Catalyst Research Center,Department of Chemical Engineering,Faculty of Engineering,Razi University,Kermanshah,Iran

Keywords:Oxidative desulfurization Hierarchical alumina-silica Silicate-1 seed-induced route Box-Behnken experimental design

ABSTRACT The catalytic performance of Mo supported on hierarchical alumina-silica(Si/Al=15)with Mo loadings of 3,6 and 15 wt%was investigated for the oxidative desulfurization(ODS)of model and real oil samples.Hierarchical alumina-silica(hAl-Si)was synthesized by economical and ecofriendly silicate-1 seed-induced route using cetyltrimethylammonium bromide(CTAB)as mesoporogen.The effect of CTAB on the structure of catalyst was studied by characterization techniques.The results revealed that 6%Mo/hAl-Si had the highest sulfur removal compared to the other catalyst loadings.The effect of operating parameters was evaluated using Box-Behnken experimental design.The optimal desulfurization conditions with the 6%Mo/hAl-Si catalyst were determined at oxidation temperature of 67 °C,oxidation time of 42 min,H2O2/S molar ratio of 8 and catalyst dosage of 0.008 g·ml?1 for achieving a conversion of 95%.Under optimal conditions,different sulfur-containing compounds with initial concentration of 1000 ppm,Dibenzothiophene(DBT),Benzothiophene(BT)and Thiophen(Th),showed the catalytic oxidation reactivity in the order of DBT ＞BT ＞Th.According to the regeneration experiments,the 6%Mo/hAl-Si catalyst was reused 4 times with a little reduction in the performance.Also,the total sulfur content of gasoline and diesel after ODS process reached 156.6 and 4592.2 ppm,respectively.

1.Introduction

SOXforms from the combustion of fuels containing sulfur,such as gasoline and diesel.These compounds contribute to air pollution,formation acid rain,catalyst poisoning,and corrosion in pipeline,pumping,and other refinery equipment[1,2].So the sulfur content in fuels is controlled by environmental regulation.To this purpose,researchers have investigated several methods to purify fuels from sulfur compounds[3].Oxidative desulfurization has been regarded to be one of the new outstanding approaches,having several advantages over the Hydrodesulfurization (HDS)method.This remediation process can be effectively accomplished under mild operating condition(at room temperature,ambient pressure,no requiring hydrogen).Also,such refractory sulfur compounds such as BT,DBT,and 4,6-Dimethyldibenzothiophene (4,6 DMDBT)are highly reactive in the ODS duo to their high electron densities [4].

The ODS process involves two main steps,the oxidation of thiophenic compounds with suitable oxidants to their corresponding sulfones in the first step followed by removing sulfones by extraction or adsorption in the second step [5,6].BT,DBT,and 4,6 DMDBT sulfur compounds have relatively large molecular size.The use of catalyst with small pore diameter leads to major drawbacks[7,8]:(I)Inaccessibility of bully reactants to active sites located in inside catalyst pores;(II)Trapping smaller molecules inside the pore structure followed by the occurrence of side reactions and coke deposition.

To overcome these disadvantages,mesoporous and hierarchical(micropores-mesoporous materials)structures are suggested.SBA-15/16 [9-11],MCM-41 [12,13],HMS [14,15],KIT-6 [16]and CMK-3[17,18]are some mesoporous materials used as supports in the ODS process.However,these supports have low selectivity toward the sulfur compounds because of low level of Lewis acid centers.Indeed,Lewis sites can promote the sulfur compound adsorption on the catalyst surface duo to interaction unbound paired electrons of sulfur compounds with free orbitals of Lewis centers.Finally reactants can access to the active sites on the surface of the catalyst easily[11].

Titania-silica and alumina-silica[19-23],compared to the mentioned mesoporous structures,have inherent Lewis acid centers of Ti4+and Al3+,respectively.Therefore,researchers have focused on the synthesis of hierarchical Titania-Silica and hierarchical aluminasilica.For example,Yang et al.[19]synthesized mesoporous TS-1 using a hybrid SiO2-TiO2xerogel combined with an organosilane precursor.Du et al.[20]synthesized hierarchical TS-1 via a novel threestep crystallization method by the use of PVA,an eco-friendly and inexpensive polymer as a mesopore template.They have compared the mesoporous TS-1 with a conventional TS-1 and reported that the mesoporous TS-1 removes sulfur compounds more effectively due to its mesoporous structure and high external surface area.

Among different methods for synthesizing hierarchical structures,surfactant-based methods are effective approaches[24].It is worth noting that surfactants are expensive,toxic,and difficult to obtain.Therefore,the silicate-1 seed-induced route is suggested.This synthesized method is economical and eco-friendly compared with the direct hydrothermal synthesis because of lower consumption of surfactants[25,26].Also,cationic surfactants are widely used because of the favorable interaction with negative charge during the synthesis process and easy dissolution in the synthesis solution.CTAB is an example of sufficient cationic surfactants[27].

Another key issue in the ODS process is the use of an active phase to further enhance the desulfurization efficiency by integration the positive effects of the support and active phase.The most commonly used active phases for the ODS process include the transition metal oxides such as Mo[3,12,13],V[6,18],W[28,29],Ti[30,31],etc.,among which molybdenum was selected since it is sufficiently available and remarkably active.Jin et al.[32]investigated activity of Ca/MoO3/Al2O3catalyst for the oxidation of 4,6-DMDBT,DBT and BT under mild conditions.Sulfur removal of these compounds reached close to 100%only in 10 min,8 min and 16 min,respectively.Yang et al.[3]used molybdenum supported on 4? molecular sieve for removing DBT.They also applied Box-Behnken experimental design to determine the optimum DBT conversion.Kang et al.[33]revealed that molybdenum supported on Ti-pillared interlayer clay(Ti-PILC)can remove nearly 100%of the sulfur content in the fuel under the mild reaction conditions.In another work,Sikarwar et al.[13]investigated the performance of Mo/MCM-41 catalyst for DBT removal.They reported that approximately 94%of DBT can remove under the optimal conduction.

Thus,the synergistic effect of the higher surface area,larger pore diameter,greater pore volume,appropriate Lewis acidity and metal active phase can promote catalytic activity for the ODS process.

Impregnation is one of the most interesting methods for preparing supported metal catalysts.The impregnation procedure involves contacting the porous support with a solution of the metal(oxide)precursor followed by evaporation of the solvent.Two main impregnation methods are distinguished,wet impregnation(WI)and incipient wetness impregnation(IWI).In the IWI method,the amount of solution is calculated to just fill the porous volume of the support,whereas in the WI method an excess amount of solution is used.IWI method is more preferable than WI because of the better dispersion of metal on the support[34].

Deferent variables such as temperature,time,catalyst dosage and H2O2/S molar ratio are effective operating factors in the ODS process.Sulfur removal can improve by optimizing these factors via one factor at a time(OFAT)and design of experiments(DOE)methods.The OFAT method involves modifying only one factor at a time,so it suffers from some disadvantages like increasing the number of experiments and not including the interactive effects of different variables [35-37],while design of experiments doesn't have these faults.The DOE method involves different techniques such as factorial design,fractional factorial design,response surface method (RSM),etc.Some recent reports published on the optimization of sulfur removal by RSM have been cited in[2,3,16,38]references.

In this study,hierarchical alumina-silica with Si/Al molar ratio of 15 as support was fabricated through silicate-1 seed-induced route using CTAB as mesoporogen.Molybdenum active site was loaded on the support by incipient wetness impregnation method with varying molybdenum contents.The structural properties of the catalysts were investigated with various characterized techniques.Next,the catalytic desulfurization process was carried out over both model and real oil samples by hydrogen peroxide and acetonitrile as the oxidant and extraction solvent,respectively.The operating parameters(temperature,time,H2O2/S molar ratio,and catalyst dosage)were optimized by Box-Behnken experimental design based on response surface methodology to obtain the maximum efficiency of desulfurization.Finally,the effect of substrate nature and the reusing ability of catalyst on the sulfur removal were examined.

2.Experimental

2.1.Materials

Tetraethylorthosilicate(TEOS,99%),aluminum nitrate nonahydrate(Al(NO3)3·9H2O,99.997%),tetrapropylammonium bromide (TPABr,98%),sodium hydroxide (NaOH,97%),ammonium heptamolybdate tetrahydrate ((NH4)6Mo7O24·4H2O,99%),cetyl trimethylammonium bromide(C19H42BrN,＞99%),Hydrogen peroxide(H2O2,35%wt),acetonitrile (CH3CN,99%),n-hexane (C6H14,99%),dibenzothiophene(C12H8S,99%,),benzothiophene (C8H6S,99%)and thiophene (C4H4S,99%)were purchased from Merck Company.gasoline and diesel with sulfur contents of 512 ppm and 8550 ppm were obtained from National Oil Products Distribution Company of Kermanshah.

2.2.Catalyst preparation

Hierarchical alumina-silica was synthesized in two steps.Silicate-1 seed solution was produced at first and then it was added to the main none-template alumina-silica.

2.2.1.Synthesis of the silicate-1 seeds

Silicate-1 seeds were synthesized as follows:0.04 g of NaOH,2.5 ml of tetrapropyl ammonium bromide,3.7 ml of TEOS and 10 ml of distilled water were mixed.The mixture was stirred for 2.5 h at room temperature.Then,0.34 g of CTAB was dissolved in to 5 ml of distilled water and added drop wise to the first solution under stirring at room temperature for 30 min.The obtained solution was aged at 140°C for 12 h to attain silicate-1 seed solution.After cooling,the obtained suspension was directly used in the next step of synthesis without further treatment[39].

2.2.2.Synthesis of hierarchical alumina-silica

The precursor of alumina-silica with Si/Al molar ratio of 15 was prepared with the following molar composition:6Na2O:3.333Al2O3:100SiO2:1500H2O.In a typical synthesis,4.841 g of Al2(NO3)3·9H2O2together 0.46 g of NaOH was dissolved in 52 ml of distilled water under stirring at room temperature.Then,43 ml of TEOS was added to the solution dropwise.Finally,2 g of pre-synthesized of silicate-1 seeds was added into the mixture.After stirring for 3 h at room temperature,the mixture was placed into the autoclave at 160°C for 8 h.The solid products were dried at 110 °C overnight and calcined in air at 550°C for 5 h.Ion exchange of Na-formed alumina-silica was carried out by 1 mol·L?1solution of NH4NO3at 80°C for 8 h under the reflux system.The product was washed by distilled water,dried at 90°C and calcined at 450°C for 3 h[40].

2.2.3.Synthesis of Mo supported on hierarchical alumina-silica

Molybdenum active site was loaded on the support via incipient wetness impregnation method.At the first,a required amount of ammonium heptamolybdate tetrahydrate was dissolved in to distilled water under stirring at room temperature.Secondly,a certain amount of support was added to the previous solution.After 1 h stirring,the solid products were dried at 60°C overnight and calcined in air at 500 °C,with a heating rate of 1 °C·min?1for 4 h [3].The catalysts were donated as xMo/hAl-Si,where x refers to Mo loading.

2.3.Characterization of catalyst

In this project,the synthesized samples were characterized by XRD,FESEM,FTIR,BET and NH3-TPD.In order to determine the sample phases,X-ray diffraction(XRD)patterns were recorded using Philips PW 1730 with a radiation of Cu kα in wavelength of 0.1540 nm.Moreover,the morphology and elemental analysis of the catalyst were investigated by applying Field Emission Scanning Electron Microscopy(FESEM)and Energy-dispersive X-ray spectroscopy(EDX)respectively using TESCAN MIRA.Textural properties such as surface area,pore diameter,and pore volume were determined by Brunauer-Emmett-Teller (BET)and BJH method using PHS-1020 (PHSCHINA)at 77 k under N2flow.Fourier-transform infrared (FT-IR)spectroscopy was performed using Bruker,ALPHA.NH3-Temperature programmed Desorption(NH3-TPD)experiments were carried out using a NanoSORD NS91 (made by Sensiran Co.,Iran)apparatus.50 mg of sample was pretreated at 300°C for 1 h and then cooled down to 110°C under N2flow.NH3gas stream was injected at the same temperature until adsorption saturation was reached.The samples were purged in a He stream at 110 °C for 30 min to remove the loosely bound ammonia.Then,they were heated again from 110 to 900°C at a heating rate of 10°C·min?1in a He flow.

2.4.Oxidative desulfurization reaction

Oxidative desulfurization experiments were performed using H2O2,acetonitrile,and xMo/hAl-Si as oxidant,extraction solvent,and catalyst,respectively.The model oils were obtained via dissolving Th,BT,and DBT in the n-hexane separately with the initial sulfur concentration of 1000 ppm.In a typical reactive cycle,5 ml of the model oil,5 ml of acetonitrile,a certain amount of catalyst (0.01 g to 0.05 g)and oxidant(molar ratio of H2O2/S of 1 to 15)were added into the roundbottomed flask(50 ml).The reactor was placed in a water bath under the oxidation temperature from 50 °C to 80 °C and oxidation times between 20 min to 60 min under vigorous stirring at a constant speed of 500 r·min?1.At the end of each reaction,the amount of sulfur in the model oil was detected by Antek instruments Model 9000F Sulfur Analyzer.The desulfurization yield(Y)was calculated using the following relationship[Eq.(1)]:

where C0and C refer to the initial sulfur concentration and final sulfur concentration after t minutes reaction time.

2.5.Design of experiments

The effect of temperature,time,catalyst dosage,and molar ratio of hydrogen peroxide to sulfur on the efficiency of desulfurization was investigated by a 3-level four factor Box-Behnken experimental design with three replicates at the center point.The statistical software Design Expert(Version 7)was applied for this aim.The input parameters with their levels are listed in Table 1.The other variables such as mixing rate and ratio of aqueous phase to organic phase were fixed at 500 r·min?1and 1,respectively.

Table 1 Range and levels of the operating parameters in the experimental design

The predicted responses were calculated from quadratic polynomial equation.This equation is presented in the following form[Eq.(2)]:

where Y refers to the response variable,Xiand Xjare independent variables and β0,βi,βii,and βijare intercept,linear,quadratic,and interaction coefficient,respectively [41].F-test and P-test recognize the statistical significance of the model and the variables.The larger F-statistic(F ＞1)and smaller P-value(p ＜0.05)show the more significant of the corresponding variables [42].

3.Results and Discussion

3.1.Structural characterization

3.1.1.XRD

The X-ray diffraction patterns of Mo supported on hierarchical alumina-silica with different Mo contents(3 wt%,6 wt%,and 15 wt%)are presented in Fig.1.The broad characteristic peak at 2θ=20°-30°(centered at 22.1°)is assigned to amorphous SiO2[43].There is no obvious peak corresponding to the metal oxide for 3%Mo/hAl-Si and 6%Mo/hAl-Si,indicating the amorphous nature of MoO3and homogeneous dispersion of molybdenum species into the framework of the supports [44].However,when the amount of Mo doping increased up to 15%,a relatively sharp peak at 2θ=27.33°appears.This peak is attributed to MoO3crystallite species[45].

Fig.1.XRD patterns of synthesized catalysts.(a)3%Mo/hAl-Si,(b)6%Mo/hAl-Si,(c)15%Mo/hAl-Si.

3.1.2.FTIR

Fig.2 presents the FTIR spectra of MoO3and xMo/hAl-Si(x=3,6,15%)samples.The bands at 618.05 cm?1,885.03 cm?1,and 991.59 cm?1in MoO3spectra are attributed to the δ(Mo--O--Mo)mode,Mo--O--Mo vibrations of Mo6+,and vibrations of vMo=O,respectively.The IR bands at 3435.35 cm?1and 1628.70 cm?1are related to the bending mode of adsorbed water[45,46].

In FTIR spectra of samples (a,b,c),the absorption bands close to 466.52,815.98,and 1093.92 cm?1are attributed to the vibration band of T--O(T=Si,Al),external symmetric stretch of T ?O ?T(T=Si,Al),and internal asymmetric stretch of T ?O ?T(T=Si,Al),respectively.The broad band located at 3428 cm?1and other at 1600 cm?1are ascribed to the vibration of hydroxyl groups of water.Furthermore,the band close to 3742 cm?1is correlated to the vibration of hydroxyl in Si--O--H group.Si--O--H groups are Bronsted acid sites located in the terminal structure of catalyst.All the absorption bands of MoO3are clearly observed in all of the samples.With increasing molybdenum loading,the bands related to molybdenum become clearer whereas those related to Si--O--H groups gradually disappear[47,48].

Fig.2.FTIR spectra of(a)3%Mo/hAl-Si,(b)6%Mo/hAl-Si,(c)15%Mo/hAl-Si,(d)MoO3.

Fig.4.EDX spectrum of 6%Mo/hAl-Si.

3.1.3.FESEM-EDX

Fig.3.FESEM images of 6%Mo/hAl-Si.

Table 2 Composition of 6%Mo/hAl-Si from EDX analysis

The FESEM images of 6%hAl-Si (low and high magnifications)are presented in Fig.3.It can be seen that the catalyst is composed of aggregated nanoparticles with size in the range of 30-35 nm.The formation of aggregated nanoparticles can be attributed to performance of the Surfactant.CTAB can stabilize the nanoparticles in the synthesis solution and decrease their surface Gibbs free energy via strong Coulomb force between CTA+(polar head group of CTAB)with nanoparticles during the hydrothermal synthesis meaning that a self-assembly process occurs.Also,the interactions between CTAB with particles can prevent their excessive growth[26,49].

The elemental analysis of 6%hAl-Si has been achieved by EDX technique.The corresponding results in Fig.4 and Table 2 show the core levels of Al,Si,O,and Mo in the catalyst structure with the absence of any impurity or other elements.

Fig.5.N2adsorption-desorption isotherms and BJH pores size distributions of(a)3%Mo/hAl-Si,(b)6%Mo/hAl-Si,(c)15%Mo/hAl-Si.

3.1.4.Textural properties

Fig.5 shows the N2adsorption-desorption isotherms of the catalyst samples.All of the samples have a type IV isotherm with H3 type hysteresis loop(at relative pressures(P/P0)between 0.6 and 0.9)that indicate the formation of a mesoporous structure.Also,the nitrogen uptake in relative pressure range of less than 0.01 corresponds to the presence of micropores on the surface of the catalysts[50,51].Therefore,it can be said that hierarchical structure was well-constructed.Mesoporosity in these samples was originated from the void space between accumulated nanoparticles[26].On the other hand,the same shape of hysteresis loop in 3%Mo and 6%Mo supported on hierarchical alumina-silica can be attributed to their similar pore structure.When Mo content increased up to 15%,the shape of hysteresis loop changed slightly due to entering the amount of molybdenum inside the porous[51].Textural properties of the synthesized catalysts are reported in the Table 3.The decrease in the BET surface area and pore volume with increasing molybdenum loading up to 15%may be related to the pore blockage[51].Also,the average pore diameter of the synthesized catalysts decreases with increasing molybdenum content as showed in the BJH pore size distribution curves.

Table 3 Textural properties of the synthesized catalysts

3.1.5.Acidity characteristics of catalyst

Fig.6 shows the NH3-TPD profile of the synthesized catalysts.Two distinct desorption peaks are seen at low-temperature(located at 200°C)and high-temperature(located at 700°C)which can be attributed to the weak (non-framework Al (Al3+))and strong acid sites(Al--O--Si groups),respectively.The total acidity value for hAl-Si,3%Mo/hAl-Si,6%Mo/hAl-Si and 15%Mo/hAl-Si is 0.734,0.726,0.638 and 0.615 mmol·g?1,respectively.The abundance of Al3+species can promote the catalytic activity via trapping the unbound paired electrons of DBT and absorbing them onto the catalyst surface,providing a favorable interaction between the target molecule and the surface active sites.There is no obvious difference between the original hAl-Si catalyst and the 3wt%Mo/hAl-Si sample.With increasing molybdenum content,peak position dependent on strong acid sites shifted to lower temperatures because of the migration of part of the molybdenum into the alumina-silica channels.On the other hand,there is no significant change in the peak positions of the weak acid sites.It can attribute to the weak interaction between MoO3species and weak acid sites[43].

Fig.6.NH3-TPD profile of a)hAl-Si,b)3%Mo/hAl-Si,c)6%Mo/hAl-Si,d)15%Mo/hAl-Si.

3.2.Catalytic studies

3.2.1.Effect of molybdenum loading(wt%)

The influence of molybdenum loading on the sulfur removal was investigated under the same conditions(oxidation temperature of 50°C,oxidation time of 60 min,H2O2/S molar ratio of 34,catalyst dosage of 0.02 g,and extraction solvent phase to organic phase ratio of 1).According to the presented results in Fig.7,6%Mo/hAl-Si had the highest efficiency of desulfurization among three catalyst samples.One can see that the desulfurization efficiency diminished as the Mo loading reached to 15%.This behavior may be related to the accumulation of Mo oxides on the catalyst surface which causes decreasing BET surface area and pore volume(according to the Table 3 and Fig.5).Indeed,6%Mo was dispersed homogeneously on the support surface of hierarchical alumina-silica compared to 15%Mo(according to the XRD patterns).Similarly,the sample containing 3%Mo loading showed the desulfurization efficiency lower than that of 6%Mo/hAl-Si,indicating the poor catalytic activity caused by the lack of active sites required to accelerate the catalytic reactions[13].

Fig.7.Influence of molybdenum loading on the conversion of DBT.

Table 4 Experimental design matrix and response results of DBT conversion

Table 5 ANOVA results of the predicted model for DBT conversion

3.2.2.The effect of solvent

In order to investigate the effect of solvent on the sulfur removal,desulfurization of model oil occurred at temperature of 50 °C under catalyst free conditions within 60 min with extraction solvent phase to organic phase ratio of 1.In extraction system,acetonitrile was used as extractant,and sulfur content of DBT by this solvent approximately reached 700 ppm which is less than the values obtained from desulfurization in the presence of a catalyst.Sulfur removal in the presence of acetonitrile can be attributed to extraction of dibenzothiophene by acetonitrile.

Fig.8.(a)Normal%probability against internally studentized residuals,(b)internally studentized residuals against run number,(c)Predicted response against actual response.

3.2.3.Statistical analysis

The proposed Box-Behnken design consisted of 27-runs at random.Table 4 shows these experimental conditions together with their corresponding experimental and predicted responses.A quadratic model was derived to explain the mathematical relationship between efficiency of desulfurization as response with input variables.This model equation in terms of coded factors is given as follows(Eq.(3)):

where X1,X2,X3and X4denote reaction temperature,oxidant/sulfur molar ratio,oxidation time and catalyst dosage,respectively.

The competence of the model was evaluated through analysis of variance(ANOVA)showed in Table 5.Based on ANOVA,for this model the F-value and P-value were 81.96 and ＜0.0001,respectively,meaning that the model has well predictability.Also,not significant lack of fit(P-value:0.1813)indicates that the model obtained is desirable.Table 5 illustrates that the variables and interactions such as X1,X2,X3,X4,X1X2,X1X3,X1X4,X2X3,X12,X22,andare significant model terms because their P-values are less than 0.05.The parameter X2with F-value of 203.13 is the most influential parameter in the model.Also,the values of R2and adjust R2are 0.9860 and 0.9739,respectively.The small difference between this two values indicated that the model has well accuracy.On the other hand,the value of adequate precision is satisfactory because this value is greater than 4.

The residual plots can be used to assess the adequacy of regression model.The plot of normal probability against internally studentized residuals is presented in Fig.8(a).This plot shows that the error terms are normally distributed closed to a straight line.Also,the internally studentized residual plot versus run number is illustrated in Fig.8(b).The dispersion of random residuals is along the line which indicated the model's validity.The predicted vs.actual value plot is depicted in Fig.8(c).According to this plot,the predicted and experimental values are very close to each other;therefore the model is fit and desirable[3,38].

3.2.4.Effects of main operating variables on catalytic ODS

The effect of operational parameters on sulfur removal efficiency is shown in Fig.9(a-d).

3.2.4.1.Effect of oxidant/sulfur molar ratio.The effect of H2O2/S molar ratio on efficiency of desulfurization is exhibited in Fig.9(a)while the temperature,time and catalyst dosage are kept at the medium values of 65 °C,40 min,and 0.03 g,respectively.According to Fig.9(a),the sulfur removal curve has an absolute extrema in terms of H2O2/S molar ratio.The increase in H2O2/S molar ratio up to 8 provides more active radicals(HO·)required for the formation of hydroperoxomolybdate species [52].By further increasing H2O2/S molar ratio,undesirable decomposition of H2O2to H2O decreases the efficiency of desulfurization[53].Indeed,according to reaction stoichiometry,1 mol DBT usually needs 2 mol H2O2for complete conversion into sulfone.But the results show that the excess amount of oxidant is required because of the undesirable decomposition of H2O2[3,53].The parallel decomposition reactions of H2O2are presented as follows[53]:

Fig.9.Effects of main variables on DBT conversion in ODS reaction.

Fig.10.Response surface and response contour plots for DBT conversion.

In the first reaction,radical species of HO·are strong agents which oxidize sulfur compounds.On the other hand,the second reaction has a negative effect on the sulfur removal due to more water production.This reaction happens at high temperatures,during the long reaction times,or in the presence of excess H2O2[53].

3.2.4.2.Effect of reaction temperature.Fig.9(b)illustrates the efficiency of sulfur removal as a function of temperature at the medium values of time,H2O2/S molar ratio,and amount of catalyst.The rising temperature up to about 67°C has a positive effect on the ODS yield.Indeed,increasing the reaction temperature causes the molecules to move faster,and the collision probability between reactants increases.Also,at the high temperatures,the most strongly adsorbed solfones would be removed from the catalyst surface easily.The excessive temperature rise decreases sulfur removal owing to decomposition of the oxidant[16,54].

3.2.4.3.Effect of reaction time.Fig.9(c)shows the effect of reaction time on the DBT conversion.The other parameters are kept at the fixed values.The sulfur removal improved by increasing reaction time up to 42 min,meaning that it needed enough time to complete the reactions between DBT and H2O2.However,further increase in reaction time from 42 min to 60 min had a negligible negative effect on the sulfur removal trend.As the result of a long-time ODS process,the decomposition of H2O2to give water as well as the pore blockage of the catalyst caused by deposition of adsorbed oxidized products on the catalyst surface may be the reasons of this observation[53,55].

3.2.4.4.Effect of catalyst dosage.Fig.9(d)shows that upon increasing catalyst dosage from 0.01 to 0.03 g at medium values of the other parameters,the efficiency of desulfurization increased slightly due to increasing the number of hydroperoxomolybdate species[56,57].The further increasing the catalyst dosage didn't appreciably improve the desulfurization efficiency.Indeed,the addition of excess amount of catalyst leads to particle agglomeration of catalyst,thereby reducing the contact area between catalyst and model oil[13].

3.2.5.Combined effect of process variables

3-D response surface and response contour plots for combined effect of(a)temperature and H2O2/S molar ratio,(b)temperature and time,(c)temperature and catalyst,(d)time and H2O2/S molar ratio on DBT conversion are presented in Fig.10.

In Fig.10(a),the reaction time and catalyst dosage were kept at the medium values of 40 min and 0.03 g,respectively.At high temperatures,the sulfur removal decreased at the low or high H2O2/S molar ratio.This phenomenon can be related to the decomposition of H2O2at high temperatures[57].The maximum sulfur removal(＞%90)was obtained at 6 ＜H2O2/S ＜12 and 56°C ＜temperature ＜77°C.

In Fig.10(b),H2O2/S molar ratio and catalyst dosage were kept at 8 g and 0.03 g,respectively.The results indicate that the maximum sulfur removal(＞%90)was achieved when reaction temperature and reaction time were kept at about 57-78°C and 30-60 min,respectively.At the lower oxidation time,the sulfur removal decreased at both high and low temperatures because the reaction does not have enough time to complete[53,55].Also,at the low temperatures,sulfur removal percentage decreased because of decreasing collision probably between the reactants[3,16].

In Fig.10(c),H2O2/S molar ratio and reaction time were fixed at 8 min and 40 min,respectively.The maximum catalytic conversion(＞%90)can be seen where reaction temperature and catalyst dosage were limited to 57-79°C and 0.02-0.03 g,respectively.When catalyst dosage increases,it is expected that the sulfur removal to be increased,but the results in Fig.10(c)showed that at low temperature and high amount of catalyst,the efficiency of sulfur removal is low.This result relates to the strong adsorption of polar products on the catalyst surface at low temperatures[16].

In Fig.10(d),temperature and catalyst dosage were kept at 65°C and 0.03 g,respectively.When reaction time is short,reactants do not have enough time to react with each other in the high and low H2O2/S molar ratio,therefore sulfur removal decreases.On the other hand,H2O2decomposes in the long time and high amount of oxidant [53,55].The highest sulfur removal(＞90%)was achieved when H2O2/S molar ratio and reaction time were about 6-13 and 19 min-60 min,respectively.

3.2.6.Optimization by Box-Behnken design

The response optimization technique has been used to determine the optimal conditions for four main parameters:H2O2/S molar ratio,catalyst dosage,oxidation temperature and time with 100%desirability.According to the results,the highest sulfur removal(97%)was obtained under the optimal conditions:oxidation temperature of 67°C,oxidation time of 42 min,H2O2/S molar ratio of 8 and catalyst dosage of 0.04 g.In order to investigate the predicted conditions,an experiment under optimal conditions was conducted three times.A 95%conversion rate was obtained which has a good agreement with the predicted value.It suggests the adequacy of the regression equation.

3.2.7.Reactivity of different sulfur containing compounds

The effect of substrate nature on the desulfurization was examined under the optimal conditions.The results in Fig.11 demonstrate that the catalytic activity of 6%Mo/hAl-Si in the presence of Th,BT,DBT decreases in the order of DBT ＞BT ＞Th.This reactivity trend is consistent with the order of electron density of S atoms because of the electrophilic mechanism of ODS process.The electron densities on sulfur atoms are 5.696,5.739 and 5.758 for Th,BT and DBT,respectively.The lower sulfur removal activity of Th compared to other substrates can be assigned to its lower electron density[58,59].

Fig.11.Sulfur removal of different sulfur compounds.

3.2.8.Oxidative desulfurization of commercial diesel and gasoline

To assess the capability of the ODS process in practical applications,the catalytic performance of the efficient sample(6%Mo/hAl-Si)was evaluated for desulfurization of both diesel and gasoline oil samples under the optimal conditions(oxidation temperature of 67°C,oxidation time of 42 min,H2O2/S molar ratio of 8 and catalyst dosage of 0.04 g).The initial sulfur content of diesel and gasoline was 8550 ppm and 512 ppm respectively.The efficiency of sulfur removal from gasoline and diesel were 69.4 and 46.29%,respectively.It can be seen from the results that the desulfurization of diesel was more difficult than that of gasoline since diesel has complex sulfur compounds with multiple alkyl side chains.Therefore,it is expected that the steric hindrance inhibits the desirable oxidation of diesel[60].

3.2.9.Reusability of catalyst

The reusability of the catalyst was investigated under the optimal conditions.After each catalytic run,the used catalyst was recovered from the reaction mixture by centrifuging and rinsed several times with methanol to ensure that the adsorbed sulfones was removed completely.Then,the recycled catalyst was allowed to dry at 80°C overnight in an oven.Next,fresh oxidant,extraction solvent and model oil together with the recycled catalyst were employed for the next experiment.Fig.12 showed that the catalytic activity decreased slightly from 95%to 86%duo to the loss of active sites and more adsorption of sulfones on the catalyst surface[12].Also,to investigate the likely changes in the structure of the catalyst sample,FTIR analysis was taken from the fresh and washed samples.This is important for study of stability and reuse ability of catalysts.As shown in Fig.13,no significant changes are observed in the catalyst structure before and after using.Although,the peak located at 970 cm?1disappeared in the washed sample because of slightly losing Mo active sites upon washing treatment.In general,it can be stated that 6%Mo/hAl-Si catalyst has the suitable stability features for ODS process.

Fig.12.Reusability of 6%Mo/hAl-Si in the ODS reaction of DBT.

Fig.13.FTIR spectra of fresh(a)and washed catalyst(b).

4.Oxidation Desulfurization Mechanism

Oxidative desulfurization reaction is a surface reaction initiated by the active species of hydroperoxomolybdate as a product of interaction between H2O2and molybdenum alumina-silica catalyst.Subsequently,the hydroperoxomolybdate species formed the peroxo species during elimination process of H2O.The sulfur compounds reacted with the generated peroxo species and formed sulfoxide and regenerated peroxo agents.In the final step,sulfoxides are also oxidized to sulfones.The presence of AL3+on the support surface causes more adsorption of DBT molecules on the catalyst surface and subsequently the favorable interaction between them and molybdenum active sites[61].

5.Conclusions

The oxidative desulfurization from model and real oil samples was carried out using Mo supported on hierarchical alumina-silica with various Mo loadings of 3 wt%,6 wt%and 15 wt%.The hierarchical aluminasilica was synthesized by silicate-1 seed-induced route in the presence of CTAB.The addition of CTAB in the structure of catalyst can form small aggregates.The void space between the aggregated nanoparticles leads to the formation of mesoporous structures.The excellent catalytic performance of 6%Mo/hAl-Si compared to 3%Mo/hAl-Si and 15%Mo/hAl-Si can be the result of more active sites,higher surface area,larger pore volume and sufficient dispersion of molybdenum on the support.Modeling and optimization of ODS process were investigated using Box-Behnken design.The values of the model fitness parameters implied that the actual results were in good agreement with the predicted results.The optimum conditions to obtain the maximum desulfurization from DBT were as follows:the oxidation temperature of 67°C,oxidation time of 41 min,H2O2/S molar ratio of 8 and catalyst dosage of 0.04 g.According to the results,oxidative reactivity of DBT was higher than those of TH and BT because of its high electron density.The recycling experiments indicate that the 6%Mo/hAl-Si catalyst can be reused for four times with appropriate stability.Also,the total sulfur content of gasoline and diesel after ODS process reached to 156.672 ppm and 4592.205 ppm,respectively.