1

FULL PAPER

Dispersed Ni and Co Promoted MoS2 Catalysts with Magnetic Greigite as a Core: Performance and Stability in Hydrodesulfurization

Seyyedmajid Sharifvaghefi [a] ,and Ying Zheng*[a,b]

Abstract. Core-shell composites MoS2/Fe3S4, CoMoS/Fe3S4, NiMoS/Fe3S4 with greigite as the core were synthesized and compared to MoS2/Fe3O4 for hydrodesulfurization (HDS). The composites with a greigite core were found to have higher hydrogenation (HYD) selectivity in HDS of dibenzothiophene (DBT) compared to the one with a magnetite core and showed higher stability during different cycles of DBT HDS test. Cobalt and Nickel as a promoter are uniformly dispersed over MoS2 and decorated on the layer-structured MoS2. NiMoS/Fe3S4 showed the highest activity in the hydrodesulfurization of DBT and 4,6-Dimethyldibenzothiophene (4,6-DMDBT) followed by CoMoS/Fe3S4 and MoS2/Fe3S4. For the promoted catalysts, the direct desulfurization (DDS) pathway was found to be more selectively enhanced than the HYD route in HDS of DBT. HYD pathway was the dominant route in HDS of 4,6-dimethyldibenzothiophene (4,6-DMDBT). Moreover, they have high stability, and the magnetic properties associated with these nano-composites makes them an excellent choice for hydroprocessing of heavy petroleum oils in slurry reactors as they can be easily separated and reused.

Introduction

During the past years, the share of heavy and extra-heavy oils has increased in the oil market and this has caused issues using the conventional fixed-bed hydrotreating reactors. The extra impurities in these heavy crudes can easily clog the catalyst bed. This will cause premature shutdown of the reactor which will have a detrimental effect on the economics of the refinery [1]. Different technologies and processes based on slurry-bed reactors are developed to resolve this problem [2]. There are many advantages associated with slurry-bed type of reactors; however, its complex and costly procedure for the separation of catalysts has kept the technologies from deployment of commercial implementation. The traditional hydrotreating catalysts are MoS2 that is typically promoted using Co or Ni. Applying these catalysts in a slurry-bed reactor in a once-through put mode is costly and uneconomical. Injecting the oil-soluble precursors of MoS2 was a solution studied by different groups [3]. However, it has been demonstrated that the technology is inefficient for in-situ promoting MoS2 catalyst [4].

Recently, we showed a novel way to incorporate Fe3O4

particles as magnetic carriers to synthesize a MoS2

composite with magnetic properties[5]. The composite showed an excellent ability in removing sulfur mainly through the direct desulfurization pathway (DDS). However, there are concerns regarding the stability of these composites used in sulfur-containing environment and as part of a composite containing sulfided catalysts. To address this issue, a more suitable magnetic core is needed. Among different magnetic carriers, greigite (Fe3S4) has recently found applications in electrochemistry, biomedicine, and water treatment [6]. Because of its magnetic properties and the presence of sulfur in its structure which can be better adapted with the sulfided catalysts used in the hydrotreating process, greigite was chosen as the substitute for magnetite as the core material.

Formation of a uniform phase with the addition of the promoter atoms to MoS2 has always been a challenge. The conventional methods based on impregnation or hydrothermal methods for the synthesis of Ni or Co-supported catalysts generally results in part of the

[a] S. Sharifvaghefi, Prof. Y. Zheng Department of Chemical Engineering

University of New Brunswick15 Dineen Drive, Fredericton, NB, E3B 5A3 (Canada)E-mail: yzheng@unb.ca

[b] Prof. Y. Zheng School of Engineering, University of Edinburgh, Mayfield

road, Edinburgh, EH9 3DW (UK)

1

23456789

1011121314151617181920212223242526

27

282930313233343536373839404142434445464748

4950515253545556575859606162636465

6667686970

1

2

promoter atom to segregate forming separate Ni or Co sulfide phases. It is found that these phases are less active than the Ni(Co)-Mo-S phase formed from the uniform dispersion of the promoter atoms [7]. Therefore, preparing the catalyst with a uniformly dispersed promoter atom can increase the activity of the catalyst.

Herein, we report the facile synthesis of these nanocomposites in which greigite is used as the core material and the outer layers consisted of (Ni/Co)MoS2

catalysts. The resulting nanocomposites are then characterized and tested for the HDS of DBT and 4,6-DMDBT as model sulfur-containing compounds. The promotional effect of Co and Ni was then studied. Finally,

the activity and stability of these composites were compared to the nanocomposites prepared with magnetite core through several cycles of the HDS test.

Results and Discussion

To prepare the MoS2 composite samples, both Fe3S4 and Fe3O4 particles were used as the core material and an outlayer of MoS2 catalyst was formed around the particles. Magnetite particles were used as a precursor to generate greigite particles and the transforming procedure was described in the experimental section. During the

Figure 1: XRD pattern of Fe3O4 and synthesised Fe3S4.

Table 1: Properties of the fresh and treated samples.

Sample Sulfur content (± 0.2 wt %)

Carbon content (± 0.2 wt %)

Specific surface area (m2/g)

Average particle size (nm)

Average number of layers

Fe3S4 - - 62 - -Fresh Mo/G 29.2 18.4 18 3.5±1.3 1.2±0.3Treated Mo/G 41.1 1.5 217 6.5±2.0 2.3±1.5Fresh CoMo/G 22.6 15.8 19 3.6±1.8 1.3±0.5Treated CoMo/G 31.1 1.3 84 6.9±2.1 2.4±1.1Fresh NiMo/G 19.8 16.9 21 3.9±1.8 1.5±0.2Treated NiMo/G 27.2 1.8 83 7.2±2.7 2.5±1.0

123456

789

10111213

141516

17

18192021222324

2526

27

28

1

3

Figure 2: TEM images of fresh (a,b) and treated (c,d) Mo/G sample.

synthesis procedure, S2- released from the

decomposition of thiourea reacts with the Fe3O4

nanoparticles to produce Fe3S4 nanoparticles. The XRD pattern of Fe3O4 particles and the as-prepared Fe3S4

product are shown in Figure 1. All the peaks in the XRD pattern can be indexed to Fe3S4 and no characteristic peaks of impurities were detected [8].Three catalyst samples were prepared and compared in this study. They were MoS2/Fe3S4 (denoted as Mo/G), CoMoS/Fe3S4 (denoted as CoMo/G), and NiMoS/Fe3S4

(denoted as NiMo/G). The freshly synthesized samples were treated following the procedure described in the experimental section to remove the surfactant from the catalysts and also to stabilize their crystal structure. The TEM images of the catalysts before and after the treatment are shown in Figure 2 and their properties are shown in Table 1. The weight ratio of sulfided metal component to greigite was kept at a constant value of 0.6 for all the samples. The amount of promoters added was calculated to reach the promoter/(Promoter+Mo) molar ratio of 0.35. The BET surface areas of the fresh samples were 17.8, 19.4, and 21.5 for catalysts Mo/G, CoMo/G, and NiMo/G, respectively. The carbon contents in the samples was due to the presence of surfactant in the catalyst structure. As shown in our previous work [5], the surfactant CTAC plays two important roles. First, it assists with anchoring MoS2 on the surface of the greigite core and second, it functions as a scaffold platform to support the MoS2 crystalline in the synthesis to form a core-shell structure. After the thermal treatment, surfactant was decomposed and the surface area of the catalyst was increased approximately 12 times for Mo/G and 4.4 times for the promoted samples. The average length and number of layers for the samples also increased by about 50% after the treatment. These

treated samples were further used for the hydrodesulfurization tests. It should be mentioned

Figure 3: XRD patterns of Mo/G, CoMo/G, and NiMo/G.

d

b

c

a

12

3456789

10111213141516171819202122232425262728293031323334353637

3839

404142

1

4

that no significant changes were observed in the crystal structure of the treated samples after the hydrodesulfurization tests. The XRD patterns of the prepared Fe3S4 core, Mo/G, CoMo/G, and NiMo/G are shown in Figure 3. All Mo-based sulfide catalysts exhibited broad diffraction peaks, indicating a nano-sized crystallized MoS2 structure, particularly when the promoter was present. No Ni or Co sulfide phase was detected in any of the samples. This shows well dispersion of the promoters in MoS2 which is also supported by the findings through TEM imaging. Figure 4 shows the magnetization curves measured at 300 K for the Fe3S4 and Mo/G samples. The curves presented no hysteresis loop, suggesting that Fe3S4

particles had paramagnetic behavior. The saturation magnetization values for the Fe3S4 and MoS2/Fe3S4 were 25 and 22 emu/g, respectively. The small difference between the saturation values suggested that the addition of the catalyst shell had little effect on the magnetic properties of the sample. The possibility for magnetic separation of samples was tested by exposing the catalysts to an external magnetic field. Mo/G dispersed in model oil after 2 h of the HDS reaction was shown in Figure 5a and it is separated afterwards with an external magnet (Figure 5b). The catalyst can be separated easily from the reaction medium within a few seconds. All of the samples showed the same magnetic effect. Therefore, this method provides an easy and efficient way to separate catalysts from the product stream by an external magnetic field.

Figure 4: Magnetization curves for Fe3S4 and MoS2/Fe3S4 measured at 300 K.

Figure 5: Mo/G composite after the HDS reaction a) dispersed inside the model oil and b) separated from the model oil after seconds of exposure to an external magnet.

DBT conversion THDBT HHDBT CHB BCH BP0

102030405060708090

54.0

6.83.4

49.6

4.0

23.8

74.0

1.4 0.8

35.5

54.4

77.0

1.3 0.8

35.3

52.7

Mo/G CoMo/G NiMo/G

Sele

ctivi

ty o

r Con

vers

ion

(%)

Figure 6: DBT conversion and product selectivity for the synthesized catalysts after 2h of reaction.

ba

123456789

10111213141516171819202122232425262728293031

323334

35

36

37

38394041

4243

4445

1

5

CPMB DCPE CHCPM0.0

1.0

2.0

3.0

4.0

5.0

6.0 5.3

1.4

5.7

3.3

1.0

3.6

4.6

1.2

4.1

Mo/G CoMo/G NiMo/G

Sele

ctivi

ty (%

)

Figure 7: The selectivity of the isomerized products over the synthesized catalysts after 2 hr of HDS of DBT.

Scheme 1 (shown in supplementary information) depicts possible reaction pathways for the HDS of DBT. There were two main desulfurization pathways: direct desulfurization (DDS) and hydrogenation (HYD). In the DDS route, DBT undergoes direct C-S bond cleavage to form bipheny (BP). In the HYD route, one typical step is to hydrogenate one benzene ring of DBT to produce tetrahydrodibenzothiophen (THDBT) which C-S bond could be further ruptured to form cyclohexylbenzene (CHB) or which C=C bond is saturated to generate hexahydrodibenzothiophene (HHDBT). CHB can then be isomerized to form (cyclopentylmethyl) benzene (CPMB). Bicyclohexyl (BCH) was mainly formed through hydrogenation and desulfurization of HHDBT. Both BCH and CPMB can isomerize to (cyclohexane)cyclopentylmethyl (CHCPM) which can undergo further isomerization to form dicyclopentylethane (DCPE). The DBT conversion and product distribution of the prepared catalysts is shown in Figure 6. The mass ratio of catalyst-to-model oil was maintained the same for all the catalysts. The selectivity of the isomerized products is shown in Figure 7. The DBT conversions for Mo/G, CoMo/G, and NiMo/G are 54, 74, and 77 percent, respectively. As expected, promoted catalysts show better activity than Mo/G in HDS of DBT with a nearly 28% increase in activity. Considering the uncertainty levels, the activity of CoMo/G and NiMo/G for DBT conversion is the same. The selectivity of the products suggested that Mo/G had a higher tendency towards the HYD pathway than the promoted catalysts. The DDS/HYD ratio (calculated by dividing the selectivity of BP by the sum of the selectivities of other products) for Mo/G was found to be 0.31. Both the CoMo/G and NiMo/G catalysts show the same selectivity for DDS and HYD reaction pathways with nearly a 50% selectivity for each route. The DDS/HYD ratio was 1.19 and 1.11 for the Ni-promoted and Co-promoted catalysts, respectively. The results suggested that the promotional effect of Co and Ni was mainly reflected by the enhancement in the activity of DDS route. The active sites for the unpromoted MoS2 are proposed to be the coordinatively unsaturated sites or exposed Mo ions with sulfur vacancies at the edges of the catalyst. These sites are active for hydrogenation and hydrogenolysis. Recently, the unsaturated sites with metallic characteristics at Mo edges of the (Co/Ni)MoS2

catalysts (called brim sites) were also found to be active

for hydrogenation reactions. For Co or Ni-promoted catalysts, the Co(or Ni)–Mo–S phase is considered to be the active phase [9]. In this model, small MoS2 crystals with the promoter molecules located at the edges of the MoS2 layers in the same plane as Mo is the building blocks for the Co(or Ni)–Mo–S structures. The promotional effect induced by Co or Ni is the result of electron donation from Ni or Co to Mo which adjusts the Mo-S bond strength to an optimal range for the HDS activity [10]. The presence of isomerized products in the samples suggests that acid sites are present on the surface of the prepared catalysts. The results show that in HDS of DBT, the selectivity of the isomerized products over Mo/G is higher than that of the promoted catalysts. The creation of acid sites over sulfided catalysts can occur through the dissociation of hydrogen over the sulfurs at the edges of the catalyst. These sites can also be created through dissociation of hydrogen sulfide at the coordinatively unsaturated (CUS) sites. Considering the significant hydrogenation ability of the promoted catalysts, the higher acidity of Mo/G can be attributed to the dissociation of hydrogen sulfide. The dissociation of H2S can increase the acidity of the catalyst by converting the CUS sites to –SH. These –SH groups are considered to be active in hydrogenation and isomerization reactions over the catalysts. The acidic nature of these groups, however, depends on H+ electron affinity and the bond strength of metal–sulfide [11]. If the metal-sulfur bond is strong (as in MoS2), the –SH groups act as Brønsted acid cites while in the presence of a weak metal-sulfur bond they act as nucleophilic centers [12]. As mentioned before, the promotion effect of Co or Ni occurs through the reduction

DMDBT Conversion

THDMDBT HHDMDBT MMCHB MCHT DMBCH DMBP0

10

20

30

40

50

60

70

80

41.0

9.24.1 2.0

25.6

2.6

34.0

68.0

10.3

3.5 2.9

25.2

5.3

19.9

75.0

7.62.7 3.5

21.3

6.7

16.2

Mo/G CoMo/G NiMo/G

Sele

ctivi

ty o

r con

vers

ion

(%)

12

3456789

1011121314151617181920212223242526272829303132333435363738394041424344454647484950

51525354555657585960616263646566676869707172737475767778798081828384

1

The selectivity of the isomerized products over the synthesized catalysts after 2 hr of HDS of 4,6-DMDBT.

: 4,6-DMDBT conversion and product selectivity for the synthesized catalysts after 2h of reaction.

6

of the Mo-sulfur bond strength. The stronger bond in Mo/G implies that the –SH sites created through H2S dissociation are acidic over this catalyst while the same sites are nucleophilic centers over the promoted catalysts. Thus the Mo/G shows better acidity than the promoted catalysts in the HDS of DBT.

Scheme 2 (shown in supporting information) depicts the possible reaction pathway for the HDS of 4,6-DMDBT. Similar to DBT, two reaction pathways are considered for the HDS of 4,6-DMDBT. Dimethyl biphenyl (DMBP) is formed through DDS route and the THDMDBT and HHDMDBT are formed as intermediates via HYD route. 1-methyl-3-(3-methylcyclohex-1-en-1-yl)benzene (MMCHB) was detected in this study which is an intermediate in the formation of MCHT. This suggests that the same compound withouth the methyl groups ((cyclohex-1-en-1-yl)benzene) is an intermediate in formation of CHB from THDBT. (cyclohex-1-en-1-yl)benzene is usually not detected in experiments due to its fast conversion rate, however, the presence of the methyl substituted compound in HDS of 4,6-DMDBT strongly suggests that (cyclohex-1-en-1-yl)benzene plays a role as an intermediate in the formation of CHB. HHDMDBT can undergo C-S bond cleavage to form MCHT and/or DMBCH after hydrogenation. MCHT and DMBCH can be isomerized to 1-(cyclohexylmethyl)-4-methylbenzene

(CHMMB) and 1-(cyclohexylmethyl)-3-methylcyclohexane (CHMMCH), respectively. CHMMB can be further isomerized to 1‐(2‐cyclopentylethyl)‐4‐ methylbenzene (CPEMB) which can be converted to form (3-cyclopentylpropyl)benzene (CPPB).

Figure 8 shows the 4,6-DMDBT conversion and the selectivity of the main products over the catalysts. The 4,6-DMDBT conversion was found highest for NiMo/G which had 81% conversion compared to CoMo/G with 68% and Mo/G with 41% conversion. The results show the accumulation of THDMDBT for all of the catalysts suggesting the the difficulty in removing sulfur from the molecules. The DDS/HYD ratio for the catalysts was calculated as 0.5, 0.25, and 0.19 for Mo/G, CoMo/G, and NiMo/G, respectively. These ratios show that the HYD pathway is favored for all of the prepared catalysts and Mo/G has the highest affinity toward the DDS pathway with 34% selectivity for DMBP. The selectivity of the isomerized products is shown in Figure 9. CHMMB is the main isomerized compound with 13.0, 17.8, and 23.7% selectivity for Mo/G, CoMo/G, and NiMo/G, respectively.

CHMMB CPEMB CPPB CHMMCH0.0

5.0

10.0

15.0

20.0

25.0

13.0

3.1 2.7 3.7

17.8

5.54.3 5.3

23.7

6.95.4 6.0

Mo/G CoMo/G NiMo/G

Sele

ctivi

ty (%

)

123456

789

10111213141516171819202122232425262728

2930313233

34353637383940414243444546474849

1

7

Studies with model compounds have shown that the DDS pathway is the preferential route in HDS of DBT with Co and Ni-promoted catalysts [13]. The addition of alkyl-substitutes in 4, 6 positions of the DBT molecule affects the HDS reaction in two different ways. First, the HDS reactivity of the molecule is reduced and second, the preferential reaction pathway is shifted toward HYD, making it the dominant route [14]. Hydrogenation of DBT and 4,6-DMDBT mainly occurs through π-bonding with the catalyst while the hydrogenolysis occurs through σ-adsorption via the sulfur atom of the molecule [13]. The

difficulty in converting the alkyl-substituted DBT compounds is known to be related to the steric hindrance of the alkyl groups positioned close to the sulfur atom which prevents the interaction of these molecules with the active sites of the catalyst through σ-adsorption. The partial saturation of one of the rings through π-bonding which is not sterically hindered in these compounds changes the spatial configuration of them, thus reducing the hindrance effect making them more flexible and accessible to the reaction sites of the catalyst [15].

Table 2: DBT conversion and selectivity for 7 cycles in HDS of DBT over MoS2/Fe3S4 catalyst.

  Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6 Cycle 7DBT conversion 54 49 56 54 49 50 51THDBT 6.8 5.4 6.1 4.4 3.9 3.5 3.4

HHDBT 3.4 2.5 2.9 2.0 2.3 1.9 2.0

CHB 49.6 48.9 48.2 48.5 49.3 49.9 49.0

BCH 4.0 4.1 4.5 4.5 4.7 5.0 5.1

BP 23.8 25.7 26.0 26.1 24.6 24.2 24.5

CPMB 5.3 5.8 4.9 6.3 6.7 6.8 7.0

DCPE 1.4 1.6 1.8 1.7 1.8 1.8 1.9

CHCPM 5.7 6.0 5.6 6.5 6.7 6.9 7.1

Table 3: DBT conversion and selectivity for 7 cycles in HDS of DBT over MoS2/Fe3O4 catalyst.

  Cycle 1 Cycle 2 Cycle 3 Cycle 4 Cycle 5 Cycle 6 Cycle 7DBT conversion 48 50 47 58 65 63 62

THDBT 1.7 1.4 1.6 3.6 2.3 2.0 1.4

HHDBT 0.0 0.0 0.0 1.4 1.4 1.1 1.1

CHB 19.4 17.1 18.3 32.0 48.6 47.2 48.4

BCH 0.0 0.0 0.0 0.3 2.4 3.1 3.2

BP 77.0 79.9 78.3 58.1 34.5 34.7 33.5

CPMB 1.9 1.5 1.8 3.5 7.0 7.5 8.0

DCPE 0.0 0.0 0.0 0.5 0.6 0.7 0.7

CHCPM 0.0 0.0 0.0 0.5 3.1 3.6 3.7

123456789

1011

12131415161718192021

22

2324

2526

27

282930

1

8

It is suggested that the change in the preferential reaction pathway by the addition of alkyl-substitutes to DBT is mainly due to the severe inhibition on the DDS pathway and that the presence of these alkyl groups hardly affects the HYD pathway. The difference between the reactivity of DBT and 4,6-DMDBT was related to the selective promoting effect on the DDS pathway for DBT [16]. While this agrees with our results for the promoted catalysts, Mo/G shows an opposite trend. The selectivity of the 4,6- DMDBT through the DDS pathway was found to be higher over this catalyst than DBT. It seems that active sites for hydrogenation of DBT and 4,6-DMDBT are different over this catalyst. This could be due to a special synergic effect between MoS2 and the greigite core. Addition of the promoter to the MoS2 enhances the rate of C-S bond cleavage which is the rate-limiting step in desulfurization of the DBT and 4,6-DMDBT. Thus, the activity of the desulfurization of 4,6-DMDBT over the promoted catalysts is significantly enhanced.

Greigite vs Magnetite core

The stability of the MoS2/Fe3S4 composite was investigated by reusing the catalyst for several cycles of DBT HDS. After each usage, the catalyst was collected, washed and then returned to the reactor to perform another round of reaction. The results of each cycle are shown in Table 2. As a comparison, similar experiment was performed using the composite containing magnetite cores (MoS2/Fe3O4). Table 3 shows the recycling results for this composite. The results for MoS2/Fe3S4 shows no significant change after 7 cycles of HDS test. When compared to the composites prepared with a magnetite core, composites containing a greigite core show higher hydrogenation selectivity in HDS of DBT for the first 3 cycles. Daage and Chianelli [17] had shown in their study that changes in the number of MoS2 layers can change the selectivity of the catalyst. However, the length and number of layers of MoS2 were similar for both catalysts. The

1

23456789

101112131415161718192021

22

23242526272829303132333435363738394041

1

9

Figure 10: XRD patterns of treated MoS2/Fe3O4 and after cycles 1, 3, 5, and 7.

difference was most likely originated from the different core materials which can cause different synergic effects with the sulfided metals. It was reported that doping Co and Ni to the Fe-Mo catalyst resulted in the formation of Ni(Co)-Fe-Mo that was found to significantly increase the hydrodenitrogenation (HDN) activity of the catalyst [18]. HDN, which was well known, favors the hydrogenation pathway. This may serve as an indirect support explanation to support the increase in the hydrogenation selectivity over the catalysts with greigite core.

For Catalyst MoS2/Fe3O4, the desulfurization pathway shifts from DDS to HYD reactions after the 3 rd cycle. At cycle 4, the DBT conversion was increased and theselectivity of hydrogenated compounds (CHB and BCH) and isomerized products was increased. On the other hand, the selectivity of BP was decreased by 25%. At cycle 5, similar increasing trend for DBT conversion and selectivity of the hydrogenated products and isomers was observed and the values only showed a slight change after the 5th cycle. After cycle 3, the catalyst had lost its

1

2

3

4 56789

1011121314

15161718192021222324

1

10

magnetic properties and could not be collected by the hand-held magnet (the catalyst was separated through centrifugation). To investigate the changes in the crystal structure of the MoS2/Fe3O4 sample, XRD tests were performed (Figure 10). After cycles 1 and 3, new peaks observed can be assigned to iron showing the reduction of part of magnetite particles. Moreover, the intensity of the magnetite peaks at the 3rd cycle is reduced. At cycle 5, the magnetite peaks are replaced by pyrrhotite (Fe7S8) indicating the change in the magnetite particles via adsorption of H2S gas produced as the result of HDS reaction. The iron peak can also be seen as a shoulder to the pyrrhotite peak at 53° (inserted figure) and as a small peak at 65°. At the same time, the selectivity of the catalyst is shifted toward HYD indicated by higher selectivity of CHB and BCH (Table 3). At cycle 7, the intensity of the pyrrhotite peaks was increased and the catalyst activity and selectivity remained mostly unchanged in favor of HYD reaction pathway. The results clearly show the higher stability of the greigite cores compared to their magnetite counterparts during our reaction conditions after 7 cycles of HDS of DBT.

ConclusionsWe have demonstrated the preparation of unpromoted and uniformly dispersed promoted MoS2-based composites with magnetic properties through a facile procedure. High activity toward the refractory sulfur-containing compounds and the magnetic properties associated with these nano-composites provides a vital advantage for them, as easy separation and reuse is facilitated, which can be very useful in the processing of heavy and extra-heavy crude oils in slurry reactors. Three catalysts were synthesized in this work, two promoted catalysts (CoMoS/Fe3S4 and NiMoS/Fe3S4) and one unpromoted catalyst (MoS2/Fe3S4). Greigite (Fe3S4) was used as the core material considering its magnetic properties and easy adaptability with the sulfided catalyst and the sulfur-containing environment of the hydrotreating units. The greigite core was covered with catalyst layers as the active phase with the aid of cetyltrimethylammonium chloride (CTAC) as surfactant. When previous techniques failed to utilize promoted sulfide catalysts in slurry reactors, the novel technique used in this study showed the possibility of promoting MoS2 and increase the activity of the catalyst. NiMo/G showed the highest activity in the HDS of DBT and 4,6-DMDBT. The activity of CoMo/G in HDS of DBT and the DDS/HYD ratio for this catalyst was found to be similar to the Ni-promoted catalyst. Mo/G showed a higher tendency toward the HYD pathway with a DDS/HYD ratio of 0.31. The selectivity of the isomerized products was found to be higher for Mo/G, an indication of higher acidity of this catalyst. This was assigned to the difference in the acidic nature of –SH groups formed through the dissociation of H2S over the CUS sites for the promoted and unpromoted catalysts. In the presence of a weak metal-sulfur bond as in promoted catalysts, the –SH groups were considered to work as nucleophilic centers while over Mo/G with a strong metal-sulfur bond, they act as Brønsted acid cites.NiMo/G showed the highest HYD activity (81% conversion) with a DDS/HYD ratio of 0.19. CoMo/G and Mo/G followed in terms of activity for HDS of 4,6-DMDBT

and showed DDS/HYD ratios of 0.25 and 0.5, respectively. The higher affinity of the Mo/G toward the DDS pathway was related to a possible synergic effect between the MoS2 and the Fe3S4 core. The preference for the HYD pathway in HDS of 4,6-DMDBT for the prepared catalysts was correlated to the hindrance effect of alkyl groups close to the sulfur atom of 4,6-DMDBT. Finally, compared to the MoS2/Fe3O4 composite, the MoS2/Fe3S4 composite showed improved stability during several cycles of HDS test and higher HYD selectivity in the first 4 cycles.

Supporting Information SummaryThe reaction routes for HDS of DBT and 4,6-DMDBT depicted as schemes 1 and 2 and details on the catalysts synthesis, characterization, treatment, and evaluation can be found in the supporting information.

AcknowledgmentsThe authors are grateful for the financial support given by the Natural Science and Technology Research Council of Canada and the Canada Research Chair program.

KeywordsGreigite, Heterogenous catalysis, Hydrodesulfurization, Magnetic properties, Ni(Co)MoS/Fe3S4.

References[1] a) S. Zhang, D. Liu, W. Deng, G. Que, Energy

Fuels 2007, 21, 3057-3062; b) E. Furimsky, Catalysts for upgrading heavy petroleum feeds, Vol. 169, Elsevier, 2007.

[2] a) K. Niemann, F. Wenzel, Fuel Process. Technol. 1993, 35, 1-20; b) A. Delbianco, S. Meli, L. Tagliabue, N. Panariti, in 19th World Petroleum Congress, World Petroleum Congress, 2008.

[3] a) H. Rezaei, X. Liu, S. J. Ardakani, K. J. Smith, M. Bricker, Catal. Today 2010, 150, 244-254; b) R. Shen, H. Zhao, C. Liu, G. Que, Pet. Process. Petrochem. 1998, 29, 10-12.

[4] G. Bellussi, G. Rispoli, A. Landoni, R. Millini, D. Molinari, E. Montanari, D. Moscotti, P. Pollesel, J. catal 2013, 308, 189-200.

[5] S. Sharifvaghefi, Y. Zheng, ChemCatChem 2015, 7, 3397-3403.

[6] a) Y.-S. Chang, S. Savitha, S. Sadhasivam, C.-K. Hsu, F.-H. Lin, J. Colloid Interface Sci. 2011, 363, 314-319; b) M. Feng, Y. Lu, Y. Yang, M. Zhang, Y.-J. Xu, H.-L. Gao, L. Dong, W.-P. Xu, S.-H. Yu, Sci. rep. 2013, 3; c) X. Wang, W. Cai, G. Wang, Z. Wu, H. Zhao, CrystEngComm 2013, 15, 2956-2965.

[7] a) B. Yoosuk, J. H. Kim, C. Song, C. Ngamcharussrivichai, P. Prasassarakich, Catal. Today 2008, 130, 14-23; b) W. Lai, Z. Chen, J. Zhu, L. Yang, J. Zheng, X. Yi, W. Fang, Nanoscale 2016, 8, 3823-3833; c) F. Niefind, J. Cruz-Reyes, M. Del Valle, L. Kienle, A. Lotnik, G. Alonso-Nunez, M. Armbrüster, W. Bensch, Catal. lett. 2012, 142, 1312-1320; d) A.-F. Lamic, A. Daudin, S. Brunet, C. Legens, C. Bouchy, E. Devers, Appl. Catal., A 2008, 344, 198-204; e) G. Leyral, S. Brillouet, J.

123456789

10111213141516171819202122232425

26272829303132333435363738394041424344454647484950515253545556575859606162636465

66676869707172737475767778

79808182

83

8485868788

8990

91

9293949596979899

100101102103104105106107108109110111112113114115116117118119120121122123124125126127

1

11

Rousseau, F. Richard, A. S. Mamede, L. Courthéoux, A. Pradel, M. Ribes, S. Brunet, Appl. Catal., A 2017, 532, 120-132.

[8] a) Y. Zhang, Y. Du, H. Xu, Q. Wang, CrystEngComm 2010, 12, 3658-3663; b) L. Kong, L. Yan, Z. Qu, N. Yan, L. Li, Journal of Mater. Chem. A 2015, 3, 15755-15763; c) F. Cui, G. Li, X. Li, M. Lu, M. Li, Catal. Sci.Technol. 2015, 5, 549-555.

[9] J. V. Lauritsen, J. Kibsgaard, G. H. Olesen, P. G. Moses, B. Hinnemann, S. Helveg, J. K. Nørskov, B. S. Clausen, H. Topsøe, E. Lægsgaard, J. Catal. 2007, 249, 220-233.

[10] R. Chianelli, G. Berhault, P. Raybaud, S. Kasztelan, J. Hafner, H. Toulhoat, Appl. Catal., A 2002, 227, 83-96.

[11] S. H. Yang, C. N. Satterfield, J. Catal. 1983, 81, 168-178.

[12] J. Lauritsen, M. Nyberg, J. K. Nørskov, B. Clausen, H. Topsøe, E. Lægsgaard, F. Besenbacher, J. Catal. 2004, 224, 94-106.

[13] M. Egorova, R. Prins, J. Catal. 2004, 225, 417-427.

[14] a) K. G. Knudsen, B. H. Cooper, H. Topsøe, Appl. Catal., A 1999, 189, 205-215; b) M. Landau, Catal. Today 1997, 36, 393-429.

[15] T. Kabe, A. Ishihara, Q. Zhang, Appl. Catal., A 1993, 97, L1-L9.

[16] F. Bataille, J.-L. Lemberton, P. Michaud, G. Pérot, M. Vrinat, M. Lemaire, E. Schulz, M. Breysse, S. Kasztelan, J. Catal. 2000, 191, 409-422.

[17] M. Daage, R. Chianelli, J. Catal. 1994, 149, 414-427.

[18] T. Ho, A. Jacobson, R. Chianelli, C. Lund, J. Catal. 1992, 138, 351-363.

123456789

101112131415161718192021222324252627282930313233343536

37

38

1

12

Entry for the Table of Contents

Steps to synthesising magnetically recyclable Ni(Co)MoS/Fe3S4 nanocomposites were shown. These composites were found to be highly active in hydrodesulfurization of dibenzothiophene and 4,6-dimethyldibenzothiophene. Composites with greigite core were found to be more stable and have higher hydrogenation selectivity in different cycles of DBT HDS, when compared with composites containing magnetite as the core.

1

2

3

456789

10

1