Efficient hydrodeoxygenation of sulfoxides into sulfides under mild conditions using heterogeneous cobalt–molybdenum catalysts

Abstract

Nitrogen-doped carbon-supported cobalt–molybdenum bimetallic catalysts (abbreviated as Co–Mo/NC) are active for the hydrodeoxygenation of sulfoxides to sulfides under mild conditions (25–80 °C and 10 bar H₂), which represents the first example of the use of heterogeneous non-noble metal catalysts for this transformation. MoO₃ with Lewis acid sites assists the hydrodeoxygenation of sulfoxides into sulfides by hydrogen over cobalt nanoparticles.

---

The selective hydrodeoxygenation of sulfoxides into sulfides is one of the most important organic transformations in organic chemistry,¹ which has attracted great interest in the fine chemical industry as well as in biochemistry. For example, sulfoxides are now widely utilized as chirons in numerous asymmetric transformations wherein the sulfoxide groups, after stereoselective introduction, are removed by deoxygenation to their corresponding sulfides followed by hydrodesulfurization.² In the early days, the hydrodeoxygenation of sulfoxides was performed by the use of excess amounts of sacrificial agents such as metal hydrides and zinc metal,^3–5 which produce large quantities of waste. There have also been some reports on metal-free deoxygenation of sulfoxides with halides, such as iodides in the presence of acids.^6,7 However, these methods produce halide by-products, which are not environmentally friendly.

Currently, catalytic hydrodeoxygenation represents an environmentally friendly method. Compared with other reducing agents such as alcohols⁸ and silanes,⁹ hydrogen is a clean and environmentally friendly reducing agent because of high atom-efficiency with only the release of water. Homogeneous catalysts were earlier utilized for the hydrodeoxygenation of dimethyl sulfoxide,^10,11 but it is difficult to recycle homogeneous catalysts and there is the purification of the products to be considered. Obviously, the use of heterogeneous catalysts can address the drawbacks associated with homogeneous catalytic systems. However, the deoxygenation of sulfoxides with H₂ is unprecedented over heterogeneous catalysts.¹² In 2014, Mitsudome and co-workers made a great contribution with the hydrodeoxygenation of sulfoxide over a Ru/TiO₂ catalyst, which was reported to be active under atmospheric H₂ pressure.¹³ Later, heterogeneous platinum catalysts (Pt–MoO_X/TiO₂ and Pt/V_0.7Cr_0.3O_X) were developed for the hydrodeoxygenation of sulfoxides, which was performed at 50–150 °C under hydrogen pressure (1–7 bar).^14,15 Although heterogeneous noble metal catalysts are effective for this transformation, the use of heterogeneous non-noble metal catalysts is highly attractive because of the lower cost.

To the best of our knowledge, no heterogeneous non-noble metal catalysts have been reported for this transformation. In recent years, nitrogen-doped carbon material-supported non-noble metal catalysts have received a great deal of attention in the replacement of noble metal catalysts. Our group recently discovered that nitrogen-doped carbon material-supported cobalt catalysts demonstrate excellent activity in hydrogenation and oxidation reactions.^16,17 Therefore, we also tried to use nitrogen-doped carbon-supported cobalt catalysts for the hydrodeoxygenation of sulfoxides, but failed. Inspired by the fact that some metals such as Re, Mo and W are oxyphilic elements, nitrogen-doped carbon material-supported Co–Mo bimetallic catalysts were constructed, and found to be active for the selective hydrodeoxygenation of sulfoxides.

ZIF-67 as one of the most investigated metal–organic frameworks was used for the preparation of Co–Mo/NC-T catalysts, where T represents the reduction temperature. In order to introduce the oxyphilic element molybdenum, (NH₄)₆Mo₇O₂₄·4H₂O was introduced into ZIF-67 via a double-solvent method, and the mixture was subjected to pyrolysis at different temperatures (see the Experimental details in the ESI†). Powder X-ray diffraction (PXRD) was performed to determine the crystal structure of the Co–Mo/NC-T catalysts. As shown in Fig. 1, three characteristic peaks at 2θ = 44.2°, 51.5° and 75.9° corresponding to metallic Co (111), (200), (220) reflections were observed for the Co–Mo/NC-T catalysts (PDF 15-0806).¹⁸ The average size of Co nanoparticles was estimated to be 9.8, 28.5 and 33.9 nm from the line broadening of the Co (111) reflection for the Co–Mo/NC-400, Co–Mo/NC-600 and Co–Mo/NC-800 catalysts, respectively. There were no XRD peaks of molybdenum species for the Co–Mo/NC-400 and Co–Mo/NC-600 catalysts, while characteristic peaks at 2θ = 34.5°, 38.1° and 39.5° were observed in the PXRD pattern of the Co–Mo/NC-800 catalyst, which were assigned to the (021), (200), and (121) reflections of Mo₂C.¹⁹ According to the reported results,¹⁹ (NH₄)₆Mo₇O₂₄·4H₂O was gradually changed into MoO₃ from 300 °C to 500 °C and MoO₂ from 500 °C to 700 °C, and then to Mo₂C beyond 775 °C. No PXRD peaks of MoO₃ or MoO₂ were observed for the Co–Mo/NC-400 and Co–Mo/NC-600 catalysts, which was possibly due to the fact that MoO₃ or MoO₂ did not form crystal structures or the size of them was too small to be visible.

Fig. 1 PXRD patterns of the Co–Mo/NC-T catalysts.

As shown in Fig. S1,† two peaks at around 1347 and 1594 cm⁻¹ were observed in the Raman spectra of the Co–Mo/NC-T catalysts, corresponding to the D and G bands, respectively. The G band represents the graphitic structure of the catalysts, while the D band can be attributed to the lattice defects.²⁰ The intensity ratios of the D and G bands (I_D/I_G) were calculated to be 1.005, 0.979 and 0.904 for the Co–Mo/NC-400, Co–Mo/NC-600 and Co–Mo/NC-800 catalysts, and the decrease in the number of defects may result from the decrease in the number of nitrogen atoms in the Co–Mo/NC-T catalysts as a result of the increase in the reduction temperature.²¹

A transmission electron microscope (TEM) image (Fig. 2a) revealed that the Co nanoparticles were homogeneously dispersed on the surface of the Co–Mo/NC-400 catalyst with an average size of 10.1 nm in agreement with the XRD results. However, the growth of cobalt nanoparticles into larger particle size as well as aggregation were observed in the TEM images of the Co–Mo/NC-600 and Co–Mo/NC-800 catalysts. As shown in high-resolution TEM (HR-TEM) images (Fig. S2 and S3†), the lattice spacing measured for the crystalline plane was 0.20 nm (Fig. 2b), which corresponded to the d values of the (111) planes for metallic cobalt nanoparticles, while no lattices for the Mo species were found in the HR-TEM image of the Co–Mo/NC-400 catalyst.

Fig. 2 TEM image (a) and HR-TEM image (b) of the Co–Mo/NC-400 catalyst.

X-ray photoelectron spectroscopy (XPS) was further used to study the valence states of the Co and Mo species in the Co–Mo/NC-T catalysts. Two peaks with binding energies at 781.4 and 796.6 eV were assigned to the Co 2p_3/2 and Co 2p_1/2 of Co²⁺, while two peaks with binding energies at 779.5 eV and 795.0 eV were attributed to the Co 2p_3/2 and Co 2p_1/2 of Co³⁺. Two peaks with binding energies at 785.7 and 802.4 eV corresponded to the Co²⁺ LMM auger peak in Co₃O₄.²² The presence of Co₃O₄ in the Co–Mo/NC-400 catalyst was due to the oxidation of surface metallic cobalt nanoparticles during storage in the air. Similarly, the surface of cobalt nanoparticles was also present in the oxidation states for Co–Mo/NC-600 and Co–Mo/NC-800 catalysts (Fig. S4†). The Mo 3d_5/2 XPS peak could be deconvoluted into two peaks with binding energies of 231.6 and 232.2 eV, which are attributed to the Mo⁵⁺ and Mo⁶⁺ oxidation states for the Co–Mo/NC-400 catalyst (Fig. 3), respectively,²³ and Mo⁶⁺ was the main oxidation state, corresponding to MoO₃. Peaks with binding energies at 231.1, 232.4 and 227.6 eV were observed for the Mo 3d_5/2 XPS peak for the Co–Mo/NC-600 catalyst, corresponding to the Mo⁵⁺, Mo⁶⁺ oxidation states and metallic Mo (Fig. S5†).²³ Two peaks with binding energies at 232.4 and 228.6 eV were the characteristic Mo 3d_5/2 XPS peaks of Mo⁶⁺ and Mo₂C of the Co–Mo/NC-800 catalyst (Fig. S5†), and Mo₂C was the main state in the Co–Mo/NC-800 catalyst, consistent with the PXRD results.

Fig. 3 XPS spectra of Co 2p and Mo 3d in the Co–Mo/NC-400 catalyst.

The hydrodeoxygenation of diphenyl sulfoxide into diphenyl sulfide was used as a model reaction to study the catalytic activity of the Co–Mo/NC-T catalysts (Table 1). Initially, the reaction was conducted in the presence of a Co/NC-400 catalyst, which was prepared by the same method as described for the Co–Mo/NC-400 catalyst without the introduction of Mo. No conversion of diphenyl sulfoxide was observed (Table 1, entry 1). In order to cleave the S=O bond, Mo as a common oxyphilic element was introduced into the nitrogen-doped carbon-supported Co catalysts to activate the S=O bond. As expected, the Co–Mo/NC-400 catalyst could smoothly promote the hydrodeoxygenation of diphenyl sulfoxide into diphenyl sulfide, affording a conversion of 77% after 2 h at 80 °C under 10 bar H₂ (Table 1, entry 2). However, the conversion of diphenyl sulfoxide sharply decreased to 24% over the Co–Mo/NC-600 catalyst (Table 1, entry 3), and no conversion was observed over the Co–Mo/NC-800 catalyst (Table 1, entry 4). The decrease in the catalytic activity of the Co–Mo/NC-T catalysts with an increase in the reduction temperature could be mainly due to the following two reasons. On the one hand, cobalt nanoparticles became larger and aggregated with an increase of the reduction temperature, which provided fewer active sites for the activation of H₂. On the other hand, the Mo species in the Co–Mo/NC-T catalysts also played a crucial role in the hydrodeoxygenation of diphenyl sulfoxide. As characterized by XPS, MoO₃ was the major component for the Co–Mo/NC-400 catalyst, while MoO₂ was the major component for the Co–Mo/NC-600 catalyst. MoO₃ with abundant oxygen vacancy sites (also called Lewis acid sites) was reported to be effective in combining with oxygen atoms in the S=O bond,²⁴ thus facilitating the hydrodeoxygenation of diphenyl sulfoxide. This was further confirmed that in the Co–Mo/NC-800 catalyst was completely inactive towards the hydrodeoxygenation of diphenyl sulfoxide, in which Mo species was Mo₂C. A Co–V/NC-400 catalyst was also prepared with a similar method as for the Co–Mo/NV-400 catalyst, but it showed no catalytic activity for the hydrodeoxygenation of sulfoxides (Table 1, entry 6). These results suggested that vanadium oxides could not activate the substrates. We prepared an active carbon-supported Co–Mo catalyst by a similar method, and used it in the hydrodeoxygenation of sulfoxide. However, it showed no activity under the same reaction temperature of 80 °C (Table 1, entry 7). These results suggested that the nitrogen atoms in the carbon materials played important roles in the hydrodeoxygenation of sulfoxides. According to previous reports,^17,26 nitrogen atoms might play the following roles in the catalyst activity: (a) nitrogen atoms act as Lewis base sites to promote the heterolytic cleavage of H₂ in coordination with the Co nanoparticles; (b) the nitrogen atoms should have a strong interaction with Co nanoparticles to improve the catalytic activity of Co nanoparticles and to stabilize them.

Table 1 Hydrodeoxygenation of diphenyl sulfoxide over different catalysts^a

Entry	Catalyst	Time (h)	Conversion (%)	Sel. (%)
a Reaction conditions: Diphenyl sulfoxide (0.5 mmol), 1,4-dioxane (10 mL), H2 (10 bar), catalysts (20 mg), 80 °C.
1	Co/NC-400	2	0	—
2	Co–Mo/NC-400	2	77	100
3	Co–Mo/NC-600	2	24	100
4	Co–Mo/NC-800	2	0	—
5	MoO3	2	0	—
6	Co–V/NC-400	2	0	—
7	Co–Mo/AC-400	2	0	—

---

According to the above results, we can conclude that metallic Co nanoparticles, nitrogen atoms and MoO₃ played a synergistic role in the hydrodeoxygenation of diphenyl sulfoxide. MoO₃ would activate the S=O groups by combination with the oxygen atom in the S=O bond. Metallic Co nanoparticles activated hydrogen molecules with the assistance of nitrogen atoms to generate the Co–H⁻ hybrid and N–H⁺ species. The H⁺ in N–H⁺ was transferred to the oxygen atoms in S=O bonds to generate –S–OH species, and the sulfur–oxygen bond was then cleaved by H⁻ to give the target product.

The reaction conditions were then optimized over the Co–Mo/NC-400 catalyst. Firstly, the effect of the reaction solvents was studied. Generally, protic solvents with strong polarity as well as water gave high to quantitative conversions of diphenyl sulfoxide (Table S1, entries 1 vs. 4†). Particularly, it is interesting to note that the conversion of diphenyl sulfoxide in alcohols slightly increased in the order iso-PrOH, ethanol, and methanol. Moderate conversions of diphenyl sulfoxide were produced in N,N-dimethylformamide (DMF) and three other kinds of solvents with moderate polarity (Table S1, entries 5 vs. 8†). The reaction in hexane with non-polarity produced the lowest conversion of diphenyl sulfoxide of 45%. The excellent catalytic performance of the Co–Mo/NC-400 catalyst in water and alcohols could be for the following two reasons. On the one hand, the alcohol solvents and water should have the ability to activate the substrates or the intermediates via hydrogen-bonding interactions, similar to the hydrogenation of carbonyl groups in ethanol.²⁵ On the other hand, solvents with strong polarity resulted in a high dispersion of diphenyl sulfoxide and the Co–Mo/NC-400 catalyst, as both of them were hydrophilic. The best results were obtained in methanol, which produced diphenyl sulfide with a 100% yield at 80 °C and 10 bar H₂. In addition, the catalyst was active even at 25 °C and 10 bar, but it took 60 h to get a high yield (Table S1, entry 10†).

The excellent catalytic performance of the Co–Mo/NC-400 catalyst at 80 °C inspired us to carry out the deoxygenation of diphenyl sulfoxide under milder conditions. At a low temperature of 40 °C, it only produced 5% conversion of diphenyl sulfoxide after 2 h under 10 bar H₂ (Fig. S6†). Then the conversion greatly increased with an increase in the reaction temperature from 50 to 80 °C. A quantitative conversion of diphenyl sulfoxide was observed at 80 °C after 2 h. Then the effect of hydrogen pressure on the hydrodeoxygenation of diphenyl sulfoxide was studied over the Co–Mo/NC-400 catalyst at 50 °C. As shown in Fig. S7,† the conversion of diphenyl sulfoxide increased with an increase of hydrogen pressure, especially from 1 to 10 bar. For example, the conversion of diphenyl sulfoxide greatly increased from 5% at 1 bar to 26% at 10 bar, and then slowly increased to 33% at 20 bar. The increase in hydrogen pressure resulted in an increase in the hydrogen concentration in the reaction solution, resulting in an increase in the reaction rate. Fig. S8† depicts the time course of the product distribution of the hydrodeoxygenation of diphenyl sulfoxide over the Co–Mo/NC-400 catalyst at 50 °C and 10 bar H₂. The percentage of unreacted diphenyl sulfoxide gradually decreased during the reaction process, and the yield of diphenyl sulfide gradually increased on prolonging the reaction time. After 10 h, a full conversion of diphenyl sulfoxide was obtained at 50 °C and 10 bar H₂ with a 100% selectivity of diphenyl sulfide. It is worth noting that the Co–Mo/NC-400 catalyst demonstrated comparable or even superior activity to the reported noble metal catalysts.¹⁴ For example, a long reaction time of 24 h was required to complete the conversion of diphenyl sulfoxide over a Pt–MoO_X/TiO₂ catalyst at 50 °C under 7 bar H₂,¹⁴ while the reaction time to get full conversion of diphenyl sulfoxide was only 12 h over the Co–Mo/NC-400 catalyst at 50 °C under 10 bar H₂.

Then, the scope of the as-prepared Co–Mo/NC-400 catalyst towards the hydrodeoxygenation of various sulfoxides was investigated. A number of aromatic, benzylic and aliphatic sulfoxides were smoothly converted into the corresponding sulfides with excellent yields (Table 2, entries 2–6). Compared with diphenyl sulfoxide, substituted diphenyl sulfoxides with large steric hindrance were less active, which required a longer reaction time to achieve high conversions (Table 2, entries 1–4). Meanwhile, a substituted diphenyl sulfoxide with electron-withdrawing groups showed lower activity than counterparts (Table 2, entry 2 vs. entries 3 and 4). Phenylalkyl sulfoxides demonstrated similar activity to diphenyl sulfoxides (Table 2, entry 2 vs. entries 3 and 4), which have small steric hindrance but low electronic density of the sulfoxide groups. Dialkyl sulfoxides were also successfully converted into the corresponding alkyl sulfides, which were performed at high reaction temperature of 80 °C, due to the low activity of the dialkyl sulfoxides with low electron density of the sulfoxide groups. Notably, the Co–Mo/NC-400 catalyst was effective for the selective dehydrogenation of sulfoxide substrates without affecting other reducible functional groups such as C=C double bonds, Cl, and Br.

Table 2 Substrate scope of the hydrodeoxygenation of sulfoxides^a

Entry	Substrate	Temperature (°C)	Time (h)	Yield (%)
a Reaction conditions: Substrates (0.5 mmol), methanol (10 mL), H2 (10 bar), Co–Mo/NC-400 catalyst (20 mg).
1		50	10	100
2		50	15	100
3		50	18	100
4		50	20	100
5		50	24	100
6		50	22	100
7		50	26	100
8		80	25	94
9		80	15	98
10		80	18	85

---

Finally, recycling experiments of the Co–Mo/NC-400 catalyst were conducted. As shown in Fig. S9,† the catalyst showed similar conversions of diphenyl sulfoxide of around 58% during 6 runs, and the selectivity of diphenyl sulfide remained at 100%. These results suggested that the Co–Mo/NC-400 catalyst showed a high stability. ICP-AES analysis of the solution after the first reaction showed that the Co and Mo contents in the solution were below the detection limits.

In conclusion, we have developed a new method for the use of heterogeneous non-noble metal catalysts for the catalytic hydrogenation of sulfoxides for the first time. Notably, the as-prepared Co–Mo/NC-400 catalyst exhibited an unprecedented activity under mild reaction conditions (25–80 °C and 10 bar H₂) and greatly broadens the applicability to a wide range of sulfoxides. Control experiments demonstrate that metallic Co and Mo in the Co–Mo/NC-400 catalyst played a synergistic role in the hydrogenation of sulfoxides with H₂. The activity of the as-prepared Co–Mo/NC-400 catalyst was comparable to or even higher than that of reported noble metal catalysts.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by National Natural Science Foundation of China (no. 21203252) and Basic Scientific Research of Central Colleges, South-Central University for Nationalities (CZY 19005).

Notes and references

1. S. Sharma, D. Bhattacherjee and P. Das, Adv. Synth. Catal., 2018, 360, 2131–2137.
2. R. Hille, Chem. Rev., 1996, 96, 2757–2816.
3. D. J. Harrison, N. C. Tam, C. M. Vogels, R. F. Langler, R. T. Baker, A. Decken and S. A. Westcott, Tetrahedron Lett., 2004, 45, 8493–8496.
4. J. Zhang, X. Gao, C. Zhang, J. Luan and D. Zhao, Synth. Commun., 2010, 40, 1794–1801.
5. N. Iranpoor, H. Firouzabadi and H. R. Shaterian, J. Org. Chem., 2002, 67, 2826–2830.
6. M. Madesclaire, Tetrahedron Lett., 1988, 44, 6537–6580.
7. H. Firouzabadi and A. Jamalian, J. Sulfur Chem., 2008, 29, 53–97.
8. Y. Mikami, A. Noujima, T. Mitsudome, T. Mizugaki, K. Jitsukawa and K. Kaneda, Chem. – Eur. J., 2011, 17, 1768–1772.
9. Y. Takahashi, T. Mitsudome, T. Mizugaki, K. Jitsukawa and K. Kaneda, Chem. Lett., 2014, 43, 420–422.
10. S. C. A. Sousa, J. R. Bernardo, M. Wolff, B. Machura and A. C. Fernandes, Eur. J. Org. Chem., 2014, 1855–1859.
11. M. Bagherzadeh and S. Ghazali-Esfahani, New J. Chem., 2012, 36, 971–976.
12. K. Ogura, M. Yamashita and G. Tsuchihashi, Synthesis, 1975, 6, 385–387.
13. T. Mitsudome, Y. Takahashi, T. Mizugaki, K. Jitsukawa and K. Kaneda, Angew. Chem., Int. Ed., 2014, 53, 8348–8351.
14. A. S. Touchy, S. M. A. H. Siddiki, W. Onodera, K. Kon and K. Shimizu, Green Chem., 2016, 18, 2554–2560.
15. T. Uematsu, Y. Ogasawara, K. Suzuki, K. Yamaguchi and N. Mizuno, Catal. Sci. Technol., 2017, 7, 1912–1920.
16. S. Xu, P. Zhou, Z. H. Zhang, C. J. Yang, B. G. Zhang, K. J. Deng, S. Bottle and H. Y. Zhu, J. Am. Chem. Soc., 2017, 139, 14775–14782.
17. P. Zhou, L. Jiang, F. Wang, K. J. Deng, K. L. Lv and Z. H. Zhang, Sci. Adv., 2017, 3, e1601945.
18. X. Dai, Z. Li, Y. Ma, M. Liu, K. Du, H. Su, H. Zhuo, L. Yu, H. Sun and X. Zhang, ACS. Appl. Mater. Interfaces, 2016, 8, 6439–6448.
19. R. Ma, Y. Zhou, Y. Chen, P. Li and Q. Liu, Angew. Chem., Int. Ed., 2015, 54, 14723–14727.
20. C. B. Lu, D. Tranca, J. Zhang, F. Rodríguez Hernández, Y. Z. Su, X. D. Zhuang, F. Zhang, G. Seifert and X. L. Feng, ACS Nano, 2017, 11, 3933–3942.
21. A. C. Ferrari and J. Robertson, Phys. Rev. B: Condens. Matter Mater. Phys., 2000, 61, 14095–14107.
22. P. Zhou, C. L. Yu, L. Jiang, K. L. Lv and Z. H. Zhang, J. Catal., 2017, 352, 264–273.
23. L. Lin, Z. K. Yang, Y. F. Jiang and A. W. Xu, ACS Catal., 2016, 6, 4449–4454.
24. T. Prasomsri, T. Nimmanwudipong and Y. Román-Leshkov, Energy Environ. Sci., 2013, 6, 1732–1738.
25. R. Ouyang and D. E. Jiang, ACS Catal., 2015, 5, 6624–6629.
26. S. G. Wang, P. Zhou, L. Jiang, Z. H. Zhang, K. J. Deng, Y. H. Zhang, Y. X. Zhao, J. L. Li, S. Bottle and H. Y. Zhu, J. Catal., 2018, 368, 207–216.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c9gc02465d

‡ These authors contributed equally to this work.

---