompanied by cubic Mo 3 N 2 . The dominant crystalline phase was anatase. A shift of the peak at about 38 • 2 theta to the right can be explained as a partial overlapping of the carbide/nitride peak with the anatase phase of the support. A similar composition of the active phases was observed on ZrO 2 , which due to the small pore volume, contained only about half of the amount of molybdenum against the TiO 2 support. The proportion of the carbide phases was identical, whereas, for nitrides, it was not possible to exclude the presence of the gamma phase. At the carbide pathway, the SBA-15 mesoporous silica provided a mixture of α-Mo 2 C and β-Mo 2 C at a ratio of about 40:60; in the case of the nitride pathway, there was a cubic Mo 3 N 4 and a part of the monoclinic MoO 2 . The diffractogram, when measuring low angles, shows that the mesoporous structure was preserved ( Figure 16). The obtained diffractograms also show the presence of reflections characteristic of microporous zeolite Beta and well-dispersed β-Mo 2 C and cubic Mo 3 N 2 of carbide and nitride samples, respectively. Diffractograms of the samples prepared on AZF support show the presence of the zeolite phase (clinoptilolite) and crystalline α-Mo 2 C and β-Mo 2 C (approximately 60:40) and β-Mo 2 N, respectively. The influence of the support acidity on the formation of carbide or nitride phases was not clear. On the more acidic supports, in the case of carbides, a mixture of αand βphases was produced, while in the mixed nitride phase, the crystalline cubic phase of Mo 3 N 2 was observed. The microphotographs from the optical and electron microscope (Figures 17 and 18) show the structure of the prepared supported samples. The composite samples on AZF had macroporous cavities filled with crystalline carbides (Figure 17). The porous structure did not change, even after heat treatment at 700 • C. On the detailed SEM image, a sponge-like structure with a large number of small pores is clearly visible. The samples prepared on the SBA-15 support are shown in Figure 18. A regular structure of the mesoporous silica SBA-15 was not disordered during the preparation of the carbides or nitrides, even at 700 • C. Its surface was formed by nitride or carbide crystals, probably shaped by precursor leakage from the cylindrical pores during heat treatment. The TEM image of carbided SBA-15 (Figure 18b) clearly shows in detail the basic mesoporous structure of SBA-15 that contained the MoC x nanocrystals inside the cylindrical pores and outside on the catalyst surface. Moreover, the dramatic decrease in the surface area of SBA-15 was confirmed by TEM, as the blockage caused by the formation of crystals inside the long cylindrical mesoporous structure of the support (Figure 18b).   The non-supported nitride sample was applied by Murzin's group in the direct hydrodeoxygenation of algal lipids extracted from Chlorella alga, and found to be a perspective catalyst [54]. Other samples have been tested in different hydrotreatment reactions under batch and flow conditions conducted at UniCRE. Further evaluation of the results are under the publishing process.
According to the literature [22,35,36,44,45,[55][56][57][58][59][60][61][62][63][64][65][66][67][68][69], the main precursor for the synthesis of supported molybdenum catalysts with the nitride or carbide active phases is considered to be AHM (MoO 3 ), and less often the mixture of HMT with AHM, while the preparation of supported catalysts by the impregnation of the HMT-AHM complex is not practically used. AHM impregnated supports are initially calcined in the air to form MoO 3 on the surface of the support. The most commonly used one is Al 2 O 3 followed by SiO 2 /SBA-15. In the case of Mo 2 C containing catalysts, carbon (activated carbon, carbon black, nanotubes, etc.) is preferably used, which contributes to the formation of the carbide phase on the support surface.

Conclusions
This study set out the possibility of molybdenum carbide and nitride synthesis using various precursors and reaction conditions. Based on the analysis of the studied supported and bulk materials, the most suitable preparation conditions to obtain the desired phases were considered. The ability to prepare carbide and nitride phases was demonstrated on the commonly used supports Al 2 O 3 , TiO 2 , ZrO 2 , SBA-15, and zeolite Beta, and also on the less common AZF support. It was investigated that the synthesis of target molybdenum nitrides strongly depends on the structure of the precursor and temperature conditions, while the synthesis of carbide samples always led to the target phase composition. Unlike the carbide samples, where the α-Mo 2 C phase was predominant during nitridation, the mixture of β-Mo 2 N and MoO 2 with a small amount of metal molybdenum was generally formed. However, using the precursor complex obtained from the mixture of hexamethylenetetramine with ammonium heptamolybdate (HMT-AHM), the pure phase of molybdenum nitride was achieved at 800 • C. At 700 • C, γ-Mo 2 N with a small amount of α-Mo 2 C was formed from the same precursor. A similar situation occurred when using the precursor synthesized by evaporation of the ammonia solution of HMT with AHM at a molar ratio of 2:1, where the resulting cubic Mo 3 N 2 phase was also accompanied by a small amount of α-Mo 2 C at 700 • C. Supported samples, even at the high molybdenum content, had the high dispersion of the phases. All carbide samples were composed of αand β-Mo 2 C mixtures. Nitrides supported on Al 2 O 3 consisted of β-Mo 2 N, and on TiO 2 and ZrO 2 consisted of βand γ-Mo 2 N mixtures, SBA-15, and BEA, despite the βand γ-phases also containing Mo 3 N 2 and Mo 3 N 4 . The composite samples prepared on the foamed AZF support included cavities filled with crystalline αand β-Mo 2 C in the case of carbides and the β-Mo 2 N crystalline phase for nitrides.

Funding:
The publication is a result of the project Development of the UniCRE Centre (LO1606), which has been financially supported by the Ministry of Education, Youth, and Sports of the Czech Republic (MEYS) under the National Sustainability Program I. The result was achieved using the infrastructure of the project Efficient Use of Energy Resources Using Catalytic Processes (LM2015039), which has been financially supported by MEYS within the targeted support of large infrastructures.