Deep oxidative desulfurization catalyzed by an ionic liquid-type peroxotungsten catalyst

Abstract

New peroxotungsten anion-based ionic liquid-type catalysts were synthesized and characterized using NMR, IR, TG, etc. Then, these salts were used as catalysts for the deep desulfurization of model oil containing dibenzothiophene (DBT) with H₂O₂ as an oxidant. The effects of the temperature, H₂O₂/DBT molar ratio, and catalyst loading on the desulfurization activity have been investigated in detail. The present catalysts showed excellent desulfurization activity under solvent-free and mild reaction conditions. Moreover, this catalyst could be recycled at least ten times without any decrease in activity. It was found that the H-bonding interaction between the H2-proton in imidazolium and the basic S atom in the DBT molecule played a very important role in enhancing the catalytic activity. On the basis of the activity measurements and characterization of the catalyst, the reaction mechanism for the oxidative desulfurization process has been suggested.

---

1 Introduction

With the development of the automotive industry, the desulfurization of fuels has been attracting more and more attention as a topic in both industry and environmental protection. The combustion of sulfur-containing compounds in transportation fuels is the main reason for the formation of smog, sour gases, and acid rain. Therefore, the development of novel and green processes to convert these sulfur-containing compounds into different innocuous forms has been attracting more and more attention around the world.^1,2

Up to now, the traditional process for the desulfurization of fuels in industry has been hydrodesulfurization (HDS), which is an efficient method for removing thiols, sulfides, and disulfides from fuels, but less effective for dibenzothiophene (DBT) and its derivatives.^3–5 Additionally, HDS requires harsh reaction conditions, such as high temperature and high pressure with a suitable catalyst. In order to overcome these problems, oxidative desulfurization (ODS) has been explored intensively over the past decades, since the reaction is highly efficient for the removal of sulfur compounds under mild conditions. Reported oxidative desulfurization catalysts include peroxometalate,^6–8 metal oxide hybrids,^9–11 ionic liquid,^12–14 molecular sieves^15–17 and so on. In particular, peroxometalate anion-based catalysts have tended to be more attractive due to their high reactivities, selectivities and stabilities for H₂O₂-based ODS in the past years.

It is known that ionic liquids possess many desirable properties, such as high thermal stability, non-volatility, and good solubility.^18,19 They also show good extraction ability for aromatic sulfur-containing compounds, and are immiscible with aliphatic liquids, such as fuel oil. Therefore, they can effectively eliminate further environmental and safety problems.^20,21 In recent years, some peroxometalate anion-based salts have been used as catalysts, and ionic liquid as the solvent or extractant, for the desulfurization reaction.^22–25 In this regard, three hexatungstates dissolved in hydrophobic 1-octyl-3-methylimidazolium hexafluorophophoric ([Omim]PF₆) ionic liquid, forming a water-in-IL emulsion system with H₂O₂, have been reported.²⁶ This desulfurization system could achieve a high conversion value with good recyclability. In addition, a lanthanide-containing polyoxometalate has also been used as a desulfurization catalyst with 30% H₂O₂ as the oxidant, and ionic liquid as an extraction solvent.^27,28 In most previous research, a hydrophobic ionic liquid was normally needed to extract sulfur-containing compounds in these systems. Up to now, there have been few reports about peroxometalate anion-based catalysts applied in the desulfurization reaction without any other extractants. In this regard, several surfactant-type polyoxometalate-based ionic liquids have been synthesized and their oxidative desulfurization of model oil by using H₂O₂ as the oxidant without other extractants has been investigated.²⁹ Besides, an amphiphilic catalyst composed of a peroxotungsten anion and quaternary ammonium cation has also been reported for the selective oxidation of S-containing molecules present in diesel to sulfones in W/O emulsion systems.³⁰ These pieces of work offered an alternative for the oxidative desulfurization of actual pre-hydrotreated fuel. Although the desulfurization reaction could be performed under mild conditions with an excellent desulfurization efficiency for dibenzothiophene and its derivatives, an emulsion was formed in these catalytic systems, which could result in difficulty in the separation of oil after the reaction. Thus, there is still much room for the improvement of oxidative desulfurization catalysts. For example, developing easier procedures for the recovery of the catalyst and methods of carrying out the desulfurization reaction under milder and solvent-free reaction conditions would be highly desirable.

Our group previously employed peroxometalate anion-based ionic liquids as catalysts for the epoxidation of olefins with good conversion and selectivity.^31–34 Based on these previous investigations, in this paper we synthesized and characterized novel ionic liquid-type catalysts, which were composed of tungsten peroxo anion complexes and alkyl imidazolium cations. Then the novel catalysts were employed in oxidative desulfurization without any other extractants. It was found that there was a hydrogen bonding interaction between DBT and the imidazolium cation, which promoted the oxidation of DBT molecules under mild reaction conditions. Moreover, the catalyst could be reused conveniently at least ten times without any significant decrease in its activity.

2 Experimental

2.1 Materials

All solvents (A.R. grade) were dried using the standard methods. Commercially available H₂O₂ (30% in water), ethyl acetate, ethanol, 1-octane, ether and 1-methyl imidazole were purchased from Sinopharm Chemical Reagent Co. Ltd. Na₂WO₄·2H₂O, 2-picolinic acid, dibenzothiophene (DBT), cetyltrimethylammonium chloride, tetrabutylammonium chloride, 1-dodecyl chloride and 1-hexadecyl chloride were purchased from Sigma-Aldrich and used without further purification.

2.2 Method

All manipulations involving air-sensitive materials were carried out using standard Schlenk line techniques under an atmosphere of nitrogen. All NMR spectra were recorded on a Bruker Avance III 400 instrument (400 MHz for ¹H) by using CDCl₃ as the solvent and TMS as the reference. Chemical shifts (δ) are given in parts per million and the coupling constants (J) in hertz. The elemental analysis of C, H and N was performed on an Elementar Vario EI III Elementa and ICP-AES analysis of W on a Vanan 710 instrument. FT-IR spectra were recorded at room temperature on a Nicolet Fourier transform infrared spectrometer (Magna 550). A Perkin Elmer Pyris Diamond was used in the current study for the thermogravimetric analysis (TGA) measurements. A constant heating rate of 10 °C min⁻¹ was used under a N₂ atmosphere. The products were analyzed using a Shimadzu GC-2014 equipped with a HP-5 column (30 m, 0.25 mm i.d.) and an FID detector.

2.3 Catalyst preparation

Synthesis of ionic liquid [C₁₂mim][Cl], [C₁₆mim][Cl]. The 1-dodecyl-3-methylimidazolium chloride ([C₁₂mim][Cl]) was synthesized according to the conventional method.³⁵ Equimolar amounts of 1-methyl imidazole and 1-dodecyl chloride were added to an autoclave with stirring at 90 °C for 48 h under a 0.7 MPa N₂ atmosphere. After the reaction, the crude products were washed with ethyl acetate and dried under vacuum at 60 °C for 2 h. ¹H NMR (400 MHz, CDCl₃) δ = 10.8 (s, H, CH), 7.5 (s, H, CH), 7.3 (s, H, CH), 4.3 (t, 2H, CH₂), 4.1 (t, 3H, CH₃), 1.9 (s, 2H, CH), 1.3 (m, 18H, CH₂), 0.9 (t, 3H, CH₃).

[C₁₆mim][Cl] was synthesized using the same procedure as that for [C₁₂mim][Cl], except that 1-dodecyl chloride was replaced with 1-hexadecyl chloride. ¹H NMR (400 MHz, CDCl₃) δ = 10.7 (s, H, CH), 7.4 (s, H, CH), 7.3 (s, H, CH), 4.3 (t, 2H, CH₂), 4.1 (t, 3H, CH₃), 1.9 (s, 2H, CH), 1.3 (m, 26H, CH₂), 0.9 (t, 3H, CH₃).

Synthesis of [C₁₂mim][W(O)(O)₂(C₅H₄NCO₂)] ([C₁₂mim][PyW]) and [C₁₆mim][W(O)(O)₂(C₅H₄NCO₂)] ([C₁₆mim][PyW]). The catalysts were prepared conveniently using a one-pot method. A solution of 2-picolinic acid (0.74 g, 6 mmol) in 1 mL of water was added to a solution of tungstic acid (1.5 g, 6 mmol) dissolved in 7.5 mL of 30% hydrogen peroxide chilled to 0 °C. The resulting solution was then stirred overnight at room temperature to afford tungsten peroxo complexes.³⁶ Then a solution of [C₁₂mim][Cl] in 3 mL of ethanol was added dropwise into the above aqueous solution. A white precipitate was formed immediately. After stirring for 24 h at room temperature, the precipitate was filtered off and washed with water and ethanol, and then dried under vacuum to afford a slightly yellow powder. (Yield: 80%). Mp = 76.5 °C. ¹H NMR (400 MHz, CDCl₃) δ = 9.1 (s, H, CH), 8.4 (d, H, CH), 8.1 (d, H, CH), 7.9 (t, H, CH), 7.5 (t, H, CH), 7.4 (s, H, CH), 7.3 (s, H, CH), 4.3 (t, 2H, CH₂), 4.1 (t, 3H, CH₃), 1.9 (s, 2H, CH), 1.3 (m, 18H, CH₂), 0.9 (t, 3H, CH₃). (Found: C, 40.63; H, 5.35; N, 6.20; W, 25.87%; C₂₂H₃₆N₃O₇W requires C, 41.39; H, 5.68; N, 6.20; W, 28.80%).

[C₁₆mim][PyW] was also synthesized in a similar way to [C₁₂mim][PyW], except that the solution of [C₁₂mim][Cl] was replaced with [C₁₆mim][Cl]. (Yield: 90%). Mp = 82.3 °C. ¹H NMR (400 MHz, CDCl₃) δ = 9.1 (s, H, CH), 8.4 (d, H, CH), 8.1 (d, H, CH), 7.9 (t, H, CH), 7.5 (t, H, CH), 7.3 (s, H, CH), 7.2 (s, H, CH), 4.3 (t, 2H, CH₂), 4.1 (t, 3H, CH₃), 1.9 (s, 2H, CH), 1.3 (m, 26H, CH₂), 0.9 (t, 3H, CH₃). (Found: C, 44.39; H, 5.72; N, 5.88; W, 23.74%; C₂₆H₄₄N₃O₇W requires C, 44.97; H, 6.39; N, 6.05; W, 26.47%).

Synthesis of [CTA][W(O)(O)₂(C₅H₄NCO₂)] ([CTA][PyW]) and [TBA][W(O)(O)₂(C₅H₄NCO₂)] ([TBA][PyW]). [CTA][PyW] was prepared by the same procedure with cetyltrimethylammonium chloride (CTAC) as the precipitant. (Yield: 90%). Mp = 84.2 °C. ¹H NMR (400 MHz, CDCl₃) δ = 8.4 (d, H, CH), 8.1 (d, H, CH), 7.9 (t, H, CH), 7.5 (t, H, CH), 3.4 (t, 2H, CH₂), 3.3 (s, 9H, CH₃), 1.8 (s, 2H, CH), 1.3 (m, 26H, CH₂), 0.9 (t, 3H, CH₃). (Found: C, 43.82; H, 6.89; N, 3.80; W, 24.80%; C₂₅H₄₇N₂O₇W requires C, 44.72; H, 7.05; N, 4.17; W, 27.38%).

[TBA][PyW] was prepared by the same procedure with tetrabutylammonium chloride (TBAC) as the precipitant. (Yield: 80%). Mp = 75.1 °C .¹H NMR (400 MHz, CDCl₃) δ = 8.4 (d, H, CH), 8.1 (d, H, CH), 7.9 (t, H, CH), 7.5 (t, H, CH), 3.3 (m, 8H, CH₂), 1.7 (m, 8H, CH₂), 1.8 (m, 8H, CH₂), 1.0 (t, 9H, CH₃). (Found: C, 37.90; H, 6.01; N, 3.94; W, 25.80%; C₂₂H₄₁N₂O₇W requires C, 41.98; H, 6.57; N, 4.45; W, 29.21%).

2.4 Typical procedure for the oxidative desulfurization and epoxidation reactions

A solution of DBT in n-octane was used as model oil with a S-content 1000 ppm. The oxidation reaction was carried out in a 50 mL Schlenk flask equipped with a reflux condenser and a thermometer. After [C₁₆mim][PyW] (0.0031 mmol) was added, 5 mL of the model oil, 30% H₂O₂ (0.064 mL), and an internal standard (n-dodecane) were charged into the Schlenk flask (n_DBT/n_catalyst/n_H2O2 = 50 : 1 : 200 molar ratio). Then, the reaction mixture was stirred at 50 °C for 2 h and cooled down to room temperature. The upper clear solution was withdrawn and characterized using GC with an internal standard. After washing with n-octane and ether respectively, and drying at 50 °C for 1 h under reduced pressure, the catalyst could be reused for the next catalytic cycle. The solid product dibenzothiophene sulfone (DBTO₂) could be separated from the bottom phase by filtration. Then, DBTO₂ was extracted by acetonitrile, and further purified by recrystallization at 0 °C to afford white needle crystals.

The cis-cyclooctene epoxidation was carried out in a 25 mL Schlenk flask equipped with a reflux condenser and a thermometer. After 0.02 mmol catalyst was added, 1 mmol cis-cyclooctene and 1.5 mmol H₂O₂ were then charged into the Schlenk flask. Then the reaction mixture was stirred at 60 °C for 6 h and then cooled down to room temperature. The products were extracted thrice with cyclohexane. The resulting organic layer was dried with MgSO₄ and then analyzed by using GC.

2.5 Adsorption of DBT on the catalysts

The adsorption of DBT on the catalysts was performed by using our similar reported method.³⁷ After DBT (0.15 mmol) was dissolved in 5 mL n-octane, the [CTA][PyW] or [C₁₆mim][PyW] (0.03 mmol) catalyst was added to the n-octane solution and the solution was stirred vigorously at 50 °C. The decrease of the DBT concentration in n-octane was monitored by using GC with n-dodecane as an internal standard the until saturation of the adsorbent was reached.

2.6 The hot filtration experiments for oxidative desulfurization

After adding 0.003125 mmol [C₁₆mim][PyW], 5 mL of the model oil, 0.064 mL 30% H₂O₂ and n-dodecane (internal standard), the mixture was heated up to the reaction temperature and stirred vigorously for 1 h. Then, the catalyst was rapidly filtered out with a hot filter funnel from the reaction mixture at the reaction temperature, and the remaining filtrate was allowed to continue reacting after the addition of fresh H₂O₂.

3 Results and discussion

3.1 Catalyst preparation and characterization

[C₁₂mim][Cl] was synthesized easily according to the previously reported method,³⁵ while the tungsten peroxo anion-based catalyst was prepared through ionic exchange (Scheme 1). Because [C₁₂mim][PyW] does not dissolve in water, HCl as a by-product can be removed easily by washing the catalyst fully. Sequentially, [C₁₆mim][PyW], [TBA][PyW] and [CTA][PyW] were obtained using a similar method. On the basis of the ¹H NMR analysis, elemental analysis of C, H and N, and ICP-AES analysis of W, the ionic liquid-type catalysts were synthesized successfully.

Scheme 1 The synthetic route of one of the ionic liquid catalysts.

The synthesized catalysts were firstly characterized using FT-IR. As shown in Fig. 1a, the main bands of [C₁₆mim][PyW] are 3430(OH) cm⁻¹, 1609(C–N) cm⁻¹, 1468 cm⁻¹, 1394(C–H) cm⁻¹ and 1167(MeN) cm⁻¹ which can be assigned to the vibrations of the imidazolium cation, and 1706 cm⁻¹, 951 cm⁻¹, 850 cm⁻¹, 583 cm⁻¹ and 556 cm⁻¹, which can be ascribed to the vibrations of C=O, W=O, O–O, W(O₂)(asym) and W(O₂)(sym), respectively. This indicates that the anion structure was retained in the ionic liquid-type catalyst. The FT-IR spectrum of the [C₁₂mim][PyW] catalyst was very similar to that of [C₁₆mim][PyW]. However, a small red-shift in the wavenumber of the band at 1706 cm⁻¹ to 1690 cm⁻¹ (C=O vibration) was observed when the imidazolium cation was replaced by the quaternary ammonium cations, probably due to the interactions between the different cations and anion (Fig. 1a–d).

Fig. 1 FT-IR spectra of (a) [C₁₆mim][PyW]; (b) [C₁₂mim][PyW]; (c) [CTA][PyW]; and (d) [TBA][PyW].

The thermal stability of the catalysts was also examined by using TGA. As shown in Fig. 1S,† for the ionic liquid-type [C₁₆mim][PyW] catalyst, the slight loss at around 100–200 °C resulted from water removal and the obvious weight loss of nearly 47% at around 200–350 °C was the result of the decomposition of the imidazolium-based cation. The further weight loss was due to the decomposition of the anion of the catalyst. The overall weight loss was in accordance with the result of the ICP-AES analysis of W. The [C₁₂mim][PyW] catalyst had a very similar decomposition trend to that of [C₁₆mim][PyW] as they contain similar cations and the same anion, but quaternary ammonium salts like [CTA][PyW] and [TBA][PyW] showed a slightly lower thermal decomposition temperature, implying that they have lower thermal stability than the imidazolium cation salts (Fig. 1S†).

3.2 Catalytic performance

Firstly, the effect of different cations in the tungsten peroxo complex on the desulfurization of DBT was investigated. In the present work, a tungsten peroxo anion [PyW]⁻ was combined with four different countercations and applied for the desulfurization of DBT with H₂O₂ as an oxidant (Scheme 2). As shown in Fig. 2, the imidazolium-based catalysts showed higher catalytic activity than the quaternary ammonium-based catalysts. After reacting for 2 h, both of the imidazolium-based catalysts afforded excellent DBT conversions, in particular [C₁₆mim][PyW], which almost provided a complete conversion. This implies that the efficiency of the sulfur removal increases with increasing alkyl chain length. This could be explained by the catalyst with a longer alkyl chain having stronger hydrophobic interactions with the substrates (DBT) through van der Waals forces, which favors the adsorption/coordination of DBT and therefore the reaction of the substrates. However, the hydrophobicity values of [CTA][PyW] and [C₁₆mim][PyW] were almost the same, yet the former afforded much lower activity than the latter. This implies that DBT molecules were more easily coordinated or adsorbed on active tungsten peroxo anions in the presence of imidazolium cations.

Scheme 2 The oxidative desulfurization of DBT over different catalysts.

Fig. 2 Time profile of the desulfurization of DBT catalyzed with different catalysts. Reaction conditions: model oil = 5 mL (S-content 1000 ppm), T = 50 °C, and n(DBT)/n(catalyst)/n(H₂O₂) = 50 : 1 : 200 (molar ratio).

The adsorption of DBT over the [CTA][PyW] and [C₁₆mim][PyW] catalysts was carried out to identify the role of the cations in the catalytic process. As shown in Fig. 3, the DBT molecule exhibited significantly different adsorption rates over the two catalysts. After adsorption for 30 min at 50 °C, the concentration of DBT in n-octane was decreased from 0.03 mmol mL⁻¹ to 0.023 mmol mL⁻¹ due to the adsorption of DBT on [C₁₆mim][PyW], while the concentration of DBT only decreased a little on [CTA][PyW]. Interestingly, the adsorption capacity of DBT on [C₁₆mim][PyW] was nearly equimolar to the amount of [C₁₆mim][PyW] used (the ratio of the adsorbed DBT to [C₁₆mim][PyW] = 1 : 1). The adsorption result indicated clearly that [C₁₆mim][PyW] was better able to adsorb/coordinate the DBT molecules, as compared with [CTA][PyW], and then promote the desulfurization reaction.

Fig. 3 The concentration changes of DBT in n-octane in the presence of different catalysts. Experimental conditions: T = 50 °C, 5 mL n-octane, 0.15 mmol DBT, and n(DBT)/n(catalyst) = 5 : 1 (molar ratio).

Subsequently, the possible interactions between the DBT molecules and imidazolium ring were further clarified through ¹H NMR measurements. As shown in Fig. 4, the proton chemical shifts at H2, H4 and H5 of the imidazolium ring in [C₁₆mim][PyW] were 9.13, 7.26 and 7.23 ppm in CDCl₃, respectively. However, after [C₁₆mim][PyW] was added with DBT in CDCl₃, the proton chemical shifts exhibited obvious changes. In particular, the proton chemical shift at the H2 position of the imidazolium ring moved downfield considerably (from 9.13 to 9.32). The resonance signals of the H5 and H4 ring protons also showed slight shifts, but the changes in their chemical shift were much smaller than that of the proton H2. The variation of the chemical shifts could be explained by several contributions, such as the aromatic ring current effect (π–π), hydrogen bonding effect, C–H–π interaction between the cation and thiophene ring, anion effect, dilution effect and electrostatic field effect. In this investigation, the H2 proton shift was moved downfield, which is different to previous research that reports the aromatic ring current effect to be the dominant factor and, therefore, the upfield shift of the H2 proton.³⁸ It is well known that the formation of a hydrogen bond will cause a proton chemical shift (H2) to move downfield. The change in chemical shift is indicative of the formation of a strong hydrogen bond.³⁹ On the basis of the above discussion, it was demonstrated that a strong H-bonding interaction between the acidic proton H2 and basic S atom existed, resulting in the fast adsorption of DBT onto the imidazolium ring and thus easy accessibility to the catalytically active tungsten peroxo anion. In contrast, it was observed that all of the resonance signals of [CTA][PyW] exhibited almost no changes under the same conditions. This strongly suggests that there was no significant interaction between the quaternary ammonium cation (CTA) and DBT molecules. This was the reason that [C₁₆mim][PyW] afforded a much higher desulfurization efficiency than [CTA][PyW].

Fig. 4 ¹H NMR signal assignment of the protons of (a) [C₁₆mim][PyW], (b) [C₁₆mim][PyW] + DBT, (a′) [CTA][PyW] and (b′) [CTA][PyW] + DBT.

The effect of temperature on the catalyst/H₂O₂ system was investigated. As shown in Fig. 5, the conversion of DBT increased with the reaction temperature at the same reaction time. It was obvious that a higher reaction temperature favored the oxidation of DBT and that the desulfurization efficiency increased sharply when the reaction temperature was over 40 °C. With a view to conserving energy and reducing the unproductive decomposition of hydrogen peroxide at high temperature, the reaction temperature of 50 °C was chosen for further investigations in the present work.

Fig. 5 Time profile of the desulfurization of DBT at different temperatures over the catalyst [C₁₆mim][PyW]. Reaction conditions: model oil = 5 mL (S-content 1000 ppm), T = 50 °C and n(DBT)/n(catalyst)/n(H₂O₂) = 50 : 1 : 200 (molar ratio).

Furthermore, DBT/catalyst molar ratios ranging from 30 : 1 to 100 : 1 were examined for the desulfurization reaction. As shown in Fig. 6A, the reaction rate increased with the rise of the molar ratio of the DBT/catalyst. The sulfur removal increased slightly when the DBT/catalyst ratio decreased from 50 : 1 to 30 : 1. In view of the desulfurization efficiency and catalyst cost, 50 : 1 was chosen as the appropriate DBT/catalyst ratio. These results showed that a suitable molar ratio of DBT to catalyst was important for the desulfurization process.

Fig. 6 (A) Time profile of the desulfurization of DBT with different DBT/catalyst molar ratios. Reaction conditions: model oil = 5 mL (S-content 1000 ppm), T = 50 °C and n(DBT)/n(H₂O₂) = 1 : 4 (molar ratio); (B) time profile of the desulfurization of DBT with different H₂O₂/DBT molar ratios. Reaction conditions: model oil = 5 mL (S-content 1000 ppm), T = 50 °C and n(DBT)/n(catalyst) = 50 : 1 (molar ratio).

H₂O₂ is a green oxidant in comparison with other organic oxidants. In order to optimize the reaction conditions and elucidate the influence of the amount of H₂O₂ on the desulfurization, different molar ratios of H₂O₂ to DBT were evaluated at a temperature of 50 °C, using 5 mL model oil (S content 1000 ppm) and a 50 : 1 molar ratio of DBT/catalyst. As shown in Fig. 6B, the reaction rate showed an obvious dependence on the H₂O₂/DBT molar ratio. As the ratio of H₂O₂/DBT increased from 3 : 1 to 4 : 1, the removal efficiency of DBT increased correspondingly, but higher H₂O₂/DBT molar ratios could not improve the removal efficiency of DBT. In contrast, if the H₂O₂/DBT molar ratio was further increased up to 6, the removal efficiency of DBT decreased considerably. For example, DBT could be almost completely removed when the molar ratio of H₂O₂/DBT was 4 : 1 or 5 : 1 after a reaction of 2 h, but the removal of DBT only reached 61.9% at the molar ratio of 6 : 1 H₂O₂/DBT under the same conditions. Because the present oxidant was 30% H₂O₂ aqueous solution, the addition of more H₂O₂ caused a larger amount of water to be introduced into the reaction system, which significantly affected the reaction environment. As the reaction system was composed of three phases, namely water, n-octane and the catalyst, when the amount of H₂O₂ was increased, the mass transfer efficiency was decreased to some extent, which might lower the catalytic activity.^23,24,40

The recycling of the [C₁₆mim][PyW] was investigated at a temperature of 50 °C, using 5 mL model oil (S content 1000 ppm) and a 200 : 50 : 1 molar ratio of H₂O₂/DBT/catalyst. As shown in Fig. 2S,† the present catalyst could be reused at least ten times without a decrease in catalytic activity in the desulfurization of DBT. The amount of leached W detected using ICP-AES was only about 5 ppm in the n-octane phase, which illustrated clearly that the present catalyst was highly leaching-resistant during the desulfurization reaction. After the reaction, the product (DBTO₂) was isolated and then characterized using ¹H NMR and FT-IR. As shown in Fig. 3S,† the ¹H NMR spectrum of the recovered DBTO₂ is the same as that of standard DBTO₂.^41,42 The specific infrared absorption bands of DBTO₂ at 1169 and 1289 cm⁻¹ were seen, which can be attributed to sulfone groups (Fig. 4S†).

In order to identify further whether the leached W species played a crucial role in the oxidative desulfurization reaction of DBT, a fast hot filtration experiment was carried out under the reaction temperature. As shown in Fig 7, after a 1 h reaction the catalyst was filtered out and the residual n-octane phase continued to react after adding fresh H₂O₂, but no further DBT was converted to the corresponding product. This result proved that the trace amount of leached W had no influence on the desulfurization activity.

Fig. 7 Time profile of the reaction and fast hot filtration test of DBT catalyzed with [C₁₆mim][PyW]. Reaction conditions: model oil = 5 mL (S-content 1000 ppm), T = 50 °C and n(DBT)/n(catalyst)/n(H₂O₂) = 50 : 1 : 200 (molar ratio).

In the next step, the [C₁₆mim][PyW] and [CTA][PyW] catalysts were applied to catalyze the epoxidation of cis-cyclooctene with 30% aqueous H₂O₂ as the oxidant under solvent-free conditions at 60 °C (Scheme 3). As shown in Fig. 8, there was almost no significant difference between the catalytic activity of [C₁₆mim][PyW] and [CTA][PyW]. After reacting for 6 h, cis-cyclooctene could be converted to the corresponding epoxide with about 80% conversion and 100% selectivity. This result was quite different to that of the oxidative desulfurization reaction, in which the imidazolium-based catalyst had a higher catalytic activity. Combined with the previous ¹H NMR characterization (Fig. 4), we can further establish that the interactions between the ionic liquid-type catalyst and substrate were very important for the desulfurization reaction. Similar activities for epoxidation were obtained because no basic S atom (compared with DBT) existed in the cis-cyclooctene molecules and thus the interactions between cis-cyclooctene and the imidazolium of [C₁₆mim][PyW] were weak. The use of the catalysts [C₁₆mim][PyW] and [CTA][PyW], with the same anion and similar hydrophobicity, led to almost the same epoxidation activity under the same reaction conditions.

Scheme 3 Epoxidation of cis-cyclooctene.

Fig. 8 Time profile of the epoxidation of cis-cyclooctene catalyzed with different catalysts. Reaction conditions: T = 60 °C, 0.02 mmol catalyst, 1 mmol cis-cyclooctene and 1.5 mmol 30% H₂O₂.

3.3 Proposed mechanism for the desulfurization reaction

On the basis of the experimental results and discussion above, the desulfurization process of DBT catalyzed by [C₁₆mim][PyW] was proposed, as shown in Scheme 4. First, the oxidant H₂O₂ could continuously provide active oxygen to the active peroxo species in the hydrophilic group of the catalyst, and DBT was adsorbed to the cation through a hydrogen bonding interaction between the imidazolium ring (H2 position) and the S atom of DBT. Then, the active peroxo species was transferred to the substrate molecules. Thus, the continuous oxidation of DBT to the corresponding product (DBTO₂) was observed. The polarity of the sulfur compound was increased after oxidation. With an increasing number of cycles, the sulfones were gradually precipitated and they accumulated in the lower phase.

Scheme 4 Proposed reaction mechanism for the desulfurization of DBT.

4 Conclusions

In summary, we successfully synthesized a new peroxotungsten anion-based ionic liquid catalyst for the deep desulfurization of DBT in model oil, using H₂O₂ as the oxidant under solvent-free conditions. The sulfur removal could reach 99.0% over the [C₁₆mim][PyW] catalyst under mild conditions. It was demonstrated that there was a hydrogen bonding interaction between the H2 position of the imidazolium cation and S atom of the DBT molecules, which played a crucial role in enhancing the desulfurization efficiency. The present catalyst [C₁₆mim][PyW] can be recycled ten times with an unnoticeable decrease in activity. In summary, this process provides a novel and effective way to remove sulfur from model oil. Due to these advantages, this ionic liquid-type catalyst can be used in the construction of a promising, green catalytic system.

Acknowledgements

The authors are grateful for support from the National Natural Science Foundation of China (21373082), Innovation Program of Shanghai Municipal Education Commission (15ZZ031), and the Fundamental Research Funds for the Central Universities.

References

1. R. Abro, A. A. Abdeltawab, S. S. Al-Deyab, G. Yu, A. B. Qazi, S. Gao and X. Chen, RSC Adv., 2014, 4, 35302–35317.
2. J. M. Campos-Martin, M. Capel-Sanchez, P. Perez-Presas and J. Fierro, J. Chem. Technol. Biotechnol., 2010, 85, 879–890.
3. A. Niquille-Röthlisberger and R. Prins, J. Catal., 2006, 242, 207–216.
4. Y. Fan, H. Xiao, G. Shi, H. Liu and X. Bao, Energy Environ. Sci., 2011, 4, 572–582.
5. O. Y. Gutiérrez and T. Klimova, J. Catal., 2011, 281, 50–62.
6. M. Craven, R. Yahya, E. Kozhevnikova, R. Boomishankar, C. M. Robertson, A. Steiner and I. Kozhevnikov, Chem. Commun., 2013, 49, 349–351.
7. E. Rafiee and S. Eavani, J. Mol. Catal. A: Chem., 2013, 380, 18–27.
8. A. Nisar, Y. Lu, J. Zhuang and X. Wang, Angew. Chem., 2011, 123, 3245–3250.
9. L. C. de Oliveira, N. T. Costa, J. R. Pliego Jr, A. C. Silva, P. P. de Souza and P. S. de O Patrício, Appl. Catal., B, 2014, 147, 43–48.
10. R. Sundararaman, X. Ma and C. Song, Ind. Eng. Chem. Res., 2010, 49, 5561–5568.
11. E. Torres-Garcia, A. Galano and G. Rodriguez-Gattorno, J. Catal., 2011, 282, 201–208.
12. H. Li, W. Zhu, Y. Wang, J. Zhang, J. Lu and Y. Yan, Green Chem., 2009, 11, 810–815.
13. W. Zhu, J. Zhang, H. Li, Y. Chao, W. Jiang, S. Yin and H. Liu, RSC Adv., 2012, 2, 658–664.
14. D. Zhao, J. Wang and E. Zhou, Green Chem., 2007, 9, 1219–1222.
15. W. Wang, G. Li, W. Li and L. Liu, Chem. Commun., 2011, 47, 3529–3531.
16. Y. Shiraishi, T. Naito and T. Hirai, Ind. Eng. Chem. Res., 2003, 42, 6034–6039.
17. D. Serrano, R. Sanz, P. Pizarro, I. Moreno and S. Medina, Appl. Catal., B, 2014, 146, 35–42.
18. R. D. Rogers and K. R. Seddon, Science, 2003, 302, 792–793.
19. R. Sheldon, Chem. Commun., 2001, 2399–2407.
20. P. S. Kulkarni and C. A. Afonso, Green Chem., 2010, 12, 1139–1149.
21. H. Zhao, S. Xia and P. Ma, J. Chem. Technol. Biotechnol., 2005, 80, 1089–1096.
22. J. Zhang, W. Zhu, H. Li, W. Jiang, Y. Jiang, W. Huang and Y. Yan, Green Chem., 2009, 11, 1801–1807.
23. M. Zhang, W. Zhu, S. Xun, H. Li, Q. Gu, Z. Zhao and Q. Wang, Chem. Eng. J., 2013, 220, 328–336.
24. H. Lü, W. Ren, H. Wang, Y. Wang, W. Chen and Z. Suo, Appl. Catal., A, 2013, 453, 376–382.
25. H. Lü, C. Deng, W. Ren and X. Yang, Fuel Process. Technol., 2014, 119, 87–91.
26. Y. Ding, W. Zhu, H. Li, W. Jiang, M. Zhang, Y. Duan and Y. Chang, Green Chem., 2011, 13, 1210–1216.
27. J. Xu, S. Zhao, Y. Ji and Y. F. Song, Chem.–Eur. J., 2013, 19, 709–715.
28. J. Xu, S. Zhao, W. Chen, M. Wang and Y. F. Song, Chem.–Eur. J., 2012, 18, 4775–4781.
29. W. Zhu, G. Zhu, H. Li, Y. Chao, M. Zhang, D. Du, Q. Wang and Z. Zhao, Fuel Process. Technol., 2013, 106, 70–76.
30. J. Gao, S. Wang, Z. Jiang, H. Lu, Y. Yang, F. Jing and C. Li, J. Mol. Catal. A: Chem., 2006, 258, 261–266.
31. Y. Qiao, Z. Hou, H. Li, Y. Hu, B. Feng, X. Wang, L. Hua and Q. Huang, Green Chem., 2009, 11, 1955–1960.
32. H. Li, Z. Hou, Y. Qiao, B. Feng, Y. Hu, X. Wang and X. Zhao, Catal. Commun., 2010, 11, 470–475.
33. L. Hua, J. Chen, C. Chen, W. Zhu, Y. Yu, R. Zhang, L. Guo, B. Song, H. Gan and Z. Hou, New J. Chem., 2014, 38, 3953–3959.
34. J. Chen, L. Hua, W. Zhu, R. Zhang, L. Guo, C. Chen, H. Gan, B. Song and Z. Hou, Catal. Commun., 2014, 47, 18–21.
35. J. G. Huddleston, A. E. Visser, W. M. Reichert, H. D. Willauer, G. A. Broker and R. D. Rogers, Green Chem., 2001, 3, 156–164.
36. S. E. Jacobson, R. Tang and F. Mares, Inorg. Chem., 1978, 17, 3055–3063.
37. C. Chen, X. Zhao, J. Chen, L. Hua, R. Zhang, L. Guo, B. Song, H. Gan and Z. Hou, ChemCatChem, 2014, 6, 3231–3238.
38. B.-M. Su, S. Zhang and Z. C. Zhang, J. Phys. Chem. B, 2004, 108, 19510–19517.
39. A. G. Avent, P. A. Chaloner, M. P. Day, K. R. Seddon and T. Welton, J. Chem. Soc., Dalton Trans., 1994, 3405–3413.
40. L. C. Caero, E. Hernández, F. Pedraza and F. Murrieta, Catal. Today, 2005, 107, 564–569.
41. X. Zhou, J. Li, X. Wang, K. Jin and W. Ma, Fuel Process. Technol., 2009, 90, 317–323.
42. S. Otsuki, T. Nonaka, N. Takashima, W. Qian, A. Ishihara, T. Imai and T. Kabe, Energy Fuels, 2000, 14, 1232–1239.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra00136f

---