Titania supported on silica as an efficient catalyst for deep oxidative desulfurization of a model fuel with exceptionally diluted H₂O₂

Abstract

Deep oxidative desulfurization has been carried out, achieving the production of a model fuel with minimal sulphur content (0–15 ppm). This desulfurization process has been developed in continuous flow, using a titanium supported heterogeneous catalyst and exceptionally low hydrogen peroxide concentration (0.77% w/w).

---

Due to the tight sulphur limits in transportation fuels, deep desulfurization of liquid fuels is becoming a critical topic in environmental sciences. The society at large is stepping on the road to zero sulphur fuel, therefore the implementation of effective desulfurization methods, able to replace the classical hydrodesulfurization (HDS), which is associated with harsh reaction conditions, is a crucial goal in technical chemistry.¹ An alternative to HDS is provided by adsorptive desulfurization, and the scope of this methodology is focused mainly on jet fuels.²

The most promising desulfurization methodology is undoubtedly the oxidative desulfurization (ODS), which consists of two steps: (i) the oxidation of sulphur compounds (sulphides to sulfones) and (ii) the extraction of oxidation products.³ Both reactions can be performed under mild conditions, using hydrogen peroxide as the oxidant;⁴ in particular, biphasic systems⁵ together with suitable catalysts⁶ have been investigated, providing satisfactory results. The selection and the preparation of the catalyst have obviously a prominent role in the development of an efficient ODS route; indeed a huge scientific effort has been made in testing different catalytic materials, ranging from classical mixed metal oxides⁷ to polyoxometalates,⁸ metal–organic frameworks⁹ and biomimetic materials.¹⁰ Excellent results have been achieved using titanium based catalysts, which is so far the most commonly employed metal in ODS.^11,12

Hydrogen peroxide is the oxidant of choice for ODS since it is a cheap, mild and environmentally benign reagent.¹³ Nevertheless, the manufacture, transport and handling of concentrated H₂O₂ is associated with safety issues and considerable logistical efforts, limiting the implementation of hydrogen peroxide routes in industry. A promising solution to this problem is the in situ generation of hydrogen peroxide¹⁴ and its direct combination with selective oxidation reactions, increasing the environmental and economic attractiveness of the use of H₂O₂ as a green oxidant.¹⁵

Since the concentration of in situ generated H₂O₂, e.g. by electrochemical synthesis, is typically rather low, the development of a catalytic material which is active in the presence of diluted H₂O₂ represents a new challenge that has been rarely addressed. The use of diluted hydrogen peroxide for the selective oxidation of thioanisole to its sulfoxide has been already reported;¹⁶ however, in order to perform the oxidation of refractory sulphides such as dibenzothiophene (DBT), an even more efficient catalyst has to be designed and tested. ODS has been recently used as a suitable example for the coupling of hydrogen peroxide synthesis and utilization in oxidation reactions.¹⁷ Indeed, some interesting results using tandem Pd/Ti (ref. 18) and Au/Ti (ref. 19) catalysts have been already reported.

Herein, the development of an efficient solid catalyst for the ODS of a model fuel with diluted hydrogen peroxide is described.

Considering the reported results on the use of sulfonic resins as heterogeneous catalysts,¹⁶ the first batch experiments were performed using Amberlite IR 120H as the catalyst for the oxidation of dibenzothiophene (DBT) with 3% H₂O₂. The poor solubility of the DBT oxidation products (dibenzothiophene sulfoxide and sulfone) in water necessitates the use of acetonitrile as a co-solvent in the polar phase in order to extract the sulphur from the fuel. A biphasic reaction mixture was prepared combining 3% w/w aqueous hydrogen peroxide, acetonitrile (final H₂O₂ concentration: 0.77% w/w) and the selected catalyst with a model fuel (DBT solution in dodecane, with a sulphur content of 1000 ppm). By varying the reaction parameters, with the use of tenfold excess hydrogen peroxide at 60 °C for 24 hours, Amberlite still gave 345 ppm residual sulphur in the model fuel solution in the form of DBT (see the ESI†). The water phase contains the oxidation products dibenzothiophene sulfoxide and sulfone, allowing a coherent mass balance.

On the basis of these results, a variety of heterogeneous catalysts, commercially available or prepared in our labs, were tested under the same reaction conditions; the results are presented in Table 1.

Table 1 Catalyst screening

Entry	Catalyst^a	Residual sulphur^b (ppm)
a Reaction conditions: 2 ml of DBT solution in dodecane (0.024 mol L^−1), 2 ml of acetonitrile, 250 mg of the catalyst, 0.533 ml of H2O2 solution at 3%, 60 °C, 24 h. b Determined by HPLC analysis.
1	—	770
2	Amberlite IR 120H	345
3	Nafion	464
4	Aquivion	432
5	TiO2 (rutile)	445
6	TiO2 (anatase)	242
7	Ti-SBA-15	335
8	TS-1	205
9	TS-1 B-doped	231
10	TiO2@SiO2	6

---

The HPLC results show that in the absence of any catalyst, hydrogen peroxide is not able to reduce the sulphur content, likely owing to the high dilution of the mixture. In a blank experiment without any catalyst, 770 ppm residual sulphur was detected, which indicates that partial desulfurization occurs solely due to the extraction of DBT by acetonitrile. Perfluorinated sulfonic resins such as Nafion and Aquivion were slightly less active than Amberlite. Switching to inorganic catalysts, interesting results were obtained using titanium-based materials: anatase performed significantly better than the polymorphous rutile, showing a residual sulphur content of 242 and 445 ppm, respectively. On the basis of these considerations, a series of catalysts with 3.5 wt% titanium content were tested.^12,20 Ti-SBA-15, a mesoporous silica incorporating titanium atoms, gave an unsatisfactory sulphur oxidation (335 ppm residual sulphur), whereas with TS-1 and its boron-doped counterpart,²¹ residual sulphur values similar to that of anatase were achieved (205 and 231 ppm, respectively). In order to obtain a more efficient catalyst, TiO₂ was anchored on the surface of silica by grafting of TiCl₄ and subsequent calcination at 450 °C; the prepared catalyst TiO₂@SiO₂ allowed almost complete desulfurization of the model fuel (6 ppm residual sulphur). These results suggest the need of well dispersed and accessible catalytic sites on the surface of the silica to promote the oxidation process. To further examine this hypothesis, a TiO₂@SiO₂ material was prepared, avoiding the final calcination step; the solid material, heated in an oven at 100 °C merely to remove the organic solvents, did not oxidize DBT at all. XRD analyses of both uncalcined and calcined TiO₂@SiO₂ materials provide clarifying results (Fig. 1). A completely amorphous material has been obtained by the grafting of Ti onto silica; nevertheless, the calcined catalyst showed the presence of the typical anatase peaks. These results show that a phase transition, which occurs during the calcination step, is responsible for the formation of titania nanocrystals dispersed on amorphous silica, representing the catalytic active sites.

Fig. 1 XRD spectra of anatase (blue), fresh (black) and calcined (red) TiO₂@SiO₂.

The desulfurization progress with time has been monitored to optimize the reaction parameters (Fig. 2). Surprisingly, the ODS process is almost complete in just 20 minutes. Considering that all the other materials tested were unable to attain complete oxidation even after 24 hours, these results clearly show the great potential of this catalyst. However, the quality of a heterogeneous catalyst is related to its reusability. To test its reusability, the catalyst was filtered off after the reaction, washed with acetonitrile, dried at 80 °C and reused in successive runs. The results for the first 6 runs are reported in Table 2.

Fig. 2 Reaction trend for the batch oxidation of DBT with 3% aqueous hydrogen peroxide catalysed by TiO₂@SiO₂ under optimized conditions.

Table 2 Recycling tests

Run	Residual sulphur^a (ppm)	Run	Residual sulphur^a (ppm)
a Determined by HPLC analysis.
1	5	4	424
2	234	5	425
3	404	6	421

---

A partial catalyst deactivation started already from the second run (234 ppm residual sulphur); later from the 4th to 6th runs, the residual sulphur content attained a plateau at around 420 ppm, revealing a marked decrease of the catalyst activity but not a complete inactivity. The hot filtration test²² and the ICP analysis of the solution obtained after catalyst filtration confirmed the absence of any leaching. Hence, the decrease in the activity of TiO₂@SiO₂ could be ascribable to the poisoning of titanium catalytic sites due to water coordination or to the presence of reaction products within the catalyst's pores.²³

To test whether it would be possible to avoid deactivation, the process was operated under continuous flow conditions;²⁴ the setup was assembled as schematized in Fig. 3.

Fig. 3 Schematic diagram of the setup for the continuous flow oxidation of DBT with hydrogen peroxide.

Preliminary test showed the need to homogenize the DBT and hydrogen peroxide solutions prior to entering the reactor in order to obtain a better oxidative extraction.²⁵ The reactor consists of a tubular glass column filled with TiO₂@SiO₂ dispersed with glass beads (≈1/2 w/w) heated at 60 °C. In a typical experiment, the reactor was filled with 1 gram of the catalyst, and the continuous reaction was run for 1 hour; the results are listed in Table 3.

Table 3 Continuous flow optimization of DBT oxidation with diluted aqueous hydrogen peroxide^a

Entry	Model fuel flow (mL min^−1)	AcCN/H2O2 flow (mL min^−1)	Residence time (min)	Residual sulphur^b (ppm)
a Reaction conditions: DBT solution in dodecane (0.024 mol L^−1), acetonitrile : 3% aqueous H2O2 3.75/1 v/v, 1 g of TiO2@SiO2, temperature: 60 °C, reaction time: 1 h. b Determined by HPLC analysis. c Reactor was washed with AcCN every 10 min for 10 min. d Reactor was washed with AcCN/H2O2 every 10 min for 10 min. e Reaction was carried out with silanized TiO2@SiO2 and the reactor was washed with AcCN/H2O2 every 20 min for 10 min. f Reaction was carried with 5 g of silanized TiO2@SiO2 and the reactor was washed with the polar phase every 20 min for 10 min; the result is the average of 2 experiments.
1	0.04	0.14	5	225
2	0.04	0.14	10	160
3	0.08	0.28	5	138
4^c	0.08	0.28	5	63
5^d	0.08	0.28	5	86
6^e	0.06	0.22	10	<LOD
7^f	0.16	0.58	10	15

---

By carrying out the process with a residence time of 5 minutes, modest results were achieved (a final sulphur content of 225 ppm); moreover, when either the residence time was increased up to 10 minutes or the polar flow was doubled, only modest improvements were detected (a final sulphur content of 160 and 138 ppm). Additionally, by prolonging the reaction time from 1 to 5 hours, a substantial worsening was observed (544 ppm residual sulphur), confirming that also under continuous flow conditions, poisoning of the active sites is occurring.

With the aim of restoring the initial activity of the catalyst, the reactor was purged with polar solutions, namely acetonitrile, or with the polar solution of the feed (3% aqueous hydrogen peroxide in acetonitrile – Table 3, entries 4 and 5). A remarkable decrease in the final sulphur content was observed: 63 and 86 ppm, respectively.

These data support the hypothesis of polar species (DBT oxidation products) as catalyst poisoning agents, and suggest that a less polar material could prove to be more robust. Therefore, the silanization of TiO₂@SiO₂ by treatment with ethyltrimethoxysilane was considered. This procedure allows the conversion of the free silanols on the catalyst's surface to silanes, drastically reducing the polarity and the hydrophilicity of the material. According to the described hypothesis, the tailored catalyst became more robust without losing its outstanding activity, delivering a sulphur-free model fuel (Table 3, entry 6) and requiring fewer washing steps with the polar phase (every 20 min instead of 10 min). The reduced final sulphur content (15 ppm) has been accomplished also in an up-scaled experiment (Table 3, entry 7).

Conclusions

A TiO₂@SiO₂ catalyst has been developed and successfully applied in the deep desulfurization of a model fuel. The presence of anatase nanocrystals appear to be decisive in producing an active catalyst, and additional silanization of the material substantially increases its lifetime. Achieving the ODS with 0.77% hydrogen peroxide provides a useful application for in situ generated H₂O₂, which, to date, can be produced only in such a range of concentrations. The realization of an ODS demonstrator for the direct utilization of H₂O₂, produced through an electrochemical process, is currently ongoing.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Fraunhofer Internal Program: Leitprojekt “Strom als Rohstoff”. The authors are grateful to Prof. D. Cauzzi and A. Polyzoidis for measuring and discussing XRD and B.E.T. data.

Notes and references

1. A. Stanislaus, A. Marafi and M. S. Rana, Catal. Today, 2010, 153, 1.
2. (a) J. M. Palomino, D. T. Tran, J. L. Hauser, H. Dong and S. R. J. Oliver, J. Mater. Chem. A, 2014, 2, 14890–14895; (b) J. M. Palomino, D. T. Tran, A. R. Kareh, C. Miller, J. M. V. Gardner, H. Dong and S. R. J. Oliver, J. Power Sources, 2015, 278, 141–148.
3. J. M. Campos-Martin, M. C. Capel-Sanchez, P. Perez-Presas and J. Fierro, J. Chem. Technol. Biotechnol., 2010, 85, 879.
4. F. M. Collins, A. R. Lucy and C. Sharp, J. Mol. Catal. A: Chem., 1997, 117, 397.
5. (a) D. Zhao, J. Wang and E. Zhou, Green Chem., 2007, 9, 1219; (b) L. Xia, H. Zhang, Z. Wei, Y. Jiang, L. Zhang, J. Zhao, J. Zhang, L. Dong, E. Li, L. Ruhlmann and Q. Zhang, Chem. – Eur. J., 2017, 23, 1920; (c) H. Lu, J. Gao, Z. Jiang, Y. Yang, B. Song and C. Li, Chem. Commun., 2007, 150.
6. (a) K.-G. Haw, W. A. W. A. Bakar, R. Ali, J.-F. Chong and A. A. A. Kadir, Fuel Process. Technol., 2010, 91, 1105; (b) Y. Hu, Q. He, Z. Zhang, N. Ding and B. Hu, Chem. Commun., 2011, 47, 12194; (c) P. Yang, S. Zhou, Y. Du, J. Li and J. Lei, RSC Adv., 2016, 6, 53860; (d) A. E. S. Choi, S. Roces, N. Dugos and M.-W. Wan, Sustainable Environ. Res., 2016, 26, 184.
7. J. L. García-Gutiérrez, G. A. Fuentes, M. E. Hernández-Terán, F. Murrieta, J. Navarrete and F. Jiménez-Cruz, Appl. Catal., A, 2006, 305, 15.
8. J. L. García-Gutiérrez, G. C. Laredo, P. García-Gutiérrez and F. Jiménez-Cruz, Fuel, 2014, 138, 118.
9. (a) S. Ribeiro, C. M. Granadeiro, P. Silva, F. A. Almeida Paz, F. F. de Biani, L. Cunha-Silva and S. S. Balula, Catal. Sci. Technol., 2013, 3, 2404; (b) X.-S. Wang, L. Li, J. Liang, Y.-B. Huang and R. Cao, ChemCatChem, 2017, 9, 971.
10. H. Liu, S. Bao, Z. Cai, T. Xu, N. Li, L. Wang, H. Chen, W. Lu and W. Chen, Chem. Eng. J., 2017, 317, 1092.
11. (a) S. Du, F. Li, Q. Sun, N. Wang, M. Jia and J. Yu, Chem. Commun., 2016, 52, 3368; (b) A. Corma, P. Esteve, A. Martinez and S. Valencia, J. Catal., 1995, 152, 18; (c) L. Kong, G. Li and X. Wang, Catal. Today, 2004, 93–95, 341; (d) K. Leng, X. Li, G. Ye, Y. Du, Y. Sun and W. Xu, Catal. Sci. Technol., 2016, 6, 7615; (e) Y. Qin, S. Xun, L. Zhan, Q. Lu, M. He, W. Jiang, H. Li, M. Zhang, W. Zhu and H. Li, New J. Chem., 2017, 41, 569; (f) C. Shen, Y. J. Wang, J. H. Xu and G. S. Luo, Green Chem., 2016, 18, 771; (g) S. Du, X. Chen, Q. Sun, N. Wang, M. Jia, V. Valtchev and J. Yu, Chem. Commun., 2016, 52, 3580.
12. R. Castillo, B. Koch, P. Ruiz and B. Delmon, J. Mater. Chem., 1994, 4, 903.
13. R. Noyori, M. Aoki and K. Sato, Chem. Commun., 2003, 1977.
14. R. Dittmeyer, J.-D. Grunwaldt and A. Pashkova, Catal. Today, 2015, 248, 149.
15. B. Puértolas, A. K. Hill, T. García, B. Solsona and L. Torrente-Murciano, Catal. Today, 2015, 248, 115.
16. R. Maggi, S. Chitsaz, S. Loebbecke, C. G. Piscopo, G. Sartori and M. Schwarzer, Green Chem., 2011, 13, 1121.
17. Positionspapier Katalytische Oxidationsreaktionen als Schlüsseltechnologie, ed. Dechema, 2015.
18. S. K. Karmee, L. Greiner, A. Kraynov, T. E. Muller, B. Niemeijer and W. Leitner, Chem. Commun., 2010, 46, 6705.
19. (a) S. Ma, G. Li and X. Wang, Chem. Eng. J., 2010, 156, 532; (b) H. Song, G. Li, X. Wang and Y. Chen, Microporous Mesoporous Mater., 2011, 139, 104.
20. A. Y. Shan, T. I. M. Ghazi and S. A. Rashid, Appl. Catal., A, 2010, 389, 1.
21. D. T. On, S. Kaliaguine and L. Bonneviot, J. Catal., 1995, 157, 235.
22. H. Lempers and R. A. Sheldon, J. Catal., 1998, 175, 62.
23. (a) A. M. Cojocariu, P. H. Mutin, E. Dumitriu, F. Fajula, A. Vioux and V. Hulea, Appl. Catal., B, 2010, 97, 407; (b) X. Kong, C. Zeng, X. Wang, J. Huang, C. Li, J. Fei, J. Li and Q. Feng, Sci. Rep., 2016, 6, 29049 , EP -.
24. H. P. L. Gemoets, Y. Su, M. Shang, V. Hessel, R. Luque and T. Noel, Chem. Soc. Rev., 2016, 45, 83.
25. S. Doherty, J. G. Knight, M. A. Carroll, J. R. Ellison, S. J. Hobson, S. Stevens, C. Hardacre and P. Goodrich, Green Chem., 2015, 17, 1559.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c7re00192d

---