The growth of the consumption of hydrocarbon resources often causes the need for refining of heavy high-sulfur crude oil [1]. High sulfur content of such feedstock gives rise to serious problems in refining, leading to poisoning of catalysts, corrosion of process equipment, etc. [2, 3]. Sulfur compounds present in oil cause environmental pollution [4]; therefore, the total sulfur content of modern motor fuels is rigorously restricted [5]. Despite the fact that the presently existing processes for reducing sulfur content, hydrotreating and hydrocracking, allow production of petroleum products with the preset sulfur content, the trend toward worsening of the quality of the extracted hydrocarbon resources makes topical both modification of the existing processes and development of new alternative approaches to reducing sulfur content without using hydrogen.

Among hydrogenless methods for reducing sulfur content, such processes should be noted as extraction [6], adsorption [7], oxidative desulfurization [3], and biodesulfurization [8]. Because the polarity of organic sulfur compounds present in hydrocarbon fuels is close to that of aromatic hydrocarbons, the extraction and adsorption show low selectivity, leading to major loss of the hydrocarbon fraction in the course of the treatment. Biodesulfurization, in turn, involves certain restrictions associated with the use of bacteria. Therefore, considerable attention is paid in the recent literature to oxidative desulfurization as an alternative hydrogenless procedure for reducing sulfur content [9, 10].

The oxidative desulfurization is a two-step process. In the first step, organic sulfur compounds are oxidized to the corresponding sulfoxides and sulfones, which are considerably more polar than other components of the hydrocarbon feedstock. This allows highly selective removal of sulfoxides and sulfones from the feedstock by extraction or adsorption.

The oxidative desulfurization has been relatively well studied [9–15]. Such oxidizing agents are used as ozone [11], oxygen [12], organic peroxides [13], and inorganic oxidants [14], including hydrogen peroxide [15]. The latter is used as an oxidant most frequently, because it allows the process to be performed under relatively mild conditions (temperature of up to 80°C, atmospheric pressure), and the only reduction by-product is water.

Hydrogen peroxide taken alone does not ensure exhaustive oxidation of organic sulfur compounds present in petroleum fractions. Therefore, catalysts are used to enhance the process efficiency. The most widely used catalysts are acids [16] or compounds of transition metals capable of formation of peroxo complexes (molybdenum, tungsten, vanadium) [17]. As a rule, these catalytic systems ensure exhaustive oxidation of, e.g., relatively inert DBT within 1–2 h. However, the development and scaling of the oxidative desulfurization process require the development of catalytic systems that ensure complete conversion of oxidation-inert heteroaromatic sulfur-containing compounds in a minimal reaction time.

It is known that combination of two types of catalytically active sites (carboxyl groups and molybdenum compounds) allows considerable enhancement of the catalyst activity, so that exhaustive oxidation of DBT in 5 min is ensured [18]. However, in [18] we used expensive ionic liquids supported on mesoporous silicates; this factor considerably hinders wide use of such catalysts.

In this study, we synthesized a catalyst based on commercially available ASKG silica gel with the surface modified with sulfo groups and ammonium heptamolybdate. Combination of the two types of catalytic sites not only considerably enhances the catalyst activity, but also allows the active phase washout to be avoided.

EXPERIMENTAL

The following chemicals were used for preparing SiO2-based catalysts: ASKG silica gel (GOST 3956-76, LenReaktiv, Russia), 3-mercaptopropyl(trimethoxy)silane (Sigma Aldrich), toluene (99.5%, Prime Chemicals Group, Russia), hydrogen peroxide (Prime Chemicals Group, 50%), sulfuric acid (H2SO4, main substance content no less than 93.6%, EKOS-1, Russia), and ammonium heptamolybdate (analytically pure grade, Prime Chemicals Group).

The following chemicals were used in oxidation experiments: dibenzothiophene (DBT, Sigma Aldrich, 98%), benzothiophene (BT, Sigma Aldrich, 98%), 4-methyldibenzothiophene (MeDBT, Sigma Aldrich, 96%), 4,6-dimethyldibenzothiophene (Me2DBT, Sigma Aldrich, 95%), dodecane (98%, EKOS-1), and hydrogen peroxide (Н2О2, Prime Chemicals Group, 50%)/

Synthesis of catalysts. The intermediate compound SiO2–SO3H was prepared by chemical modification of silica gel with 3-mercaptopropyl(trimethoxy)silane. In the first step, we prepared 3-mercaptopropyl–SiO2 according to [19]. To this end, we prepared a solution of 734 μL of 3-mercaptopropyl(trimethoxy)silane (4 mmol, 1.057 g/mL) in 30 mL of toluene; then, 2 g of SiO2 was added, and the mixture was refluxed for 24 h. The obtained 3-mercaptopropyl–SiO2 was filtered off and dried at 100°C for 5 h.

In the second step, the grafted thiol groups were oxidized: 2 g of 3-mercaptopropyl–SiO2 was added to 20 mL of a mixture of 30% Н2О2 and concentrated H2SO4 (0.031 g, 0.32 mmol), and the mixture was stirred at room temperature for 20 h. The precipitate was filtered off on a Büchner funnel and washed with distilled water to neutral reaction of the medium. To make sure that all the sulfo groups have been protonated, the precipitate was additionally allowed to stand in 20 mL of 0.05 М H2SO4 for 5 h. Then, the precipitate was again filtered off and washed to neutral reaction of the medium. The product obtained, SiO2 containing sulfo groups, was dried in air at 110°C for 5 h.

The modification of SiO2–SO3H and SiO2 was performed by impregnation with a solution of 0.046 g of (NH4)6Mo7O24·4H2O in distilled water. 0.5 g of SiO2–SO3H or SiO2 was placed in 5 mL of an aqueous solution of ammonium heptamolybdate for 2 h for impregnation with continuous stirring on a rotary evaporator at 40°С, after which the temperature was elevated to 60°C and the mixture was stirred for an additional 2 h. Then, the catalysts were dried in an oven at 80°C in the stepwise mode with 10°C steps and 4-h keeping in each step.

The samples were analyzed by Fourier transform IR spectroscopy with a Nicolet IR2000 device (Thermo Scientific) using the multiple attenuated total reflection technique with a Multireflection HATR attachment containing a ZnSe 45° crystal for different wavelength ranges with 4 nm resolution. The nitrogen adsorption–desorption isotherms were recorded at 77 K with a Gemini VII 2390 (V1.02t) device (Micromeritics). Prior to measurements, the samples were degassed at 120°С for 6 h. The surface area was calculated using the Brunauer–Emmett–Teller (BET) method from the data on the adsorption at relative pressures in the range Р/Р0 = 0.05–0.2. The total pore volume was determined from the amount of adsorbed nitrogen at the relative pressure P/P0 = 0.95.

The elemental composition was determined by X-ray fluorescence (XRF) analysis with a Thermo ARL Perform’x Sequential XFR device equipped with a 2500 V X-ray tube. Prior to analysis, 200-mg samples were pelletized with boric acid. The amount of sulfo groups was determined by back-titration.

Oxidation of model fuel. A reactor equipped with a helical reflux condenser and a magnetic stirrer and preheated to the reaction temperature was charged with 5 mL of a model fuel consisting of the sulfur compound dissolved in dodecane (sulfur content of the model mixture 500 ppm), 0.0094–0.0375 g of the catalyst (0.25–1.00 wt %), and 3–27 μL of a 50% hydrogen peroxide solution (Н2О2 : S molar ratio from 1 : 1 to 8 : 1; S is the sulfur-containing compound being oxidized). The reaction was performed for 10–120 min; the stirring rate in all the experiments was 700 rpm. The concentrations of the substrate and reaction products (wt %) were determined by gas chromatography with a Crystal 2000M gas chromatograph equipped with a flame ionization detector and a Zebron column (L = 30 m, d = 0.32 mm; ZB-1 liquid phase).

The experimental uncertainty did not exceed 5%; the data presented below in Figs. 3–8 are mean values of three replicate experiments with agreeing results.

RESULTS AND DISCUSSION

Physicochemical Characteristics of Catalysts

The catalyst compositions were studied by two methods (Table 1): acid–base titration to determine the amount of sulfo groups and elemental analysis to determine the amounts of sulfur and molybdenum. By chemical modification with 3-mercaptopropyl(trimethoxy)silane, we prepared silica gel containing approximately 9 wt % sulfo groups. The actual molybdenum content of the catalyst correlates with the calculated value. It should be noted that the data on the sulfur content of the catalyst, obtained using back-titration and elemental analysis, agree with each other, which confirms exhaustive oxidation of thiol groups to sulfo groups.

Table 1. Composition and textural characteristics of the support and catalysts
Full size table

According to the data of low-temperature hydrogen adsorption–desorption, modification of the silicon oxide surface leads to a decrease in the amount of the adsorbed nitrogen; i.e., the modification influences the textural characteristics of the samples (Fig. 1, Table 1). Primary chemical modification of ASKG with 3-mercaptopropyl(trimethoxy)silane, followed by oxidation of thiol groups to sulfo groups, exerts a stronger effect than the impregnation with ammonium heptamolybdate does. This may be due, first, to the large amount of the introduced sulfo groups and, second, to the occurrence of a chemical reaction between the silanol groups of silica gel and methoxy groups of the silane.

Fig. 1.
figure 1

Isotherms of low-temperature nitrogen adsorption–desorption for the support and catalysts obtained.

Full size image

In the IR spectra of all the samples synthesized (Fig. 2), there are strong peaks at 1055 and 792 cm–1, corresponding to the vibrations of the Si–O–Si bonds in silica gel, and weak peaks at 953 cm–1, corresponding to Si–OН vibrations [20, 21]. The molybdenum-containing catalyst exhibits an additional peak at 914 cm–1, corresponding to the Mo–O bond [22]. A certain increase in the intensity of the plateau at 1150–1300 cm–1 for the modified materials is due to the presence of sulfo groups [23].

Fig. 2.
figure 2

IR spectra of the synthesized samples. Assignments: 792 cm–1, Si–O–Si (stretching); 905 cm–1, Мо–O–Мо (stretching); 953 cm–1, Si–OН (bending); 1055 cm–1, Si–O–Si (bending).

Full size image

Catalytic Experiments

The activity of the catalysts was evaluated using a model mixture containing dibenzothiophene as a substrate. The substrate choice was governed by its relative inertness to oxidation and by wide occurrence in petroleum fractions. The catalyst activities are compared in Fig. 3.

Fig. 3.
figure 3

Comparison of the catalyst activities in dibenzothiophene oxidation. Oxidation conditions: H2O2 : S = 6 : 1 (mole/mole), catalyst amount wcat = 0.5%, 80°С.

Full size image

According to the results obtained, in the presence of unmodified silica gel the DBT conversion is ~15%, which is associated with the dibenzothiophene adsorption in support pores. Ammonium heptamolybdate supported on silica gel (AHM/SiO2) shows only slightly higher catalytic activity than the blank sample does, which may be due to washout of ammonium heptamolybdate from the support surface in the presence of hydrogen peroxide. The SiO2–SO3H catalyst, on the other hand, shows high performance and allows oxidation of more than a half of the substrate in 30 min. The activity of this catalyst is due to the ability of sulfo groups to form an active per acid in the presence of hydrogen peroxide. The possibility of using the catalysts containing sulfo groups for the oxidation of sulfur-containing compounds was demonstrated previously [24]. Combination of two types of catalytic sites (ammonium heptamolybdate and sulfo groups) allows considerable enhancement of the catalyst activity, so that complete oxidation of DBT in the presence of AHM/SiO2–SO3H can be reached already in 30 min.

Figure 4 shows the time dependence of the DBT conversion in the oxidation at different temperatures. With a decrease in temperature, the substrate conversion drastically decreases, becoming less than 20% at 40°C. Such trend is typical of many known catalysts [3, 17] and is associated with hindered diffusion of a nonpolar substrate into pores of a polar catalyst.

Fig. 4.
figure 4

Time dependence of the DBT conversion in the oxidation at different temperatures. Oxidation conditions: H2O2 : S = 6 : 1 (mole/mole), wcat = 0.5%, catalyst AHM/SiO2–SO3H.

Full size image

With an increase in the hydrogen peroxide amount, the DBT conversion passes through a maximum (Fig. 5). This pattern may be due to the fact that, at low oxidant dosages, the reaction decelerates, and at hydrogen peroxide to sulfur molar ratios higher than 4 : 1 the hydrophilic catalyst particles undergo aggregation due to excess aqueous phase, which leads to a decrease in the DBT conversion. Similar trends have been reported previously [25].

Fig. 5.
figure 5

DBT conversion as a function of the oxidant amount. Oxidation conditions: wcat = 0.5%, 10 min, 80°С, catalyst AHM/SiO2–SO3H.

Full size image

Figure 6 shows how the DBT conversion depends on the catalyst amount. With a decrease in the catalyst dosage from 1 to 0.25 wt % relative to the model mixture, the DBT conversion monotonically decreases. Such a decrease in the reaction rate may be due to a decrease in the absolute amount of the catalyst active sites.

Fig. 6.
figure 6

DBT conversion in the oxidation as a function of the catalyst amount. Oxidation conditions: H2O2 : S = 4 : 1 (mole/mole), 10 min, 80°С, catalyst AHM/SiO2–SO3H.

Full size image

To determine the possibility of the repeated use of the AHM/SiO2–SO3H catalyst, we performed oxidation of a large volume of the model mixture (60 mL) with the preservation of the catalyst and oxidant dosages under the conditions indicated in the caption to Fig. 7. After each cycle, the catalyst was separated from the reaction mixture by centrifugation. Then this catalyst was used for the oxidation of the fresh portion of the model fuel. Five cycles were performed, after which the spent catalyst was washed with acetonitrile to remove the residual adsorbed sulfone and was dried at reduced pressure and a temperature of 80°C for 4 h. According to the data of elemental analysis and low-temperature nitrogen adsorption–desorption, the molybdenum content of the catalyst and its textural properties changed insignificantly; i.e., the catalyst was stable under the conditions of the oxidation process.

Fig. 7.
figure 7

Repeated use of the catalyst. Oxidation conditions: H2O2 : S = 4 : 1 (mole/mole), wcat = 0.5%, 10 min, 80°С, catalyst AHM/SiO2–SO3H.

Full size image

The results of the oxidation of model mixtures with other sulfur-containing substrates are shown in Fig. 8. As can be seen, the presence of alkyl substituents in the DBT molecule leads to a decrease in the conversion by approximately 20%. Low conversion in the case of benzothiophene is associated with lower electron density on the sulfur atom in this substrate [3].

Fig. 8.
figure 8

Oxidation of different sulfur-containing organic substrates. Oxidation conditions: H2O2 : S = 4 : 1 (mole/mole), wcat = 1%, 30 min, 80°С, catalyst AHM/SiO2–SO3H.

Full size image

CONCLUSIONS

Catalysts on ASKG silica gel support, containing two types of active sites, were prepared. The catalyst was examined by a set of physicochemical methods: low-temperature nitrogen adsorption/desorption, IR spectroscopy, elemental analysis.

Comparative evaluation of the catalyst activity with oxidation of the model DBT-containing mixture as an example shows that combination of two types of active sites, sulfo groups and heptamolybdate anion, allows considerable improvement of the results compared to the catalysts containing only a single type of catalytic sites. It should be noted that the presence of sulfo groups allows the ammonium heptamolybdate washout from the support surface to be reduced to a minimum, so that the catalyst activity is preserved in five oxidation cycles. The major factors influencing the oxidation were studied, namely, the oxidant and catalyst amounts, temperature, reaction time, and kind of the sulfur-containing substrate. Under the optimum conditions (80°C, Н2О2 : sulfur molar ratio 6 : 1, AHM/SiO2–SO3H catalyst weight fraction 0.5 wt %), complete oxidation of dibenzothiophene is reached already in 30 min.