Hydroconversion of Aromatic Hydrocarbons over Bimetallic Catalysts

: Bimetallic catalysts (BMC) for hydroconversion of aromatic hydrocarbons (ArH) have been designed by modiﬁcation of Ni / Al 2 O 3 with chromium(0) compounds and phosphoromolybdic heteropolyacid (HPA). Catalysts were tested in hydrogenation of benzene and toluene, in hydrodemethylation of pure toluene and they were shown to possess a high activity, selectivity and sulfur tolerance under conditions of the processes above. The activity of BMC in these processes was much higher as compared with that of two-component (Ni-Cr, Ni-HPA) or conventional Ni / Al 2 O 3 catalysts. Using BMC, hydrogenation of benzene and toluene proceeds with activity increased (up to 34–38 mol / kg · h) and toluene hydrodemethylation may be performed with improved selectivity (90.3%) and benzene yield (81%). The high sulfur tolerance of BMC was demonstrated by performing hydrodemethylation of toluene containing up to 500 ppm S.


Introduction
Aromatic hydrocarbon fractions C 6 -C 8 (referred to as BTX) are usually produced from catalytic processes in refineries, such as fluid catalytic cracking, naphtha reforming, and thermal cracking various petroleum fractions. Benzene is the most valuable component in BTX for producing such chemicals as cyclohexane and ethylbenzene that are intermediates in caprolactam and styrene production, respectively [1][2][3]. Unlike benzene, toluene is of limited application, so conversion of the latter (and also C 7 -C 9 aromatics) to benzene is desirable for producing petrochemicals [4,5]. On the other hand, in refineries, motor fuel hydroprocessing including dearomatization is a large-scale process, which scope permanently increases, as environmental limitations become more severe [6][7][8].
Main characteristics of catalysts for hydrogenating aromatics are activity and sulfur tolerance. Catalysts based on Pt [9,10] and Pd [11,12] are the most active, however they have a rather low sulfur tolerance. For example, Pt-Ru significantly lose their hydrogenating activity in the presence of thiophene: 2-6 times for 57 ppm S and dozen times for 560 ppm S [10]. To increase the sulfur tolerance, a combination of Pd/zeolite A-Pd/zeolite Y was proposed [11] or the use of beta-zeolite [12] as a support for Pd. However, these techniques provide only partial problem solving, and catalytic activity decreases, e.g., for naphthalene and tetralin hydrogenation losses of activity are about 40% (for 500 ppm S) [11] and 10-90% (for 200 ppm S) [12]. The new approach comprises obtaining active sulfide phase in situ during the reaction inside porous aromatic frameworks (PAF) as supports. The NiWS phase obtained has high hydrogenating activity and sulfur tolerance, but the preparation process of such catalysts is a very complicated and time-consuming task. Firstly, PAF materials are synthesized by Suzuki cross-coupling reaction between tetrakis-[p-bromophenyl]methane and To elucidate peculiarities of chromium distribution, the magnetization and chemisorption of oxygen were measured for alumina supported samples, including mono-and bimetallic compositions. Table 1 presents data of surface for catalysts examined. As shown, all catalysts possess about the same BET surface area (128-135 m 2 /g), but consumption of O 2 by Ni-Cr samples sharply decreases, and Ni specific surface falls by an order of magnitude (cf. data of Samples 6 and 7).
The effect of organochromium compounds is even more pronounced, and the number of Ni atoms exposed toward oxygen being greatly decreased (cf. data of . It should be noted that neither hydrogen, nor oxygen reacts with CrA/Al 2 O 3 sample. One may propose that decomposition of chromium compound is localized on the surface of nickel crystallites [26]. Due to the low diffusivity of Cr in the Ni lattice [27], all Cr-containing species (CrA, CrCO) cover Ni particles. As a result, the Ni particles are almost unavailable to both O 2 and H 2 .
The specific distribution of chromium provides unusual magnetic properties of Ni-Cr system. The data of measurements ( Figure 1) show that the addition of chromium results in a great increase of magnetization, which conflicts with the data reported for conventional Ni-Cr alloys [28]. It is generally known that metallic Ni and Cr may form an alloy and intermetallide Ni 3 Cr. As for magnetization (B) of these compounds, it decreases to zero at ratio Cr/Ni~0.08 for solid solution and B~0 for Ni 3 Cr, because of paramagnetic nature thereof. As opposed to the known compounds, magnetization of Ni-Cr samples, obtained by decomposition of CrCO (or CrA) on the Ni metal surface, significantly increases in magnetic field. This unusual behavior may be explained by very high dispersivity of Cr particles, with facile domains orientation in magnetic field [29], according to TEM data. As seen from Figure 2, Ni and Cr belong to the same region of nanoparticles having the size of 2-10 nm and with greater predominance for the size of 4 nm. It should be noted, however, that the size of Cr particles is much lower than that of Ni particles (cf. Cr and Ni pictures in Figure 2A). Based on data presented, one may infer that Ni-Cr samples have a core-shell structure being observed in Ni-Cr nanoparticles [30].  Figure 3A shows the XRD profile from 10%Ni-1.5CrСОAl2O3. The metallic Ni peaks are a primary phase in the sample (44.4, 51.82, 76.38° 2θ) [31], but the wide peak at 37.28° 2θ and shoulder in the range of 60.6-63.24° 2θ closely match the most prominent peaks for NiAl2O4 spinels [31,32]. Peaks at 19.5 and 32.66° correspond to nickel aluminum oxide phases as the impurity phase. The close interaction between NiO and Al2O3 at high temperatures can cause the formation of nickel aluminate [32].
In addition, three broad peaks at 2θ = 37.5, 45.62 and 66.8° in the XRD curve indicated the presence of γ-Al2O3 [33]. It was suggested that the catalysts prepared by decomposition of Cr(CO)6 or organochromium compounds are composites of metallic Ni and NiAl2O4 spinels with impurity of nickel aluminum oxides.
The presence of NiAl2O4 was corroborated by TPR experiments. Since the interaction between NiAl2O4 and the support is stronger than that of between NiO and the support, the spinel exhibits a higher temperature of reduction [33]. This fact is clearly indicated by TPR profile of 10%Ni-1.5CrСО/Al2O3 sample ( Figure 3B). We observed a broad peak in the range of 620-1150 K with three maxima at 735, 800 and 1015 K. The first one corresponds for NiO nanoparticles supported on alumina, the next peak to the surface NiAl2O4 phase, and third to the bulk spinels in the cavities of alumina with an areas ratio = 30:55:15 respectively [34,35].  Figure 3A shows the XRD profile from 10%Ni-1.5CrCOAl 2 O 3 . The metallic Ni peaks are a primary phase in the sample (44.4, 51.82, 76.38 • 2θ) [31], but the wide peak at 37.28 • 2θ and shoulder in the range of 60.6-63.24 • 2θ closely match the most prominent peaks for NiAl 2 O 4 spinels [31,32]. Peaks at 19.5 and 32.66 • correspond to nickel aluminum oxide phases as the impurity phase. The close interaction between NiO and Al 2 O 3 at high temperatures can cause the formation of nickel aluminate [32].
In addition, three broad peaks at 2θ = 37.5, 45.62 and 66.8 • in the XRD curve indicated the presence of γ-Al 2 O 3 [33]. It was suggested that the catalysts prepared by decomposition of Cr(CO) 6 or organochromium compounds are composites of metallic Ni and NiAl 2 O 4 spinels with impurity of nickel aluminum oxides.
The presence of NiAl 2 O 4 was corroborated by TPR experiments. Since the interaction between NiAl 2 O 4 and the support is stronger than that of between NiO and the support, the spinel exhibits a higher temperature of reduction [33]. This fact is clearly indicated by TPR profile of 10%Ni-1.5CrCO/Al 2 O 3 sample ( Figure 3B). We observed a broad peak in the range of 620-1150 K with three maxima at 735, 800 and 1015 K. The first one corresponds for NiO nanoparticles supported on alumina, the next peak to the surface NiAl 2 O 4 phase, and third to the bulk spinels in the cavities of alumina with an areas ratio = 30:55:15 respectively [34,35].
It may be expected, therefore, that these samples will also possess very interesting performance as the catalysts. Table 2 shows the results obtained on testing Ni-containing catalysts in hydrogenation of benzene and toluene. It is obvious from Table 2 that after surface modification of nickel-containing catalysts with chromium their bulk hydrogenation activity significantly increases from 1.22 to 2 times (cf. sample No. 1-2 and 3-4). If the decrease of the active surface SNi after modification is considered, the specific activity (A/SNi) increases more than tenfold. It should be stressed, however, that both CrA/Al2O3, and HPA/Al2O3 are inactive over temperatures studied. It can be inferred, therefore, that the increase of catalytic activity observed for modified Ni-Cr catalysts is due to formation of new sites with finest distribution of chromium sub-nanoparticles over Ni/Al2O3 (Figure 2A). It is worth mentioning the sharp increase in activity of samples containing phosphoromolybdic heteropolyacid (cf. Samples 1-2 and 8-9). The effect is well observable both in bulk and specific hydrogenation activity, and being as It may be expected, therefore, that these samples will also possess very interesting performance as the catalysts. Table 2 shows the results obtained on testing Ni-containing catalysts in hydrogenation of benzene and toluene. It is obvious from Table 2 that after surface modification of nickel-containing catalysts with chromium their bulk hydrogenation activity significantly increases from 1.22 to 2 times (cf. sample No. 1-2 and 3-4). If the decrease of the active surface S Ni after modification is considered, the specific activity (A/S Ni ) increases more than tenfold. It should be stressed, however, that both CrA/Al 2 O 3 , and HPA/Al 2 O 3 are inactive over temperatures studied. It can be inferred, therefore, that the increase of catalytic activity observed for modified Ni-Cr catalysts is due to formation of new sites with finest distribution of chromium sub-nanoparticles over Ni/Al 2 O 3 (Figure 2A). It is worth mentioning the sharp increase in activity of samples containing phosphoromolybdic heteropolyacid (cf. Samples 1-2 and 8-9). The effect is well observable both in bulk and specific hydrogenation activity, and being as high as 5 times. This synergistic effect of activity observed for modified Ni-Cr-HPA may be explained by activation of adsorbed aromatics with HPA, because the reactivity of ArH in hydrogenation is correlated with the concentration of the medium strength acid sites, as known for supported Ru/ [5] and Au-Pd/silica-alumina [36] catalysts. It is noteworthy that hydrogenating activity of modified Ni-Cr-HPA catalyst based on active metal -0.014 mole/(min·g) is very close to that of Pt 90 Ni 10 /MCM41 -0.015 mole/(min·g) [9] at 370 K, but our catalyst is free of noble metal.
Another feature of Ni-CrA/Al 2 O 3 catalyst is a stable operation even at low ratios H 2 /ArH such as 3.5. Usually, Ni-containing catalysts suffer from self-poisoning with ArH during hydrogenation thereof. Kinetically, this effect results in a negative reaction order with respect to aromatics [3,37]. Therefore, to maintain the catalytic activity, high ratios H 2 /ArH are required. In contrast to Ni/Al 2 O 3 , where intensive retardation of benzene hydrogenation occurs at ratios H 2 / benzene <10, in case of Ni-CrA/Al 2 O 3 , this ratio may be as low as 4.0 [38]. The effect may be explained by decrease of strong benzene-Ni interaction due to core-shell structure [39] of Ni-Cr catalysts derived from Cr(CO) 6 or CrA (see Figure 2A). As a result, self-poisoning with ArH is excluded and hydrogenation of aromatics may be performed at lower ratios of H 2 /ArH.
In the case of toluene hydroconversion at higher temperatures, the second pathway is possible, namely, hydrodemethylation. The main factor of toluene hydrodemethylation is the selectivity. It is known that the selectivity of conventional nickel-containing catalysts is poor because an intensive aromatic ring opening proceeds [21].
For this reason, in determining the range of shifting the reaction pathway, we have applied the Ni-CrA catalyst and monitored toluene hydroconversion up to 750 K. Figure 4 presents the result obtained under pressure of 0.75 MPa and WHSV = 1.5 h −1 .
Catalysts 2019, 9, x FOR PEER REVIEW 6 of 12 high as 5 times. This synergistic effect of activity observed for modified Ni-Cr-HPA may be explained by activation of adsorbed aromatics with HPA, because the reactivity of ArH in hydrogenation is correlated with the concentration of the medium strength acid sites, as known for supported Ru/ [5] and Au-Pd/silica-alumina [36] catalysts. It is noteworthy that hydrogenating activity of modified Ni-Cr-HPA catalyst based on active metal -0.014 mole/(min·g) is very close to that of Pt90Ni10/MCM41 -0.015 mole/(min·g) [9] at 370 K, but our catalyst is free of noble metal. Another feature of Ni-CrA/Al2O3 catalyst is a stable operation even at low ratios H2/ArH such as 3.5. Usually, Ni-containing catalysts suffer from self-poisoning with ArH during hydrogenation thereof. Kinetically, this effect results in a negative reaction order with respect to aromatics [3,37]. Therefore, to maintain the catalytic activity, high ratios H2/ArH are required. In contrast to Ni/Al2O3, where intensive retardation of benzene hydrogenation occurs at ratios H2/ benzene <10, in case of Ni-CrA/Al2O3, this ratio may be as low as 4.0 [38]. The effect may be explained by decrease of strong benzene-Ni interaction due to core-shell structure [39] of Ni-Cr catalysts derived from Cr(CO)6 or CrA (see Figure 2A). As a result, self-poisoning with ArH is excluded and hydrogenation of aromatics may be performed at lower ratios of H2/ArH.
In the case of toluene hydroconversion at higher temperatures, the second pathway is possible, namely, hydrodemethylation. The main factor of toluene hydrodemethylation is the selectivity. It is known that the selectivity of conventional nickel-containing catalysts is poor because an intensive aromatic ring opening proceeds [21].
For this reason, in determining the range of shifting the reaction pathway, we have applied the Ni-CrA catalyst and monitored toluene hydroconversion up to 750 K. Figure 4   As Figure 4 shows, toluene hydrogenation proceeds up to 620 K with selective formation of methylcyclohexane (>99%). The hydrodemethylation process starts at 630 K with formation of benzene and methane. Data presented in Table 3 show that modification of nickel by chromium provides a great improvement of selectivity and stability thereof. Whereas toluene is totally cracked over Ni/Al2O3 at 680 K, after modification by chromium, the Ni-CrA/Al2O3 catalyst acquires a good selectivity of 85%. In the case of this process, however, we have not found that acid modification of Ni-CrA/Al2O3 improves the catalyst performance, and the only effect being some decrease of selectivity. As Figure 4 shows, toluene hydrogenation proceeds up to 620 K with selective formation of methylcyclohexane (>99%). The hydrodemethylation process starts at 630 K with formation of benzene and methane. Data presented in Table 3 show that modification of nickel by chromium provides a great improvement of selectivity and stability thereof. Whereas toluene is totally cracked over Ni/Al 2 O 3 at 680 K, after modification by chromium, the Ni-CrA/Al 2 O 3 catalyst acquires a good selectivity of 85%. In the case of this process, however, we have not found that acid modification of Ni-CrA/Al 2 O 3 improves the catalyst performance, and the only effect being some decrease of selectivity. For understanding the scheme of hydrodemethylation and ring opening processes, the set of experiments was performed at higher toluene conversions (X T 40-90%) with variation of contact time (WHSV range of 1.5-4.5 h −1 ). The results obtained are set out in Figure 5 as a dependence of benzene selectivity (S B ) from toluene conversion. For understanding the scheme of hydrodemethylation and ring opening processes, the set of experiments was performed at higher toluene conversions (XT 40-90%) with variation of contact time (WHSV range of 1.5-4.5 h -1 ). The results obtained are set out in Figure 5 as a dependence of benzene selectivity (SB) from toluene conversion. As Figure 5 shows, benzene selectivity is actually constant up to toluene conversion of 60%. It means that hydrodemethylation and ring opening proceed as parallel reactions, and the ratio of their intensity (k1/k2) being ≥10/1. However, at toluene conversions more than 60%, benzene is also decomposed by consecutive reaction, with its selectivity being dropped. k2 k1 k3 CH4← Toluene → Benzene → 6 CH4 One may suppose that Reactions 1, 3, and 2 proceed over different sites of catalyst, and benzene hydrocracking to methane (Reaction 3) being proceed on acidic sites. In this case, some decrease of catalyst acidity may suppress the ring opening process, and as a result, benzene selectivity will be improved.
The results obtained for catalyst modified by mono-Li salt of HPA are presented in Table 3 and Figure 5. These data confirm the assumption above on the role of acidic sites in benzene hydrocracking. As a matter of fact, the use of LiHPA salt instead of heteropolyacid suppresses Reaction 3 and benzene selectivity increases from 62% to 90.3%. Therefore, we may conclude that Ni-CrA-LiHPA catalyst developed has high activity (~3.8 h -1 ) that is at least dozen times more that of As Figure 5 shows, benzene selectivity is actually constant up to toluene conversion of 60%. It means that hydrodemethylation and ring opening proceed as parallel reactions, and the ratio of their intensity (k 1 /k 2 ) being ≥10/1. However, at toluene conversions more than 60%, benzene is also decomposed by consecutive reaction, with its selectivity being dropped.
One may suppose that Reactions 1, 3, and 2 proceed over different sites of catalyst, and benzene hydrocracking to methane (Reaction 3) being proceed on acidic sites. In this case, some decrease of catalyst acidity may suppress the ring opening process, and as a result, benzene selectivity will be improved.
The results obtained for catalyst modified by mono-Li salt of HPA are presented in Table 3 and Figure 5. These data confirm the assumption above on the role of acidic sites in benzene hydrocracking. As a matter of fact, the use of LiHPA salt instead of heteropolyacid suppresses Reaction 3 and benzene selectivity increases from 62% to 90.3%. Therefore, we may conclude that Ni-CrA-LiHPA catalyst developed has high activity (~3.8 h −1 ) that is at least dozen times more that of commercial catalyst Cr 2 O 3 /Al 2 O 3 , and high selectivity that is actually the same as for Rh-catalysts known in the art. Moreover, in contrast to the most active catalytic systems based on Rh, the catalyst developed has a much higher sulfur tolerance (see Table 4 and Figure 6). Actually, during 33 h testing of toluene hydrodealkylation in the presence of thiophene (up to 500 ppm S) we observe only a minor decrease of toluene conversion (about 2%), and benzene selectivity being practically the same, in the range of 90-91%. various sulfur concentrations (as a thiophen). As Figure 6 shows, addition of thiophen (250 ppm S) has no effect on toluene conversion, but it slightly decreases (at most 2%) at higher thiophen concentration (500 ppm S). It should be noted that this effect is a reversible one, and the catalyst activity being restored on operation with sulfur-free toluene in 3 h. We may conclude, therefore, that Ni-CrA-LiHPA catalyst has high sulfur tolerance.

Catalyst Preparation
The initial catalyst, Ni/Al 2 O 3 , was prepared by the standard impregnation procedure [40] starting from microspherical γ-Al 2 O 3 and nickel nitrate followed by calcination (450 • C, 6 h), and reduction (H 2 , 500 • C, 3 h). Bimetallic catalysts were prepared by decomposition of hexacarbonyl chromium or bis(arene)chromium compounds (CrA) in hydrogen flow over prereduced Ni/Al 2 O 3 or over alumina (for comparison) as described earlier [26]. Decomposition of Cr(CO) 6 or CrA was monitored by chromatography and mass-spectrometry, analyzing the concentrations of methane and benzene in the effluent H 2 . Catalysts modified with heteropolyacid (HPA) were prepared by impregnation of bimetallic catalysts with ethanol solution of H 4 PMo 11 VO 40 or its monolithium salt (LiHPA). The latter was prepared by exchange reaction of HPA with LiNO 3 as described in article [41]. Finally, all the modified catalysts were treated in hydrogen flow at 500 • C for 3 h.

Analyses and Instrumentations
Transmission electron microscopy (TEM) analysis was performed using a Jem-2100 microscope (Jeol, Akishima, Japan) with electron tube voltage of 100 kV. The count of particle sizes was performed by processing microimages obtained using Image-Pro Plus 6.0 program.
The content of each element contained in catalysts was determined by atomic absorption spectroscopy (Perkin-Elmer AAS, Waltham, MA, USA) after mineralization of samples. All percentages are indicated as weight percentages.
The surface area of catalysts was measured by the standard BET method using Gemini VII 2390 setup (Micromeritics, Norcross, GA, USA). For evaluation of the specific surface of Ni (S Ni ) hydrogen-oxygen titration was performed on Micromeritics AutoChem 2950HP. The procedure was as follows: at first, all samples were treated with H 2 (673K, 2 h) followed by He flow (298K, 0.5 h) to remove gaseous H 2 from the system. Then adsorbed hydrogen was titrated with oxygen pulses (3% O 2-97% He) in He flow. The amount of oxygen reacted was determined with catharometer, and water formed being analyzed with QMS 200 mass-spectrometer (Stanford Research Systems, Sunnyvale, CA, USA). This amount was used to calculate the S Ni value (in m 2 /g) using the stoichiometry O:Ni = 1:1.
Temperature-programmed reduction (TPR) was performed on Micromeritics AutoChem 2950HP analyzer (Micromeritics, Norcross, GA, USA). Before the analysis, 100 mg samples were pretreated at 673 K for 30 min under Ar flow in order to remove water and other contaminants, and then cooled to 333 K. The reduction step was performed under the 30 ml/min flow of 8 vol.% H 2 -92 vol.% Ar mixture in the range from 333 to 1173 K with heating rate of 10 K/min. The consumption of H 2 was calculated using the AutoChem HP V2.04 program.
The static magnetization of samples was measured by the Faraday method using a vacuum balance "Setaram" in fields from 10 to 15,000 Oe at 20 • C taking into account a correction for diamagnetism of quartz ampoule and support. For magnetization measurements a sample (30 mg) was placed in ampoule, reduced under standard conditions, evacuated to 10-4 Pa at 400 • C for 1 h, then ampoule was sealed.
The catalytic activity of the freshly reduced catalysts was tested in a flow type setup at temperatures 370-750 K, pressures 0.1-1.0 MPa, and H 2 /ArH ratio of 3.5 to 90. All the experiments were performed in the kinetic-controlled range [38]. The conversion and reaction rates were calculated based on product concentrations that were determined by on-line GC (20 m × 0.15 mm capillary column with OV-17 phase, Sigma-Aldrich, St Louis, MO, USA). The conversion, X, was determined as the amount of reacted aromatics (ArH) divided by the amount of ArH introduced in the reactor. The hydrogenation selectivity, S, to cyclohexane (methylcyclohexane) was determined as the amount of cyclohexane (methylcyclohexane) formed divided by the amount of reacted benzene (toluene) respectively.
In case of toluene hydrodemethylation, the benzene selectivity was determined by formula: where S M and S B are methane and benzene peak areas in the chromatogram respectively and n = 4.875 -chromatography calibration coefficient. The benzene yield, Y B , was calculated as the product of toluene conversion and the benzene selectivity.
The sulfur tolerance in toluene hydrodemethylation was studied in the presence of thiophene at 753 K and hydrogen pressure of 0.7 MPa. A sulfur-free feedstock was initially fed for 3 h, until steady state was established. Then sulfur poisoning experiments were performed using thiophene solution in toluene (equivalent of total sulfur content of 250 or 500 ppm (v/v)). Finally, the sulfur-free feedstock was used again for 3 h, to check for any change of catalyst activity.

Conclusions
Bimetallic supported nickel-containing catalysts modified with chromium(0) compounds and phosphoromolybdic heteropolyacid were prepared. Based on data obtained by chemisorption, magnetization, TEM, and XRD, the surface state of catalysts may be presented as Ni nanoparticles (2-10 nm) covered with chromium nanoparticles (Ni-CrA). The properties of Ni-CrA are quite different from conventional Ni-Cr-catalysts: (1) as opposed to the latter, magnetization of Ni-Cr is not decreased by chromium, but significantly increased; (2) it scarcely chemisorbs oxygen; and (3) demonstrates the synergistic activity in hydrogenation of aromatics.
The new catalysts were tested in hydrogenation of benzene, toluene and hydrodemethylation of toluene and they were shown to possess a high activity, selectivity, and sulfur tolerance in the processes above. For toluene hydrodemethylation, their activity was at least ten times more that of commercial catalyst Cr 2 O 3 /Al 2 O 3 . By selectivity and activity, they are close to Rh-catalysts, but not so expensive, and having much higher sulfur tolerance.
Therefore, the three-component Ni-CrA-LiHPA catalyst was shown to be very promising for hydroconversion of aromatic hydrocarbons and fuels containing thereof and may be scaled-up for industrial applications.