Probing the Mechanism of Oxidative Desulfurization of Crude Oils by Mass Spectrometry Employing Synthesized Sulfur-Containing Imidazolium Salts

Abstract

Extractive and catalytic oxidative desulfurization (ECODS) has emerged as an attractive alternative to hydrodesulfurization for producing ultra-low-sulfur fuels under mild conditions. Although methyltrioxorhenium (MTO)/H₂O₂ systems effectively facilitate sulfur removal from aromatic sulfur compounds, the detailed reaction mechanism for S-removal remains largely inferred just from product analysis. Here, we synthesized a charge-tagged thiophene derivative and used it as a mechanistic probe to study the MTO-catalyzed oxidation of thiophene by hydrogen peroxide. Real-time monitoring with electrospray ionization high-resolution mass spectrometry (ESI-HRMS) enabled us to directly intercept and characterize key intermediates formed during the reaction. Sulfoxide and sulfone derivatives, as well as homoand heterodimeric cycloadducts produced through Diels-Alder reactions, were detected and structurally characterized using accurate-mass measurements and tandem mass spectrometry. Time-resolved analysis showed rapid conversion of oxidized thiophene intermediates into cycloadducts, followed by sulfur extrusion via SO and SO₂ eliminations, leading to desulfurized products. Based on the experimentally characterized intermediates, we propose a revised mechanism that extends the mechanism previously reported., now incorporating direct evidence for transient oxidation and cycloaddition species in ECODS systems. These findings deepen our understanding of oxidative desulfurization processes and could support the development of more efficient technologies and catalysts to produce cleaner fuels.

Keywords:
oxidative desulfurization; atmospheric pollution; reaction mechanism; mass spectrometry; reaction monitoring

Introduction

Petroleum is still a major yet non-renewable chemical energy source, being mainly composed of a complex mixture of fossil-derived hydrocarbons.¹,² Hydrocarbon combustion leads to CO₂ and H₂O, but the combustion of other petroleum derivative components poses serious threats to the environment since it releases various air pollutants to the atmosphere. Aromatic sulfur molecules, such as thiophene (T), benzothiophenes (BT), dibenzothiophene (DBT), benzonaphthothiophene (BNT) and its alkylated derivatives, stand out as major atmospheric pollutants originating from crude oil.³-⁵

Currently, the petroleum industry is investing in new sulfur removal technologies to reduce costs while improving desulfurization efficiency to meet increasingly stringent fuel sulfur specifications (< 15 ppm sulfur).⁵

Several desulfurization methods are available for liquid fuels, but hydrodesulfurization (HDS) is the most widely used and classic.⁵ However, HDS operates at high pressure (50 atm) and temperature (350 ºC), generating hydrogen sulfide (H₂S) (as a degradation product), a highly toxic, corrosive, and environmentally harmful by-product.⁵

Oxidative desulfurization increases the polarity of sulfur compounds, facilitating their separation by extraction with polar solvents.⁶,⁷ Oxidation of the sulfur atom increases polarity, allowing separation using polar extractants.⁶,⁷

Zhou et al.⁸ demonstrated the use of MTO as an effective catalyst for the extractive and catalytic oxidative desulfurization (ECODS) of model crude oils and gasoline. Using 30% H₂O₂ as the oxidant to generate the key oxidant species (Scheme 1) and an ionic liquid as the solvent and extractant, the authors were able to remove organic sulfur at temperatures below 60 ºC. Several other oxidation catalysts employing vanadium,⁹ rhenium,¹⁰ tungsten,¹¹ and molybdenum¹² plus H₂O₂, have also been used for desulfurization.

Scheme 1
Reaction between methyltrioxorhenium (MTO) and H₂O₂.

In this reaction medium, monoperoxorhenium and bisperoxorhenium complexes are proposed to form (via cycloaddition) and act as catalysts for the oxidation of aromatic sulfur compounds, including thiophene (Scheme 2), thioethers, and alkyl thiophenes.

Scheme 2
Oxidation of thiophene by H₂O₂ catalyzed by MTO.

Scheme 3 illustrates the proposed bicyclic catalytic cycle for desulfurization, as suggested by Zhou et al.⁸ Because thiophene is distributed between the oil and ionic liquid (IL) phases, its oxidation can occur in both media. The resulting oxidized products, owing to their higher polarity, are preferentially extracted into the IL phase. In the presence of excess H₂O₂, thiophene 1-oxide is further oxidized to thiophene 1,1-dioxide, which subsequently undergoes cycloaddition to form the products shown in Scheme 2.

Scheme 3
Zhou et al.⁸ mechanistic proposal for thiophene oxidation by MTO/H₂O₂.

Given its importance in reducing a major environmental threat, and to further improve the key reaction proposed in Scheme 3, or even to find better alternatives, it is essential to elucidate its reaction mechanism. Here, a mass spectrometric (MS) investigation of the MTO-catalyzed desulfurization process using a charge-tagged thiophene derivative as a mechanistic probe is presented. Mass spectrometry, particularly when coupled with electrospray ionization (ESI), has become a powerful tool for mechanistic studies.¹³-¹⁷ Its power in this field arises from its unique ability to provide continuous snapshots of the ionic composition of reaction solutions. Metal-catalyzed organic reactions have been extensively investigated, with catalytic intermediates successfully characterized by ESI-MS.¹⁸-²⁵

The proposed mechanism by Zhou et al.⁸ (Scheme 3) therefore to fit well with ESI-MS monitoring using charge-tagged reagents, mainly because of the expected ionic nature of all key intermediates, which should facilitate their ejection from the reaction solution to the gas phase. This charge-tagged monitoring should provide a comprehensive view of the whole process, including key intermediates.

Experimental

Materials and reagents

The following equipment was used: round-bottom flasks (5, 100, and 250 mL), magnetic stir bars, separatory funnels, vacuum drying vessels, temperature-controlled magnetic hot-plate stirrer, analytical balance, micropipettes with pipette tips, and glass vials, and NMR tubes. 2-Methyl-1-(2’-thiophenylmethyl)imidazolium hexafluorophosphate was synthesized according to the procedure described in the Supplementary Information (SI) section. Methyltrioxorhenium (98%), hydrogen peroxide (30 wt.% in water), methanol (HPLC grade), and formic acid (95%) were used as received, 2-formylthiophene (≥ 98%), sodium borohydride (≥ 98%), thionyl chloride (≥ 99%), triethylamine (≥ 99%), 1-methylimidazole (≥ 99%), potassium hexafluorophosphate (≥ 98%), deuterated chloroform (CDCl₃, 99.8 atom% D) and deuterated dichloromethane (CD₂Cl₂, 99.8 atom% D).

Oxidation reaction conditions

A 5 mL round-bottom flask was charged with 2-methyl-1-(2’-thiophenylmethyl)imidazolium hexa-fluoro-phosphate (6.2 × 10^-5 mol), MTO (98%, 2.46 × 10^-6 mol), and a 30 wt.% aqueous hydrogen peroxide solution (0.6 mL). The reaction mixture was stirred at 60 °C for 3 h. During this period, aliquots were collected at 30 min intervals. Each aliquot (1.0 μL) was diluted in methanol containing 0.1% formic acid and directly analyzed by positive-ion ESI-MS.

NMR experiments

Nuclear magnetic resonance (NMR) spectra were recorded on a Bruker Avance III 500 MHz spectrometer (Bruker, Germany) equipped with a 5 mm broadband observe (PABBO BB) probe. The operating frequencies were 500 MHz for ¹H NMR and 125 MHz for ¹³C NMR.

Spectra were acquired at 298 K using tetramethyl-silane (TMS) as the internal reference (d 0.00 ppm). Samples were prepared by dissolving 5-10 mg of the compounds in 0.6 mL of deuterated solvent and transferred to standard 5 mm NMR tubes. Compounds 1 and 2 were analyzed in CDCl₃ (99.8 atom% D), while compound 4 was analyzed in CD₂Cl₂ (99.8 atom% D). ¹H and ¹³C NMR spectra were acquired using standard Bruker pulse sequences (zg30 and zgpg30, respectively). Spectral processing included phase and baseline correction, zero-filling, and exponential multiplication (line broadening of 0.30 Hz for ¹H and 1.0 Hz for ¹³C NMR). Chemical shifts were referenced to TMS or residual solvent signals.

MS experiments

Mass spectrometry (MS) analyses were performed using a Q-Exactive™ mass spectrometer (Thermo Scientific, Bremen, Germany), a hybrid quadrupole-orbitrap (Q-Orbitrap) instrument equipped with an electrospray ionization (ESI) source. Spectra were acquired at a resolving power of 140,000 (full width at half maximum (FWHM) at m/z 200). Major operating conditions were as follows: spray voltage of 5 kV, sheath gas flow rate of 10 arbitrary units, auxiliary gas temperature of 60 °C, capillary temperature of 320 °C, sample infusion flow rate of 5 μL min-¹, automatic gain control (AGC) target of 1 × 10⁶, and S-lens RF (radio frequency) level of 70. The instrument was operated in full-scan mode over the m/z range 50-750.

Tandem mass spectrometry (ESI(+)-MS/MS) experiments were performed using an isolation window of 0.4 Da, collision energies ranging from 10 to 40 eV, and argon as the collision gas.

Samples were prepared by diluting 1.0 μL of the reaction mixture aliquot or solution of synthesized compound 4 in 1.0 mL of methanol containing 0.1% formic acid. The resulting solutions were directly introduced into the mass spectrometer by direct infusion under positive ion mode. Spectra processing was carried out using Xcalibur™ software (Thermo Scientific).

Results and Discussion

Thiophene 1 was synthesized in four steps (Scheme 4). First, 2-formyl thiophene was reduced to 2-(hydroxymethyl)thiophene, followed by chlorination and subsequent nucleophilic substitution of the chloride by methyl imidazole, with concomitant exchange of the Cl^- counter ion by PF₆^-. The identity of compound 1 was confirmed by ESI(+)-HRMS analysis (Figure S7a, in the SI section). The spectrum of the synthesized compound showed the expected molecular ion at m/z 179.0638, in agreement with the calculated exact mass (Figure S7b, in the SI section), confirming the successful formation of product 1.

Scheme 4
Steps for the synthesis of labeled thiophene using PF₆^- as a counter ion.

Then, the oxidation of product 1 by H₂O₂ in the absence of MTO was monitored by ESI(+)-MS. No significant spectral changes were observed, and the ion at m/z 179.06 remained the predominant species throughout the 3 h monitoring period. These findings indicate that charge labeled thiophene 1 is inert toward H₂O₂ under the reaction conditions employed.

Next, the MTO-catalyzed reaction (Figure 1) was monitored. A typical ESI(+)-MS spectrum was recorded at t = 60 min, in which three main reaction intermediates were detected and characterized by MS analysis. The singly charged ions at m/z 195.1 and 211.1 correspond to the sulfoxide 2 (Scheme 5) and sulfone derivative 3, whereas the doubly charged ion at m/z 203.1 corresponds to the heterodimer 6, that is, [4 + 2] Diels-Alder cycloadduct between compounds 2 and 3.

Figure 1
ESI(+)-MS spectra of the reaction solution at (a) t = 0 min and (b) t = 60 min.

Figure 2
Trends in relative abundances of reagent and products for the desulfurization reaction of charge-tagged probe 1 with H₂O₂ catalyzed by MTO. Relative abundances were calculated by normalizing the summed abundances of all monitored ions at each time point to 100%.

The doubly charged homodimers 7 (m/z 195.1) and 4 (m/z 211.1) are not readily apparent in the ESI(+)-MS shown in Figure 1 because exhibit the same nominal m/z values as their singly charged precursors (Scheme 5). However, because they are doubly charged, their ¹³C isotopologs display the characteristic 0.5 Da spacing relative to the corresponding ¹2C isotopologs, allowing their unequivocal identification.

An additional highly abundant product ion 5 (m/z 179.063), isobaric with compound 1, was also detected (Figure 1). This ion results from SO₂ loss through a retro-Diels-Alder reaction of the doubly charged heterodimer 4 (m/z 179.071) and could be distinguished from the starting reagent 1 (m/z 179.063), as shown in the expanded region of Figure S8 (in the SI section). The structural distinction between these two isobaric ions was further confirmed by MS/MS experiments performed at t = 0 min and t = 120 min (Figures S11 and S12, in the SI section, respectively). The distinct dissociation patterns were consistent with the assigning of ion 5 as the SO₂^- eliminated cycloadduct rather than the starting reagent 1.

By monitoring the reaction from t = 0 to t = 180 min, the relative abundances of the starting reagent and the intercepted products were also plotted as a function of time (Figure 2). It should be noted that ion abundances observed by ESI(+)-MS do not necessarily correlate with their actual concentrations in solution. However, the charge tags of the ions are expected to promote to promote rapid and relatively uniform ionization, resulting in similar ionization yields for all species.

Scheme 5
Proposed mechanism for the oxidation of thiophene by MTO/H₂O₂ (as revealed by its charge tagged analogue 1) and ESI(+)-MS monitoring.

As Figure 2 shows, the relative abundances of the first reaction intermediates, sulfoxide 2 (m/z 195.1) and sulfone 3 (m/z 211.1), remain low and nearly constant throughout the 180 min reaction-monitoring period. This trend suggests a high rate of conversion of compound 2 to 3, and then, to the homodimers 4 (m/z 211.1) and 7 (m/z 195.1) and the heterodimer 6 (m/z 203.0) (Scheme 5). Indeed, this is the expected, since oxidation of thiophenes leads to the loss of aromaticity, thereby increasing their reactivity. The relatively high concentration of compound 6 (m/z 203.0), as well as the sharp increase in compound 5 (m/z 179.0), indicates that the SO₂^- eliminated cycloadduct 5 is the final product, whereas oxidation of compound 6 appears to be the rate-limiting step. The formation of the homodimers (Figures S13 and S15, in the SI section) and heterodimers (Figure S14, in the SI section) was further supported by MS/MS performed on the corresponding doubly charged ¹3C-isotopologues of 7 (m/z 195.6) and 4 (m/z 211.6), which exhibited diagnostic dissociation chemistries consistent with dimeric structures.

Based on the intercepted intermediates for the reaction of the charged-tagged thiophene 1, a new and detailed mechanism for the desulfurization of its analogue thiophene with hydrogen peroxide catalyzed by MTO was proposed in Scheme 5.

In Scheme 5, thiophene (as represented by compound 1) is first converted to its sulfoxide 2, and then to its sulfone derivative 3. From compound 3, two divergent pathways, involving either compound 4 or compound 6, as intermediates, subsequently converge to form the same final desulfurized product, compound 5. For both compounds 4 and 6, the final product 5 is formed by retro-Diels-Alder reactions with the elimination of SO₂ from compound 4 or SO from compound 6.²⁵ A third pathway to compound 5 also diverges from compound 2 through a [4 + 2] Diels-Alder cycloaddition, forming the homodimer 7, which subsequently undergoes a retro-Diels-Alder reaction with loss of SO to afford compound 8 (m/z 171). It should be noted that compound 8 (m/z 171.1) was also intercepted at t = 60 min as a minor ion (Figure 1), in agreement with the known reactivity of oxidized thiophene systems in cycloaddition and fragmentation pathways.²⁶ According to Tashiro and co-workers,²⁷ the sulfoxide heterodimer 7 is inert towards oxidation, hence it may form the final product via SO loss followed by oxidation. Although the charge-tagged probe 1 was used as a model, similar mechanisms are expected to operate for other aromatic sulfur compounds.

In contrast to the findings of Zhou et al.,⁸ trimers were not detected in our ESI(+)-MS monitoring experiments. This discrepancy may be attributed to the use of the charge-tagged reagent 1, despite the advantages conferred by the charge tag for ESI-MS monitoring. Charge-tagged trimers would be expected to be disfavored due to the strong Coulombic repulsion associated with triply charged species.

Conclusions

Using the charge-tagged thiophene 1 and ESI(+) MS monitoring, a detailed mechanism for the oxidation of thiophene by MTO/H₂O₂ is proposed, substantially expanding the mechanism previously suggested by Zhou et al.⁸ Whereas the mechanism of Zhou et al.⁸ was based largely on the structures of final oxidation products, the mechanism proposed here is primarily based on key intermediates intercepted directly from the reaction solution by ESI(+) and further characterized by MS and MS/MS experiments. These species include key transient intermediates, such as sulfoxides, sulfones, and cycloadducts, providing experimental evidence for sequential oxidation-cycloaddition-desulfurization sequences. They also reveal the coexistence of multiple reaction channels before sulfur extrusion. Our study also highlights the extreme value of charge-tagged probes for elucidating reaction mechanisms by ESI-MS.

Supplementary Information

Supplementary Information (spectra and additional experimental data) is available free of charge at http://jbcs.sbq.org.br as PDF file.

Supplementary PDF

Supplementary PDF

Acknowledgments

B. R. S. P., M. N. E., P. H. V., and T. L. acknowledge FAPESP, CNPq, MackPesquisa and CAPES for financial support.

Data Availability Statement

The data supporting the findings of this study are available within the article and its Supporting Information. The raw mass spectrometry data files (.raw) generated during this study are available from the corresponding author upon reasonable request.

References

• ¹ Gary, J. H.; Handwerk, G. E.; Petroleum Refining: Technology and Economics, 4^th ed.; Marcel Dekker: New York, USA, 2001, p. 7.
• ² Syntyhaki, E.; Karonis, D.; Anal. Lett. 2020, 54, 1470. [Crossref]
  » Crossref
• ³ Mello, M. I. S.; Sobrinho, E. V.; Pergher, S. B. C.; Perspectiva 2012, 36, 85. [Link] accessed in July 2026
  » Link
• ⁴ Kropp, K. G.; Anderson, J. T.; Fedorak, P. M.; Environ. Sci. Technol. 1994, 28, 1348. [Crossref]
  » Crossref
• ⁵ Samokhvalov, A.; Catal. Rev. 2012, 54, 281. [Crossref]
  » Crossref
• ⁶ Adam, F.; Muller, H.; Al-Hajji, A.; Bourane, A.; Koseoglu, O.; Energy Fuels 2015, 29, 2312. [Crossref]
  » Crossref
• ⁷ Eseva, E. A.; Akopyan, A. V.; Anisimov, A. V.; Maksimov, A. L.; Pet. Chem. 2020, 60, 979. [Crossref]
  » Crossref
• ⁸ Zhou, M.; Meng, W.; Li, Y.; Wang, Q.; Li, X.; Zang, S.; Energy Fuels 2013, 28, 516. [Crossref]
  » Crossref
• ⁹ Ge, J. H.; Zhou, Y. M.; Yang, Y.; Xue, M. W.; China Pet. Process. Petrochem. Technol. 2012, 14, 25. [Link] accessed in July 2026
  » Link
• ¹⁰ Shojaei, A. F.; Rezvani, M. A.; Loghmani, M. H.; Fuel Process. Technol. 2014, 118, 1. [Crossref]
  » Crossref
• ¹¹ Jiang, X.; Li, H. M.; Zhu, W. S.; He, L. N.; Shu, H. M.; Lu, J. D.; Fuel 2009, 88, 431. [Crossref]
  » Crossref
• ¹² Zhu, W. S.; Zhu, G. P.; Li, H. M.; Chao, Y. H.; Chang, Y. H.; Chen, G. Y.; Han, C. R.; J. Mol. Catal. A: Chem. 2011, 347, 8. [Crossref]
  » Crossref
• ¹³ Santos, L. S.; Reactive Intermediates: MS Investigations in Solution; Wiley-VCH: Weinheim, Germany, 2010.
• ¹⁴ Santos, L. S. J. Braz. Chem. Soc. 2011, 22, 1827. [Link] accessed in July 2026
  » Link
• ¹⁵ De la Mora, J. F.; Van Berkel, G. J.; Enke, C. G.; Cole, R. B.; Martinez-Sanchez, M.; Fenn, J. B.; J. Mass Spectrom. 2000, 35, 939. [Crossref]
  » Crossref
• ¹⁶ Moraes, M. C. B.; do Lago, C. L.; Quim. Nova 2003, 26, 556. [Crossref]
  » Crossref
• ¹⁷ Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M.; Mass Spectrom. Rev. 1990, 9, 37. [Crossref]
  » Crossref
• ¹⁸ Dias, H. J.; Santos, W. H.; Filho, L. C. S.; Crevelin, E. J.; Mcindoe, J. S.; Vessecchi, R.; Crotti, A. E. M.; J. Mass Spectrom. 2024, 59, e5062. [Crossref]
  » Crossref
• ¹⁹ Huang, M. Z.; Lee, C. Y.; Chen, K. L.; Lin, S. Y.; Wu, C. M.; Shiea, J.; Green Anal. Chem. 2025, 13, 100269. [Crossref]
  » Crossref
• ²⁰ Ai, W. P.; Song, S. Y.; Bai, Y.; Liu, H. W.; Chem. J. Chinese Univ. 2020, 42, 2598. [Crossref]
  » Crossref
• ²¹ Zhu, W.; Yuan, Y.; Zhou, P.; Zeng, L.; Wang, H.; Tang, L.; Guo, B.; Chen, B.; Molecules 2012, 17, 11507. [Crossref]
  » Crossref
• ²² dos Santos, M. R.; Gomes, A. F.; Gozzo, F. C.; Suarez, P. A. Z.; Neto, B. A. D.; ChemSusChem 2012, 5, 2383. [Crossref]
  » Crossref
• ²³ Santos, L. S.; Metzger, J. O.; Rapid Commun. Mass Spectrom. 2008, 22, 898. [Crossref]
  » Crossref
• ²⁴ Termopoli, V.; Torrisi, E.; Famiglini, G.; Palma, P.; Zappia, G.; Cappiello, A.; Vandergrift, G. W.; Zvekic, M.; Krogh, E. T.; Gill, C. G.; Anal. Chem. 2019, 91, 11916. [Crossref]
  » Crossref
• ²⁵ Treiber, A.; Dansette, P. M.; El Amri, H.; Girault, J.-P.; Ginderow, D.; Mornon, J.-P.; Mansuy, D.; J. Am. Chem. Soc. 1997, 119, 1565. [Crossref]
  » Crossref
• ²⁶ Nakayama, J.; Sugihara, Y. In Topics in Current Chemistry, vol. 205; Page, P. C. B., ed.; Publisher NameSpringer: Berlin, Heidelberg 1999, p. 131-195. [Crossref]
  » Crossref
• ²⁷ Li, Y.; Thiemann, T.; Sawada, T.; Mataka, S.; Tashiro, M.; J. Org. Chem.1997, 62, 7926. [Crossref]
  » Crossref

Edited by

• Editor handled this article:
  Giovanni Wilson Amarante (Executive)

Publication Dates

• Publication in this collection
  10 Aug 2026
• Date of issue
  2026

History

• Received
  20 Apr 2026
• Accepted
  21 July 2026