In situ synthesis of the MoS₂–GO catalyst and unveiling its potential for deep hydrogenation desulfurization

Abstract

The MoS₂ catalyst shows great potential in deep hydrodesulfurization (HDS) but is limited by high metal usage and low active site utilization. A MoS₂–GO composite catalyst with trace amounts of graphene oxide (GO) was synthesized via an in situ solvothermal method. Owing to its high polarity, deionized water acts as an effective dispersant for GO, ensuring uniform dispersion while preserving its sheet-like morphology. The Mo precursor, bearing organic functional groups, is homogeneously anchored onto the oxygen functionalities of GO sheets, resulting in a densely packed monolayer MoS₂ structure with abundant, highly exposed HDS edge sites across the layered GO surface. Combined X-ray photoelectron spectroscopy (XPS) and density functional theory (DFT) analyses reveal that MoS₂ forms a heterostructure with GO through interactions between S atoms and the surface oxygen functionalities of GO. In the HDS reaction, it achieves 98.3% dibenzothiophene (DBT) conversion at 280 °C and exhibits high hydrogenation desulfurization (HYD) selectivity (S_(HYD/DDS) up to 12.8). Notably, it demonstrates excellent activity for sterically hindered 4,6-dimethyldibenzothiophene (4,6-DMDBT, 80.7% conversion at 300 °C) and a high HYD pathway selectivity (S_(HYD/DDS) up to 13.9). Raman spectroscopy coupled with DFT calculations reveals that the MoS₂–GO catalyst features extensive Mo–S–O(GO) electron-transport networks, which facilitate H₂ dissociation and drive continuous hydrodesulfurization of sulfur-containing species. This study provides insights into the preparation of heavy oil hydrocracking catalysts and the regulation of hydrogenation pathway selectivity.

---

1. Introduction

Energy is the lifeblood that drives societal development. Despite the global consensus on transitioning to low-carbon energy to effectively address climate change and environmental degradation, fossil fuels derivates (such as gasoline, diesel, and jet fuel) still account for the largest share of global energy consumption.^1,2 With the increasing proportion of heavy crude oil each year, the traditional refining industry faces even greater challenges in developing clean fuels and achieving their green, efficient utilization.³ HDS, as a critical process in refining to obtain clean fuels, plays a vital role in environmental protection.^4,5 HYD and direct desulfurization (DDS) are the two pathways in HDS; higher hydrogenation selectivity (HYD/DDS) can not only reduce energy consumption and coking during crude oil hydrocracking but also enhance the hydrogenation selectivity of diesel.^5–7

Non-supported catalysts, owing to their high active component content, are considered to have enormous potential for hydrogenation desulfurization.^6,8 In particular, non-supported MoS₂ catalysts exhibit superior HYD/DDS selectivity, distinguishing them from the other deep hydrogenation desulfurization catalysts and aligning with emerging hydrogenation requirements.^9–11 Based on the “rim-edge” model proposed by Daage and Chianelli,¹² it has been suggested that the edge sites of MoS₂ stacks—specifically those on the top and bottom layers—participate in both the HYD and DDS pathways, while the edge sites in the intermediate layers predominantly favor the DDS pathway. The active sites for HDS, also referred to as S vacancies, may be located at either the Mo edge or the S edge of molybdenum-based catalysts.^13–15 Strategies for optimizing the performance of unsupported MoS₂ catalysts involve not only tailoring the relative proportions of different active sites but also engineering structures with highly exposed active sites to maximize the utilization efficiency of active metals.

The low active metal utilization is a critical factor limiting the performance enhancement and large-scale industrialization of MoS₂ catalysts, and it has long been a primary focus of research.¹⁶ Currently, researchers often employ strategies such as constructing porous or layered structures to achieve high exposure of active metals. This involves adding templates or structure-directing agents during synthesis. For example, some researchers have used mesoporous carbon,¹⁷ carbon nitride,¹⁸ and silica¹⁹ as templates to synthesize porous Mo-based catalysts via hydrothermal methods, significantly enhancing catalyst activity. Wang et al. selected nickel–aluminum layered double hydroxide (NiAl-LDH) as a structure-directing template to synthesize a multimetallic NiAlMoW catalyst.²⁰ This catalyst exhibits intrinsic catalytic activity twice that of commercial CoNiMoW/Al₂O₃ catalysts. Although methods to optimize MoS₂ catalyst structures can enhance the atomic utilization of non-supported molybdenum-based catalysts to some extent, they are not easy to develop. The use of templates necessitates their removal—often with strong acids or bases—which may lead to pore collapse during subsequent sulfidation or calcination, resulting in poor catalyst stability.^17–20 Secondly, although porous or layered structures can fully expose active sites, the selectivity of desulfurization pathways has not been effectively regulated.^19,20 Therefore, further investigation into microscale structures remains a key research focus.

Building on our previous research, single-layer MoS₂ nanosheets have been successfully synthesized using organic solvents, enhancing the utilization of active sites and mass transfer efficiency.²¹ Further optimization of the desulfurization activity and pathway selectivity of unsupported MoS₂ catalysts is required. Under solvothermal conditions, non-metallic additives can be introduced in situ to further modulate the microstructure and morphology of non-supported MoS₂, thereby increasing the abundance of MoS₂ edge sites and enhancing metal utilization. Evidence from previous studies indicates that oxygen-containing groups on GO readily anchor and couple with Mo precursors bearing organic groups in organic media, thereby ensuring high dispersion of Mo precursors.^22–24 However, for MoS₂–GO composites, the key challenge is that this two-dimensional material tends to aggregate or tightly stack during the composite process due to van der Waals forces or strong π–π interactions.^25,26 This leads to reduced dispersion, surface area, and GO defect sites, thereby limiting the improvement of catalytic activity in the composite material. Therefore, achieving the dispersion of GO in the synthesis solution and the anchoring and uniform dispersion of MoS₂ on GO sheets has emerged as a key research focus.

In this study, MoS₂–GO composite sheet cluster catalysts were synthesized by the in situ solvothermal method. GO readily exfoliates into single sheets in highly polar deionized water. Its abundant surface oxygen functionalities anchor Mo-based organic precursors, enabling this method to successfully achieve the effective dispersion and oriented alignment of monolayer MoS₂ on GO sheets. In the HDS reaction, it achieves 98.3% dibenzothiophene (DBT) conversion at 280 °C and exhibits high HYD desulfurization selectivity (S_(HYD/DDS) up to 12.8). The HYD pathway exhibits selectivity 7 times higher than that of solvothermally synthesized MoS₂ catalysts and 6 times higher than that of the ethanol-dispersed MoGO-E catalyst. Notably, it demonstrates excellent activity for sterically hindered 4,6-dimethyldibenzothiophene (4,6-DMDBT, 80.7% conversion at 300 °C) and a high HYD pathway selectivity (S(HYD/DDS) up to 13.9). The integration of Mo–S–O(GO) electron-transfer pathways with GO defect sites constructs a robust interfacial network that promotes H₂ activation and selectively enhances the HYD route.

This work is a classic case of engineering precision control of hydrogenation desulfurization MoS₂ catalysts to achieve high activity, high selectivity, and high stability catalyst utilization.

2. Experimental

2.1 Materials

Thiourea (AR, 99%), cyclohexylamine (AR, 99%), absolute alcohol (AR, 99%), ammonia solution (AR, 25%–48%), ammonium molybdate tetrahydrate (AR, 99%), and ammonium sulfide solution (AR) were purchased from Sinopharm Chemical Reagent Co., Ltd, China. GO dispersion (0.94 wt%) was purchased from Changzhou Sixth Element Materials Technology Co., Ltd. Deionized water was self-made.

2.2 Synthesis of catalysts

Solvothermal synthesis of MoS₂–GO composite catalysts. At room temperature, two portions of 2.6 g ammonium tetrathiomolybdate (ATTM) (0.1 mol) were weighed and each dissolved in 40 mL of cyclohexylamine.²¹ Concurrently, two portions of 2.0 g of a GO dispersion were prepared by initially dispersing them in 2 mL of deionized water (W) and anhydrous ethanol (E), respectively. These dispersions were then added to the ATTM-cyclohexylamine solutions to obtain two mixtures of the organic molybdenum precursor and GO. Next, 0.76 g of thiourea was added to each mixture and stirred until uniform. The resulting solutions were transferred into 50 mL stainless steel autoclaves lined with Teflon, sealed, and placed in a preheated oven at 200 °C for 10 hours. After cooling to room temperature, the black products were centrifugally washed six times with deionized water and anhydrous ethanol, then dried in a vacuum oven at 60 °C for 12 hours. This process yielded two GO-doped solvothermal MoS₂ catalysts, namely MoGO-W (prepared using deionized water dispersion) and MoGO-E (prepared using anhydrous ethanol dispersion).

Using an identical synthesis procedure and the same molybdenum precursor dosage, MoS₂–GO composite catalysts were prepared with GO doping amounts of 1 g, 1.5 g, and 2.5 g, respectively; in each case, the GO (0.94 wt%) was dispersed in 2 mL of deionized water. The catalysts were designated as MoGO1, MoGO1.5, MoGO2 (MoGO-W) and MoGO2.5 according to the GO doping level.

Solvothermal synthesis of MoS₂ catalyst. According to our previous reports,²¹ ATTM was dissolved in cyclohexylamine at room temperature, with thiourea added under stirring to form a uniform red solution. This mixture was then transferred to a Teflon-lined autoclave and heated at 180 °C for 10 h. After cooling, the black precipitate was centrifuged, washed with deionized water and anhydrous ethanol, and dried at 60 °C in a vacuum oven for 12 h to yield Solvo-MoS₂.

2.3 Characterization of catalysts

X-ray diffraction (XRD) patterns were collected on a Panalytical X'Pert PRO MPD diffractometer using Cu Kα radiation (λ = 1.54056 Å, 40 kV, 40 mA). Morphology, particle size and elemental mapping were examined by scanning electron microscopy (SEM, JEOL JSM-7900F) and transmission electron microscopy (TEM, JEOL JEM-2100F). Textural properties (specific surface area, pore volume and pore size) were determined from N₂ adsorption–desorption isotherms measured on a Quantachrome Autosorb IQ. Thermogravimetric–mass spectrometry (TG–MS) was performed on a Netzsch STA449F5 under N₂ (50–900 °C, 2 °C min⁻¹). Fourier-transform infrared (FT-IR) spectra were recorded on a Thermo Nicolet NEXUS spectrometer. Raman spectra were acquired on a WITec alpha300 R confocal Raman microscope (633 nm excitation, ∼20 mW). X-ray photoelectron spectroscopy (XPS) for elemental composition and oxidation states was carried out on a Kratos AXIS Ultra DLD using Al Kα radiation (1486.6 eV).

The average slab length and the average number of stacked layers were statistically calculated by HRTEM images according to the following formula.²¹

N _i is the stacking number of the layers, and n_i is the number of stacked units with a stacking number of N_i.²⁰

L _i is the length of the layer, and n_i represents the number of layers with a length L_i.

DFT calculations were performed with the Vienna Ab Initio Simulation Package (VASP) using spin-polarized GGA–PBE.²⁷ The PAW method and a plane-wave cutoff of 400 eV were employed. Geometry optimizations used a force convergence criterion of 0.03 eV Å⁻¹ and an SCF energy threshold of 1 × 10⁻⁴ Ha, with a 1 × 1 × 1 k-point mesh. Slab models comprised a three-layer GO sheet (derived from graphene (001)) and the MoS₂ (002) plane, each represented by p (3 × 3) supercells and separated by a 15 Å vacuum. Three models—MoS₂-1Layer, MoS₂–GO and MoS₂-2Layer—were constructed to evaluate the adsorption energies and Gibbs free energies.

The adsorption energy (E_ads) of hydrogen molecules on the surface is calculated using the following equation (eqn (3)):^15,27

E_ads = E_{(adsorbate+surface)} − E_surface − E_gas (3)

In this equation, E_surface represents the total energy of the surface without adsorption, E_gas is the energy of the free gas molecules in the vacuum, and E_{adsorbats+surface} is the total energy of the system after adsorption. A negative value of E_ads indicates that the process is exothermic, and adsorption can occur spontaneously. An example of this application is reported in SI.

2.4 HDS evaluation of catalysts

HDS tests were performed in a high-pressure fixed-bed microreactor using DBT and 4,6-DMDBT (1 wt% in n-heptane) over 2 mL of the catalyst (reactor volume 5 mL, quartz sand-packed). Reactions were carried out at 240–320 °C (20 °C steps) with LHSV = 3.0 h⁻¹, P = 2.0 MPa, feed = 6 mL h⁻¹ and H₂/oil ≈ 300 (v/v). Catalysts were pretreated under N₂ at 120 °C (10 °C h⁻¹, 2 h), heated to 240 °C, switched to H₂, and fed; samples were taken after 4 h and every 2 h thereafter. Products were analyzed by GC (Agilent 7890N) using internal-standard normalization.

The main products of the DBT HDS reaction are biphenyl (BP), cyclohexylbenzene (CHB), and bicyclohexyl (BCH). Biphenyl is primarily produced via the DDS pathway, while BP and CHB are generated through the HYD pathway. The selectivity of the HYD pathway can be obtained using²¹

In this equation, C_BP, C_CHB, and C_BCH represent the concentrations of BP, CHB, and BCH in the reaction products, respectively.

To compare the activity per unit mass of the catalysts, the activation energy (E_a) of the HDS of DBT and 4,6-DMDBT over all the catalysts was calculated according to the Arrhenius formula. At each temperature, the apparent rate constant was first calculated from the experimental conversion based on the reaction order. Since the desulfurization reaction with sulfides (DBT or 4,6-DMDBT) can be approximated as a first-order process, the integrated rate equation:

X = 1 − e^−kt (5)

The formula:

Here, X denotes the conversion of DBT or 4,6-DMDBT in HDS, t is the reaction time, k refers to the rate constant of DBT or 4,6-DMDBT in HDS (mol g⁻¹ h⁻¹), T refers to the reaction temperature (K), R refers to the gas constant (8.314), and C refers to the constant.

3. Results and discussion

3.1 Catalyst structure characterization

The MoS₂–GO catalysts were synthesized by the in situ solvothermal method (Scheme 1). The MoGO-W catalyst, synthesized with water as the dispersant, preserves the sheet-like architecture of GO, with its surface densely covered by well-aligned MoS₂ nanosheets. In contrast, employing anhydrous ethanol as the dispersant yields the MoGO-E catalyst, which displays an aggregated morphology composed of spherical nanoparticles.

Scheme 1 Scheme of the synthesis of the MoS₂–GO catalysts.

Due to the incorporation of GO, all the MoS₂–GO catalysts exhibited a Mo loading of approximately 36 wt%, which is lower than the 47.9 wt% observed for Solvo-MoS₂. MoGO-W exhibits high dispersion and the lowest specific surface area (<13 m² g⁻¹, Fig. S1 and Table S1), indicating effective suppression of MoS₂ nanosheet stacking and the formation of additional intercrystalline mesopores.^27,28 The GO composite had no significant effect on the phase composition or crystallization of Solvo-MoS₂ (Fig. 1a). From the infrared spectra, it can be observed that the solvothermal series catalysts (MoGO and Solvo-MoS₂) exhibited saturated C–H bond stretching vibration peaks at 2858 cm⁻¹ and 2930 cm⁻¹, indicating the presence of organic compounds (Fig. 1b).^21,28 Thermogravimetric analysis further confirms that the weight loss primarily arises from the release of organic species (Fig. S2). After GO incorporation, the MoGO-W and MoGO-E catalysts showed vibration peaks at 1101 cm⁻¹, 1220 cm⁻¹, and 1618 cm⁻¹, which correspond to C–O–C, C–O, and C–OH bonds,^25,29,30 respectively, indicating the successful incorporation of GO components.

Fig. 1 (a) XRD, (b) IR and (c and d) Raman spectra of the MoGO-W, MoGO-E and Solvo-MoS₂ catalysts.

Raman spectroscopy was used to analyze the composition and structure of the MoS₂–GO catalysts. In the 200–500 cm⁻¹ range, all three solvothermal catalysts showed Raman peaks at 234 cm⁻¹, 280 cm⁻¹, and 333 cm⁻¹, corresponding to the characteristic modes of monolayer MoS₂ (1T/MoS₂) (Fig. 1c).^31,32 The J₂ and J₃ peaks of the MoGO-W and MoGO-E catalysts shifted to lower wavenumbers compared to solvothermal MoS₂, indicating GO incorporation.^31,33 The MoGO-W and MoGO-E catalysts also displayed GO-related Raman peaks, including the D band at 1348 cm⁻¹ (the vibrations of disordered, edge, and defect sites of sp³-hybridized carbon atoms in graphene) and the G band at 1595.9 cm⁻¹ (the vibrations of sp³ carbon atoms in an ideal graphene layer).^34,35 The I_D/I_G ratio was higher for MoGO-W (1.40) than for MoGO-E (1.13), indicating more defects (Fig. 1d). These results indicate the successful integration of monolayer MoS₂ with GO. Furthermore, under solvothermal and thiourea-sulfurization conditions, the removal of oxygen-containing groups from GO leads to the exposure of additional defect sites.^36,37

XPS analysis was performed to investigate the elemental valence states in the MoS₂ catalysts doped with GO. Upon analysis, GO had no significant effect on the binding energy of Mo⁴⁺ (Fig. 2a). The Mo 3d_3/2 and Mo 3d_5/2 peaks for all three samples appeared near 231.5 eV and 228.3 eV, respectively, with a slight increase in the sulfurization degree of Mo (Table S2). XPS S 2p spectra of the three catalysts show that the S 2p peak of Solvo-MoS₂ appears at 160.3 eV, whereas those of MoGO-W and MoGO-E shift to 161.0 and 161.4 eV, respectively, consistent with the defect sites in GO depleting the electron density around MoS₂-associated S atoms and thus increasing the S 2p binding energy (Fig. 2b). Concurrently, the C 1s feature assigned to C=C/C–C shifts from 284.6 eV in Solvo-MoS₂ to 284.1 eV in the MoGO samples (Fig. 2c). Together, these shifts indicate electronic interaction between MoS₂ and GO and support the formation of a MoS₂–GO composite.^38,39

Fig. 2 XPS spectrum of MoGO-W, MoGO-E and Solvo-MoS₂ (a. Mo 3d, b. S 2p and c. C 1s).

The MoGO-W catalyst retains the sheet-like morphology of graphene oxide, forming densely aligned MoS₂ nanosheets on the surface (Fig. 3a). In contrast, MoGO-E, dispersed in anhydrous ethanol, exhibits spherical-like nanoparticle morphology resembling that of Solvo-MoS₂ (Fig. 3b and c). The difference in the dispersion stability of GO in deionized water and ethanol primarily arises from the variation in hydrogen-bonding interactions between the solvent molecules and GO functional groups, the differences in the dielectric constants of water and ethanol, and the resulting changes in the charge state of GO sheets, which affect the interlayer electrostatic repulsion. The high polarity and strong hydrogen-bonding capacity of water enable GO to become highly hydrated and charged, thereby forming a stable dispersion. In contrast, the relatively hydrophobic nature and low dielectric constant of ethanol weaken the surface charge screening of GO, allowing van der Waals forces to dominate and leading to aggregation.^40–42 The sheet-like graphene oxide, rich in surface oxygen groups, offers abundant anchoring sites for Mo precursors, thereby driving the nucleation and growth of MoS₂ nanosheets on its surface.^43,44 This intimate interfacial interaction produces dense MoS₂ lattice fringes and yields MoGO-W with an interlayer spacing of only 0.86 nm—distinctly smaller than that of Solvo-MoS₂ (0.93 nm ) and much smaller than that of MoGO-E (1.67 nm) (Fig. 3d, e and Fig. S3(a)). Water promotes uniform GO dispersion and stabilizes its morphology by forming strong hydrogen bonds and enhancing electrostatic repulsion, enabling dense vertical anchoring of monolayer MoS₂ on GO sheets with abundant exposed edge sites, which in turn enhances catalytic activity.

Fig. 3 SEM pictures of (a) MoGO-W, (b) MoGO-E and (c) Solvo-MoS₂. TEM pictures of the (d1–d3) MoGO-W and (e1–e3) MoGO-E catalysts.

3.2 Optimization and formation mechanism of MoS₂–GO

GO enables the ordered arrangement of high-density monolayer MoS₂. The variation in GO content did not significantly affect the phase composition of the MoS₂–GO composite structure (Fig. S4). Increasing GO content weakens the 14.4° (002) peak (Fig. S4a), indicating reduced MoS₂ stacking and enhanced dispersion.^46,47 The sulfurization degree of Mo in the four MoGO catalysts were similar, slightly higher than that of the Solvo-MoS₂ (Fig. S4c, d and Table S2).

The microstructural morphology of the MoS₂–GO catalysts with varying GO contents was characterized. For MoGO1 and MoGO1.5, the samples are composed of micron-sized spheres and sheets, with spheres being predominant (Fig. 4a and b). Upon magnification, the spherical particles and sheet surfaces are clearly visible, exhibiting a fuzzy texture. These correspond to the solvothermally synthesized MoS₂ particles and MoS₂–GO composite sheet clusters. As the GO doping amount increased, MoGO2 and MoGO2.5 were completely composed of micron-sized sheet clusters with fuzzy surfaces (Fig. 4c and d). This indicated that the Mo precursor can be fully anchored in the solvothermal environment, forming an expanded sheet structure with a high density of MoS₂ distributed on the surface, by adjusting the GO content. However, excessive GO content in MoGO2.5 results in incomplete MoS₂ coverage on the sheet surface, likely leading to insufficient active edge sites and reduced catalytic activity (Fig. 4d). Therefore, the appropriate ratio of Mo precursor to GO is crucial for achieving a uniform composite sheet cluster morphology. In the MoGO1 and MoGO1.5 catalysts, the MoS₂ lattice fringes are mainly observed at the particle edges (Fig. S5). In contrast, in MoGO2 and MoGO2.5, the fringes are uniformly distributed across the GO surface, with a wider interlayer spacing observed. As the GO content increases, the MoS₂ sheet length also significantly increases, from 3–4 nm in MoGO1 to 6–8 nm in MoGO2.5 (Fig. 3d and S5c). The interlayer spacing of the MoS₂–GO catalysts was greater than the 0.62 nm of Solvo-MoS₂, indicating an expanded monolayer structure.

Fig. 4 SEM pictures of the MoGO series catalysts with different GO contents (a1–a3: MoGO1, b1–b3: MoGO1.5, c1–c3: MoGO2, and d1–d3: MoGO2.5 ).

All MoGO catalysts with varying GO contents exhibited Raman vibrational peaks characteristic of the 1T/MoS₂ structure, indicating that they were monolayer structures (Fig. 5a). By comparing the peak intensity ratios of the D band and G band for each catalyst, it was found that the MoGO2 catalyst had the highest I_D/I_G ratio (MoGO1: 1.13, MoGO1.5: 1.33, MoGO2: 1.40, MoGO2.5: 1.32), reflecting its abundance of defect sites (Fig. 5b).³⁷ Combining this with HRTEM characterization, it can be seen that as the GO content increases, the MoS₂ sheet length in the MoS₂–GO composite catalysts gradually increases, and the density of GO defect sites in the sheets improves. However, excessive aggregation of the sheets and poor expansiveness lead to a reduction in the number of GO defect sites.

Fig. 5 Raman spectra of MoGO series catalysts with different GO contents (a: 100–6000 cm⁻¹, b: 1000–3000 cm⁻¹).

3.3 Performance of catalysts

The HDS performance of the MoS₂-based catalysts was systematically assessed. Under low-temperature conditions (≤280 °C), MoGO-W outperformed MoGO-E and Solvo-MoS₂ in DBT conversion (Fig. 6a). At 280 °C, MoGO-W attained a DBT conversion of 98.3% and a HYD/DDS ratio of 12.8, markedly surpassing MoGO-E (92.8%, 2.1) and Solvo-MoS2 (86.5%, 2.2). Increasing the temperature to 300 and 320 °C, the catalysts show similar conversion due to the higher temperature, but a decrease in S_(HYD/DDS). This superior HYD selectivity is ascribed to the oriented assembly of MoS₂ on flake-like GO, exposing additional HYD-active edge sites. The catalysts were subsequently tested for HDS of sterically hindered 4,6-dimethyldibenzothiophene (4,6-DMDBT). At 300 °C, MoGO-W exhibited the highest conversion (80.7%) and HYD/DDS ratio (13.9), outperforming Solvo-MoS₂ (51.1%, 8.3) and MoGO-E (58.9%, 10.8) (Fig. 6b). Compared with previously reported catalysts, MoS₂-GO demonstrates a pronounced advantage in selectivity for HDS (Table S3). Kinetic analysis (Fig. S6) identified MoGO-W as exhibiting the lowest apparent activation energies—59.9 kJ mol⁻¹ for DBT and 60.2 kJ mol⁻¹ for 4,6-DMDBT—underscoring its superior intrinsic activity. These findings confirm that MoGO-W possesses a high density of monolayer MoS₂ active sites within the conductive graphene oxide network, and that the highly dispersed MoS₂ nanosheets facilitate π-adsorption of bulky thiophene substrates at the edge sites.^12,13,45

Fig. 6 HDS conversion and S_(HYD/DDS) of DBT (a and c) and 4,6-DMDBT (b and d) catalyzed by the MoS₂-based catalysts.

We systematically evaluated the DBT HDS performance of MoGO catalysts bearing four different graphene oxide loadings. Under identical reaction conditions, DBT conversion and HYD selectivity exhibited a volcano-type trend as the GO loading increased. MoGO2 delivered the highest desulfurization activity (Fig. 6c). At temperatures below 280 °C, all four catalysts exhibited comparable HDS activity for 4,6-DMDBT. However, at higher temperatures, MoGO2's advantage became evident, achieving a 79.3% conversion at 300 °C (Fig. 6d). MoGO2 also exhibited superior HYD selectivity, with a HYD/DDS ratio of 12.8 for DBT HDS at 280 °C—significantly higher than those of the other catalysts (<10, Fig. 6c)—and a peak ratio of 14.8 for 4,6-DMDBT HDS at 320 °C (Fig. 6d). Kinetic analysis (Fig. S7) revealed that MoGO2 exhibited the lowest apparent activation energies—59.9 kJ mol⁻¹ for DBT and 60.2 kJ mol⁻¹ for 4,6-DMDBT—among the four catalysts, suggesting a larger Arrhenius pre-exponential factor and a higher density of active sites.⁴⁸ These findings indicate that highly exfoliated graphene oxide sheets ensure uniform dispersion of MoS₂, significantly enhancing active-site utilization and underpinning the exceptional HDS activity of MoGO2. Moreover, MoGO2 maintained high stability over 80 h of continuous operation, with DBT and 4,6-DMDBT conversions sustained at ∼90% and ∼79%, respectively, underscoring its exceptional catalytic durability and practical HDS potential (Fig. S8). In particular, the main product of the HDS of DBT is CBH, followed by BP.

3.4 Study on microelectronic environment and reaction mechanism

Differential charge density mapping (Fig. S9) reveals pronounced electronic coupling between GO and MoS₂ driven by electron donation from GO oxygens, which depletes electron density on surface S atoms and enriches catalytically active electrons. Comparison across MoS₂-2Layer and MoS₂–GO interfaces (Fig. S9d–f and h) shows that GO modification induces heterogeneous charge redistribution—unlike the uniform Mo–to–S interlayer transfer in multilayer MoS₂—thereby strengthening Mo–S bond polarization and boosting catalytic activity.

The density of states (DOS) analysis for MoS₂-1Layer, MoS₂-2Layer, and MoS₂–GO structures reveals that all three interfaces exhibit continuous energy band distributions near the Fermi level (EF), indicating good electron transfer capability (Fig. 7). Further examination of the projected density of states (PDOS) for Mo shows that the d-band center follows the order: MoS₂-2Layer (−0.994 eV) < MoS₂-1Layer (−0.799 eV) < MoS₂–GO (−0.348 eV) (Fig. 7d). In the MoS₂-2Layer structure, the strong interlayer interaction between the MoS₂ layers causes electron transfer toward S atoms, shifting the d-band center further away from the Fermi level. In contrast, for the MoS₂–GO structure, the d-band center is much closer to the Fermi level, indicating that the incorporation of GO enhances the electronic structure of Mo. Compared to MoS₂-1Layer (4.940), the Bader charge of MoS₂-2Layer (4.919) decreases, while that of MoS₂–GO (4.965) increases, which is consistent with the d-band distribution results of PDOS (Table S4). This indicated that in the MoS₂–GO composite structure obtained via the solvothermal method, GO influenced the electron density of monolayer MoS₂, facilitating catalytic electron transfer.

Fig. 7 Density of states of (a) MoS₂-1Layer, (b) MoS₂-2Layer, and (c) MoS₂–GO, and the projected DOS (d) of the above three structures.

DFT calculations were conducted to investigate the adsorption and dissociation of H₂ on three MoS₂-based surfaces (Fig. 8a). The results showed that the adsorption energies of H₂ on both MoS₂-2Layer and MoS₂–GO surfaces are similar, indicating chemisorption. Additionally, the dissociation energy of H₂ was lower on both the MoS₂-1Layer and MoS₂–GO surfaces.

Fig. 8 (a) Gibbs free energy for H₂ adsorption and dissociation on different metal surfaces and the energy barriers for HYD paths of DBT HDS catalyzed by (b) MoS₂–GO, (c) MoS₂-1Layer, and (d) MoS₂-2layer.

Given that the MoS₂–GO catalysts exhibit higher selectivity for the HYD pathway, simulation calculations were performed to determine the energy barriers for DBT desulfurization via the hydrogenation pathway on the three MoS₂ catalysts (Fig. 8b–d). In this pathway, DBT was initially hydrogenated to form tetrahydrodibenzothiophene (THDBT) (T1) and hexahydrodibenzothiophene (HHDBT) (T2). This was followed by desulfurization to yield CHB (T3), and further hydrogenation produces BCH (T4). The energy barriers for this reaction pathway were as follows: MoS₂–GO (8.90 eV) < MoS₂-1Layer (9.37 eV) < MoS₂-2Layer (9.51 eV), confirming that MoS₂–GO demonstrates the best catalytic performance for desulfurization through the HYD pathway.

4. Conclusions

A simple solvothermal one-pot method was used to synthesize MoS₂ catalysts doped with trace amounts of GO. When dispersed in deionized water, GO readily forms sheet-like structures that uniformly anchor Mo precursors bearing organic groups, resulting in composite sheet clusters characterized by high density, expanded interlayer spacing, and abundant GO defect sites. The optimal amount of GO in the catalysts is close to 2. Catalytic performance evaluations indicate that these MoS₂–GO catalysts exhibit superior desulfurization activity and enhanced selectivity for the hydrogenation pathway in DBT HDS reactions, as well as excellent catalytic activity toward sterically hindered compounds, demonstrating significant potential for deep desulfurization. The plentiful GO defect sites provide an efficient electron transport network, while the high-density rim sites of the MoS₂ monolayers create an active network. Together with the highly expandable sheet structure and uniform MoS₂ arrangement, these features enhance the utilization of active sites, facilitating the flat adsorption of sulfur-containing compounds and their subsequent desulfurization via the hydrogenation pathway. The study is of great importance for the development of catalysts for ultra-deep desulfurization and provides a theoretical foundation for future research.

Author contributions

Xianglong Meng and Kun Sun carried out the design and operation of the experiment, collation of the data, data processing, and completion of the initial manuscript. Xuyu Zhao and Xianglong Meng conducted the literature review, data supplementation, and theoretical research. Hailing Guo, Girolamo Giordano, Yongming Chai and Chenguang Liu provided valuable guidance and writing assistance. Hailing Guo contributed to writing reviews, content editing, experimental supervision, and financial support. All authors discussed the results and provided comments on the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI), and comprise, but are not limited to, XRD, TEM, SEM, and N₂ adsorption–desorption characterization results. See DOI: https://doi.org/10.1039/d5qi01825k.

Acknowledgements

This work was supported by the National Key Research and Development Program of China of the Ministry of Science and Technology (2024YFE0206800), National Natural Science Foundation of China (22175200 and 22475237), Shandong Province: “Double-Hundred Talent Plan” on 100 Foreign Experts and 100 Foreign Expert Teams Introduction (WSR2023056), and Fujian Province-Science and Technology Program, Innovation Fund (202411002).

References

1. J. A. Melero, J. Iglesias and A. Garcia, Biomass as renewable feedstock in standard refinery units. Feasibility, opportunities and challenges, Energy Environ. Sci., 2012, 5, 7393–7420.
2. E. Wang, F. Yang, M. Song, G. Chen, Q. Zhang, F. Wang, L. Bing, G. Wang and D. Han, Recent advances in the unsupported catalysts for the hydrodesulfurization of fuel, Fuel Process. Technol., 2022, 235, 107386.
3. H. D. Velázquez, R. C. Camacho, M. L. M. Mondragón, J. G. H. Cortez, J. A. M. Fuente, M. L. H. Pichardo, T. A. B. Oviedo and R. M. Palou, Recent progress on catalyst technologies for high quality gasoline production, Catal. Rev., 2023, 65, 1079–1299.
4. A. Tanimu and K. Alhooshani, Advanced hydrodesulfurization catalysts: A review of design and synthesis, Energy Fuels, 2019, 33, 2810–2838.
5. A. S. Walton, J. V. Lauritsen, H. Topsøe and F. Besenbacher, MoS₂ nanoparticle morphologies in hydrodesulfurization catalysis studied by scanning tunneling microscopy, J. Catal., 2013, 308, 306–318.
6. R. Candia, B. S. Clausen and H. Topsøe, The origin of catalytic synergy in unsupported CoMo HDS catalysts, J. Catal., 1982, 77, 564–566.
7. H. Liu, C. Yin, H. Li, B. Liu, X. Li, Y. Chai, Y. Li and C. Liu, Synthesis, characterization and hydrodesulfurization properties of nickel–copper–molybdenum catalysts for the production of ultra-low sulfur diesel, Fuel, 2014, 129, 138–146.
8. P. Afanasiev, The influence of reducing and sulfiding conditions on the properties of unsupported MoS₂-based catalysts, J. Catal., 2010, 209, 269–280.
9. A. Olivas, D. H. Galván, G. Alonso and S. Fuentes, Trimetallic NiMoW unsupported catalysts for HDS, Appl. Catal., A, 2009, 352, 10–16.
10. G. Alonso, G. Berhault, A. Aguilar, V. Collins, C. Ornelas, S. Fuentes and R. R. Chianelli, Characterization and HDS activity of mesoporous MoS₂ catalysts prepared by in situ activation of tetraalkylammonium thiomolybdates, J. Catal., 2002, 208, 359–369.
11. A. N. Varakin, A. V. Mozhaev, A. A. Pimerzin and P. A. Nikulshin, Comparable investigation of unsupported MoS₂ hydrodesulfurization catalysts prepared by different techniques: Advantages of support leaching method, Appl. Catal., B, 2018, 238, 498–508.
12. M. Daage and R. R. Chianelli, Structure-function relations in molybdenum sulfide catalysts: The “Rim-Edge” model, J. Catal., 1994, 149, 414–427.
13. F. Besenbacher, M. Brorson, B. S. Clausen, S. Helveg, B. Hinnemann, J. Kibsgaard, J. V. Lauritsen, P. G. Moses, J. K. Nørskov and H. Topsøe, Recent STM, DFT and HAADF-STEM studies of sulfide-based hydrotreating catalysts: Insight into mechanistic, structural and particle size effects, Catal. Today, 2008, 130, 86–96.
14. P. G. Moses, B. Hinnemann, H. Topsøe and J. K. Nørskov, The hydrogenation and direct desulfurization reaction pathway in thiophene hydrodesulfurization over MoS₂ catalysts at realistic conditions: A density functional study, J. Catal., 2007, 248, 188–203.
15. M. Šarić, J. Rossmeisl and P. G. Moses, Modeling the adsorption of sulfur containing molecules and their hydrodesulfurization intermediates on the Co-promoted MoS₂ catalyst by DFT, J. Catal., 2018, 358, 131–140.
16. V. Hetier, D. Pena, A. Carvalho, L. Courthéoux, V. Flaud, E. Girard, D. Uzio, S. Brunet, P. L. Desmazes and A. Pradel, Influence of pluronic® P123 addition in the synthesis of bulk Ni promoted MoS₂ catalyst. Application to the selective hydrodesulfurization of sulfur model molecules representative of FCC gasoline, Catalysts, 2019, 9, 793.
17. G. Chen, W. Xie, Q. Li, W. Wang, L. Bing, F. Wang, G. Wang, C. Fan, S. Liu and D. Han, Three-dimensionally ordered macro–mesoporous CoMo bulk catalysts with superior performance in hydrodesulfurization of thiophene, RSC Adv., 2020, 10, 37280–37286.
18. D. Ryaboshapka and P. Afanasiev, Carbon nitride used as a reactive template to prepare mesoporous molybdenum sulfide and nitride, RSC Adv., 2021, 11, 21678–21684.
19. S. E. Skrabalak and K. S. Suslick, Porous MoS₂ synthesized by ultrasonic spray pyrolysis, J. Am. Chem. Soc., 2005, 127, 9990–9991.
20. L. Wang, Y. Zhang, Y. Zhang, P. Liu, H. Han, M. Yang, Z. Jiang and C. Li, Hydrodesulfurization of 4,6-DMDBT on a multi-metallic sulfide catalyst with layered structure, Appl. Catal., A, 2011, 394, 18–24.
21. K. Sun, H. Guo, F. Jiao, Y. Chai, Y. Li, B. Liu, S. Mintova and C. Liu, Design of an intercalated nano-MoS₂ hydrophobic catalyst with high rim sites to improve the hydrogenation selectivity in hydrodesulfurization reaction, Appl. Catal., B, 2021, 286, 119907.
22. G. M. Hingangavkar, S. A. Kadam, Y. R. Ma, S. S. Bandgar, R. N. Mulik and V. B. Patil, MoS₂-GO hybrid sensor: A discerning approach for detecting harmful H₂S gas at room temperature, Chem. Eng. J., 2023, 472, 144789.
23. T. Huang, Y. Luo, W. Chen, J. Yao and X. Liu, Self-assembled MoS₂-GO framework as an efficient cocatalyst of CuInZnS for visible-light driven hydrogen evolution, ACS Sustainable Chem. Eng., 2018, 6, 4671–4679.
24. Z. Wang, J. Kou, M. Li, X. Zhang, Y. Hua, Q. Fang, M. Tian, M. Cao, Z. Shao and X. Wu, Enhancement and sustained uranium removal of 2D transition metal sulfide-graphene oxide composite/carbon cloth cathodes in capacitive deionization system, Desalination, 2025, 605, 118745.
25. P. Kuang, M. He, H. Zou, J. Yu and K. Fan, 0D/3D MoS₂-NiS₂/N-doped graphene foam composite for efficient overall water splitting, Appl. Catal., B, 2019, 254, 15–25.
26. S. Mao, G. Lu and J. Chen, Three-dimensional graphene-based composites for energy applications, Nanoscale, 2015, 7, 6924–6943.
27. C. Feng, Q. Gao, G. Xiong, Y. Chen, Y. Pan, Z. Fei, Y. Li, Y. Lu, C. Liu and Y. Liu, Defect engineering technique for the fabrication of LaCoO3 perovskite catalyst via urea treatment for total oxidation of propane, Appl. Catal., B, 2022, 304, 121005.
28. K. Sun, H. Guo, C. Feng, F. Tian, X. Zhao, C. Wang, Y. Chai, B. Liu, S. Mintova and C. Liu, One–pot solvothermal preparation of the porous NiS₂//MoS₂ composite catalyst with enhanced low-temperature hydrodesulfurization activity, J. Colloid Interface Sci., 2024, 659, 650–664.
29. Z. Wei, Z. Guan, F. Liu, Y. Xue, N. Shan, Y. Zhao, L. Fu, Z. Huang, J. Xu, M. G. Humphrey and C. Zhang, Covalently linked graphene oxide-transition metal disulfide quantum dots nanocomposites featuring enhanced nonlinear optical absorption, Mater. Today Phys., 2023, 38, 101261.
30. J. Xu, J. Liang, Y. Zou, F. Xu, Q. Chen, C. Xiang, J. Zhang and L. Sun, Layer-by-layer self-assembled GO-MoS₂Co₃O₄ three-dimensional conducting network for high-performance supercapacitors, J. Energy Storage, 2021, 43, 103195.
31. L. Wang, X. Zhang, Y. Xu, C. Li, W. Liu, S. Yi, K. Wang, X. Sun, Z. S. Wu and Y. Ma, Tetrabutylammonium-intercalated 1T-MoS₂ nanosheets with expanded interlayer spacing vertically coupled on 2D delaminated MXene for high-performance lithium-ion capacitors, Adv. Funct. Mater., 2021, 31, 2104286.
32. Y. Zhao, K. Chang, Q. Gu, B. Yang, J. Xu, Y. Zhang, C. Pan, Z. Wang, Y. Lou and Y. Zhu, Noble metal-free 2D 1T-MoS2 edge sites boosting selective hydrogenation of maleic anhydride, ACS Catal., 2022, 12, 8986–8994.
33. R. Naz, M. Imtiaz, Q. Liu, L. Yao, W. Abbas, T. Li, I. Zada, Y. Yuan, W. Chen and J. Gu, Highly defective 1T-MoS₂ nanosheets on 3D reduced graphene oxide networks for supercapacitors, Carbon, 2019, 152, 697–703.
34. Y. E. Shin, Y. J. Sa, S. Park, J. Lee, K. H. Shin, S. H. Joo and H. Ko, An ice-templated, pH-tunable self-assembly route to hierarchically porous graphene nanoscroll networks, Nanoscale, 2014, 6, 9734–9741.
35. W. Qin, L. Han, H. Bi, J. Jian, X. Wu and P. Gao, Hydrogen storage in a chemical bond stabilized Co₉S₈–graphene layered structure, Nanoscale, 2015, 7, 20180–20187.
36. M. T. Manzoor, J. E. Kim, J. H. Jung, C. Han, S. B. Choi and I. K. Oh, Two-dimensional rGO-MoS₂ hybrid additives for high-performance magnetorheological fluid, Sci. Rep., 2018, 8, 12672.
37. X. Liu, G. Liu, C. Yu, L. Xie and Y. Chen, A unique hierarchical composite with auricular-like MoS2 nanosheets erected on graphene for enhanced lithium storage, J. Solid State Electrochem., 2019, 23, 2759–2770.
38. Y. J. Tang, Y. Wang, X. L. Wang, S. L. Li, W. Huang, L. Z. Dong, C. H. Liu, Y. F. Li and Y. Q. Lan, Molybdenum disulfide/nitrogen-doped reduced graphene oxide nanocomposite with enlarged interlayer spacing for electrocatalytic hydrogen evolution, Adv. Energy Mater., 2016, 6, 1600116.
39. Y. Liu, X. Xu, J. Zhang, H. Zhang, W. Tian, X. Li, M. O. Tade, H. Sun and S. Wang, Flower-like MoS₂ on graphitic carbon nitride for enhanced photocatalytic and electrochemical hydrogen evolutions, Appl. Catal., B, 2018, 239, 334–344.
40. H. Ma, Z. Shen and S. Ben, Understanding the exfoliation and dispersion of MoS₂ nanosheets in pure water, J. Colloid Interface Sci., 2018, 517, 204–212.
41. H. Ma, Z. Shen and S. Ben, Surfactant-free exfoliation of multilayer molybdenum disulfide nanosheets in water, J. Colloid Interface Sci., 2019, 537, 28–33.
42. S. Karunakaran, S. Pandit, B. Basu and M. De, Simultaneous exfoliation and functionalization of 2H-MoS₂ by thiolated surfactants: Applications in enhanced antibacterial activity, J. Am. Chem. Soc., 2018, 140, 12634–12644.
43. W. Song, D. Wang, X. Yue, C. Jin, Y. Wu, Y. Shi, J. Liu, A. Wu, C. Tian and H. Fu, Cluster-like Mo₂N anchored on reduced graphene oxide as an efficient and high-performance catalyst for deep-degree oxidative desulfurization, Inorg. Chem. Front., 2025, 12, 1303–1314.
44. L. N. Long, P. T. Thi, P. T. Kien, P. T. Trung, M. Ohtani, Y. Kumabe, H. Tanaka, S. Ueda, H. Lee, P. B. Thang and T. V. Khai, Controllable synthesis of MoS₂/graphene low-dimensional nanocomposites and their electrical properties, Appl. Surf. Sci., 2020, 504, 144193.
45. F. Fan, Z. Chen, A. Zhou, Z. Yang, Y. Zhang, X. He, J. Kang and W. Zhou, Theoretical investigation on the inert pair effect of Ga on both the Ga-Ni-Mo-S nanocluster and the direct desulfurization of 4,6-dimethyldibenzothiophene, Fuel, 2023, 333, 126351.
46. H. Wu, A. Duan, Z. Zhao, D. Qi, J. Li, B. Liu, G. Jiang, J. Liu, Y. Wei and X. Zhang, Preparation of NiMo/KIT-6 hydrodesulfurization catalysts with tunable sulfidation and dispersion degrees of active phase by addition of citric acid as chelating agent, Fuel, 2014, 130, 203–210.
47. N. Li, Y. Chai, Y. Li, Z. Tang, B. Dong, Y. Liu and C. Liu, Ionic liquid assisted hydrothermal synthesis of hollow vesicle-like MoS₂ microspheres, Mater. Lett., 2012, 66, 236–238.
48. S. Li, X. Liu, J. Ma, F. Xu, Y. Lyu, S. Perathoner, G. Centi and Y. Liu, Develop high-performance Cu-based RWGS catalysts by controlling oxide–oxide interface, ACS Catal., 2025, 15, 3475–3486.

---

Footnotes

† This paper is dedicated to Professor Zifeng Yan (School of Chemistry and Chemical Engineering, China University of Petroleum — East China) on the occasion of his 60th birthday.

‡ Both of the authors contributed equally to this work.

---