On Catalytic Behavior of Bulk Mo₂C in the Hydrodenitrogenation of Indole over a Wide Range of Conversion Thereof

Abstract

The catalytic activity of bulk molybdenum carbide (Mo₂C) in the hydrodenitrogenation (HDN) of indole was studied. The catalyst was synthesized using a temperature-programmed reaction of the respective oxide precursor (MoO₃) with the carburizing gas mixture of 10 vol.\% CH₄/H₂. The resultant material was characterized using X-ray diffraction, CO chemisorption, and nitrogen adsorption. The catalytic activity was studied in the HDN of indole over a wide range of conversion thereof and in the presence of a low amount of sulfur (50 ppm), which was used to simulate the processing of real petroleum intermediates. The molybdenum carbide has shown high activity under the tested operating conditions. Apparently, the bulk molybdenum carbide turned out to be selective towards the formation of aromatic products such as ethylbenzene, toluene, and benzene. The main products of HDN were ethylbenzene and ethylcyclohexane. After 99% conversion of indole HDN was reached (i.e., lack of N-containing compounds in the products was observed), the hydrogenation of ethylbenzene to ethylcyclohexane took place. Thus, the catalytic behavior of bulk molybdenum carbide for the HDN of indole is completely different compared to previously studied sulfide-based systems.

1. Introduction

The petroleum industry is under constantly increasing pressure exerted by legislators in terms of improving the quality of diesel fuel towards reducing the toxicity of its exhaust emissions. In recent decades, the Environmental Protection Agency (EPA) has established several regulations targeted at the lowering of sulfur emission in diesel fuel from 50 down to 10 ppm [1,2,3]. Indeed, the decrease in sulfur content in diesel fuels has an additional advantage—it allows the use of the catalytic converters in the engines so that nitrogen oxide emissions may also be reduced. However, it is well known that organic compounds containing nitrogen are inhibitors of the hydrodesulfurization (HDS) process. That is why the permissible content of such N-compounds should also be reduced to a level below 60 ppm in order to make the very deep HDS process feasible (less than 10 ppm of S) and also reduce the emission of NOx during the combustion of fuel in an engine [4,5,6].

Nowadays, transition metal sulfides are commonly used as industrial catalysts of hydrotreating processes, mainly HDS and hydrodenitrogenation (HDN). However, recently transition metal nitrides and carbides (especially molybdenum and tungsten carbides and nitrides) have attracted increasing attention of researchers due to their promising potential as catalysts for use in commercial hydroprocessing and other reactions. For example, in addition to HDN and HDS [7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23], their activity has been extensively studied in other processes, i.e., hydrogenolysis, isomerization, aromatization, hydrogenation/dehydrogenation [24,25,26,27,28,29], hydrodechlorination [30,31], water gas shift (WGS) and steam reforming [32,33,34,35], hydrodeoxygenation (HDO) of bio-oils [36,37,38,39,40,41], methanation reaction and hydrogen production [42,43], as well as ring-opening [44].

The temperature-programmed reaction as a method of synthesis of high specific surface area bulk nitrides and carbides powders was originally reported by Volpe and Boudart [45,46]. In brief, it relies on a high-temperature reaction of relevant oxides with NH₃/H₂ (nitridation) or CH₄/H₂ (carburization) using a very slow heating rate. After synthesis, these materials are pyrophoric, and a passivation step is required to prevent the bulk oxidation thereof when exposed to air. A facile method used for this purpose is treatment with a 1 vol.\% of O₂/He mixture at room temperature (pulse method), which leads to the formation of an oxygen surface monolayer. Subsequently, these superficial oxygen atoms can react with hydrogen atoms incorporated into the bulk material during the nitridation or carburization processes. These hydrogen atoms can diffuse from the bulk to the surface, which can lead to the formation of –OH surface groups. As a result, two types of sites exist on the surface of nitrides or carbides prepared at high temperature after exposure to oxygen, i.e., metallic sites (transition metal atoms modified by C and/or N) and acidic ones (Brønsted groups) [47].

The nitrogen compounds present in diesel fuel are mainly non-basic-type compounds such as indoles and carbazoles, which account for approximately 70% of the nitrogen content in atmospheric gas oil (ca. 300 ppm), while the next 30% are basic-type compounds (namely quinolines) [48]. Therefore, many authors have used indole, even recently, as a model compound for the catalytic study on HDN over various catalytic systems [2,8,47,49,50,51,52,53,54,55,56,57,58,59,60,61,62].

The main goal of the present research was to investigate the catalytic behavior of bulk molybdenum carbide Mo₂C in the reaction of hydrodenitrogenation of indole. The catalytic runs were carried out at a constant volume H₂/feed ratio of 600, at a total pressure of 60 bar, and at the temperature of 613 K. The contact time varied between 0.13 and 1.07 s. It was found that the conversion of indole governs the preferential reaction pathway. Namely, low indole conversion favors the direct denitrogenation route, while in the case of high conversion, changes the reaction path towards full hydrogenation. This brings important advantages in terms of industrial purposes, i.e., by proper adjustment of the processing parameters, the consumption of hydrogen in HDN could be significantly reduced.

2. Results and Discussion

2.1. Characterization of the Catalyst

The specific surface area (S_g) of the fresh passivated Mo₂C was determined to be 33 m² g⁻¹. The XRD patterns (Figure 1) collected for a passivated catalyst before and after HDN reaction may be assigned to the lattice planes of the hexagonal compact phase (hcp) β-Mo₂C structures.

The sharp peak at 2θ = 36° of a weak intensity may be assigned to the characteristic diffraction line of the residues of silicon carbide SiC (the reflection ascribed to (1 1 1) Miller index), which was used to dilute the catalyst in the catalytic bed (see Section 3.4). As seen, no alteration of the crystallographic structure of the Mo₂C catalyst was observed after the test of HDN of indole. The experimental CO uptake of the fresh sample was equal to 142 μmol g⁻¹ and decreased down to 7.69 μmol g⁻¹ after the HDN reaction. This change may be due to the coking of the catalyst’s surface. On the other hand, it should be noted that the catalyst was exposed to contact with air after it was removed from the reactor for the purpose of studying the phase composition (XRD). This, in turn, could lead to passivation and a significant reduction in the number of active sites adsorbing CO. For this reason, care must be taken when analyzing the result of CO sorption for the spent catalyst. More details are provided in our former work [7].

2.2. HDN of Indole over Mo₂C

The possible reaction scenarios of indole HDN over the bulk Mo₂C with all detected compounds are presented in Scheme 1.

As seen, notwithstanding the further reaction pathways, the first step thereof is the hydrogenation of indole towards indoline, followed by the opening of the N-containing ring to o-ethylaniline (OEA). The latter one can undergo either direct denitrogenation (called DDN route or route #1 in this work) to the ethylbenzene or full hydrogenation of the ring before denitrogenation (called HYD or route #2) to o-ethylaniline. The third possibility is OEA dealkylation to aniline (route #3).

It is pertinent to mention that in our research, no intermediate compound such as octahydroindole (OHI) was detected in the reaction product mixture. This is in agreement with the results presented by Bunch et al. [53] for the HDN of indole over sulfides as hydrotreating catalysts. They stated that in the presence of H₂S, indole undergoes transformation to OEA because full hydrogenation of indoline into OHI was not observed. It may be concluded that in case of using molybdenum carbide, the preferred route is the formation of OEA. Benzene present in the products originates from the subsequent dealkylation of ethylbenzene (EB) and toluene (T), as well as from the hydrogenolysis of C_(sp2)–N of aniline. Similar results were reported by Li et al. [8], who carried out the indole HDN process over Mo carbides and nitrides.

The presence of traces of ethylcyclohexene (ECHen) for the tests with short contact time undoubtedly indicates that it was formed by the extraction of a hydrogen atom from the carbon atom in the β position of the o-ethylcyclohexylamine molecule (OECHA). Indeed, the hydrogenolysis of the C_(sp3)–N bond of OECHA is significantly easier compared with the C_(sp2)–N bond scission of OEA.

The conversion of OEA occurs following two main competitive reaction routes. The first one (route #1) results in the formation of ethylbenzene (EB), toluene (T), and benzene (B) (hydrogenolysis of C_(sp2)–N bond resulting in the formation of ammonia and EB). Route #2 leads to the formation of OECHA, ECHen, and ethylcyclohexane (ECH) (hydrogenation of aromatic ring to OECHA) followed by hydrogenolysis of C_(sp3)–N bond resulting in the elimination of NH₃ and formation of ethylcyclohexene (ECHen) and, ultimately, ethylcyclohexane (ECH) [47,56]. The aforementioned hydrogenolysis of C–N bond in OEA molecules occurs according to the β-elimination mechanism (Hoffman type), although it may also follow the alternative mechanism of nucleophilic substitution that would result in the same products [63]. The formation of small amounts of benzene and toluene is the result of hydrogenolysis of the C–C bond, which takes place on the metallic centers of molybdenum carbide [47].

The hydrogenation of EB towards ECH is inhibited because of the presence of nitrogen-containing compounds being stronger adsorbed on the active sites. Since measurable amounts of aniline (A) were also observed in the reaction products, an additional pathway of OEA dealkylation (route #3 in Scheme 1) could not be excluded. Aniline is formed after the hydrogenolysis of the C–C bond between the ethyl and phenyl group in the OEA molecule. However, for the sake of simplicity and taking into account the absence of OHI and the negligible amounts of aniline (<0.5%), only two pathways for the HDN of indole were considered (routes #1 and #2) as shown below and discussed hereinafter.

Table 1. Distribution of products (mole %) during the indole HDN over Mo₂C catalyst at 613 K, 6 MPa, H₂/feed = 600, and at a contact time of 0.37 s.

HDN Products	Share [Mole %]
1—Nitrogen compounds [%]:   Indole   Indoline   o-ethylaniline (OEA)   Aniline (A)	0 Traces 20 Traces
2—Direct denitrogenation (DDN, route #1, Scheme 1) [%]:   Ethylbenzene (EB)   Toluene (T)   Benzene (B)	34 11 2
3—Hydrogenation (HYD, route #2, Scheme 1) [%]:   Ethylcyclohexane (ECH)   Ethylcyclohexene (ECHen)	32 Traces
Total conversion of indole [%] (1 + 2 + 3)	99
HDN conversion [%] (2 + 3)	79

displays the shares of indole, all the intermediates and final products, as well as total conversion, and HDN conversion over bulk Mo₂C for a selected contact time of 0.37 s.

As seen, the main reaction products are ethylbenzene, ethylcyclohexane, and compounds resulting from ethylbenzene dealkylation, i.e., toluene and benzene. Moreover, nitrogen-containing compounds such as indoline, o-ethylaniline, and aniline were also present in the products mixture. Additionally, traces of ethylcyclohexene have been detected. The data collected in Table 1 as well as in Figure 2 indicate that molybdenum carbide shows relatively high conversion under the tested operating conditions. Indeed, for a contact time of 0.37 s, the HDN conversion (i.e., degree of HDN) over Mo₂C reached 79%, while the total conversion of indole was equal to 99%.

The selectivity towards route #1 (DDN route) and #2 (HYD route) was calculated for different conversions and is presented in

Table 2. Selectivity between route #1 and #2 for Mo₂C catalyst during indole HDN carried out at 613 K, 6 MPa of H₂, H₂/feed = 600, S = 50 ppm.

Total Conversion over Mo2C [Mole %]	$\mathbf{Selectivity}\frac{\mathbf{EB}+\mathbf{T}+\mathbf{B}}{\mathbf{ECH}}$
99	1.4
33	0.7
18	0.2

. Interestingly, the favored reaction pathway strongly depends on the conversion of indole (and, consequently, on contact time). Namely, in the case of low conversions (<33%), the selectivity data indicate that the HYD route (route #2) is favored, whereas for the highest conversion, the DDN route (route #1) takes over with a selectivity of 1.4 at 99% conversion. With this, it appears evident that for the highest conversion, the DDN scenario is the preferred one.

2.3. Kinetic Study of the HDN of Indole over Bulk Mo₂C

Figure 3 represents the distribution of the products (as mole %) of the HDN of indole over Mo₂C vs. contact times (t_c).

Apparently, three regions can be distinguished. In the first one (region I: t_c < 0.27 s), the HYD route prevails with the presence of indoline and OEA to some extent. The second region (region II: 0.27 < t_c < 0.37 s) is characterized by a sudden decrease in indole mole fraction along with the increase in OEA mole fraction as well as a significant increase in EB and ECH concentrations. The linear drop in indole content in regions I and II indicates that the indole extinction process is of zero order. In the second region, the real HDN reaction of OEA occurred (in this region, OEA is clearly the most abundant intermediate).

When the contact time is longer than 0.37 s (region III), indole is consumed almost quantitatively (i.e., the total conversion reaches 99%). The mole fraction of OEA decreases and takes over the role of reactants leading to HDN products such as EB and ECH (Figure 3). The simultaneous increase in both EB and ECH shares indicates that the reaction occurs according to both pathways (route #1 and #2); however, it is worth noting that route #1 is the favored one. These results are contradict the results reported by Ozkan et al. [54]. Therein, the authors have shown that the presence of H₂S could invert the selectivity towards route #2. In our work, the presence of sulfur (dimethyl disulfide (DMDS) as a source of H₂S), did not switch the reaction pathway. The alteration in selectivity between ECH and EB observed for t_c > 0.8 s (Figure 3) is only due to the progressive hydrogenation of EB to ECH, which can occur without appreciable limitations because of the absence of N-containing compounds as shown elsewhere [57,64,65,66,67,68,69,70]. To the best of our knowledge, such specific behavior (i.e., the privileged pathway of aromatics formation) is a unique feature of the molybdenum carbide catalyst.

Over Mo₂C, aromatic products are favored because the nucleophilic substitution S_N2 is the dominant mechanism governing the HDN reactions [65]. On the other hand, it is well known that the presence of indole and indoline strongly inhibits the conversion of ethylaniline [66] and, as a consequence, it inhibits the whole HDN process [67]. The formation of ethylbenzene as the dominant HDN product of indole was also reported by Piskorz et al. [68]. Based on the density functional theory (DFT) simulations, they found out that the selectivity of the reaction depends on the manner of the H: attack on the aromatic ring or the amino group of the OEA intermediate. The difference in the energy barriers between the DDN and HYD molecular paths calculated by the authors was equal to 33.5 kJ mol⁻¹ (8 kcal mol⁻¹) [68]. This justifies the higher amounts of the aromatic compounds (i.e., EB) compared to aliphatic ones (such as ECH).

At 99% indole conversion, the dominating products are EB and ECH. When the contact time is longer than 0.54 s, there are no more nitrogen-containing products. Interestingly, a change can be seen in the preferred route of the reaction—ECH is produced with increased efficiency, while the amount of formed EB decreases. While a significant number of nitrogen-compounds are still present in the feed, hydrogenation of the unsaturated products is inhibited.

In the case of the contact time t_c > 0.53 s (no N-containing compounds), the effect of the total change in the selectivity was observed. This may be associated with the hydrogenation of EB to ECH. This is in accordance with the results reported formerly by Mordenti [69], who indicated that OEA (as the main N-containing intermediate product comprising nitrogen) is responsible for blocking the catalytic centers. Consequently, the hydrogenation of ethylbenzene to ethylcyclohexane is impeded. Similar research has been published previously by Sayag [70], who presented in his work evidence supporting the proposed inhibitory effect of o-propylaniline (OPA) in isopropyl benzene hydrogenation to propylcyclohexane.

The same researcher with co-workers [57] studied the HDN of indole over Mo₂C supported on activated carbon black composites (Mo₂C/CBC). They reported a maximum HDN conversion (100%) reached over all the studied catalysts notwithstanding the contact time. The authors found out that the increase in saturated products comes only from the hydrogenation of the unsaturated ones. Indeed, while N-compounds are still present in the feed in abundant amounts, hydrogenation of the unsaturated products is inhibited because of a competitive strong adsorption of these molecules onto the Mo₂C surface [57,67].

A similar phenomenon was observed by Da Costa et al. [64], who studied the influence of 4,6-dimetylodibenzotiophene (4,6-DMDBT) on hydrogenation of biphenyl (BPh) over Mo₂C supported on Al₂O₃. When 4,6-DMDBT was present in the reaction system, the route of hydrogenation of BPh to cyclohexylbenzene (CHB) was impeded because of strong adsorption of S-compounds on the surface of the catalyst. Furthermore, a similar effect was observed in the case of HDS of dibenzothiophene (DBT) [16]; the CHB was produced from tetrahydrodibenzothiophene because of the strong adsorption of S-compounds. Furthermore, Mamede et al. [71] have observed a similar behavior when carrying out the tests of hydrogenation of toluene over Mo₂C in the presence of thiophene.

For the sake of comparison,

Table 3. Selectivity between route #1 and #2 over different catalysts tested in HDN of indole. Data from literature compared with the results reported herein.

Catalyst/Test Conditions	Temperature of Reaction [K]	Pressure [bar]	Selectivity ECH/EB	Reactor	Sulfidation Conditions	Ref.
CoMoS/Al2O3	593	50	300	Autoclave	–	[67]
NiMoS/Al2O3	593	70	15	Fixed-bed reactor	673 K; 10% H2S; 10 h	[53]
NiMoS/Al2O3	593	70	25	Fixed-bed reactor	673 K; 10% H2S; 10 h	[53]
NiMoS/Al2O3	593	70	40	Fixed-bed reactor	673 K; 10% H2S; 10 h	[53]
NiMoS CoMoS	613	35	7.5	Fixed-bed reactor	673 K; 10% H2S; 2 h	[73]
Mo2C/CBC no sulfur H2/feed = 600	623	50	1.25	Fixed-bed reactor	–	[74]
Mo2C/CBC H2/feed = 200	623	50	2	Fixed-bed reactor	–	[74]
Mo2C/C in situ	613	50	2	Fixed-bed reactor	–	[69,75]
Bulk Mo2C with 50 ppm of S	613	60	0.7	Fixed-bed reactor	–	This paper

reports the chosen literature data on HDN of indole over carbides and sulfides. Sulfide catalysts feature a high selectivity towards the saturated product (ECH), in opposition to observation over molybdenum carbide. This effect is due to the enhancement of the C_(sp3)–N bond cleavage in the presence of sulfur [72] instead of the C_(sp2)–N bond cleavage. The selectivity of HDN of indole over Mo₂C found in majority of works is shifted towards the aromatic product (i.e., EB). This is consistent with our results (see last line in Table 3). It can be concluded that, despite the presence of 50 ppm of S, the behavior of Mo₂C is completely different than that of the sulfide catalysts. The undoubted advantage of Mo₂C is the possibility of significant reduction in hydrogen consumption for the process. As such, bulk Mo₂C may be a promising alternative for commercial HDN catalysts used on a technological scale.

3. Materials and Methods

3.1. Materials

MoO₃ (≥99.5%, Honeywell Fluka^TM, Charlotte, NC, USA) was used as the precursor for the catalyst synthesis. The gases used were: hydrogen (H₂, Air Liquide, grade C, purity > 99.995%), helium (He, Air Liquide, grade C, purity > 99.995%), argon (Ar, Air Liquide, grade C, purity > 99.995%), methane (CH₄, Air Liquide, grade N30, purity > 99.9%), and oxygen (O₂, Air Liquide, grade C, purity > 99.5%). Indole (>99%, Honeywell Fluka^TM, Charlotte, NC, USA) was used as model compound for the kinetic study. Dimethyl disulfide (DMDS, ≥98%, Honeywell Fluka^TM, Charlotte, NC, USA) dissolved in decaline (≥98%, Honeywell Fluka^TM, Charlotte, NC, USA) was used as a sulfur source.

3.2. Synthesis of Bulk Mo₂C

Molybdenum carbide (Mo₂C) was prepared using a modified procedure reported by Volpe and Boudart [46]. The temperature profile used for the synthesis is presented in Figure 4.

Molybdenum trioxide was put on a sintered holder placed in a quartz reactor and was subjected to carburization with a mixture of 10 vol.\% CH₄/H₂ at a flow rate of 13.2 L g⁻¹ h⁻¹. The temperature was raised linearly from 300 to 973 K at a heating rate of 53 K h⁻¹ and then the material was held at 973 K for the next 1 h. Eventually, the sample was cooled to room temperature (RT) under the constant flow of the carburizing mixture. In the case when exposure to air was needed, the sample was passivated using 1 vol.\% O₂/He (pulse method) in order to form a protective oxide layer on its surface.

3.3. Characterization Techniques

Structural characterization of the Mo₂C catalyst was carried out by means of X-ray diffraction (XRD) using a SIEMENS D-500 diffractometer (Siemens, Berlin, Germany) working with the CuKα monochromatized radiation (λ = 1.54184 Å). The XRD patterns were analyzed using HighScore (Malvern Panalytical, United Kingdom) software. The degrees of crystallinity of the catalyst were estimated based on the intensity of the reflections recorded in the angular range of 2θ = 10–90°.

A Quantachrome–Quantasorb Jr apparatus (Quantachrome Instruments, Boynton Beach, FL, USA) was employed for the S_g measurement in dynamic conditions. The S_g of the catalyst was calculated from the N₂ adsorption/desorption at different partial pressures of nitrogen. The standard S_g measurement using the Brunauer-Emmett-Teller (BET) model is based on a three-point analysis based on the adsorption of mixtures containing 10, 20, and 30 vol.\% of N₂ in He. Prior to the measurement, the material (100 mg) was outgassed under flowing nitrogen at 623 K for 2 h.

CO adsorption is a classic technique commonly used for estimating the number of surface metallic sites. It is also employed to titrate the active centers on the surfaces of the transition metal carbides or nitrides [47,60,61,62]. The CO titration was performed in situ, i.e., in the synthesis reactor without exposing the fresh material to air or to the passivating mixture after synthesis. The pulses of given quantities of CO (17 μmol) were injected to the reactor containing Mo₂C with regular intervals at RT under He flow (40 mL min⁻¹) purified before by an oxygen trap (Oxisorb, Messer Griesheim GmbH, Bad Soden am Taunus, Germany). After each pulse, the number of CO molecules not chemisorbed was analyzed using a conventional detection system equipped with a TCD. The CO injections were continued until the probe molecule quantitatively saturated the catalyst’s surface. The acquired data were processed and the number of micromoles of the chemisorbed CO per gram of the sample was calculated.

3.4. Kinetic Study of HDN of Indole

The HDN tests of indole were carried out using a downflow fixed-bed microreactor working in a high-pressure flow system. The catalytic tests were performed over 0.8 g of catalyst (diluted with SiC at a mass ratio SiC/Mo₂C = 5:1) in the presence of sulfur (50 ppm S), evolved by the decomposition of 70 ppm dimethyl disulfide (DMDS) in the feed. The addition of sulfur was employed in order to simulate the HDN process of real petroleum intermediates that always contain certain amounts of S-compounds. The liquid feed containing 2.5 wt. % of indole dissolved in decaline was supplied to the flow manifold by a high-pressure piston pump (Gilson model 307, Artisan Technology Group, Mercury DriveIL, Champaign, IL, USA). Hydrogen flow was controlled by a mass flow controller (Brooks 5850TR, Brooks Instruments, PA, USA), while the total pressure was adjusted by a back-pressure regulator (Brooks 5866, Brooks 5850TR, Brooks Instruments, PA, USA). The temperature in the reactor oven was adjusted using a temperature controller (Hanyoung MX9, Hanyyoung, Vietnam).

The kinetic study of the HDN process was performed at a total pressure of 60 bar, and a temperature of 613 K using the H₂/feed volume ratio of 600 and contact times (t_c) between 0.13 and 1.07 s. The contact time (t_c) was calculated as follows:

$$t_c\left[\mathrm{s}\right]=\frac{\mathrm{Catalyst}\mathrm{volume}\left({\mathrm{cm}}^3\right)}{{\mathrm{H}}_2/\mathrm{feed}\mathrm{flow}\left({\mathrm{cm}}^3{\mathrm{s}}^{-1}\right)}$$

(1)

The liquid products were collected with constant intervals of 1 h and analyzed via manual injections (0.8 μL) to a gas chromatograph (HP 4890, Agilent Technologies, Santa Clara, CA, USA) equipped with a capillary column (DB1: 30 m × 0.25 mm × 0.25 µm, Agilent Technologies, Santa Clara, CA, USA) and a FID detector. The temperature of the injector was kept at 523 K, and the split ratio was 1:50. High-purity argon was used as the carrier gas. The oven temperature program was the following: the temperature was increased from 323 K to 453 K at a rate of 5 K min⁻¹, and then held at 453 K for 1 min. Next, the temperature was increased from 453 K to 493 K at 10 K min⁻¹ and held at this temperature for 2 min. The total time of the analysis was 33 min.

The carbide catalyst was stabilized in situ for 10 h using a mixture of 70 ppm of DMDS solution in decaline in hydrogen (60 mL min⁻¹) as a carrier gas, at the processing temperature under the total pressure of 60 bar.

The total conversion of indole, degree of HDN, and selectivity were calculated according to Equations (2)–(4):

$$\mathrm{Conv}.\mathrm{of}\mathrm{indole}[\%]=\frac{{\mathrm{N}}_{\mathrm{N}}}{{\mathrm{N}}_{\mathrm{N}}+{\mathrm{N}}_{\mathrm{indole}}}·100\%;$$

(2)

$$\mathrm{Degree}\mathrm{of}\mathrm{HDN}[\%]=\frac{{\mathrm{N}}_{\mathrm{H}}}{{\mathrm{N}}_{\mathrm{N}}+{\mathrm{N}}_{\mathrm{indole}}}·100\%;$$

(3)

$$\mathrm{Selectivity}[-]=\frac{\mathrm{EB}+\mathrm{T}+\mathrm{B}}{\mathrm{ECH}};$$

(4)

where:

• N_N is the mole % of all products of decomposition of indole (including N-containing compounds except indole);
• N_indole is the mole % of indole in the liquid product after reaction;
• N_H is the mole % of non-nitrogen products from the decomposition of indole;
• EB + T + B are mole % of ethylbenzene, toluene, and benzene, respectively;
• ECH means mole % of ethylcyclohexane.

4. Conclusions

Molybdenum carbide was found to be active in HDN of indole in the presence of a low amount of sulfur (50 ppm). It was found that the preferred HDN reaction path depends on the indole conversion. Namely, in the case of low conversion, direct denitrogenation is the favored reaction route with EB being the main product. Contrarily, in the case of the highest conversion, the privileged pathway is the full hydrogenation towards ethylcyclohexane. The preferential direct denitrogenation route is opposite to the mechanism, which was hitherto observed over sulfide catalyst, which in general leads mainly to the hydrogenated products. Thus, the indole HDN mechanism over the bulk molybdenum carbide is different compared to that of sulfide catalysts. Over the studied Mo₂C material, the hydrogenated products (such as ECH) become the main products only in the case where N-containing products are eradicated from the reactant mixture. Simultaneously, a minor dealkylation route towards aniline was observed. This brings certain significant advantages, i.e., the consumption of hydrogen in HDN processes may be significantly reduced. To the best of our knowledge, herein we report on this peculiar behavior of the Mo₂C catalyst for the first time.