Sulfur oxides are major components of pollution in the environment that are mainly released during the combustion process of gasoline and diesel fuels. More stringent fuel specifications with regard to sulfur content have been issued owing to the increased attention to environmental problems [1]. To meet the stricter sulfur content regulations, highly refractory sulfur molecules such as benzothiophene (BT), dibenzothiophene (DBT), 4, 6-dimethyldibenzothiophene (4, 6-DMDBT), and similar derivatives need to be desulfurized [2]. Various desulfurization technologies have been commercialized, such as hydrodesulfurization (HDS), biological desulfurization, and adsorption [3]. As one of the most important techniques, HDS is a highly efficient process and effective at a large scale [4]. Furthermore, the design and development of novel HDS catalysts with high selectivity and activity become more significant in HDS technology.

Mesoporous materials are the most ideal candidates as catalyst supports owing to their tunable pore sizes and structures and excellent connectivity. Furthermore, these mesoporous materials exhibit high surface areas, large pore volumes, and good hydrothermal stability. Until now, many mesoporous materials such as MCM-41 [5], HMS [6], and SBA-15 [7] have been employed as supports for hydrodesulfurization catalysts. However, both MCM-41 and SBA-15 have a typical two dimensional (2D) hexagonal mesostructure [8]. In addition, the pore sizes of MCM-41 and SBA-15 are relatively small, approximately 3 and 6.5 nm, respectively. Therefore, mesoporous materials with large pore sizes and three-dimensional (3D) mesostructures are urgently needed to eliminate the diffusion resistance in the reactions that involve large reactant molecules. Mesostructured cellular silica foams (MCFs), which were prepared using a microemulsion droplet templating approach [9], display ultra large 3D pore sizes (20-50 nm), high surface areas (approximately 1000 m²/g), large pore volumes (approximately 2.6 cm³/g), and uniform size ranges [9-13]. The ability to tune the texture properties make MCFs promising candidates for catalytic support materials and separation sorbents in which large molecules are involved. However, compared with microporous zeolites, pure silicon materials such as MCFs have much weaker acidity because of their electrically neutral Si frameworks [14], which limits their applications to acid-catalyzed reaction such as HDS technology. The incorporation of heteroatoms such as Al, Ti, and Zr by using various post synthesis methods has been used to moderate the acid properties and improve the acidity of the silicon materials, and thus, increase the catalytic performance.

Schmidt et al. [15] first synthesized aluminum-incorporated MCM-41 mesoporous materials. The application of Al-MCM-41 as a catalyst has been reported by many researchers [16-20]. However, the poor hydrothermal stability and the small pore size of Al-MCM-41 are the key obstacles [19] that limit the application of this material. Many efforts have been made to synthesize heteroatom-substituted SBA-15 with a larger pore size. For example, Cheng et al. [21] reported the incorporation of aluminum into SBA-15 using a post synthesis method and the aluminum atoms were successfully incorporated into the framework of SBA-15, which resulted in a new kind of mesoporous acid catalyst. Luan et al. [22] synthesized SBA-15 molecular sieves and then incorporated with aluminum to produce Al-SBA-15 through three post synthesis procedures. Yue et al. [23] reported the first direct synthesis of Al-SBA-15 under acidic conditions and found that the catalytic activity in cumene cracking was higher than that of Al-MCM-41 owing to the large pore size and good hydrothermal stability. To the best of our knowledge, there are very few reports of Al-modified MCFs [24-26].

In this research, a series of Al-MCFs materials with different Si/Al ratios were successfully synthesized using pluronic P123 triblock copolymer as the template and 1, 3, 5-triethylbenzene (TMB) as the swelling agent and used as the supporting materials for the DBT HDS catalysts [27, 28]. Furthermore, the corresponding Al-MCFs-supported NiMo catalysts were prepared through a two-step incipient wetness impregnation method. Furthermore, the physicochemical properties of the supports and catalysts were characterized by various analytical methods and the performance of the HDS was evaluated by using DBT as the probe molecule. The effect of the amount of Al incorporated in the MCFs on the catalytic performance of the DBT HDS were investigated systematically.

Triblock copolymer P123 (EO₂₀-PO₇₀-EO₂₀; EO = ethylene oxide; PO = propylene oxide, average mass weight: 5800 g/mol) was purchased from Sigma-Aldrich Company Ltd. (USA). Tetraethyl orthosilicate (TEOS), 1, 3, 5-trimethylbenzene (TMB), aluminum isopropoxide, isopropanol, ammonium heptamolybdate, nickel nitrate, and hydrochloric acid (HCl) were obtained from Sinopharm Chemical Reagent Beijing Co., Ltd. (China). All the chemicals were used as received without any further purification.

P123 (2 g) was dissolved in 60.0 mL of aqueous HCl solution (2.0 mol/L) at room temperature. The solution was transferred to a water bath at a temperature of 35 ℃ and stirred for 4 h. Then, 4.25 g of TEOS and 2.5 g of TMB were added separately. The solution was maintained at 35 ℃ for 24 h. The as-synthesized material was isolated by filtering, washing, and drying at 80 ℃ for 12 h. Finally, the sample was calcined under air at 550 ℃ for 6 h to remove the template material.

Aluminum isopropoxide was used as the aluminum source and isopropanol was used as a solvent. The typical synthetic procedure for the post synthesis of Al-MCFs was as follows: MCF (1 g) was mixed with isopropanol (100 mL) at room temperature, the mixture was stirred for 0.5 h, then a certain amount of aluminum isopropoxide was added. The mixed solution was stirred for 12 h, filtered, washed with isopropanol, and dried at 80 ℃ in an oven for 10 h. Finally, the obtained product was calcined at 550 ℃ for 5 h. The amount of aluminum isopropoxide used in the grafting procedure per gram of pure siliceous MCFs corresponded to Si/Al molar ratios of 10, 20, 30, 40, and 50. The Al-containing materials were referred to as Al-MCFs(x), in which x is the Si/Al molar ratio.

The NiMo supported catalysts were prepared by a two-step incipient wetness impregnation method by the stepwise impregnation of ammonium heptamolybdate and nickel nitrate. The impregnation of the Mo precursor (MoO₃ 15.5 wt%) was carried out first, followed by the Ni precursor (NiO 3.5 wt%). After each impregnation step, the sample was dried at 100 ℃ for 24 h and calcined at 550 ℃ for another 4 h in air. All the five obtained catalysts were crushed into 0.3-0.5 mm particles and denoted as NiMo/Al-MCFs(10), NiMo/Al-MCFs(20), NiMo/Al-MCFs(30), NiMo/Al-MCFs(40), and NiMo/Al-MCFs(50). For the reference catalyst of NiMo/γ-Al₂O₃, pseudo-boehmite (Shandong Zibo, China) was used as the raw material for obtaining the γ-Al₂O₃ support, and the active metals loaded on the support were Mo and Ni precursor with a corresponding content of 15.5 and 3.5 wt%, respectively.

The typical physicochemical properties of different materials were measured by Nitrogen adsorption-desorption isotherms at -196 ℃ using a Micromeritics Tristar 3020 system. The specific surface area was determined by the Brunauer-Emmett-Teller (BET) method. The pore diameter and window pore sizes were determined by the Broekhoff-de Boer method with the Frenkel-Halsey-Hill equation (BdB-FHH) model, because this method is generally used to measure spherical mesopores. The total pore volumes were estimated from the adsorbed amount at a relative pressure p/p₀ of 0.98. Small-angle X-ray scattering (SAXS) patterns were recorded on a Nano STAR Small-Angle X-ray scattering system (Bruker, Germany) using Cu K_α radiation (40 kV, 35 mA). The UV-Vis diffuse reflectance spectroscopy (DRS) experiments were performed on a UV-Vis spectrophotometer (HitachiU-4100) with the integration sphere diffuse reflectance attachment. Raman spectra were recorded with a Renishaw Micro-Raman System 2000 spectrometer. The laser wavelength was 325 nm. The laser spot size was approximately 1-2 mm with a power of 8 mW. The type and amount of surface acids on the samples were determined by pyridine-FTIR (Py-FTIR) spectroscopy on a MAGNAIR 560 FTIR instrument with a resolution of 1 cm^-1. H₂ temperature-programmed reduction (H₂-TPR) analysis was carried out on a Quantachrome apparatus (Auto-sorb-iQ USA) in 10% H₂-90% Ar at a flow rate of 40 mL/min. Temperature-programmed desorption of ammonia (NH₃-TPD) was performed in a conventional flow apparatus. The sample was pretreated with high purity N₂ (40 mL/min) at 500 ℃ for 1 h, then saturated with high purity anhydrous ammonia at 100 ℃ for 1 h and subsequently flushed at the same temperature for 1 h to remove ammonium. Finally, the TPD operation was carried out from 100 to 700 ℃ at a heating rate of 10 ℃/min. X-ray photoelectron spectroscopy (XPS) was performed on a PerkinElmer PHI-1600 ESCA spectrometer using a Mg K_α (hv = 1253.6 eV) X-ray source. The binding energies were calibrated using the C 1s peak of contaminant carbon (BE = 2 84.6 eV) as an internal standard.

The catalytic activities of the catalysts were evaluated using DBT in cyclohexane as the model compound. The HDS reaction was carried out in a continuous fixed-bed Inconel reactor (8 mm inner diameter and 400 mm length) with 0.5 g of catalyst. All the fresh catalysts were presulfided with a mixture of 2 wt% CS₂ in cyclohexane solution at 360 ℃ for 4 h and a H₂ pressure of 4 MPa. After sulfidation, the reaction temperature was decreased to 320 ℃, and a reactant liquid with a DBT concentration of 500 × 10^-6 was fed to the reactor at 320 ℃, 4 MPa, and 200 mL/mL. The sulfur content in the feedstock and products were determined with a sulfur and nitrogen analyzer (RPP-2000SN, Taizhou Central Analytical Instruments Co. Ltd., China).

The average slab length (L_av) and stacking layer number (N_av) were calculated according to the following equation:

(1)

in which x_i is the number of slabs or stacks and M_i represents the length of the slab or the stacking layer number [29].

The catalytic performance (HDS conversion (%)) of the catalysts was calculated by Equation (2)

(2)

in which "S content in product" and "S content in feed" are the sulfur in the product and sulfur in the feed.

The SAXS patterns of the prepared Al-containing mesocellular silica foam materials with different Si/Al atios are shown in Fig. 1. The patterns exhibit three peaks at low q values (q = 4πsinθ/λ), indicating that the original mesostructures of the materials are retained after the incorporation of Al using the grafting technique [30, 31].

The N₂adsorption-desorption isotherms of the Al-MCFs(x) materials with different Si/Al ratios and their pore size distribution (PSD) are displayed in Fig. 2. All the adsorption-desorption isotherms (Fig. 2(a)) show a type-Ⅳ curve with a H1 hysteresis loop and a sharp capillary condensation step in the p/p₀ range of 0.65-0.95, indicating that a series of typical mesoporous materials with narrow PSDs were formed. The textural and structural characteristics of the Al-containing MCFs supports (surface area A_BET, total pore volume V_p, pore diameter D_p, and window pore size D_w) are summarized in Table 1. All the Al-MCFs(x) materials show a narrow distribution of mesopores with a maximum pore diameter of approximately 21-25 nm. The Al-MCFs(20) sample displayed the largest pore diameter (25.2 nm) and window pore size (12.4 nm), suitable surface area (537 m²/g), and high pore volume (2.15 cm³/g).

The morphologies of all the prepared Al-MCFs supports were investigated by field emission scanning electron microscopy (SEM). The results are shown in Fig. 3. All the Al-MCFs(x) materials showed highly monodispersed particles with respect to their original spherical morphologies, which indicated that there was no significant change of the morphologies of the MCFs after the incorporation of the Al atoms. Therefore, the post synthetic approach is an alternative method for the incorporation of Al atoms in MCFs. However, when the amount of Al incorporated is high (x = 10), amorphous aluminum can be detected (Fig. 3(a)). The morphologies of the samples were more uniform (Fig. 3(b-e)) as the amount of aluminum decreased. Furthermore, some adhesion structures appeared owing to cross linking between different hydroxy silicones, as indicated by the arrows (Fig. 3(b-e)).

The ²⁷Al MAS NMR spectra of the Al-MCFs materials with Si/Al = 20 and 40 were obtained (Fig. 4) to analyze the state of the Al species in the as-synthesized samples. From the spectra, three obviously intense peaks at 0, 33, and 54 ppm were observed. The peak at 54 ppm was assigned to aluminum in a tetrahedral environment (AlO₄ structural unit), in which the Al atom is covalently bonded to four Si atoms through oxygen bridges. The chemical shift at 0 ppm was attributed to octahedral aluminum (AlO₆ structural unit) [32]. Aluminum located at the interface between the tetrahedral alumino silicate framework and the octahedral alumina phase is responsible for pentahedral coordination. The presence of pentahedral aluminum indicates that the local arrangement of Al atoms is different from the bulk γ-Al₂O₃, which excludes the formation of a separate alumina phase [33]. The ratio of the Al species present can be evaluated from the relative intensities of the corresponding peaks in the ²⁷Al MAS NMR spectra. The distribution ratios of tetrahedral, pentahedral, and octahedral Al species in Al-MCFs(20) and Al-MCFs(40) were determined to be approximately 1:1.5:3.2 and 1:1.4:1.8, respectively. The proportion of octahedral Al sites was significantly high at a low Si/Al molar ratio. Thus, when the aluminum amount was low, it exhibited a preferential tetrahedral coordination, whereas the formation of Al_xO_yproceeded when the amount of aluminum increased [34].

The SAXS patterns of the prepared NiMo/Al-MCFs(x) catalysts are presented in Fig. 5. Compared with the SAXS patterns of Al-MCFs(x) in Fig. 1, all the NiMo/Al-MCFs(x) catalysts show the same signals, indicating that the mesostructures were retained after the loading of the active metal.

The N₂adsorption-desorption isotherms of all the NiMo/Al-MCFs(x) catalysts and their corresponding PSDs are displayed in Fig. 6. It is clear that the type-Ⅵ hysteresis loops (Fig. 6(a)) of the catalysts are well retained after loading of the active metals (Ni and Mo), which confirms the preservation of the mesoporous structures of the Al-MCFs(x) supports even after the impregnation of the active metals. It is consistent with the results of the SAXS patterns. The textural and structural characteristics of NiMo/Al-MCFs(x) catalysts (surface area A_BET, total pore volume V_p, pore diameter D_p, and window pore size D_w) are summarized in Table 1. Compared with the Al-MCFs(x) supports, the pore diameter and the window size of the corresponding catalysts are not obvious, indicating that the mesopore structure is maintained well even after the impregnation of the active metal, which is consistent with the SAXS result.

UV-Vis DRS was performed in the wavelength ranges of 200-800 nm to investigate the aggregation state of the Mo species in the NiMo/Al-MCFs(x) catalysts and the characterization results are shown in Fig. 7. The absorption bands between 200 and 400 nm were assigned to the ligand to metal charge transfer transitions (O^2- → Mo⁶⁺) [35, 36]. None of the spectra of the NiMo/Al-MCFs(x) catalysts exhibited absorption bands at wavelengths of 350 nm, which suggested that the catalysts did not contain aggregated MoO₃ species but highly dispersed Mo oxides, which are the precursors of sulfide active phases and are favorable for improvement of HDS conversion [37, 38]. In addition, the absorption bands between 220 and 270 nm were attributed to tetrahedral configuration, and the bands between 270 and 290 nm were attributed to the octahedral geometry of isolated molybdenum oxide centers [39]. The spectra clearly show that the bands of the Mo oxide species change if the amount of aluminum in the catalyst changes. The adsorption peaks associated with octahedral coordination (between 260 and 330 nm) widened with increased aluminum doping content, which indicated that the incorporation of Al atoms into the MCFs framework improved the dispersion of Mo species. In Fig. 7(b), the energy band gap of the Mo oxides species in the NiMo/Al-MCFs(20) catalyst was larger than all the other studied catalysts, demonstrating that a decrease in the average domain size of the Mo species occurs, which indicates a lower aggregation level and a higher dispersion state of MoO_x [40]. Consequently, this can improve the HDS activity, which is consistent with the analysis result of the Raman spectra and the DBT HDS activity testing results, which will be discussed in the following section.

Raman spectroscopy was extensively used to characterize the active metal oxides (Ni and Mo) in the framework structure and the supports. The Raman spectra of the prepared NiMo/Al-MCFs(x) catalysts with different Si/Al ratios are shown in Fig. 8. For all the NiMo/Al-MCFs(x) catalysts, five Raman bands at 388, 718, 828, 894, and 959 cm^-1 were observed. The peaks at 894 and 828 cm^-1 can be ascribed to the NiMoO₄ [41, 42] and the peak at 828 cm^-1is associated with the asymmetric stretching mode of the Mo-O-Mo bridge band of the octahedral molybdate species. Furthermore, the peak at 959 cm^-1 can be attributed to the symmetric and antisymmetric M=O terminal stretches of various surface polymolybdate species, indicating the presence of highly dispersed octahedral Mo⁶⁺ surface species [43]. The broad band at approximately 388 cm^-1 was attributed to the bending mode of the terminal Mo=O bond. The peak at 718 cm^-1 was attributed to the α-NiMoO₄ phase, which is the precursor of the Ni-Mo-S active phase and favors the HDS reaction [44]. Furthermore, the absence of well-defined peaks at 667 and 994 cm^-1, which are the characteristics of MoO₃ crystallites, indicates a good dispersion of Mo oxide species on all the prepared NiMo/Al-MCFs(x) catalysts [45, 46]. The Raman analysis results are consistent with the UV-Vis DRS analysis.

H₂-TPR characterization (Fig. 9) was performed to study the metal support interaction (MSI). For all the NiMo/Al-MCFs(x) catalysts, two main peaks were detected in the temperature range of 400-650 and 700-900 ℃. The low-temperature peak is generally associated with the reduction of Mo⁶⁺ to Mo⁴⁺ (Mo⁶⁺ +2e^- → Mo⁴⁺) of polymeric octahedral Mo species. The high temperature peak is assigned to the complete reduction (Mo⁴⁺ +2e^- → Mo⁰) of polymeric octahedral, tetrahedral, and bulk crystalline MoO₃ [47, 48]. Compared with the NiMo/MCFs catalyst, a significant decrease in the reduction temperature is obtained for all the Al-modified catalysts, indicating a weaker interaction of the Al-O-Mo linkage. As shown in Fig. 9, the peak locations at a low temperature increased in the following order: NiMo/Al-MCFs(20) < NiMo/Al-MCFs(10) < NiMo/Al-MCFs(30) < NiMo/Al-MCFs(40) < NiMo/Al-MCFs(50) < NiMo/MCFs, which indicated that the NiMo/Al-MCFs(20) catalyst possessed the best redox properties. As a result, the Ni-Mo-S-Ⅱ active sites are easy to form, thus improving the exposure of the edge and rim-active sites, which enhance the HDS efficiency.

The acidity of the NiMo/Al-MCFs(x) catalysts was studied using temperature-programmed desorption of ammonia (NH₃-TPD). The strength of the acid sites can be determined by the temperature at which the adsorbed NH₃ desorbs. The desorption peaks at lower temperature ( < 150 ℃) indicate weak acid sites and the peaks at a higher temperature ( > 400 ℃) indicate strong acid sites. The desorption peak at 150-400 ℃ represents the medium acid sites [49]. From Fig. 10, all the catalysts showed weak and medium strength acid sites. The peak in the NH₃-TPD profiles was attributed to the Brӧnsted acid sites with moderate acidity. As expected, the acidity increased as the Si/Al molar ratio decreased. There was a small signal obtained at a Si/Al ratio of 50, indicating weak acidity of NiMo/Al-MCFs(10) catalyst. The acidity was the highest at a Si/Al molar ratio of 20. The increase of partial medium Brӧnsted acid sites was generated by the isomorphous substitution of Si⁴⁺ions by Al³⁺ cations leading to the formation of bridging Si-OH-Al groups [49]. This was consistent with the Py-FTIR result.

The acidity properties of the NiMo/Al-MCFs(x) catalysts were investigated by the Py-FTIR method and the results are summarized in Fig. 11 and Table 2. The total amount of acid sites was determined by Py-FTIR spectroscopy by degassing at 200 ℃ (Fig. 11(a)) and the amounts of medium and strong acid sites were determined Py-FTIR spectroscopy by degassing at 350 ℃ (Fig. 11(b)). The bands at 1450 and 1622 cm^-1 were attributed to pyridine adsorbed onto Lewis acid sites, and the bands at 1540 and 1648 cm^-1 were attributed to pyridine adsorbed onto Brӧnsted acid sites, whereas the band at 1490 cm^-1 was attributed to a combination of Brӧnsted and Lewis acid sites [50-52]. The acid strength distribution and acid quantity of the catalysts are listed in Table 2. A weak acidity has a negative effect on the dispersion of active metals, resulting in the aggregation of active metals. The incorporation of Al atoms gives rise to a small amount of moderate-strong Brӧnsted acid sites in the form of Si-O(H)-Al bridges [49]. The intensity of the acidity increases significantly with increasing Al doping amount. The acidity is the strongest when the Si/Al molar ratio is 20. Furthermore, it has the suitable acid species and strength. The acidity does not increase when the Si/Al ratio is 10, which is probably because the substitution of Si⁴⁺ ions by Al³⁺ ions reaches saturation [53].

The moderate acid strength and an appropriate acid-type distribution of NiMo/Al-MCFs(20) are expected to be responsible for enhancing the synergistic effect between B and L acids, which means that Brӧnsted acids can be enhanced by Lewis acids nearby, thus increasing their catalytic activity.

Based on the above characterization results, the formation mechanism of Al-MCFs material was proposed, as shown in Scheme 1. The mesostructured cellular silica foams were formed by a microemulsion droplet templating approach. P123 (EO₂₀-PO₇₀-EO₂₀) is a potential amphiphilic surfactants that is commonly used as a template to synthesize mesoporous materials. Furthermore, it is believed that P123 can drive the formation of micelles in which the hydrophobic poly (propylene oxide) (PPO) chains can act as the core and the hydrophilic poly (ethylene oxide) (PEO) chains will be the corona [54, 55]. When it was added to the acidic solution, the surfactant P123 formed large complex micelles. Then, TMB and TEOS were slowly dropped into the above solution. The mixture was stirred and microemulsion droplets of TMB/P123 were formed, in which the TMB acted as a swelling agent and penetrated and expanded the core of the surfactant micelles. Moreover, the EO chains of the P123 micelles attracted the cationic silica oligomer and catalyzed the silica polymerization to form silicate micelles. The micelles interacted with each other and formed mesostructured cellular silica foams through self-assembly in the oil-in-water microemulsion system [56]. After the incorporation of aluminum, Lewis acid sites were successfully formed through the replacement of Si⁴⁺ with Al³⁺, which was consistent with the pyridine-FTIR results. Furthermore, Si-O-Al bonds were also constructed through Al atoms substituting the hydrogen atoms in the silanol groups and silicon, demonstrating that most of the Al species were successfully incorporated into the pure silicon frameworks. The result is in agreement with the ²⁷Al MAS NMR spectra in Fig. 4.

To further study the sulfidation degree of the active metals, XPS analysis was performed and the Mo 3d XPS spectra and their deconvolution results of the sulfide NiMo/Al-MCFs(x) catalysts are shown in Fig. 12. The related parameters are summarized in Table 3. The standards for the fitting are as follows: the binding energies of Mo 3d_5/2 and Mo 3d_3/2 of Mo⁴⁺ (MoS₂) are 229.0 ± 0.1 and 232.1 ± 0.1 eV, respectively; those for Mo⁵⁺ (MoS_xO_y) are 230.1 ± 0.1 and 233.2 ± 0.1 eV, and those for Mo⁶⁺ (MoO₃) are 232.7 ± 0.1 and 235.8 ± 0.1 eV [57]; the binding energy of S 2s is 225.9 ± 0.1 eV [58]. The sulfidation degree of the oxidic Mo species, which is defined as the atomic ratio of Mo⁴⁺ to the sum of Mo⁴⁺, Mo⁵⁺, and Mo⁶⁺, first increased then decreased as the Si/Al mass ratio increased (Table 3). It is apparent that Al affects the distribution of the Mo species. The change in the distribution of the Mo species is small if the amount of Al is relatively low (i.e., Si/Al = 50). Meanwhile, the Mo⁴⁺/Mo_total ratio significantly increased when the Al doping content was high (Si/Al = 20), whereas the ratio decreased when the Si/Al ratio reduced to 10.

HRTEM studies were performed to compare the dispersion degree of the sulfide metal species on the NiMo/Al-MCFs(x) catalysts [29]. The representative HRTEM micrographs and the statistical results of the sulfide NiMo/Al-MCFs(x) catalysts are exhibited in Fig. 13. The stacking numbers and slab length were obtained through statistical analyses based on 250-300 MoS₂ slabs.

In Fig. 13(b), the MoS₂ species of the sulfide NiMo/Al-MCFs(20) catalyst is less stacked, with the highest number of stacking layers of approximately 2-4. This may be caused by the strong acidity of the catalyst, which favors the dispersion of the active metals and can increases the MSI; thus, the ratio of high stacking layers and lengths are lower [59]. The results of the average length and number of layers of the MoS₂ slabs are displayed in Table 4. The average length of the MoS₂ slabs on the five catalysts follow the order: NiMo/Al-MCFs(20) (5.5 nm) < NiMo/Al-MCFs(10) (5.7 nm) < NiMo/Al-MCFs(30) (6.1 nm) < NiMo/Al-MCFs(40) (6.6 nm) < NiMo/Al-MCFs(50) (6.8 nm). The average number of MoS₂ slabs on the five catalysts follows the order: NiMo/Al-MCFs(20) (3.2) < NiMo/Al-MCFs(10) (3.4) < NiMo/Al-MCFs(30) (4.0) < NiMo/Al-MCFs(40) (4.5) < NiMo/Al-MCFs(50) (4.6), indicating that the NiMo/Al-MCFs(30), NiMo/Al-MCFs(40), and NiMo/Al-MCFs(50) catalysts have larger values of average length and stacking number. The high values of average length and stacking will decrease the exposure ratios of edge and rim active sites [45]. These results are consistent with the UV-Vis and Py-FTIR results.

In the present study, the DBT HDS catalytic activities of the sulfide NiMo/Al-MCFs(x) catalysts with different Si/Al ratios were evaluated in a microreactor plant and the commercial NiMo/γ-Al₂O₃ catalyst was used as a reference. As shown in Fig. 14, the activity order is NiMo/Al-MCFs(20) ≈ NiMo/Al-MCFs(10) > NiMo/γ-Al₂O₃ > NiMo/Al-MCFs(30) > NiMo/Al-MCFs(40) > NiMo/Al-MCFs(50). The DBT conversion over different catalysts decreased with decreasing WHSV value. At a low WHSV value of 20 h^-1, the HDS efficiency of the NiMo/Al-MCFs(20) catalyst was as high as 96%. The catalytic activity exhibited nonlinear tendency with increasing Al doping amount in the support. The catalytic activities increased when the Si/Al molar ratios decreased from 50 to 20, but there was little effect on the activity of the catalyst when the Al doping content increased further (Si/Al molar ratio of 10). The changes in the catalytic behaviors can be attributed to the following two aspects. First, the incorporation of aluminum into the MCFs material provides better dispersion to the deposited oxidic and sulfided Ni and Mo species and introduces a new acidic functionality into the catalysts. Furthermore, the synergistic effects of Brӧnsted and Lewis acid sites also contributed to the catalytic activities. This is consistent with the Raman and Py-FTIR results. Second, too strong MSI, as in the case of the aluminum-rich NiMo/Al-MCFs(10) catalyst, makes the reduction and sulfidation of Ni and Mo species more difficult, which is supported by the results of the H₂-TPR and XPS analyses. The strength of the MSI also explains the result shown in Fig. 14, which shows that the catalytic activities are the maximum at a Si/Al molar ratio of 20 and maintain the same level with further increase of the Al content (Si/Al molar ratio of 10). The catalytic activity of the NiMo/Al-MCFs catalysts is better than that of the commercial NiMo/γ-Al₂O₃ catalyst. Among all the studied catalysts, the DBT HDS catalytic activity of the NiMo/Al-MCFs(20) catalyst was the best, which is closely related to its moderate acidity, large pore size, high surface area, and high sulfidation degree of the active metals.

In the present work, the effects of different Si/Al ratios (10, 20, 30, 40, and 50) of Al-modified MCFs (Al-MCFs(x)) and the performance of the corresponding NiMo/Al-MCFs(x) catalysts in the DBT HDS were systematically studied. The typical physicochemical properties of all the studied samples were characterized by SAXS, Raman spectroscopy, N₂ adsorption-desorption, ²⁷Al MAS NMR, and UV-Vis methods. The results demonstrated that the MCFs retained their mesostructures after the incorporation of Al through a grafting method. The incorporation of aluminum into the framework of MCFs contributed to an increase in the acidity of the pure MCFs and was favorable for the improvement of the DBT HDS activity. In addition, the interaction of the Ni and Mo species with the support became stronger with increasing Al doping amount and reached maximum when the Si/Al ratio was 20. The activity of the NiMo/Al-MCFs(x) catalysts for the DBT HDS followed the order: NiMo/Al-MCFs(20) > NiMo/Al-MCFs(10) > NiMo/Al-MCFs(30) > NiMo/Al-MCFs(40) > NiMo/Al-MCFs(50), which was in good agreement with the acidities and redox properties of the catalysts. Among the catalysts, the NiMo/Al-MCFs(20) catalyst exhibits the highest HDS efficiency (96% at a high WHSV value of 20 h^-1), which was attributed to the good dispersion of the Mo species and moderate acidity, especially the synergistic effects of the Brӧnsted and Lewis acids.