Performance Exploration of Ni-Doped MoS2 in CO2 Hydrogenation to Methanol

The preparation of methanol chemicals through CO2 and H2 gas is a positive measure to achieve carbon neutrality. However, developing catalysts with high selectivity remains a challenge due to the irreversible side reaction of reverse water gas shift (RWGS), and the low-temperature characteristics of CO2 hydrogenation to methanol. In-plane sulfur vacancies of MoS2 can be the catalytic active sites for CH3OH formation, but the edge vacancies are more inclined to the occurrence of methane. Therefore, MoS2 and a series of MoS2/Nix and MoS2/Cox catalysts doped with different amounts are prepared by a hydrothermal method. A variety of microscopic characterizations indicate that Ni and Co doping can form NiS2 and CoS2, the existence of these substances can prevent CO2 and H2 from contacting the edge S vacancies of MoS2, and the selectivity of the main product is improved. DFT calculation illustrates that the larger range of orbital hybridization between Ni and MoS2 leads to CO2 activation and the active hydrogen is more prone to surface migration. Under optimized preparation conditions, MoS2/Ni0.2 exhibits relatively good methanol selectivity. Therefore, this strategy of improving methanol selectivity through metal doping has reference significance for the subsequent research and development of such catalysts.


Introduction
In the process of rapid human development, the excessive emission of carbon dioxide has led to a series of natural disasters such as global warming, continuous rise in sea levels, agricultural production reduction, and growing ocean acidification [1][2][3]. It is noteworthy that CO 2 can be described as a cheap and abundant C1 material if carbon dioxide is captured by a reasonable means and formed into chemical products with high industrial application value (methanol, aromatics, etc.) [4][5][6]. Among the abundant chemicals mentioned above, methanol is the preferred choice for the direct catalytic reaction of CO 2 gas to oxygenates because it can be considered an excellent precursor for some chemicals and a fuel substitute [7,8]. The efficient catalysis of CO 2 to CH 3 OH is exothermic, and low temperatures favor equilibrium shifts in the product direction. However, the activation of the C-O inert bond in CO 2 requires a high temperature [9]. Therefore, considering the thermodynamic and kinetic contradictions of this reaction, it is extremely important and urgent to prepare a reasonable catalyst for the conversion of CO 2 and hydrogen into methanol under low-temperature conditions.
The catalyst systems mainly include Cu-based metal catalysts [10][11][12], precious metal catalysts [13,14], and metal oxide catalysts [15][16][17]. As is well known, with the birth of the industrialized Cu/ZnO/Al 2 O 3 catalyst, copper-based catalysts are the most mature and deeply explored in the research field of CO 2 to methanol. However, the active phase is poisoned and inactivated due to the H 2 O generated by the competitive side reaction of

Exploration of Various Characterization Results
As depicted in Figure 1, the crystalline phase structure of pure MoS 2 and Ni-doped samples is confirmed by XRD patterns. The pure MoS 2 catalyst exhibits a stable hexagonal 2H-type crystal structure when the metal Ni is undoped, and the individual diffraction peaks corresponding to the (002) (100) (103) (110) crystal planes of PDF#89-1495 are visible. All illustrations are presented in the form of broad reflections of lower intensity, indicating the poor crystallinity of the coherent scattering in the nanometer range. Ni doping makes the (002) crystal plane of MoS 2 broaden obviously, and the diffraction peak related to Ni is not observed, indicating that the excellent dispersibility of Ni doping is conducive to the reduction of catalyst grain size. The reduction of layers number in the MoS 2 stack after metal doping can be explained by this important result [24]. In other words, the growth of MoS 2 grains can be effectively suppressed by metal doping. As the doping amount of Ni increases to 0.2 mmol, the diffraction peak of MoS 2 is enhanced, and a peak of NiS 2 appears, which is because of the rapid binding of Ni ions with S 2− that is produced by the decomposition of thiourea. With the further increase in Ni doping amount, the sharp diffraction peaks and smaller half widths of NiS 2 and MoS 2 indicate that their crystallinities become stronger and the grain sizes are increased. Different from Ni doping, because the electronegativity of the Co element is weaker than Ni in the same period, the chemical forces between Co and MoS 2 are relatively weak, resulting in a worse dispersibility of Co doping than Ni, so the (002) crystal plane of MoS 2 is higher than Ni doping ( Figure S1).
2H-type crystal structure when the metal Ni is undoped, and the individual diffraction peaks corresponding to the (002) (100) (103) (110) crystal planes of PDF#89-1495 are visible. All illustrations are presented in the form of broad reflections of lower intensity, indicating the poor crystallinity of the coherent scattering in the nanometer range. Ni doping makes the (002) crystal plane of MoS2 broaden obviously, and the diffraction peak related to Ni is not observed, indicating that the excellent dispersibility of Ni doping is conducive to the reduction of catalyst grain size. The reduction of layers number in the MoS2 stack after metal doping can be explained by this important result [24]. In other words, the growth of MoS2 grains can be effectively suppressed by metal doping. As the doping amount of Ni increases to 0.2 mmol, the diffraction peak of MoS2 is enhanced, and a peak of NiS2 appears, which is because of the rapid binding of Ni ions with S 2− that is produced by the decomposition of thiourea. With the further increase in Ni doping amount, the sharp diffraction peaks and smaller half widths of NiS2 and MoS2 indicate that their crystallinities become stronger and the grain sizes are increased. Different from Ni doping, because the electronegativity of the Co element is weaker than Ni in the same period, the chemical forces between Co and MoS2 are relatively weak, resulting in a worse dispersibility of Co doping than Ni, so the (002) crystal plane of MoS2 is higher than Ni doping ( Figure S1). To further demonstrate the apparent structures of pure MoS2 and the new catalysts formed by metal doping, scanning electronic microscopy (SEM) is analyzed. In MoS2 structure, six S atoms are tightly packed up and down to form a sixfold triangular prism, and the triangular prism voids are formed. Mo atoms are filled into these voids, resulting in a two-dimensional sheet. Monolayers of 2D flakes can stack through van der Waals forces and weak coordination bonds and form different types of interlayer voids. Therefore, as shown in Figure 2a,b, MoS2 is composed of abundant lamellar nanoflowers, and there are certain gaps between the layers, which is beneficial for deep adsorption of carbon dioxide gas and hydrogen with the catalyst. As depicted in Figure 2c, compared to pure MoS2, a small amount of doped Ni ions is quickly combined with the S 2− that is generated To further demonstrate the apparent structures of pure MoS 2 and the new catalysts formed by metal doping, scanning electronic microscopy (SEM) is analyzed. In MoS 2 structure, six S atoms are tightly packed up and down to form a sixfold triangular prism, and the triangular prism voids are formed. Mo atoms are filled into these voids, resulting in a two-dimensional sheet. Monolayers of 2D flakes can stack through van der Waals forces and weak coordination bonds and form different types of interlayer voids. Therefore, as shown in Figure 2a,b, MoS 2 is composed of abundant lamellar nanoflowers, and there are certain gaps between the layers, which is beneficial for deep adsorption of carbon dioxide gas and hydrogen with the catalyst. As depicted in Figure 2c, compared to pure MoS 2 , a small amount of doped Ni ions is quickly combined with the S 2− that is generated by the decomposition of thiourea to form larger NiS 2 particle cubes, with a trend of concave surfaces along the centerline. As the reaction progresses, the raw materials continue to react with these large cubes, forming newly exposed surfaces from the centerline of the cube surface and gradually extending towards the interior of the cube, so the smaller NiS 2 volume cube particles are formed and separated. by the decomposition of thiourea to form larger NiS2 particle cubes, with a trend of concave surfaces along the centerline. As the reaction progresses, the raw materials continue to react with these large cubes, forming newly exposed surfaces from the centerline of the cube surface and gradually extending towards the interior of the cube, so the smaller NiS2 volume cube particles are formed and separated. When the Ni doping amount increases to 0.2 mmol, only irregular small particles are observed and are just filled at the outer edge of the MoS2 layer, resulting in an increasing in interlayer spacing and a reduction of layers number, so the particles are reduced ( Figure  2d), which are consistent with the XRD results. On the one hand, the stacking height and voids are increased, and the multilayer stacking structure is more stable [25]. On the other hand, the benign increase in interlayer spacing is conducive to the deep adsorption of carbon dioxide and hydrogen molecules between the layers, the location of NiS2 leads to the exposure of less edge S vacancies, and, finally, the CH3OH selectivity is effectively promoted. With further doping of Ni, the sample exhibits unseparated cubic blocks or irregular particles, a large amount of NiS2 is accumulated and covers the surface of MoS2, and the CH3OH selectivity is reduced (Figure 2e,f). Therefore, the formation of NiS2 particles is an anti-Ostwald ripening evolution model of large particle splitting and small particle growth separation [26]. An appropriate amount of metal doping can effectively improve the CH3OH selectivity. On the contrary, excessive metal addition will cause a large area of NiS2 accumulation in the material surface layer, which will affect the catalytic performance. Similar situations have also occurred with Co doping ( Figure S2).
The chemical states of elements of the best-performing MoS2/Ni0.2 catalyst are measured through X-ray photoelectron spectroscopy (XPS) for further study. The XPS measured spectra are shown in Figure 3, and all surface composition values are obtained from the spectral results. As displayed in Figure 3a, the Mo, S ratio in MoS2 conforms to the chemical formula, while the ratio of these two elements in MoS2/Ni0.2 is lower than the chemical formula. The spectrum of Ni 2p is composed of spin-orbit peaks at 853.2 eV and 870.8 eV of Ni 2p3/2 and Ni 2p1/2, respectively. An oscillatory satellite peak occurs from the Ni element, displaying the presence of Ni 2+ (Figure 3b) [27]. The XPS spectrum related to Mo 3d region is displayed in Figure 3c. The binding energies at 232.0 eV and 229.1 eV are attributed to Mo 3d3/2 and Mo 3d5/2, respectively, proving that Mo appears with Mo 4+ valence state in MoS2/Ni0.2. The characteristic peak at 226.1 eV corresponds to S 2s. This phe- When the Ni doping amount increases to 0.2 mmol, only irregular small particles are observed and are just filled at the outer edge of the MoS 2 layer, resulting in an increasing in interlayer spacing and a reduction of layers number, so the particles are reduced (Figure 2d), which are consistent with the XRD results. On the one hand, the stacking height and voids are increased, and the multilayer stacking structure is more stable [25]. On the other hand, the benign increase in interlayer spacing is conducive to the deep adsorption of carbon dioxide and hydrogen molecules between the layers, the location of NiS 2 leads to the exposure of less edge S vacancies, and, finally, the CH 3 OH selectivity is effectively promoted. With further doping of Ni, the sample exhibits unseparated cubic blocks or irregular particles, a large amount of NiS 2 is accumulated and covers the surface of MoS 2 , and the CH 3 OH selectivity is reduced (Figure 2e,f). Therefore, the formation of NiS 2 particles is an anti-Ostwald ripening evolution model of large particle splitting and small particle growth separation [26]. An appropriate amount of metal doping can effectively improve the CH 3 OH selectivity. On the contrary, excessive metal addition will cause a large area of NiS 2 accumulation in the material surface layer, which will affect the catalytic performance. Similar situations have also occurred with Co doping ( Figure S2).
The chemical states of elements of the best-performing MoS 2 /Ni 0.2 catalyst are measured through X-ray photoelectron spectroscopy (XPS) for further study. The XPS measured spectra are shown in Figure 3, and all surface composition values are obtained from the spectral results. As displayed in Figure 3a, the Mo, S ratio in MoS 2 conforms to the chemical formula, while the ratio of these two elements in MoS 2 /Ni 0.2 is lower than the chemical formula. The spectrum of Ni 2p is composed of spin-orbit peaks at 853.2 eV and 870.8 eV of Ni 2p 3/2 and Ni 2p 1/2 , respectively. An oscillatory satellite peak occurs from the Ni element, displaying the presence of Ni 2+ (Figure 3b) [27]. The XPS spectrum related to Mo 3d region is displayed in Figure 3c. The binding energies at 232.0 eV and 229.1 eV are attributed to Mo 3d 3/2 and Mo 3d 5/2 , respectively, proving that Mo appears with Mo 4+ valence state in MoS 2 /Ni 0.2 . The characteristic peak at 226.1 eV corresponds to S 2s. This phenomenon is in agreement with the spectral phenomena of the Mo 3d region of molybdenum disulfide without Ni doping, which proves that the doping of Ni does not result in a fundamental change in the valence state of molybdenum and the successful formation of MoS 2 in MoS 2 /Ni 0.2 [28]. It is worth mentioning that the appearance of the Ni-Mo chemical bond confirms that there is a strong chemical force between MoS 2 surfaces and Ni in the MoS 2 /Ni 0.2 catalyst, which may weaken the Mo-S bond. XPS shows two strong S 2p peaks corresponding to binding energies at 163.2 eV as well as 161.8 eV for the S 2p 1/2 and S 2p 3/2 binding energies of S 2− in the MoS 2 /Ni 0.2 sample [29][30][31], which is attributed to the low surface coordination of sulfide ions and the lattice S of metal-sulfur bonds in NiS 2 (Figure 3d) [32]. Therefore, the doping of Ni leads to the charge-density redistribution, the density of edge S vacancies is decreased, and the CH 3 OH selectivity is improved [33].
Molecules 2023, 28, x FOR PEER REVIEW 5 of of MoS2 in MoS2/Ni0.2 [28]. It is worth mentioning that the appearance of the Ni-Mo che ical bond confirms that there is a strong chemical force between MoS2 surfaces and Ni the MoS2/Ni0.2 catalyst, which may weaken the Mo-S bond. XPS shows two strong S peaks corresponding to binding energies at 163.2 eV as well as 161.8 eV for the S 2p1/2 a S 2p3/2 binding energies of S 2− in the MoS2/Ni0.2 sample [29][30][31], which is attributed to t low surface coordination of sulfide ions and the lattice S of metal-sulfur bonds in N ( Figure 3d) [32]. Therefore, the doping of Ni leads to the charge-density redistribution, t density of edge S vacancies is decreased, and the CH3OH selectivity is improved [33]. The chemical adsorption and effective catalysis of carbon dioxide are related to t alkalinity of the MoS2 surface [34,35]. To further explore the role of doped metals in t CO2 adsorption characteristics, the CO2-TPD results of MoS2/Nix (x = 0.1, 0.2, 0.3, 0.5) a shown in Figure 4. By comparing the desorption peak areas and peak positions of the fo different Ni-doped catalysts, these materials have a certain adsorption capacity for acid CO2 molecules. The CO2-TPD curve of MoS2/Nix (x = 0.3, 0.5) catalysts mainly contain m dium-strength alkaline sites. The MoS2/Nix (x = 0.1, 0.2) catalyst exhibits three CO2 desor tion regions, and the MoS2/Ni0.2 has the strongest alkaline site. The strong alkaline si are closely related to methanol selectivity, which may lead to the MoS2/Ni0.2 catalyst ha ing the best methanol selectivity [36]. As the amount of Ni added further increases, t desorption peak slowly becomes flat and wider, indicating that the bonding ability b tween CO2 and the catalyst is gradually weakened, and the catalytic performance d creased accordingly. This result indicates that an appropriate amount of Ni is doped MoS2, with the effective acid-base interaction between the basic S-Mo and S-Ni function groups of the catalyst and the acidic CO2 molecules. In addition, the dipole-dipole che ical interaction between carbon dioxide and the polar sites related to S functional grou also plays a key role in the adsorption level of CO2 gas, which has a positive effect improving product selectivity [37]. The chemical adsorption and effective catalysis of carbon dioxide are related to the alkalinity of the MoS 2 surface [34,35]. To further explore the role of doped metals in the CO 2 adsorption characteristics, the CO 2 -TPD results of MoS 2 /Ni x (x = 0.1, 0.2, 0.3, 0.5) are shown in Figure 4. By comparing the desorption peak areas and peak positions of the four different Ni-doped catalysts, these materials have a certain adsorption capacity for acidic CO 2 molecules. The CO 2 -TPD curve of MoS 2 /Ni x (x = 0.3, 0.5) catalysts mainly contain medium-strength alkaline sites. The MoS 2 /Ni x (x = 0.1, 0.2) catalyst exhibits three CO 2 desorption regions, and the MoS 2 /Ni 0.2 has the strongest alkaline site. The strong alkaline sites are closely related to methanol selectivity, which may lead to the MoS 2 /Ni 0.2 catalyst having the best methanol selectivity [36]. As the amount of Ni added further increases, the desorption peak slowly becomes flat and wider, indicating that the bonding ability between CO 2 and the catalyst is gradually weakened, and the catalytic performance decreased accordingly. This result indicates that an appropriate amount of Ni is doped on MoS 2 , with the effective acid-base interaction between the basic S-Mo and S-Ni functional groups of the catalyst and the acidic CO 2 molecules. In addition, the dipole-dipole chemical interaction between carbon dioxide and the polar sites related to S functional groups also plays a key role in the adsorption level of CO 2 gas, which has a positive effect on improving product selectivity [37].  However, excessive Ni doping results in ineffective contact between CO 2 and catalyst sites. Interestingly, the peak shapes and peak areas of materials are also different with different Ni doping amounts, indicating that the CO 2 adsorption is regulated by both the active component content and the MoS 2 support. Similarly, the CO 2 -TPD curves of MoS 2 /Co catalysts prepared under the same conditions with different Co doping levels are shown in Figure S3. From the comparison chart of the adsorption level of the CO 2 molecule, it can be seen that the four samples all have the corresponding desorption peaks in the intermediate and high-temperature regions. Although the strong alkaline site of MoS 2 /Co 0.2 moves towards the high-temperature zone, the medium alkaline site moves towards the low-temperature zone. Thus, through CO 2 -TPD analysis, it can be deduced that MoS 2 /Ni 0.2 is an ideal catalyst for carbon dioxide hydrogenation to methanol, which must be confirmed by activity evaluation.
The physical adsorption of hydrogen molecules on the catalyst surface is replaced by effective chemical adsorption at a certain temperature. Therefore, as shown in Figure 5, the reduction performance of the catalyst is studied through H 2 -TPR experiments. The active hydrogen rapidly migrates between adjacent S atoms to form H-S-H and the Mo or S edge is gradually reduced through "dissociation diffusion" [38,39]. Because of different particle shapes and dispersion levels in MoS 2 , different Ni doping differs in adsorption temperature and intensity. For MoS 2 /Ni 0.1 , one main peak appears at 324 • C [40]. With the suitable increase in metal doping, the double active sites are gradually enhanced, and the reduction energy consumption decreases, so that the reduction peaks of H 2 -TPR shift to the low-temperature direction. After the Ni doping amount is increased to 0.2 mmol, two well-separated peaks appear at 282 • C and 362 • C; the first peak significantly moves to lower reduction temperatures, and MoS 2 /Ni 0.2 has the lowest reduction temperature, indicating that the metal sulfur bonds energy is reduced.
Combining XRD and SEM analysis, it is found that the strong interaction between doped metals and MoS 2 is conducive to the dispersion of the catalyst and the reduction of catalyst particle size. The increased interlayer spacing of MoS 2 results in an increase in Ni/Mo edge site. Hydrogen participates in the surface reaction, and the sulfur atom is reductively removed, resulting in an irreversible change of MoS 2 structure. Thus, the Mo-S bond strength is weakened, the reducibility of MoS 2 is enhanced, and a good reaction basis for the carbon dioxide hydrogen to methanol is provided [40]. The doping amount of Ni increased to 0.3 and 0.5 mmol, and the excessive and irregular accumulation of NiS 2 is covered on the MoS 2 surface. The reduction temperature is increased, and the reduction ability of the catalyst is weakened. Compared with MoS 2 /Co x catalysts, the MoS 2 /Co 0.2 catalyst also has a good downward shift of the first peak position, but the reduction temperature is still higher than MoS 2 /Ni 0.2 ( Figure S4). Thus, the catalytic performance of Co-doped MoS 2 may be lower than that of MoS 2 /Ni X . Combining XRD and SEM analysis, it is found that the strong interaction between doped metals and MoS2 is conducive to the dispersion of the catalyst and the reduction of catalyst particle size. The increased interlayer spacing of MoS2 results in an increase in Ni/Mo edge site. Hydrogen participates in the surface reaction, and the sulfur atom is reductively removed, resulting in an irreversible change of MoS2 structure. Thus, the Mo-S bond strength is weakened, the reducibility of MoS2 is enhanced, and a good reaction basis for the carbon dioxide hydrogen to methanol is provided [40]. The doping amount of Ni increased to 0.3 and 0.5 mmol, and the excessive and irregular accumulation of NiS2 is covered on the MoS2 surface. The reduction temperature is increased, and the reduction ability of the catalyst is weakened. Compared with MoS2/Cox catalysts, the MoS2/Co0.2 catalyst also has a good downward shift of the first peak position, but the reduction temperature is still higher than MoS2/Ni0.2 ( Figure S4). Thus, the catalytic performance of Codoped MoS2 may be lower than that of MoS2/NiX.
The specific surface area and pore size distribution of the material play an important role in the performance of the catalyst. The MoS2/Ni catalyst with Ni doping amount of 0.2 mmol and MoS2 catalyst are analyzed by N2 physical adsorption and desorption technology. The N2 adsorption and desorption trend and pore size distribution characteristics are listed in Figure 6 and  The specific surface area and pore size distribution of the material play an important role in the performance of the catalyst. The MoS 2 /Ni catalyst with Ni doping amount of 0.2 mmol and MoS 2 catalyst are analyzed by N 2 physical adsorption and desorption technology. The N 2 adsorption and desorption trend and pore size distribution characteristics are listed in Figure 6 and Table 1. The adsorption isotherms of the MoS 2 catalyst and MoS 2 /Ni 0.2 catalyst belong to class VI isotherms. According to the standard classification of IUPAC, the characteristics of such isotherms reflect the characteristics of mesoporous materials (Figure 6a,b). When the relative pressure of P/P 0 of MoS 2 is 0.45~1.0, and the relative pressure of P/P 0 of MoS 2 /Ni 0.2 is 0.8~1.0, the adsorption capacity is improved. The pore size distribution ranges of the involved materials are shown in Figure 6c,d, which are obtained from the analysis of the pore size distribution curve. The highest point corresponding to the pore radius is greater than 2 nm, which further proves that both MoS 2 and MoS 2 /Ni 0.2 catalysts are mesoporous materials. Compared with MoS 2 , the MoS 2 /Ni 0.2 catalyst exhibits a relatively smaller specific pore diameter as well as pore volume. It can be seen that the excellent CH 3 OH selectivity of MoS 2 /Ni 0.2 may be due to its relatively small pore volume (0.090 cm 3 ·g −1 ) and pore size (10.13 nm). be seen that the excellent CH3OH selectivity of MoS2/Ni0.2 may be due to its relatively small pore volume (0.090 cm 3 ·g −1 ) and pore size (10.13 nm).  Under the condition settings of 260 °C, 5 MPa, and 12,000 mL h −1 gcat −1 of gas hourly space velocity (GHSV), all catalytic evaluations are performed by a fixed-bed reactor, and the experimental results after testing are listed in Figure 7. As a comparative catalyst, the property of 2H-type MoS2 is studied first. The result shows that the CO2 conversion rate is 3.36% ( Figure S5) and the methanol selectivity is 11.34%, and the main byproducts are methane and carbon monoxide. The catalytic characteristic of MoS2 for CO2 to product is well demonstrated by the experimental results, but the conversion and selectivity are expected to be improved. With the increase in Ni and Co doping amount, the CH3OH selectivity is also increased. When the Ni and Co doping amount is 0.2 nmol, the CH3OH of the catalyst MoS2/Ni0.2 and MoS2/Co0.2 reach the highest 83.73% and 73.82% (Figure 7). However, when the doping amount is further increased from 0.2 mmol to 0.3 mmol, the methanol selectivity shows a downward trend, because the appropriate proportion of Ni and Co addition leads to the blocking effect of edge S vacancies. On the contrary, excessive Ni and Co metals are doped, a large amount of NiS2 and CoS2 are stacked and covered on the reaction surface of the catalyst, the exposed sulfur vacancy active sites are blocked, and the catalytic performance is reduced. By comparing the CH3OH selectivity of Ni and Co-doped catalysts, it is determined that the catalytic property of the Ni-doped MoS2 cat-  Under the condition settings of 260 • C, 5 MPa, and 12,000 mL h −1 g cat −1 of gas hourly space velocity (GHSV), all catalytic evaluations are performed by a fixed-bed reactor, and the experimental results after testing are listed in Figure 7. As a comparative catalyst, the property of 2H-type MoS 2 is studied first. The result shows that the CO 2 conversion rate is 3.36% ( Figure S5) and the methanol selectivity is 11.34%, and the main byproducts are methane and carbon monoxide. The catalytic characteristic of MoS 2 for CO 2 to product is well demonstrated by the experimental results, but the conversion and selectivity are expected to be improved. With the increase in Ni and Co doping amount, the CH 3 OH selectivity is also increased. When the Ni and Co doping amount is 0.2 nmol, the CH 3 OH of the catalyst MoS 2 /Ni 0.2 and MoS 2 /Co 0.2 reach the highest 83.73% and 73.82% (Figure 7). However, when the doping amount is further increased from 0.2 mmol to 0.3 mmol, the methanol selectivity shows a downward trend, because the appropriate proportion of Ni and Co addition leads to the blocking effect of edge S vacancies. On the contrary, excessive Ni and Co metals are doped, a large amount of NiS 2 and CoS 2 are stacked and covered on the reaction surface of the catalyst, the exposed sulfur vacancy active sites are blocked, and the catalytic performance is reduced. By comparing the CH 3 OH selectivity of Ni and Co-doped catalysts, it is determined that the catalytic property of the Ni-doped MoS 2 catalyst slightly surpasses the Co-doped MoS 2 catalyst.

Determination of Intermediates through In Situ Infrared Spectroscopy
The key adsorbents and intermediates included in the formation of methanol on the MoS 2 /Ni 0.2 catalyst with the best performance are shown in Figure 8. The characteristic peak of CO 2 gas appears at approximately 2350 cm −1 [41]. The peaks at 720 cm −1 and 907 cm −1 are attributed to O* characteristic and the peak intensity is obvious [21]. The vibration peak at 1080 cm −1 is attributed to the methoxy group (CH 3 O*) [42]. In situ DRIFTS experimental results show that the chemical bond of CO 2 adsorbed on the MoS 2 /Ni 0.2 surface is broken and forms CO* and O*, then CO* and dissociated H* react to form HCO* species. The generation of methoxy and methanol is due to subsequent gradual hydrogenation.

Determination of Intermediates through In Situ Infrared Spectroscopy
The key adsorbents and intermediates included in the formation of methanol on the MoS2/Ni0.2 catalyst with the best performance are shown in Figure 8. The characteristic peak of CO2 gas appears at approximately 2350 cm −1 [41]. The peaks at 720 cm −1 and 907 cm −1 are attributed to O* characteristic and the peak intensity is obvious [21]. The vibration peak at 1080 cm −1 is attributed to the methoxy group (CH3O*) [42]. In situ DRIFTS experimental results show that the chemical bond of CO2 adsorbed on the MoS2/Ni0.2 surface is broken and forms CO* and O*, then CO* and dissociated H* react to form HCO* species. The generation of methoxy and methanol is due to subsequent gradual hydrogenation.  lations. The chemical adsorption of H 2 and CO 2 on the catalyst surface is determined as a prerequisite for subsequent reactions. Figure 9a shows the optimal adsorption structure for CO 2 on the catalyst. The CO 2 angle is from 180 • to 120.20 • and the O-C bond is elongated from 1.25 Å to 1.38 Å. The corresponding differential charge density (Figure 9b) confirms that Mo electrons are transferred to the adsorbed CO 2 . The corresponding PDOS analysis shows that the CO 2 molecular orbital is not only shifted to a lower energy level after being adsorbed on the catalyst surface (Figure 9c), but also overlaps with the Mo 4d state after the CO 2 gas becomes an adsorbed state on the MoS 2 surface, proving the CO 2 activation. As shown in Figure 9e, Ni doping leads to more unsaturated exposure of Mo, and the formed active hydrogen is more prone to surface migration. The interaction between Ni and the charge-rich center S in NiS 2 gives it an appropriate affinity and bonding range. The local action of negative charge on the S atom and H 2 occurs at the highest spin density position. The activation process involves the transfer of negative charge from the S atom to the H 2 antibonding orbital σ*, the electron cloud distribution and orbital energy of the hydrogen molecule are changed, the four-center transition state is formed, the S-S and H-H bonds are weakened and heterocleaved, and the S-H bond is strengthened, which eventually leads to the breaking of the former and the formation of the latter. A larger range of orbital hybridization near the Fermi level (Figure 9f) makes the H 2 bond break more thoroughly (Figure 9d). Combined with in situ infrared spectroscopy, the effective dissociation of hydrogen gas is conducive to the gradual hydrogenation of CO*, so the methanol selectivity is more highly catalyzed by MoS 2 /Ni 0.2 .

Determination of Mechanisms Properties through DFT Calculation
To provide a deeper explanation of metal doping in improving methanol selectivity the reaction mechanisms of MoS2 and Ni/MoS2 are investigated by DFT theoretical calcu lations. The chemical adsorption of H2 and CO2 on the catalyst surface is determined as prerequisite for subsequent reactions. Figure 9a shows the optimal adsorption structur for CO2 on the catalyst. The CO2 angle is from 180° to 120.20° and the O-C bond is elon gated from 1.25 Å to 1.38 Å. The corresponding differential charge density (Figure 9b confirms that Mo electrons are transferred to the adsorbed CO2. The corresponding PDOS analysis shows that the CO2 molecular orbital is not only shifted to a lower energy leve after being adsorbed on the catalyst surface (Figure 9c), but also overlaps with the Mo 4d state after the CO2 gas becomes an adsorbed state on the MoS2 surface, proving the CO activation. As shown in Figure 9e, Ni doping leads to more unsaturated exposure of Mo and the formed active hydrogen is more prone to surface migration. The interaction be tween Ni and the charge-rich center S in NiS2 gives it an appropriate affinity and bonding range. The local action of negative charge on the S atom and H2 occurs at the highest spin density position. The activation process involves the transfer of negative charge from th S atom to the H2 antibonding orbital σ*, the electron cloud distribution and orbital energy of the hydrogen molecule are changed, the four-center transition state is formed, the S-S and H-H bonds are weakened and heterocleaved, and the S-H bond is strengthened which eventually leads to the breaking of the former and the formation of the latter. A larger range of orbital hybridization near the Fermi level (Figure 9f) makes the H2 bond break more thoroughly (Figure 9d). Combined with in situ infrared spectroscopy, the ef fective dissociation of hydrogen gas is conducive to the gradual hydrogenation of CO*, so the methanol selectivity is more highly catalyzed by MoS2/Ni0.2.  The chemically adsorbed CO 2 is considered to have two reduction pathways: the formate pathway and CO hydrogenation. According to in situ DRIFTS spectra, CO 2 hydrogenation catalyzed by MoS 2 /Ni tends to an oxidation-reduction pathway, and the optimal adsorption structures of different intermediates as well as barrier energy are shown in Figure 10. CO 2 and H 2 are directly dissociated into CO*, O*, and 2H* through the energy barrier of 0.63 eV. The large adsorption energy indicates that the interaction between CO* and MoS 2 is very strong (E ads = −1.27 eV), and is more likely to undergo further reaction on the support without being released and forming CO gas. The calculation results indicate that CO* can be hydrogenated to CHO* by crossing only a potential barrier of 0.69 eV. It is evident that the carbon involved in CO is more susceptible to H proton attack than the oxygen atom, and forms CHO anchoring at the active site for subsequent hydrogenation reaction. The second hydrogenation step of CHO to CH 2 O needs a reaction barrier of 1.22 eV. The CH 2 O is further hydrogenated to CH 3 O, which just needs a barrier energy of 0.79 eV, and this hydrogenation process is exothermic, indicating that the CH 3 O is a beneficial intermediate. In the fourth hydrogenation reaction, CH 3 O can be reduced to CH 3 OH through an energy barrier of 1.33 eV (the decisive step of the reaction system) and easily desorbed from the catalyst surface. Therefore, based on the above mechanism exploration, methanol is the final product catalyzed by MoS 2 /Ni. Finally, the reaction mechanism is summarized as CO 2 * → CO* → CHO* → CH 2 O* → CH 3 O* → CH 3 OH.

OR PEER REVIEW 11 of 18
The chemically adsorbed CO2 is considered to have two reduction pathways: the formate pathway and CO hydrogenation. According to in situ DRIFTS spectra, CO2 hydrogenation catalyzed by MoS2/Ni tends to an oxidation-reduction pathway, and the optimal adsorption structures of different intermediates as well as barrier energy are shown in Figure 10. CO2 and H2 are directly dissociated into CO*, O*, and 2H* through the energy barrier of 0.63 eV. The large adsorption energy indicates that the interaction between CO* and MoS2 is very strong (Eads = −1.27 eV), and is more likely to undergo further reaction on the support without being released and forming CO gas. The calculation results indicate that CO* can be hydrogenated to CHO* by crossing only a potential barrier of 0.69 eV. It is evident that the carbon involved in CO is more susceptible to H proton attack than the oxygen atom, and forms CHO anchoring at the active site for subsequent hydrogenation reaction. The second hydrogenation step of CHO to CH2O needs a reaction barrier of 1.22 eV. The CH2O is further hydrogenated to CH3O, which just needs a barrier energy of 0.79 eV, and this hydrogenation process is exothermic, indicating that the CH3O is a beneficial intermediate. In the fourth hydrogenation reaction, CH3O can be reduced to CH3OH through an energy barrier of 1.33 eV (the decisive step of the reaction system) and easily desorbed from the catalyst surface. Therefore, based on the above mechanism exploration, methanol is the final product catalyzed by MoS2/Ni. Finally, the reaction mechanism is summarized as CO2* → CO* → CHO* → CH2O* → CH3O* → CH3OH.

Materials
All the materials contained in the experiments are shown in Table 2.

Materials
All the materials contained in the experiments are shown in Table 2.

Experimental Apparatus
All the instruments and types of equipment for experiments are shown in Table 3.  (35 mL). To completely dissolve the reactants and form a homogeneous solution, the beaker containing the mixed reaction liquid should be placed on a magnetic suspension stirrer and stirred for 30 min. Different contents of nickel nitrate hexahydrate or cobalt nitrate hexahydrate are weighed and placed into the above solution to be completely dissolved. The thoroughly dissolved solution is poured into a hydrothermal kettle with Teflon-lined stainless steel, and the Teflon lid is tightly closed. The hydrothermal kettle is placed into a vacuum drying box, then reacted at 210 • C for 24 h. After the completion of catalyst preparation, the hydrothermal kettle is taken out and the temperature drops to cool down. The reacted substance is washed with deionized water 3 times and then washed again with anhydrous ethanol for the same process. Finally, the washed and centrifuged material is effectively dried by a drying oven, the drying temperature is 80 • C and the time is 12 h, and the final product sample is obtained. For experimental comparison, MoS 2 is prepared by dissolving ammonium molybdate tetrahydrate (1 mmol) and thiourea (14 mmol) using the same preparation process as MoS 2 /Ni and MoS 2 /Co.

Characterization Methods
The crystal phase composition of the material is determined by a German BmkerD8 advanced X-ray diffractometer (XRD). The radiation source is Cu K α rays, the tube voltage is 40 kV, and the tube current is 40 mA. The scanning range is 3-85 • , and the scanning speed is 8 • /min.
The specific surface area and the pore size distribution of the materials, are characterized by a Quanta Autosorb IQ automatic physical and chemical adsorption instrument. The adsorbate is N 2 , and the physical adsorption test is carried out under the condition of vacuum liquid nitrogen (−196 • C). Before the adsorption test, the samples needed to be degassed at a temperature of 200 • C for 6 h. The specific surface area is obtained by linear regression of the multipoint BET (Brunauer-Emmett-Teller) model, and the pore size distribution is obtained by the DFT model.
The microscopic morphology of the material is characterized by scanning electron microscopy (SEM) (Carl Zeiss Company, Jena, Germany) NanoPorts QUANTA250/QUANTA430.
The H 2 -TPR test is carried out on a chemisorption apparatus with automatic temperature (AutoChem II 2920) from Micromeritics Company, Georgia, USA. A 50-100 mg sample is weighed and placed in reaction tubes, and the test temperature rises from 25 • C to 300 • C with a rate of 10 • C/min for drying pretreatment, then cooled to 50 • C under He gas flow (30-50 mL/min) for 1 h. A 10% H 2 /Ar mixture is passed into the reaction tube with a fixed gas flow rate (30-50 mL/min) for 1 h to be saturated, and then switched to Ar with the same gas flow to eliminate weak physical adsorption of H 2 . Finally, the tested catalyst is heated to 500 • C in Ar atmosphere (heating rate is 10 • C/min) and desorbed, and the desorbed gas is detected by TCD.
The CO 2 -TPD test is carried out on a chemical adsorber (AutoChem II 2920) from Micromeritics, USA, which is a fully automatic temperature program. A 50-100 mg sample is weighed and placed in reaction tubes, and the test temperature is increased from 25 • C to 300 • C (heating rate 10 • C/min) for drying pretreatment, then cooled to 50 • C under He gas flow (30-50 mL/min) for 1 h. A 10% CO 2 /He mixture is passed into the reaction tube at a certain gas flow rate is 30-50 mL/min for 1 h to be saturated, and then switched to He at a certain gas flow rate of 30-50 mL/min for 1 h to eliminate weak physical adsorption of carbon dioxide on the surface. The last step involves heating the catalyst to 700 • C at a heating rate (10 • C/min) in He atmosphere and desorbed, and the desorbed gas is detected by TCD.
A Thermo Scientific K-Alpha spectrometer (Thermo Fisher Scientific, Waltham, MA, USA) used for the XPS test, and the content and valence state of catalyst surface substances are analyzed under the excitation source of Al Kα rays (hv = 1486.6 eV). A suitable amount of the pressed catalyst is connected to the corresponding sample tray in a standardized manner and placed in the sample detection chamber of the Thermo Scientific K-Alpha XPS instrument. The sample is sent to the analysis chamber when the indoor pressure of the sample chamber reaches the standard value (less than 2.0 × 10 −7 bar). The important test conditions include spot size (400 µm), work voltage (12 kV), and filament current (6 mA).
Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) analysis is performed by Thermo Nicolet iS20 (Thermo Fisher Scientific, Shanghai, China) which has a liquid-nitrogen-cooled mercury cadmium-telluride (MCT) detector. Firstly, the catalyst is reduced at room temperature for 30 min under a pure N 2 (20 mL/min) atmosphere. The background spectra are obtained under N 2 blowing. Six batches of DRIFTS characterization are performed on the pretreated catalyst. Mixed CO 2 and H 2 (CO 2 :H 2 = 1:3) are introduced into the chamber at 150 • C, 180 • C, 200 • C, 220 • C, 240 • C, and 260 • C for 30 min.

Catalytic Performance Test
As shown in Figure 11, the catalytic characteristic for the reaction of CO 2 to CH 3 OH is measured on fixed-bed equipment. A certain amount of catalyst (0.2 g) and 0.4 g of quartz sand (40-60 mesh) are mixed uniformly and loaded into the constant temperature zone in the middle of the reaction tube (8 mm internal diameter). Before the performance is evaluated, the catalyst is pretreated in situ with 10 mL/min H 2 /N 2 for 3 h under test conditions of 0.1 MPa and 300 • C. After reduction, the feed gas (H 2 :CO 2 = 3:1) is transmitted through the reactor and fully contacts the catalyst. The subsequent reactions are performed under 5 MPa, a temperature range from 180 • C to 260 • C, and GHSV is maintained at 12,000 mL h −1 g cat −1 . The gas phase product is detected by Agilent 8890 (Agilent Technologies (China) Co., Ltd, Beijing, China) online chromatography, and the composition and content of the product are determined according to the peak time of the product in the chromatographic column. The specific amounts of CO 2 , CO, CH 4 , and H 2 included in the reaction product are monitored online by MolSieve 5A (Agilent Technologies (China) Co., Ltd, Beijing, China) packed column (TCD). Organic compounds such as CH 4 and CH 3 OH in the reaction product are detected online by HP-PLOT Q-packed column (FID).

Catalytic Performance Test
As shown in Figure 11, the catalytic characteristic for the reaction of CO2 to CH3OH is measured on fixed-bed equipment. A certain amount of catalyst (0.2 g) and 0.4 g of quartz sand (40-60 mesh) are mixed uniformly and loaded into the constant temperature zone in the middle of the reaction tube (8 mm internal diameter). Before the performance is evaluated, the catalyst is pretreated in situ with 10 mL/min H2/N2 for 3 h under test conditions of 0.1 MPa and 300 °C. After reduction, the feed gas (H2:CO2 = 3:1) is transmitted through the reactor and fully contacts the catalyst. The subsequent reactions are performed under 5 MPa, a temperature range from 180 °C to 260 °C, and GHSV is maintained at 12,000 mL h −1 gcat −1 . The gas phase product is detected by Agilent 8890 (Agilent Technologies (China) Co., Ltd, Beijing, China) online chromatography, and the composition and content of the product are determined according to the peak time of the product in the chromatographic column. The specific amounts of CO2, CO, CH4, and H2 included in the reaction product are monitored online by MolSieve 5A (Agilent Technologies (China) Co., Ltd, Beijing, China) packed column (TCD). Organic compounds such as CH4 and CH3OH in the reaction product are detected online by HP-PLOT Q-packed column (FID). Figure 11. The fixed-bed reaction evaluation device.

Calculation Methods for Reactant Conversion as Well as Product Selectivity
The reactant conversion is calculated using the normalization method, and the formulas for CO2 conversion and selectivity of all products are shown in Equations (1) and (2).

Calculation Methods for Reactant Conversion as Well as Product Selectivity
The reactant conversion is calculated using the normalization method, and the formulas for CO 2 conversion and selectivity of all products are shown in Equations (1) and (2).
The correction factor values for each component in tail gas are shown in Table 4.

Calculation Methods
The Vienna Ab Initio Simulation Package (VASP) is used to calculate the relevant properties of the MoS 2 /Ni system [21]. Exchange correlation is measured through Perdew-Burke-Ernzerhof (PBE) method [43]. The plane-wave cut-off energy is 400 eV and the Monkhorst-Pack k-point sampling is 1 × 1 × 1 for the structural optimization involved in this system, while the electronic performance is calculated with the larger k-point of 3 × 3 × 1 [44]. A reasonable initial model is constructed as a 5 × 5 supercell with a vacuum thickness of 15 Å, and the size of the calculation model is 15.83 × 15.83 Å. Structural convergences are completed with Hellmann-Feynman residual force convergences between each atom below 0.02 eV/Å. A CI-NEB method is used to search the transition state involved in each elementary reaction [45]. The adsorption energy of the reactant on the catalyst surface (E ads ) is represented by Equation (4): where E AB is the energy value of the reaction system where small reactant molecules are adsorbed on the material surface, E A is the energy of material structure before adsorption, and E SM is the energy possessed by reactant molecules adsorbed on the catalyst surface.

Conclusions
In conclusion, the MoS 2 catalyst is prepared by hydrothermal synthesis and doped with different amounts of Ni and Co metal to obtain MoS 2 /Ni and MoS 2 /Co composite catalyst materials. The catalytic properties of these materials in CO 2 hydrogenation to CH 3 OH are investigated.
Compared with 2H-MoS 2 , under the optimal reaction conditions, the MoS 2 /Ni 0.2 composite catalyst shows excellent product selectivity for methanol. Thus, the high methanol selectivity from CO 2 and H 2 gas is achieved.
Through the analysis of XRD, XPS, BET, H 2 -TPR SEM, and CO 2 -TPD characterization results of MoS 2 , Ni/MoS 2 , and Co/MoS 2 composite catalysts with different metal loading ratios, it is confirmed that the addition of Ni, Co metals are beneficial for blocking more edge S vacancies, and, thus, the CH 3 OH selectivity is higher. Through metal doping, the H 2 consumption and CO 2 adsorption are improved, indicating that the ability to activate H 2 and CO 2 is improved.
In situ DRIFTS spectra and DFT calculation confirmed that the hydrogenation of CO 2 to methanol on MoS 2 /Ni catalyst follows an oxidation-reduction pathway. CO 2 is activated into CO* and gradually hydrogenated to form methoxy groups; ultimately, methanol is generated. The improvement of CH 3 OH selectivity by metal doping involved in this paper has insight significance for subsequent related research.