Highly Active Low Cobalt Content-Based Bulk MoS 2 Hydrodesulfurization Catalysts with a Unique Impact of H 2 S

Chemistry Department, Faculty of Science, Mansoura University, Mansoura 35516, Egypt Department of Material Process Engineering, Graduate School of Engineering, Kyushu University, Motooka 744, Fukuoka 819-0395, Japan Geology Department, Faculty of Science, Damietta University, Damietta 34517, Egypt Physical Chemistry Department, National Research Center, 12622 Dokki, Cairo, Egypt Chemical Engineering Department, Universiti Teknologi Petronas, 32610 Seri Iskandar, Perak, Malaysia


Introduction
Significant research efforts have been directed towards the development of catalysts for better reducing the sulfur contents of petroleum fuel fractions [1][2][3][4].Dwindling oil supplies, especially those extracted under vacuum distillation, and recent stringent environmental regulations aimed at limiting the sulfur contents of transportation fuels have resulted in a strong demand for improved hydrothermal treatment techniques.Clean fuels, such as sulfur-and nitrogen-free fuels, not only have the advantage of being environmentally benign transportation fuels but can also be used in several other emerging energy-related fields, including fuel cells [5].
e main catalysts used in hydrothermal treatment processes are Ni-and/or Co-containing Mo-based Al 2 O 3 catalysts.Such catalysts are proved to be highly active for thiophenes and benzothiophenes sulfur elimination.However, they are not sufficiently active for desulfurizing compounds such as dibenzothiophene (DBT), and it is analogous.is is particularly problematic because these compounds are the major sulfur-containing species remaining in the middle distillate fractions and in the atmospheric or heavy residues.e MoS 2 phases are the main catalysts in hydrothermal treatment reactions such as hydrodesulfurization (HDS).
e addition of Co to supported MoS 2 catalysts has been studied intensively, and its role as a promoter in HDS reactions is well established.MoS 2 -based catalysts with no promoter have very low activities towards HDS reactions.e significant enhancement in the HDS performance of these catalysts following the incorporation of Co has been attributed to the synergy between the Co and Mo phases [6][7][8][9].e details of this synergy have been investigated extensively in previous occasions [10][11][12].However, the influence of the incorporation of Co on the selectivity of such MoS 2 -based catalysts towards the HDS reactions is far from being fully understood [13][14][15][16][17]. e nature of the interactions between the Mo support and the substrate can have a significant impact on the efficiency of the synergy caused by Co and/or Ni promoters.On the contrary, the function of the promoter in the bulk MoS 2 in the HDS reaction remains beyond a complete coverage.Catalysts of this type have high activities in HDS reactions [18].Although a large number of reports are available for CoMo-based catalysts, relatively very little information has been published pertaining to the unsupported CoMo catalysts.A better understanding of how Co or Mo modifies the activity of bulk MoS 2 or Co 9 S 8 catalysts could provide valuable information for the development of new Mo-based catalysts for hydrothermal treatment processes.In this study, the DBT HDS reactions using bulk MoS 2 , Mo-Co 9 S 8 , and Co-MoS 2 catalysts containing various amounts of Co, especially in low cobalt concentration range, were investigated.e potential of Co as a promoter of the bulk MoS 2 catalyst is dealt with.e effects of the hydrogen sulfide matrix on the catalytic performance of the HDS reaction of DBT were also evaluated.e bulk MoS 2 catalysts, AHS and ATS, were prepared by heat annealing the corresponding Mo precursors (AH and AT) in the presence of a concurrent flow of a 1 : 9 (v/v) mixture of H 2 S/H 2 gas at 830 °C and 400 °C, respectively.Co 9 S 8 (CS) was also synthesized from the heat annealing of Co-acetate tetrahydrate at 400 °C following the same procedure as before.e details of such processes have been described elsewhere [19].A sample of the AHS material (obtained by sulfiding the ammonium heptamolybdate tetrahydrate precursor) was comminuted using a mill with an inner volume of 100 mL equipped with a media of zirconia beads, which was purged with He prior to being used.e resulting material was denoted as AHS-G.e BET surface areas of AHS, AHS-G, and ATS, which were measured using automatic Micromeritics ASAP 2010 instrument by N 2 adsorption-desorption technique at −196 °C, were approximately >10, 115, and 65 m 2 /g, respectively.e AHS-G and ATS sulfide samples were used as preliminary supports for the addition of the Co precursor.

Experimental
e Mo-sulfide phase was impregnated with a 1 : 1 (v/v) mixture of water/alcohol containing a specific amount of cobalt acetate tetrahydrate at an ambient temperature.e solution was then subjected to sonication for 3 h, followed by drying in air and thereafter heating in a vacuum oven at 120 °C overnight.All catalysts were thereafter subjected to sulfidation with a stream of a 1 : 9 (v/v) mixture of H 2 S/H 2 gases (5 ml/min) at 400 °C.
e resulting samples were denoted as C-AHSG-and C-ATS-, I to III based on their Co loading.Another sample was synthesized in which the prepared Co 9 S 8 phase was impregnated with a certain amount of an alcoholic solution of molybdenum acetylacetonate following the typical procedure mentioned before until obtaining the sulfide form of the catalyst.is sample (denoted as MPC) represents the Mo-promoted cobalt sulfide catalyst of ca.0.4 of the Co/(Co + Mo) atomic ratio.e commercial CoMo/Al 2 O 3 catalyst was also studied for comparison.All Mo-sulfide samples exhibited a hexagonal molybdenite-2H structure (JCPDS# 65-0160) according to the XRD patterns of AHS, AHS-G, and ATS catalysts shown in Figure 1, which is in agreement with literature [20].
No phases other than MoS 2 were detected.On the contrary, the diffraction patterns of the synthesized Cosulfide phase matched well with the JCPDS# 73-1442 of Co 9 S 8 .e AHS-G and ATS catalysts had MoS 2 crystallite sizes of approximately ≈5 nm.Transmission electron microscopy images (obtained from TEM; JEOL-2000EX) for C-ATS-I were depicted in Figure 2. Five to 10 layers of MoS 2 is obviously noted.Interlayer spacing of ca.0.62 nm was determined from the (002) XRD peak at 2θ of 14.2.

Activity Measurements.
e catalysts were investigated for the DBT HDS reaction.DBT was selected as a model compound for these reactions because it is a representative of some of the sulfur-containing refractory compounds found in middle distillates and heavy residue.All tests were implemented under a 3 MPa of H 2 pressure and at 340 °C.
e experiments were conducted using a stainless steel batch microautoclave reactor (100 mL) equipped with a magnetic stirrer and stainless steel filter, which allowed convenient withdrawal of small samples from the reaction mixtures at regular time intervals.A decane solution of DBT (1 wt.%) was used as the reaction feedstock.Some reaction runs were conducted in the presence of Cu powder (ca.0.7 g), which was used as a scrubber for the H 2 S produced as a by-product during the HDS reaction.A blank run was also conducted with Cu in the absence of a catalyst, which confirmed that Cu did not exhibit any activity towards the DBT HDS reaction.
e effect of the reaction matrix was evaluated by investigating the reaction over the present catalysts in both with and without the existence of H 2 S in the feedstocks.Immediately before the reaction test, the catalyst was once more subjected to a sulfidation step with a mixture of H 2 S/H 2 gases.
e catalyst and the reaction mixture (a typical of 15 mL of 1 wt.%DBT in decane) were loaded in situ into the reactor, which was subsequently pressurized with H 2 and heated to 340 °C under continuous stirring at 1000 rpm.Small samples of the reaction mixture (0.1-0.2 mL) were withdrawn from the reactor periodically for analysis to determine the rate of conversion.Gas chromatography (Agilent HP 6890) and GC-mass spectrometry (GC-MS) equipped with an Agilent HP 5970 MS were used to analyze the reactions.All GC analyses were conducted on a methylsiloxane capillary column (0.32 mm × 50 m).

Catalyst Performances.
e main HDS reaction products obtained under the operating terms described above over the present catalysts were cyclohexylbenzene (CB), biphenyl (BP), 1,2,3,4-tetrahydrodibenzothiophene (H4-DBT), and H 2 S. Trace quantities of the di erent isomers of partially hydrogenated DBT were also detected.Based on these products, we have proposed two reaction pathways for DBT HDS, which are depicted in Scheme 1.
e rate constants for these transformations were estimated by tting the experimental data using a nonlinear least square analysis on the supposition that the HDS reaction behaves kinetically as pseudo-rst-order [3,4,8,9].Figure 3 shows the t curves for the DBT HDS over the C-ATS-I catalyst under di erent levels of H 2 S.
e reaction system was classi ed into two reaction routes, including (1) hydrogenation (HYD), which would involve the initial formation of H4-DBT and its isomers by partial hydrogenation of DBT, followed by further reduction of these intermediates to give CB and (2) direct desulfurization (DDS), which would produce BP directly by the C-S bond scission.ese data were treated kinetically and evaluated according to the model that was recently developed for consecutive-parallel reactions [21].Based on this approach, we used the following di erential equations to calculate the catalytic constants: where k 0 k 1 K 1 + k 2 K 2 and k 1 and k 2 denote the DDS and the HYD intrinsic kinetic rate constants, respectively.K 1 and K 2 point to the constants of DBT adsorption at equilibrium onto the DDS and HYD reaction sites, respectively.Journal of Chemistry e kinetic parameters were adjusted for each compound according to the following formula: k 0 n k n K n , where K n and k n are the equilibrium adsorption and intrinsic kinetic rate constants of each compound, respectively.is treatment provided a better estimation of the individual contributions of the hydrogenation and direct desulfurization reaction pathways to the HDS reaction.Furthermore, this process allowed more reliable quanti cation of the individual reaction rates according to the contribution of each reaction pathway.e experimental data were solved and tted to the model using the Mathcad program.e estimated activity data are listed in Table 1.
Figures 4 and 5 show the nonlinear curves that were tted with the product yields from the HDS reaction over the C-ATS-1 catalyst in the presence and in the absence of H 2 S, respectively, using the model described above.All catalysts tested in this study provided a reasonable t to the data exposed from the HDS reaction.
e C-ATS-I catalyst displayed much higher level of reaction activity than the ATS, AHS-G, C-AHS-G, MPC, and commercial CoMo/Al 2 O 3 catalysts.However, DBT HDS proceeded to a lower activity over the C-AHS-G catalysts than it did over the bare commuted MoS 2 catalyst.e di erence in the activity of these two systems could be attributed to the delicate structure of the commuted MoS 2 as a primary support, which led to di erences in the speci c surface areas of these catalysts (the surface area (BET) of the AHS-G catalyst was 110 m 2 /g, and therefore, it is much higher than that of the C-AHS-G-II catalyst, which was less than 5 m 2 /g).
us, commuted MoS 2 may not be a recommended preliminary support for Co promotion.Taken together, these results indicate that the selectivity pro les of the catalysts need to be analyzed in a greater detail to develop a better view of the changes in the activity following the impregnation of the catalysts with Co.In this study, the selectivity was estimated according to the determined ratio of the HYD to the DDS rate constant, that is, k HYD /k DDS .One should be careful when considering the catalytic selectivity according to the product distributions because the selectivity in this case underwent signi cant mobile changes depending on the level of conversion (Figure 4).When the bare MoS 2 catalysts were used, the HDS of DBT proceeded preferentially by the HYD route, as shown in Table 1.
e relatively high selectivity observed in this case for the products resulting from the HYD route is contrary to the common view that DBT HDS proceeds mostly by a DDS route.
ese results therefore highlight the exibility of the HDS reaction, in the sense that it can proceed by these routes with no path limitation to either of them.
e impregnation of Co onto the ATS catalyst (C-ATS series of catalysts) led to a signi cant development in the overall HDS activity.One may notice the obvious pioneer Co promotion in the HDS of these catalysts, especially for the C-ATS-I catalyst.
is catalyst showed a remarkable tendency towards the DDS pathway (ca.1∼2-fold higher activity towards the DDS route over the HYD route (Table 1)) for the HDS of DBT, whilst the bare MoS 2 catalysts exhibited a preference for the HYD pathway.e change in the activity of the Co-containing catalyst was accompanied by an obvious shift in its selectivity towards the BP product. is tendency towards the DDS pathway has been reported previously for the conventional MoS 2 -based catalysts as a result of the impregnation of a Co promoter [5].ese results therefore demonstrate that the HYD and DDS reaction pathways occur in parallel and can be independently manipulated.Notably, all Co catalysts prepared in this study underwent the HDS reaction preferentially by the DDS pathway, and a clear correlation between the selectivity and activity of the HDS reactions could be established (Figure 6(a)).It is noteworthy mentioning that the trend in the selectivity of these reactions was in accordance with similar reports from the literature for related systems [3]. Figure 6(b) displays the association between the reaction activity of the C-ATS catalysts and the Co/(Co + Mo) ratio.It is obvious that the C-ATS-I catalyst, which had 0.05 Co/(Co + Mo) ratio, exhibited the highest HDS activity of all of the catalysts belonging to this series.Interestingly, such ratio is far lower than those commonly reported for conventional Co-Mo-based catalysts, where the average is generally around 0.3.However, the present results are generally in a trend with other studies where low Co concentration catalysts were reported to be active for hydrogenation and hydrogenolysis reactions [22,23].e selectivity for the HYD reaction was clearly e ected by the amount of Co incorporated into the existing catalyst.e results in Figure 6(b) show that the tendency towards the HYD reaction decreased with the increase in Co content.
ese results therefore may indicate that only a limited number of active Mo and S sites can be further activated by the inclusion of Co. ese sites were assumed to be in the type II CoMoS phase.
e active sites of HDS catalysts are coordinately unsaturated sites.e incorporation of promoters such as Co and Ni can lead to improvements in the supply of spillover hydrogen.is may lead to an increase in the DDS and HYD activities [9,12,15,16,21].It is therefore envisaged that these catalysts must contain di erent types of active sites to allow them to catalyze multiple reactions in parallel.e Co atoms incorporated into the MoS 2 catalysts could occupy the active sites on their surface.e enlargement in the catalytic activity of the Co-containing MoS 2 catalysts (i.e., C-ATS catalysts) compared with the bare MoS 2 catalyst could be attributed to the occurrence of a large synergistic e ect (Table 2) between Co and Mo atoms.A synergistic e ect of this type could potentially result in the formation of two di erent Co-based active sites for the DDS and HYD reactions.
e Co atoms could therefore occupy di erent positions on the surface (i.e., edges and corners) of the unsaturated MoS 2 sites, resulting in new sites.However, not all of these sites would be accessible for the HDS reaction of the substrate.
e results therefore imply that the higher activity of these compounds could be attributed to an increase in the e ectiveness of the active sites.e corner-edge  model [24][25][26] can be used to explain the di erences in the selectivity pro les of promoted and unpromoted MoS 2 catalysts for the DDS route.e Mo atoms at the corner sites of the stacked cluster would probably be unsaturated because of the steric con guration of the hexagonal crystals in the catalyst.ere could also be a high proportion of coordinately unsaturated sites in this case.It can be concluded that the selectivity of the catalysts for the DDS reaction correlates highly with the number of MoS 2 layers.Nikulshin et al. [27,28] in studies for the HDS over CoMo-based catalysts showed that the (Co/Mo) edge ratio is directly proportional with the catalyst HDS activity.ey further revealed that, with increasing the cobalt content, the HDS selectivity turns slightly towards HYD but still the predominant route is the DDS.To date, our preliminary results have shown that the low content of the Co-promoted MoS 2 catalyst, especially the C-ATS catalysts, and led to an increase in the overall activity.On the contrary, MPC catalyst exhibited low HDS performance.Furthermore, the reaction followed pseudo-zero-order kinetics (Figures S1-S3).

Impact of H 2 S.
e H 2 S matrix could also have a signi cant impact on the behaviors of the catalysts developed in this study for the HDS reaction.e activities and selectivities of the current catalysts towards DBT HDS varied considerably depending on the existence of H 2 S in the reaction medium.
e parameters derived from the kinetic treatments of these experimental results are presented in Table 1.Several clear and interesting trends can be observed.For example, the presence of H 2 S resulted in a slight rise in the formation of BP (DDS route product).However, the inclusion of H 2 S also drove to a remarkable positive change in the tendency of the reaction towards the HYD route and the production of CB.Table 2 shows that the presence of H 2 S led to a rise in the activity of the present catalysts by increasing the rates of the HYD and DDS reactions by approximately 2.5-to 4-fold and 1.4-to 1.7-fold, respectively.e promotional trends of H 2 S towards the DDS reaction were very similar for all of the prepared catalysts.For instance, the enhancement in the activity resulting from the inclusion of H 2 S was much more pronounced for the HYD reaction route, especially in case of the bare MoS 2 catalyst.
e HYD pathway contributed to more than 80% of DBT HDS over AHS-G and/or ATS catalysts.H 2 S in uence on the HDS reaction has been discussed extensively in various studies [29][30][31][32][33][34][35], which consensually suggested that H 2 S suppresses the DDS route, whilst having very little impact on the HYD route.In contrast, several research groups, including our own, have reported that some transition metal sul de catalysts can e ectively promote the HYD reaction when they are carried out in the presence of H 2 S [36][37][38].Guernalec et al. [38] correlated this positive behavior to the increase of the -SH concentration on the catalyst surface.However, our obtained results represent a unique catalytic   1 and 2. Interestingly, this reaction (impacted by H 2 S) was found to be reversible indicating that H 2 S may have caused no permanent changes to the structure of the catalyst.Taken together, these results demonstrate that our collective understanding of H 2 S inuence on the selectivity of the HDS reaction is rather incomplete.Crystallite sizes in the extent of 4-5 nm for the MoS 2 crystallites appeared to exert a speci c catalytic performance, which de nitely depends on how such catalyst is synthesized.
e nature of substrate, catalyst, and reaction conditions therefore appear to be of considerable importance for quantitative evaluation of this phenomenon.

Synergy E ect.
e results of this study clearly show that the incorporation of Co had a decisive in uence on the catalytic performance of the unsupported MoS 2 catalysts (series C-ATS).It is noteworthy mentioning that the results are consistent with the well-known promoting e ects of conventional supported MoS 2 catalysts.Taken together, these results suggest that the HDS reaction resulted from divergent active sites.e synergy between the Co and Mo phases appeared to be on line with that claimed for the wellknown Co-Mo-S type II materials, such as the supported Co-Mo catalysts [10], which are assumed to be responsible for catalyzing HDS reactions.All C-ATS catalysts prepared in the present study exhibited some degree of synergy between the Co and Mo atoms. is synergistic interaction enhanced the HDS reaction by promoting the DDS route to a higher extent than that of the HYD route. is result indicated that the Co-based sites were more e ective for the C-S bond scission.us, it can be stated that the incorporation of Co not only improved the activity of the catalyst but also modi ed the selectivity along the two pathways.Figure 7 shows the relation between the synergy factor (SF, activity ratio of (Co-MoS 2 /MoS 2 )) and the Co-added ratio.Calculations based on the density functional theory (DFT) for such system have predicted that the sulfur and the metal edges of the catalyst may conduct the DDS reactions [34,35].
ese DFT studies have also shown that the Co atoms can be accommodated on the rims of the MoS 2 crystallites especially on the rims of the sulfur atoms.
It has been suggested that the HYD reactions most likely occur on the brim sites and the metal edges.In the present study, CB was isolated as the major product from the reactions catalyzed by the bare MoS 2 catalyst, which indicated that the HYD pathway was favored under these conditions.However, Table 1 shows that the share from the DDS route was higher than that of the HYD route when the HDS reaction was conducted over the C-ATS-I catalyst.e overall HDS activities of the ATS and C-ATS catalysts in the presence of H 2 S were much higher than those in case of the absence of H 2 S. e major enhancement was in the HYD reactions of DBT.
e MoS 2 and Co-containing MoS 2 catalysts contain acidic sites on their surfaces of two di erent types, including (i) Lewis acidic sites, that is, the sulfur vacancies on the Mo and Co atoms, and (ii) Brønsted acidic sites.e latter of these two sites would contain the -SH and -SH 2 groups.e -SH sites may act to eliminate sulfur from DBT and the HYD reactions (under certain circumstances), whilst the -SH 2 sites would favorably be involved in hydrogenation reactions.e involvement of these active sites in the HDS reaction would be dependent on their availability with the appropriate geometry to interact with the substrate [8,35].Given that the Co-S bond is relatively weaker than the Mo-S bond, the Co sites would be more acidic than the Mo sites [39]. is diversity in the acidity of the two sites could explain the higher DDS activity of the Co-containing MoS 2 catalysts compared with the unpromoted MoS 2 catalyst.e cleavage of the S-C bonds is most likely to occur with greater ease over the Co-promoted catalyst.is suggestion is also consistent with the results obtained for the commercial CoMo/Al 2 O 3 catalyst (Table 1).It has therefore been suggested that the inclusion of H 2 S may result in new active sites through interactions with vacant sulfur sites on the MoS 2 and Co-containing MoS 2 catalysts.e quality of these sites for promoting or suppressing the HYD and hydrogenolysis reactions would probably be dependent not only on their con guration and concentration but also on the structure of the substrate layers.
e data shown in Table 2 provide a summary of the e ects of Co and H 2 S on the HDS reaction.e main points that could be drawn from the data are as follows:

Conclusions
e catalytic results of the comminuted MoS 2 , ATS, and Cocontaining MoS 2 catalysts for the reaction of DBT showed that the nature of the preliminary MoS 2 support had a critical role on the activity of the Co-promoted one.e hydrogenation route was most predominant in the reaction over the bare MoS 2 .H 2 S inclusion enhanced the rates of both possible reaction routes but promoted the HYD pathway to a much greater extent.e activity of the MoS 2 catalysts increased significantly when the Co promoter was added.All catalysts synthesized in the current study exhibited a close trend towards the influence of H 2 S in the HDS reaction.e C-ATS and C-AHSG catalysts showed a strong preference for the DDS pathways, with BP being produced as a major product.e boost in the activity observed with the Cocontaining catalyst (C-ATS) was attributed to the obvious enhancement in the DDS route and to the significant increase in the HYD route reactions.
e results from the kinetic analysis of the HDS reactions over the MoS 2 and MoS 2 promoted by Co catalysts suggested the existence of different discrete active sites, which were dependent on the identity of the promoter synergy and the reaction matrix.
is study therefore provides important information for controlling the selectivity and activity characteristics of MoS 2 -based catalysts by tailoring their properties through synthetic manipulation and promoter ratio and adapting proper reaction matrices.

Figure 3 : 3 Scheme 1 :
Figure 3: Pseudo-rst-order plots for the HDS reactions of DBT over the C-ATS-I catalyst in the absence and presence of H 2 S.

Figure 4 :Figure 5 :
Figure 4: Transformation of dibenzothiophene over the C-ATS-I catalyst in the presence of H 2 S.

Figure 6 :
Figure 6: E ect of the Co content on the HDS activity (a) and the relationship between the activity and the selectivity (b).

( 1 )
e incorporation of Co (low concentration range) into the unsupported MoS 2 catalyst led to a signicant activity boost in the DBT HDS reaction, as well as resistance to the inhibition phenomena caused by H 2 S. (2) e Co atoms preferentially may occupy the edge sites of the MoS 2 catalyst [34, 35].(3) H 2 S inclusion in the HDS reaction feedstock led to an enhancement in the hydrogenation activities of all the studied catalysts.

Figure 7 :
Figure 7: Cobalt content as a function of synergy promotion.

Table 1 :
Activities of the catalysts towards the HDS of dibenzothiophene.
aLevel of H 2 S: low (L) and high (H) denote pressures of ca. 3 and 20 kPa, respectively.b Apparent rate constant for the direct desulfurization route, 1st order.c Apparent rate constant for the hydrogenation route, 1st order.d Normalized to the metal sul de content.e Commercial catalyst Co, 3.2%, and Mo, 13.7%, in terms of the weight percent.f Pseudo-zero-order rate constant, × 10 17 molecule/(g•S).e uncertainty was within ±5-10%.

Table 2 :
Changes in the selectivity of the HDS reaction for the MoS 2 -based catalysts caused by Co and H 2 S. Comparison of the average fold change (increase or decreases) in the activity of the ATS catalyst due to the inclusion of Co the use of H 2 S may promote the HDS reaction over unsupported Co-promoted MoS 2 catalysts.e increase observed in the catalytic activity in presence of H 2 S reported here is contrary to the well-known observation that H 2 S severely inhibits the HDS reactions that are conducted over conventional catalysts.is is evident from the data highlighting the inhibition of the commercial CoMo/Al 2 O 3 catalyst by H 2 S presented in Tables