Among recent developments in catalysis, "clean oxidation catalysis", involving catalysis and H2O2, is receiving much attention because of the importance of oxidation reactions in the manufacture of fine chemicals. These catalytic systems promote the oxidation of organic substrates because of their oxygen content, and are low cost, safe to store and use, and, importantly, the use of H2O2 is environmentally friendly. These advantages have encouraged the development of practical procedures for the selective oxidation of organic sulfur compounds, not only because of the extensive use of sulfoxides and sulfones as synthetic intermediates in the pharmaceutical industry [1], but also because the oxidative desulfurization (ODS) process is an effective method for removing sulfur compounds from fuels and industrial effluents, to satisfy new environmental legislation [2, 3, 4, 5].
ODS is an alternative or complementary technology to hydrodesulfurization for deep desulfurization [6]. It has several advantages such as mild reaction conditions (ambient pressure and low temperature), high selectivity, and the potential for desulfurization of sterically hindered sulfur compounds such as dibenzothiophene (DBT) [7]. In the ODS process, aromatic sulfur compounds such as benzothiophene (BT), DBT, and their alkylated derivatives are oxidized to the corresponding sulfone or sulfoxide; these are highly polar and can be removed using organic extractants [8, 9, 10, 11, 12, 13, 14].
Various catalytic systems have been reported for ODS [15, 16, 17, 18, 19, 20, 21, 22]. Among these, polyoxometalates (POMs) have higher sulfur removal efficiencies because of their unusual properties. The main obstacle to industrial applications of homogeneous POMs is that they are difficult to recover from the reaction medium. These materials are therefore usually impregnated on suitable materials [23, 24, 25, 26, 27, 28, 29].
Recently, progress has been achieved using amphiphilic POM catalysts, which are synthesized from a combination of quaternary ammonium cations and POM anions. These compounds can form emulsion droplets, which act as homogeneous catalysts at the interface of two immiscible liquids, giving high activities during oxidative processes [30, 31, 32].
We synthesized and characterized a novel, more effective, magnetically separable catalyst by the immobilization of 12-tungstophosphoric acid (PW) on the surfaces of silica-coated CoFe2O4 magnetic nanoparticles (CoFe@Si). This catalyst (denoted by CoFe@Si-PW) was characterized using various techniques. Selective oxidation of sulfur compounds and desulfurization of model oil were investigated using an H2O2/CoFe@Si-PW catalytic system. The effects of reaction temperature, catalyst loading, and oxidant to sulfur (O/S) molar ratio were studied to optimize the reaction conditions. The ODS of model oil was investigated under the optimized reaction conditions. The effect of nitrogen-containing compounds on ODS were investigated by adding quinoline and indole to the oxidation system.
All materials were commercial reagent grade and obtained from Merck, Aldrich, and Fluka, and used without further purification.
The CoFe@Si-PW catalyst was synthesized according to the procedure shown in Scheme 1. The CoFe2O4 nanoparticles were prepared using a coprecipitation method. CoCl2·6H2O (25 mL, 0.1 mol/L) and FeCl3·6H2O (25 mL, 0.2 mol/L) were dissolved in distilled water and kept at 70 °C. NaOH solution (24 mL, 3.0 mol/L) was added to the salt solution. The dark brown product was washed several times with distilled water and redispersed in distilled water to give a stable brown magnetic dispersion. CoFe2O4 nanoparticles (0.04 g) were dispersed in ethanol (160 mL). This dispersion was homogenized ultrasonically in a water bath for 10 min. Finally, water (40 mL), tetraethylorthosilicate (1 mL), and ammonia solution (5 mL) were slowly added to this dispersion. The obtained CoFe@Si nanoparticles were separated using a permanent magnet, washed with distilled water, and dried under vacuum. The synthesis of 40 wt% PW on CoFe@Si was performed by dissolving PW (0.5 g) in dry methanol (5 mL). This solution was added drop-wise to a suspension of CoFe@Si (1.0 g in 50 mL methanol) with dispersion by sonication. The mixture was stirred to obtain CoFe@Si-PW magnetic nanoparticles. The catalyst was collected using a permanent magnet and dried under vacuum overnight.
Fourier-transform infrared (FT-IR) spectroscopy was performed on KBr pellets using an FTIR-spectrometer (Alpha). X-ray diffraction (XRD) was performed using a D8 ADVANCE diffractometer (Bruker-AXS, Germany). The magnetic properties of CoFe@SiO2-PW were determined using a BHV-55 (Riken, Japan) vibrating sample magnetometer (VSM). The morphology of the supported catalyst was examined using scanning electron microscopy (SEM; XL30, Philips). Transmission electron microscopy (TEM) was performed using a Zeiss EM900 transmission electron microscope with an accelerating voltage of 80 kV. The size distributions and zeta potentials of the samples were determined using a laser particle size analyzer (HPPS5001, Malvern, UK). The total S contents of the model oils were determined using an elemental analyzer (Analytik Jena AG-multi EA® 3100). The W content of the catalyst was determined using inductively coupled plasma atomic emission spectroscopy (ICP-AES; Spectro Ciros CCD spectrometer).
Liquid sulfides were oxidized to the corresponding sulfones by stirring a solution of the sulfide (1 mmol) and the catalyst (0.04 g) in n-heptane (4 mL). A certain amount of H2O2 (30% aq.) was added as the oxidant. The mixture was stirred for a specified time at room temperature. After completion of the reaction, the catalyst was separated from the reaction mixture using an external magnet. The corresponding sulfone products were extracted from the reaction mixture with Et2O. The solvent was evaporated to generate the crude product. The crude product was purified by column chromatography on silica gel using hexane/ethyl acetate as the eluent (method a).
Solid sulfides were oxidized to the corresponding sulfones by stirring a solution of the sulfide (1 mmol) and the catalyst (0.15 g) in n-heptane-ethanol (v/v, 4:2). Then a certain amount of H2O2 (30% aq.) was added as the oxidant. The mixture was stirred for a specified time at 60 °C, and the reaction was monitored using thin-layer chromatography. After completion of the reaction, the catalyst was separated from the reaction solution using an external magnet. The corresponding sulfone products were separated from the reaction mixture. The solvent was evaporated to generate the crude product. The crude product was purified by column chromatography on silica gel using hexane/ethyl acetate as the eluent (method b).
Model oils were prepared by dissolving various amounts of S and N compounds in n-heptane. The compositions of the model oils are shown in Table 1. The oxidation reaction was investigated using the model oil (5 mL, feeds 1-3) and catalyst in ethanol (1 mL) as an extracting solvent. A certain amount of H2O2 (30% aq.) was added to the mixture. The mixture was stirred at 60 °C for 4 h. The catalyst was separated from the reaction mixture using an external magnet. The S content of the model oil was determined using an elemental analyzer.
The effect of nitrogen compounds on the ODS of model oil was investigated by combining the model oil (5 mL, feeds 3 and 4) and catalyst with H2O2 (O/S: 12) and ethanol (1 mL). The mixture was stirred at 60 °C for 4 h. The catalyst was separated from the reaction mixture using an external magnet. The S content of the model oil was determined using an elemental analyzer.
The FT-IR spectra of CoFe2O4, CoFe@Si, and CoFe@Si-PW are shown in Fig. 1. All samples show two intense and broad peaks at around 3395 and 1622 cm−1, associated with the stretching vibrations of free or adsorbed water remaining in the sample. The presence of Co-O and Fe-O bonds in CoFe2O4 is confirmed by the peak at 579 cm−1, which shifts to 596 cm−1 after coating with silica [33]. The bands at 1087 cm−1, along with the shoulder at 1200 cm−1, correspond to the stretching mode of the Si-O-Si bond in silica, and the band at 464 cm−1 is assigned to Si-O-Si or O-Si-O bending modes [34, 35]. The FT-IR spectrum of the CoFe@Si-PW particles has peaks at 890 and 981 cm−1, which are attributed to the W-O-W and W=O stretching modes, respectively, of PW [36, 37]. The band at 1079 cm−1 corresponds to the P-O stretching mode of PW.
The XRD patterns of CoFe@Si, CoFe@Si-PW, and PW are shown in Fig. 2. There are no peaks from crystalline SiO2, because it remains amorphous because of the low-temperature synthetic method used. When PW is immobilized on the surfaces of the CoFe@Si magnetic particles, the characteristic peaks of PW are observed for CoFe@Si-PW, which implies retention of the crystalline character of PW [38].
The morphology of the CoFe@Si-PW catalyst was examined using SEM (Fig. 3(a)). It shows that the obtained particles are agglomerated as a consequence of strong magnetization and the high particle surface energy. Wavelength-dispersive X-ray (WDX) images, which mapped Fe, Si, Co, and W elements individually in a cross-section, are shown in Fig. 3(b). The WDX analysis shows excellent uniform distribution of PW on CoFe@Si magnetic nanoparticles.
The TEM images indicate that the CoFe@Si-PW nanoparticles are almost spherical (Fig. 4). The size distribution of the CoFe@Si-PW catalyst particles was measured using a laser particle size analyzer. The results (Fig. 5) show particles of mean diameter 137 nm with a narrow size distribution. This mean diameter is lower than that determined using laser particle size analysis. This is mainly because of the different sample preparation processes. TEM gives the size in the dry state, whereas laser particle size analysis gives the size of the hydrated sample.
The magnetic properties of the catalyst were studied using a VSM, with a peak field of 10 kOe. As shown in Fig. 6, the CoFe@Si-PW catalyst was ferromagnetic. The Ms value was lower than that reported for bulk samples [39]. This decrease in the Ms is mainly attributed to the small particle surface effect, which becomes more dominant as the particle size decreases [40], and may arise from greater disorder of surface spins and the presence of SiO2 (which forms a diamagnetic matrix, which dilutes the magnetization of the otherwise compacted ferrite particles).
The effects of various reaction parameters on the oxidation of a model substrate, namely methyl phenyl sulfide (MPS), were investigated (Fig. 7) to optimize the reaction conditions for the liquid sulfides. The results show that increasing the reaction temperature had no effect on the yield, but decreased the reaction time (Fig. 7(a)). Subsequent experiments were therefore performed at room temperature. As shown in Fig. 7(b), use of an 8:1 O/S molar ratio gave methyl phenyl sulfone (MPSO2) as the sole product, in 98% yield. When the catalyst loading was decreased from 0.04 to 0.02 g, the yield of MPSO2 decreased. Increasing the catalyst weight from 0.04 to 0.08 g did not significantly increase the MPSO2 yield (Fig. 7(c)).
The best result was therefore obtained using 0.04 g of catalyst and an 8:1 O/S molar ratio at room temperature. These optimized reaction conditions were used to oxidize a series of liquid sulfides to their corresponding sulfones (Table 2, entries 1-7). The results in Table 2 show that aromatic and aliphatic sulfones were mostly obtained with 100% selectivity.
Optimization of the conditions for oxidation of solid sulfur compounds was performed using DBT as a model substrate. In this study, ethanol was used as the cosolvent for extractive desulfurization of the product, because it can be derived from agricultural products, is renewable and biologically less harmful in the environment, and has low toxicity compared with other cosolvents. The oxidation of DBT did not proceed at room temperature, therefore the influence of the O/S molar ratio (Fig. 8) and catalyst loading (Fig. 9) on DBT oxidation was investigated at 60 °C. The results show that when 0.15 g of catalyst and a 12:1 O/S molar ratio were used, the corresponding sulfone (DBTO2) was produced with 100% selectivity.
Various solid substrates were oxidized to the corresponding sulfones under these optimized conditions, with high selectivities (Table 2, entries 8-11). The reactivities of the sulfur compounds were influenced by two main factors, i.e., the electron density on the S atom and the steric hindrance of the sulfur compound. The electron density on the S atom of thiophene (entry 7) is lower than that on the BT sulfur (entry 9), therefore its reactivity is lower. The reactivity of DBT was lower than that of BT because of the higher steric hindrance.
The ODS of model oil containing various S compounds (feed 1) was performed using the optimized reaction conditions. The S removal from the model oil reached 90% after 4 h. The desulfurization efficiency of the catalyst was investigated using ODS of the model oil with a low concentration of DBT (feed 2). The results show that S removal from the model oil reached 85% after 4 h. This shows that the catalyst is efficient even at low DBT concentrations.
Activity tests were carried out on the model oils containing nitrogen compounds (feeds 3 and 4) to ascertain the effect of N compounds on ODS. Fig. 10 shows the ODS conversions for the model feeds containing indole and quinoline as a function of reaction time. It shows that the ODS activity decreased more in the presence of indole than in the presence of quinoline. These results suggest that conversion of the nitrogen compounds occurred before that of DBT, which implies that sulfur and nitrogen compounds are competitive in the oxidation. The higher activity of indole compared with that of quinoline is ascribed to the higher electron density on the nitrogen atom in indole [45]. Indole oxidation under mild reaction conditions has been reported [46], but studies of quinoline oxidation have been performed at high reaction temperatures (>200 °C) [47].
The reusability of the catalyst was investigated for oxidation of MPS. When the reaction was complete, the catalyst was separated from the mixture using a magnet, dried, and reused in a second run with fresh substrates under the same reaction conditions. The data shown in Fig. 11 indicate that the catalytic oxidation ability of CoFe@Si-PW decreased slightly. After four cycles, the yield of MPSO2 decreased by only 10%. This loss is attributed to leaching of PW species from the solid. To test this assertion, after the first cycle, the catalyst was separated, and the reaction between MPS and H2O2 was carried out using the clear solution. In this case, 5% of MPS was converted to the sulfone, suggesting that leaching of PW species was negligible. The weight percentages of PW in the fresh and recovered catalysts after four runs, determined using ICP-AES, were 31.8 and 29.2 wt%, respectively. This means that 8.2 wt% of PW had leached into the reaction media after four successive runs. The weight of recovered catalyst after each run was also determined, and only 7.5 wt% of the catalyst was lost.
The proposed mechanism is shown in Scheme 2. Immobilized PW anions react with H2O2 to generate peroxytungstate species [48,49]. This species facilitates transfer of the electrophilic oxygen to the sulfide, to yield the corresponding sulfone. The sulfone can be easily removed by extraction with a polar solvent.
Table 3 compares the catalytic activity of the CoFe@Si-PW catalyst with those of other reported catalysts for sulfoxidation reactions. The CoFe@Si-PW catalyst shows higher activity than the other catalysts and is magnetically recyclable.
PW-functionalized magnetic nanoparticles CoFe@Si-PW were prepared by anchoring PW on the surfaces of silica-coated CoFe2O4 magnetic nanoparticles. The nanoparticles were successfully used in the oxidation of sulfur-containing compounds. Various sulfides were selectively oxidized to their corresponding sulfones under mild reaction conditions. The CoFe@Si-PW catalyst also showed high activity for ODS of model oils containing DBT, BT, and thiophene, using H2O2 as the oxidant. The effect of nitrogen compounds on the ODS activity of the catalyst was also studied. The results suggest that quinoline inhibits ODS more than indole does. This is attributed to competitive adsorption between sulfur and nitrogen compounds for catalytic sites. This catalyst showed excellent catalytic activity and high selectivity, and can be reused at least four times, with only a slight decrease in catalytic activity.
2015, Vol. 36 Issue (8): 1342-1349
PDF (4645 KB)
2015, Vol. 36 Issue (8): 1342-1349
PDF (4645 KB)