In recent years, air pollution caused by exhaust gases from diesel engines has become a serious problem. Environmental legislation is placing increasingly severe restrictions on the sulfur contents of fuels. Research on efficient deep hydrodesulfurization (HDS) catalysts is therefore a hot topic. Dibenzothiophene (DBT) and its alkyl-substituted derivatives are the most difficult components to remove from the diesel fraction because of steric hindrance [1]. In the last decade, traditional HDS catalysts have usually been based on Mo (W) sulfides promoted by Co (Ni) supported on γ-Al₂O₃ [2, 3, 4, 5]. Other mesostructured materials such as MCM-41, SBA-15, HMS, KIT-1, and KIT-6 are also used as supports [6, 7, 8, 9, 10, 11, 12]. The properties of these materials are different from those of Al₂O₃. KIT-1 is a three-dimensional mesoporous molecular sieve with a uniform pore size and high surface area. In particular, disordered KIT-1 is hydrothermally more stable than ordered MCM-41 [13]. The catalytic activity of KIT-1-supported MoO₃ and/or NiO catalysts in thiophene HDS is higher than that of catalysts supported on MCM-41 and NaY zeolites [11].

These mesoporous materials have appropriate physicochemical properties. However, their practical application is limited because their pore walls are amorphous, resulting in relatively low acidity and thermal stability. The design and development of novel catalyst materials with large pores and suitable acidities are therefore needed. Large pores eliminate the diffusion resistance of reactant molecules, and increased acidity improves the hydrogenation and hydrogenolysis reactions of C-S bonds in sulfides. Some studies have shown that because of the high adsorption potential of micropores, some micro-mesoporous molecular sieves can adsorb H₂ molecules effectively. This reduces the energy cost and improves the catalyst efficiency under mild reaction conditions [14]. Micro-mesoporous materials, including Y-MCM-41, β-MCM-41, mesoporous ZSM-5, mesoporous β, β-KIT-6, and L-SBA-15, have been synthesized and tested in HDS reactions [7, 15, 16, 17, 18, 19]. Sun and Prins [19] reported high activities in the HDS of 4,6-dimethyldibenzothiophene (4,6-DMDBT) over noble metals supported on mesoporous ZSM-5 zeolites. The catalytic activity of β-MCM-41-supported NiW HDS catalyst was higher than that of the Al₂O₃-supported catalyst [15]. Li et al. [7] prepared MCM-41/HY composite materials (MY); NiMo/MY catalysts showed higher HDS activities and reducibilities than NiMo/MCM-41. We previously reported the direct hydrothermal synthesis of ZSM-5/KIT-1 micro-mesoporous molecular sieves (denoted by ZK-1) by one-step crystallization using dual templates [20, 21]. No precursor crystallization is required, which prevents overgrowth of the precursors before the final crystallization step. The framework structure, porosity, and morphology of ZK-1 can be well controlled by simply varying the crystallization temperature (90-170℃). The surface area, pore volume, hydrothermal stability, and acidity of ZK-1 are higher than those of KIT-1.

In the present work, ZK-1-supported CoMo catalysts were prepared and tested in the HDS of DBT. The effects of the support properties on the catalytic performance were clarified by comparison with other catalysts, namely CoMo catalysts supported on ZSM-5, AlKIT-1, and commercial γ-Al₂O₃. The catalysts were characterized using X-ray diffraction (XRD), high- resolution transmission electron microscopy (HRTEM), N₂ adsorption, temperature-programmed NH₃ desorption (NH₃- TPD), temperature-programmed H₂ reduction (H₂-TPR), and ultraviolet-visible diffuse reflectance spectroscopy (UV-Vis- DRS). HDS of DBT was used as a test reaction for comparison of the catalytic performance of six different catalysts.

ZK-1 was synthesized using a previously reported direct method [20]. The experimental procedure was as follows. (1) Aluminum isopropoxide was added to an aqueous solution of tetrapropylammonium hydroxide (TPAOH). The mixture was stirred at 0℃ until a clear solution was formed, and then tetraethylorthosilicate was added. The mixture was stirred at room temperature for several hours. (2) A mixture of cetyltrimethylammonium bromide (CTAB) and deionized water was added to the aluminosilicate precursor. The molar composition was 60(or 80)SiO₂:Al₂O₃:15TPAOH:15CTAB:1000H₂O. The resulting mixture was transferred to a stainless-steel autoclave. (3) After hydrothermal crystallization for 48 h at 130℃, the solid product was removed by filtration, washed with water, and dried at 80℃ for 12 h. (4) The samples were calcined at 550℃ for 10 h to give the final products ZK-1(30) (Si/Al = 30) and ZK-1(40) (Si/Al = 40).

AlKIT-1 (Si/Al = 30) was prepared using a method described in the literature [13]. γ-Al₂O₃ and ZSM-5 (Si/Al = 35) are industrial catalysts. The support Mix was a mechanical mixture of equal amounts of AlKIT-1 and ZSM-5.

H-type AlKIT-1 and ZSM-5 materials were obtained by ion exchange with NH₄NO₃ aqueous solution (0.4 mol/L) for 6 h (three times). The products were washed and dried at 100℃, and calcined at 500℃ for 4 h. The corresponding catalysts on different supports (ZK-1, γ-Al₂O₃, ZSM-5, AlKIT-1, and Mix) were prepared by co-impregnation with ammonium molybdate and cobalt nitrate. The samples were dried overnight at 100℃ and calcined at 500℃ for 3 h. The resulting catalysts, containing 20 wt% MoO₃ and 6.94 wt% CoO, were denoted by CoMo/ZK-1(30), CoMo/ZK-1(40), CoMo/γ-Al₂O₃, CoMo/AlKIT-1, CoMo/ZSM-5, and CoMo/Mix.

XRD patterns of the supports and catalysts were obtained using a Rigaku D/Max 2400 diffractometer with a Ni-filtered Cu K_α X-ray source at a scanning rate of 0.02°/s. The 2θ range was 5°-80° for wide-angle scans and 0.5°-10° for small-angle scans.

N₂ adsorption-desorption isotherms were obtained using a Quantachrome AUTOSORB-1-MP apparatus at liquid N₂ temperature (-196℃). The samples were degassed at 350℃ prior to analysis. The mesoporous structure was determined from the adsorption branch of the isotherms using the Barrett- Joyner- Halenda model. The microporous distribution was calculated using the Horvath-Kawazoe model.

NH₃-TPD was performed using a ChemBET 3000 chemisorption instrument (Quantachrome). The effluent stream was monitored continuously using a thermal conductivity detector (TCD) to determine the rate of NH₃ desorption.

H₂-TPR was performed using a ChemBET 3000 chemisorption instrument. The amount of sample used for each measurement was 0.1 g. The samples were pretreated with Ar at 20 mL/min, heated at 5℃/min to 500℃, and held at this temperature for 1 h. The temperature was lowered to 80℃, the Ar flow was switched to a 10% H₂/Ar flow, and the samples were heated to 1000℃ at 10℃/min. The effluent gas was analyzed using a TCD.

UV-Vis spectra were obtained from 190 to 800 nm using a Jasco UV-550 spectrometer; pure BaSO₄ was used as a reference.

HRTEM images of the sulfided catalysts were obtained using a Philips Tecnai G2 F20 transmission electron microscope operated at an accelerating voltage of 200 kV. The samples were suspended and dispersed on a grid. The solid samples were ultrasonically dispersed in ethanol, and test samples were prepared by dropping the dispersed suspensions on carbon-coated copper grids.

The catalytic activity of the catalysts were investigated using DBT as the probe reactant, because HDS of DBT is widely used as a model reaction for studying deep HDS of diesel fuels. The reactions were performed in a continuous fixed-bed reactor using 0.5 g of catalyst. All the catalysts were presulfided in situ with 3.0 wt% CS₂- cyclohexane and H₂ mixture at 290℃ and 3 MPa. The activity tests were performed at 320℃ and 3.0 MPa, with H₂/oil ratio of 300. The sulfur contents of the feed and product were determined using a ZWK-2001 PC sulfur chlorine analyzer (Jiangyan Gaoke Analysis Instruments Ltd., China). The catalytic activity was estimated as the HDS efficiency, which is defined as: HDS efficiency = [(S_f - S_p)/S_f] × 100%, where S_f and S_p are the sulfur concentrations in the feed and product, respectively.

The XRD patterns of the catalysts are shown in Fig. 1. Fig. 1(a) shows the small-angle XRD patterns. Samples (1), (2), (4), and (6) give peaks at 2θ = 2°-3° and 2θ = 4.5°, corresponding to the (100) and (200) planes of KIT-1, respectively [20]. The presence of these peaks indicates that the mesopores of KIT-1 remained after impregnation and calcination.

Fig. 1. Small-angle (a) and wide-angle (b) XRD patterns of the catalysts. (1) CoMo/ZK-1(30); (2) CoMo/ZK-1(40); (3) CoMo/γ-Al2O3; (4) CoMo/AlKIT-1; (5) CoMo/ZSM-5; (6) CoMo/Mix.

Fig. 1(b) shows the wide-angle XRD patterns. The ZSM-5 zeolite and Mix catalyst patterns both have peaks at 2θ = 7.7° and 8.6°, which are characteristic of zeolite ZSM-5 [22]. However, these peaks are absent from the patterns for samples (1) and (2), suggesting the absence of segregated microporous ZSM-5 crystals or that the amount is below the detection limit. Samples (1), (2), and (4) give weak peaks at 2θ = 27.3°, indicating small amounts of MoO₃ crystalline phases. The peaks at 2θ = 23.4° and 26.5° are assigned to β-CoMoO₄ crystallites [23]. For the CoMo/Al₂O₃ catalyst, the peaks at 2θ = 37.9°, 45.9°, and 66.8° are associated with γ-Al₂O₃ [24]. No obvious peaks at 2θ = 23.4°, 26.5°, and 27.3° from CoMoO₄ and MoO₃ crystals are detected in the pattern of the CoMo/Al₂O₃ catalyst, implying that MoO₃ and CoO are highly dispersed on the support. Samples (5) and (6) both give an intense peak corresponding to crystalline MoO₃, indicating that the distribution of MoO₃ is not homogeneous.

The N₂ adsorption-desorption isotherms of the supported catalysts are shown in Fig. 2(a). The adsorption and desorption curves of samples (3) and (4) show typical type IV isotherms. Sample (5) has a type I isotherm, corresponding to a microporous material. Samples (1), (2), and (6) have characteristics similar to type IV and I isotherms. The initial increase in the adsorbed volume at low pressure is the result of monolayer adsorption in micropores. Samples (1) and (2) show steep increases in their curves at a relative pressure p/p₀ < 0.01; this is caused by filling of the micropores. A further increase in the adsorbed volume occurs at p/p₀ = 0.2-0.3 for samples (1), (2), (4), and (6); this indicates capillary condensation. In the isotherms of sample (3), the inflection appears at a higher p/p₀. This suggests that the average mesopore size of sample (3) is larger, which is consistent with the mesoporous pore size distribution results. The pore size distributions obtained from the adsorption isotherms are shown in Fig. 2(b) and (c). There are no mesopores in sample (5), which is supported on ZSM-5. The average mesopore size of samples (1) and (2) is 2.3 nm, which is slightly smaller than that of 2.5 nm for samples (4) and (6).

Fig. 2. N2 adsorption-desorption isotherms (a), mesopore size distributions (b), and micropore size distributions (c) of the catalysts. (1) CoMo/ZK-1(30); (2) CoMo/ZK-1(40); (3) CoMo/γ-Al2O3; (4) CoMo/AlKIT-1; (5) CoMo/ZSM-5; (6) CoMo/Mix.

Table 1 shows the textural and structural properties of the supported catalysts. Samples (1) and (2) have the highest surface areas (~700 m²/g), followed by samples (4) (522 m²/g) and (6) (417 m²/g). The pore size distributions of samples (1) and (2) are special because the ZK-1 support has the mesoporous structure of KIT-1 and zeolite building units [19]. However, ZK-1 is not a mixture of KIT-1 and ZSM-5. The specific surface areas and pore volumes of samples (3) and (5) are smaller than those of the other samples. This is because of the smaller surface areas and pore volumes of the supports, i.e., γ-Al₂O₃ and ZSM-5.

Table 1 Textural properties of the catalysts. Sample a Surface area (m2/g) Micropore surface area (m2/g) Micropore size (nm) Mesopore size (nm) Pore volume (cm3/g) (1) (2) (3) (4) (5) (6) 699 730 230 522 299 417 574 586 0 395 265 324 0.6-1 0.6-1 — — 0.5-0.6 0.6-1 2.3 2.3 10 2.5 — 2.5 0.76 0.85 0.55 0.72 0.21 0.56 a (1) CoMo/ZK-1(30); (2) CoMo/ZK-1(40); (3) CoMo/γ-Al2O3; (4) CoMo/AlKIT-1; (5) CoMo/ZSM-5; (6) CoMo/Mix.

Table 1 Textural properties of the catalysts.

Fig. 3 shows NH₃-TPD profiles of the supported catalysts. The peak temperature represents the acidic strength and the peak intensity represents the amount of acidic centers. The TPD profiles of the catalysts contain a single peak at about 250℃. The total acidities, derived from the NH₃-TPD profile areas, decrease in the order CoMo/ZK-1(30) > CoMo/AlKIT-1 > CoMo/ZK-1(40) > CoMo/Mix > CoMo/γ-Al₂O₃ > CoMo/ZSM-5. The area of the desorption peak of CoMo/ZK-1(30) is clearly larger than that of CoMo/AlKIT-1, which has the same Si/Al ratio. This is because the introduction of microporous structures into the mesoporous material increases the amount of acidic sites and the acidic strength of the catalyst.

Fig. 3. NH3-TPD curves of the catalysts. (1) CoMo/ZK-1(30); (2) CoMo/ZK-1(40); (3) CoMo/γ-Al2O3; (4) CoMo/AlKIT-1; (5) CoMo/ZSM-5; (6) CoMo/Mix.

TPR was used to obtain information on the interactions between the supported phase and the carrier. The TPR profiles of the calcined samples are shown in Fig. 4. The profiles of samples (1) and (2) (CoMo/ZK-1) have three peaks, at 543, 604, and 810℃, with the same H₂ consumption distributions. According to previous studies [24, 25], the two overlapping peaks in the 500-650℃ region indicate the coexistence of octahedral Mo species with different degrees of agglomeration. The low- temperature reduction peaks are assigned to the first step of Mo reduction (Mo⁶⁺ + 2e⁻ → Mo⁴⁺) in polymeric octahedral species. The high-temperature reduction peak at around 810℃ is associated with complete reduction (Mo⁴⁺ + 2e⁻ → Mo⁰) of polymeric octahedral and tetrahedral Mo species, or reduction of Mo species that strongly interact with the support. Sample (4) has similar characteristics to samples (1) and (2), but the low-temperature reduction peak shifts to lower temperature (from 604℃ to 575℃). The profiles of samples (3) and (6) have two peaks, in the ranges 450-650℃ and 700-900℃. The reduction temperature of the low-temperature peak is lower and the hydrogen consumption in the high-temperature peak region is higher for CoMo/γ-Al₂O₃ than for the other samples. This can be explained as follows. The reduction of octahedral Mo⁶⁺ to Mo⁴⁺ is easier, but the strong interactions via Al-Mo linkages make it more difficult to completely reduce and sulfide Mo species on γ-Al₂O₃.

Fig. 4. TPR profiles of the catalysts. (1) CoMo/ZK-1(30); (2) CoMo/ZK-1(40); (3) CoMo/γ-Al2O3; (4) CoMo/AlKIT-1; (5) CoMo/ZSM-5; (6) CoMo/Mix.

The changes in the electronic spectra of the samples reflect the changes in the coordination structure of Co and Mo. According to the literature [26, 27], tetrahedral Mo species show absorption at 220-250 nm, whereas octahedral Mo species show transitions at 250-330 nm. Absorption at about 400 and 750 nm is characteristic of octahedral Co²⁺ species, and a broad band centered at about 600 nm is associated with tetrahedral Co ions [23, 28, 29]. An octahedral component indicates the presence of CoMoO₄ species, whereas the presence of tetrahedral coordinated Co²⁺ species indicates the formation of Co₂AlO₄ or Co₂SiO₄ via interactions of Co species with the support [27].

The UV-Vis DRS of the catalysts are shown in Fig. 5. The spectra of samples (1) and (2) (CoMo/ZK-1) show absorption from tetrahedral Mo and octahedral Mo species at 220-330 nm. No visible absorption from cobalt oxide is present at 400-700 nm, but a weak band is observed at 750 nm. This octahedral component indicates the presence of CoMoO₄ species, as confirmed by the XRD results. Similar results are observed for Mo supported on AlKIT-1 (sample (4)). Sample (4) also shows relatively weak absorption by Co²⁺ at about 600 nm. Compared with these three catalysts, catalysts (3), (5), and (6) have wider and stronger absorptions at 280-330 nm. This may be caused by formation of a large amount of polymeric molybdate species on the supports. Fig. 5 clearly shows that the intensities of the absorptions of the tetrahedral Co²⁺ species at about 600 nm are larger for samples (3) and (5). This is probably the result of the formation of large amounts of catalytically inactive Co₂AlO₄ or Co₂SiO₄ on the γ-Al₂O₃ and ZSM-5 surfaces.

Fig. 5. UV-Vis spectra of the catalysts. (1) CoMo/ZK-1(30); (2) CoMo/ZK-1(40); (3) CoMo/γ-Al2O3; (4) CoMo/AlKIT-1; (5) CoMo/ZSM-5; (6) CoMo/Mix.

HRTEM is the most effective technique for examining the morphology of active phases, by visualizing MoS₂ crystallite slabs on the support. Figs. 6 and 7 show representative HRTEM micrographs and the statistical results, respectively, for various sulfided catalysts. All the images show thread-like fringes corresponding to MoS₂ slabs with interplanar distances of 0.61 nm [30]. The particles of promoter Co species on the sulfided catalysts are too small to be visualized, because of the low loading of Co species. A quantitative comparison was made by determining the lengths and layer numbers of the MoS₂ slabs on the sulfided catalysts by statistical analyses based on about 30 micrographs including at least 300 slabs taken from different parts of each catalyst. The average slab length and stacking number were calculated according to Eq. (1):

Fig. 6. HRTEM images of the sulfided catalysts. (a) CoMo/ZK-1(30); (b) CoMo/ZK-1(40); (c) CoMo/γ-Al2O3; (d) CoMo/AlKIT-1; (e) CoMo/ZSM- 5; (f) CoMo/Mix.

where M_i is the slab length or stacking layer number of a stacked MoS₂ unit, and x_i is the number of slabs or stacks in a certain range of length or stacking layer number [30]. For the supported MoS₂ HDS catalysts, Eq. (2),

which is widely used to estimate the relative proportions of Mo atoms per crystallite, gives a good estimate of the turnover frequency [31]. In Eq. (2), n_i is the number of Mo atoms along one side of a MoS₂ slab, determined from its length (= 0.32(2n_i − 1), nm), and t is the total number of slabs shown in the HRTEM micrograph. It is assumed that all the edge sites are active in HDS, therefore the real value of f_Mo could be higher than that estimated from Eq. (2).

Fig. 7. Distributions of lengths (a) and layer numbers (b) of MoS2 slabs on various catalysts.

It can be observed that the MoS₂ phases on different supports differ significantly from each other. MoS₂ is more evenly dispersed on ZK-1 than on other supports. Good dispersion of the active metals creates a large number of accessible active sites. The catalysts CoMo/Al₂O₃ and CoMo/ZSM-5 have high degrees of stacking and long lengths because of the small surface areas of the supports. Based on the statistical results in Table 2, the values of the MoS₂ slabs of different catalysts increase in the order CoMo/ZK-1(30) < CoMo/ZK-1(40) < CoMo/AlKIT-1 < CoMo/Mix < CoMo/Al₂O₃ < CoMo/ZSM-5; the order of the values is CoMo/ZK-1(40) < CoMo/ZK-1(30) < CoMo/AlKIT-1 < CoMo/Mix < CoMo/Al₂O₃ < CoMo/ZSM-5. The trend in the f_Mo values is CoMo/ZSM-5 < CoMo/Al₂O₃ < CoMo/Mix < CoMo/AlKIT-1 < CoMo/ZK-1(30) < CoMo/ZK-1(40). The shortest length, lowest stacking number, and largest f_Mo value for MoS₂ particles are obtained over CoMo/ZK-1, confirming that CoMo/ZK-1 shows better MoS₂ dispersion and has more active edge sites. Compared with CoMo/ZK-1(30), the high stacking number (2.7 layers) and relatively short slab length (3.88 nm) of the MoS₂ crystallites on CoMo/ZK-1(40) result in the creation of more available edge and corner sites, enabling the incorporation of more promoter atoms into the structure of the type II Co-Mo-S phase. This is because of the different aluminum contents in the ZK-1 framework, which results in varying degrees of metal-support interactions. Suitable metal-support interactions improve dispersion and sulfidation of oxidic Mo, and this improves the HDS activity of catalysts. Although the specific surface area of CoMo/ZSM-5 (299 m²/g) is greater than that of CoMo/Al₂O₃ (230 m²/g), CoMo/ZSM-5 has higher values of and because of the small external surface area of ZSM-5. Highly stacked MoS₂ crystallites are unfavorable for HDS reactions. In addition, MoS₂ crystals are homogeneously distributed in CoMo/Al₂O₃ because of the strong interaction between oxidic Mo and Co species and Al-OH on the support, as confirmed by the H₂-TPR results.

Table 2 Average lengths () and average stacking layer numbers () of MoS2 in series of CoMo sulfided catalysts. Catalyst (nm) fMo CoMo/ZK-1(30) CoMo/ZK-1(40) CoMo/γ-Al2O3 CoMo/AlKIT-1 CoMo/ZSM-5 CoMo/Mix 4.03 3.88 6.43 4.66 7.67 5.42 2.51 2.73 3.32 2.78 4.54 3.17 0.27 0.28 0.20 0.24 0.16 0.21

Table 2 Average lengths () and average stacking layer numbers () of MoS2 in series of CoMo sulfided catalysts.

The catalytic performance of a series of CoMo catalysts on different support materials, namely ZK-1, γ-Al₂O₃, AlKIT-1, and ZSM-5, in HDS of DBT was evaluated. The results are shown in Fig. 8.

Fig. 8. HDS of DBT over the catalysts. (1) CoMo/ZK-1(30); (2) CoMo/ZK-1(40); (3) CoMo/γ-Al2O3; (4) CoMo/AlKIT-1; (5) CoMo/ZSM-5; (6) CoMo/Mix. Reaction conditions: temperature 320℃, pressure 3 MPa, weight hourly space velocity 5 h−1, time on stream 60 h.

Fig. 8 shows that the order of the desulfurization rates of the catalysts is CoMo/ZK-1(40) > CoMo/ZK-1(30) > CoMo/γ-Al₂O₃ > CoMo/ZSM-5 > CoMo/Mix > CoMo/AlKIT-1. The catalysts supported on ZK-1 give the best catalytic performance, especially when the Si/Al ratio is 40. These results indicate that ZK-1 effectively combines the superior acidic properties of ZSM-5 building units and the mesoporous structure of KIT-1. ZK-1 provides a larger surface area and promotes the dispersion of Co and Mo phases on the support. Suitable interactions between the Co and Mo phases and the support improve reduction and sulfidation of the metallic phases. ZK-1 consists of a mesoporous KIT-1 structure containing ZSM-5 building units. Rozanska et al. [32] reported that zeolites alone can hydrodesulfurize DBT through direct sulfur atom elimination and hydrogenation desulfurization routes. Shen’s group [33, 34] prepared a NiW-based catalyst containing USY, which showed high activity in HDS of 4,6-DMDBT and fluid catalytic cracking of diesel oil. The presence of ZSM-5 building units enhances the hydrothermal stability and acidity of mesoporous molecular sieves compared with those of A1KIT-1 [21, 22]. CoMo/ZK-1 has a larger amount of weak acid sites than the other catalysts; this decreases the hydrocracking activity and increases the HDS activity. Furthermore, micro-mesoporous channel connectivity is a major factor in significantly improving the catalytic activity.

The desulfurization rate over the catalyst supported on γ-Al₂O₃ is 90%, indicating that it is a suitable support for HDS catalysts. The activity of CoMo/ZSM-5 is lower than those of CoMo/ZK-1 and CoMo/γ-Al₂O₃. The reasons are as follows. (1) The small pore channels in the ZSM-5 zeolite hinder access of the reactant molecules to the active sites. (2) The lower external surface area (34 m²/g) induces polymerization of the MoO₃ phases, which leads to the formation of more large stacked MoS₂ crystallites, as shown in the HRTEM micrographs. The lower f_Mo (0.16) of CoMo/ZSM-5 means that its DBT HDS activity is lower than those of the other samples.

Micro-mesoporous ZK-1 molecular sieves were used as supports for HDS catalysts. The characterization results show that CoMo/ZK-1 catalysts have large surface areas, large pore volume, and micro-mesoporous structure; these promote the dispersion of Co and Mo oxide phases on the ZK-1 support. CoMo/ZK-1 has a larger number of acidic sites than CoMo/AlKIT-1, which have the same Si/Al ratio. Suitable interactions between the Co and Mo phases and ZK-1 promote reduction and sulfidation of the metallic phases. HRTEM images show shorter lengths, lower stacking numbers, and larger f_Mo values for MoS₂ particles in CoMo/ZK-1, confirming that CoMo/ZK-1 has better MoS₂ dispersion and more active edge sites. The catalytic activity of the catalysts supported on ZK-1 in HDS of DBT is higher than that of catalysts supported on γ-Al₂O₃, AlKIT-1, ZSM-5, and a mixture of AlKIT-1 and ZSM-5.