Fu Yang, Ruyi Wang, Shijian Zhou, Xuyu Wang, Yan Kong,*, Shuying Gao,*

1 School of Environmental and Chemical Engineering, Jiangsu University of Science and Technology, Zhenjiang 212003, China

2 State Key Laboratory of Materials-Oriented Chemical Engineering, College of Chemical Engineering, Nanjing Tech University, Nanjing 210009, China

Keywords:Nanomaterials Moleclar sieves Oxidative desulphurization Radicals Synergetic effect Surface

ABSTRACT The oxidative desulphurization (ODS) has become mainly popular by rapid catalytic oxidation of dibenzothiophene (DBT) relied on efficient heterogeneous catalyst.V-based catalytic active species were regarded as the potential option in the activity-preferred ODS systems.Herein, we reported the redispersion of vanadium oxide (VOx) on the mesoporous silica modified with manganese oxide (Mn3O4)through one progressive insertion approach of metal oxides in the silica.Impressively, mesoporeencaged vanadium-manganese oxides in the silica (VMn-MS) as the admirable output of excellent ODS catalyst was demonstrated compared to other monometal-modified counterparts and one-pot implanted one.The characterization results revealed the post-implanted VOx species not only deposited around the pre-covered Mn3O4 on the mesoporous surface but also inserted the surface layer of Mn3O4 inducing the amorphous evolution of aggregated Mn3O4 and the reconstruction of final active sites.This integrated approach made the reconstructed active species afford more exposed catalytic sites and the tailored surface redox cycles owing to the electronic communication of V-Mn.The catalytic results demonstrated the excellent catalytic desulphurization efficiency (～100%) during 60 min at 80 °C, which made the sulphur content reduce to 6 mg·L-1,remarkably superior to other comparative samples.The outstanding catalytic performance of VMn-MS catalyst can be ascribed to the synergistic effect of V-Mn dual metals rendering two different reaction pathways, which includes free-radical reaction and ring-forming reaction, where Mn site acted as active center triggering reactive free radicals which could be further optimized by surrounded V sites around Mn sites to promote the ODS process.

1.Introduction

As a consequence of the ongoing demand for improvements of the fuel oils quality, the development of catalytic deep oxidative desulfurization process(ODS)has in recent years become the focus of a significant number of studies, by merit of mild and simple reaction condition as well as the high efficiency in removing stubborn aromatic thiophenes including benzothiophene (BT),dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene(4,6-DMDBT) [1-3].

In general, ODS technology using catalytic process is mainly dependent on active transition metal oxides catalysts [4-8], in which vanadium-based catalysts show superior catalytic desulfurization activity.Specifically, Cedeno-Caeroet al.[9]developed a series of V2O5catalysts supported on Al2O3, TiO2, CeO2, NbOx,and SiO2applied to the oxidesulfurization process and demonstrated that the oxidation activity of DBTs depends on the support used.Rivoiraet al.[10]reported that vanadium and titanium oxide supported on mesoporous CMK-3 respectively for DBT oxidation,achieved a rational DBT elimination efficiency using hydrogen peroxide(H2O2)as oxidant under mild conditions.Zhu group[11]fabricated 2D-2D V2O5/BNNS nanocomposites, and 99.6% removal of dibenzothiophene (DBT, ≤2 mg·L-1) could be achieved under atmospheric pressure.

In recent years, vanadium-based catalysts have received wide attention in the field of deep oxidative desulfurization process due to their abundant coordination modes, low cost andetc.Improving the vanadium content on the carriers had been proved an effective strategy for enabling high ODS catalytic activity.However,upon loading high vanadium concentration,the active species trended to aggregate into bulk crystals,resulting in a low atom utilization efficiency.More seriously,a highly-loaded catalyst tends to cause a lower catalytic activity owing to the aggregation of active metal species and the blockage of diffusion channel to the reactant.Large surface area carriers could provide more surface regions to accommodate more active metal species and make them more dispersed to achieve the higher catalytic activity.Therefore, many porous carriers such as mesoporous silica [12,13], alumina [14],and carbon[15]have been devoted to enlarge the loading and dispersity of metal through various developed loaded means.However, these approaches would still cause above mentioned problems at a slightly higher loading of metal.Also, vanadium oxide species are toxic heavy metals with high costs, therefore,reducing the used dosage of vanadium precursor also became urgent.

To improve the dispersity of vanadium species on the carrier and reduce the usage dosage of vanadium species,many researchers consider introducing a second active metal to improve the catalytic performance through the synergistic effect between the two metals to achieve the purpose of deep desulfurization.Wanget al.[16]designed and prepared VOx-Ga-SBA-15 catalyst by incorporating gallium as heteroatom into silica framework,and then obtained VOxsupported Ga-SBA-15 catalyst by the wet dipping method.They found that the incorporation of gallium improved the dispersion of vanadium species and that the resulted composites were very active for most refractory sulfur compounds in the catalytic process.Teimouri group [17]studied a series of MoO3/V2O5/MCM-41 and achieved 99%conversion of dibenzothiophene under the optimum conditions.However, currently reported bimetallic supported catalysts are usually introduced simultaneously through some traditional coprecipitation method,so the resulted metal-tometal contact and interaction may be accompanied by a certain degree of randomness which is not easily controlled, and the ratio of two sets of metallic species with synergistic cooperation is not enough [18-20].Therefore, how to precisely spatiotemporal control the contact of two metal active species on the support remains one big challenge in the synthesis of synergetic catalyst.

Based on the above thoughts, we further designed and synthesized vanadium-manganese bimetal oxide-based mesoporous silica by a two-step progressive insertion method for deep oxidative desulfurization reaction.Both the structural characteristics and the chemical status of V and Mn were systematically characterized by XRD, FT-IR, UV-vis, H2-TPR, SEM, XPS, and so on.Furthermore,catalytic influence factors including reaction temperature, reaction time, and the durable catalytic performance were meticulously examined to identify the optimal reaction conditions.Finally,the possible deep oxidation mechanism was proposed and proved accordingly.

2.Experimental

2.1.Reagents and materials

Dodecyl amine (DDA, ＞98%), tetraethoxysilane (TEOS), vanadium sulfate tetrahydrate (VOSO4·4H2O), and manganese chloride tetrahydrate (MnCl2·4H2O) were purchased from Aladdin.Tertbutyl hydroperoxide (TBHP, 65% (mass)), ethanol anhydrous, and tetrachloromethane(CCl4)were purchased from Sinopharm Chemical Reagent Co., Ltd.n-octane,n-tetradecane, dibenzothiophene(DBT) (98%), 4-methyldibenzothiophene (4-MDBT, 99%) and dimethyl dibenzothiophene (4,6-DMDBT, 99%) were purchased from Sigma-Aldrich.All chemicals were used without further purification.

2.2.Catalyst design and preparation

The Mn-MS sample was prepared in advance according to our previously developed method[21].First,DDA was dissolved into a mixed solution consisting of 62.5 ml water and 25 ml ethanol and mixed uniformly at 45 °C, which was defined as solution A.An appropriate amount of MnCl2·4H2O precursor was added to the solution under stirring which was changed from milky white to brown translucent for the micelle solution.Then, TEOS was dropwise added to the solution and the stirring was kept for 4 h and static 18 h at 45°C.The resulting samples were obtained by centrifugation and washing with water and ethanol,and then calcined at 550°C in the air for 6 h in programmed temperature control.

Vanadium precursor was controlled to introduce into MnOxcovered mesoporous channel of silica by the post-impregnation method.Specifically, an ethanol solution of VOSO4·4H2O was used as a vanadium source.The amounts of VOSO4·4H2O were calculated to obtain controlled molar ratios of vanadium/manganese(0.2, 0.4, 0.6, 0.8, and 1).The above vanadium-bearing ethanol solution was stirred with Mn-MS sample at room temperature for 2 h, after that, rotating-evaporating until their complete dryness,then calcined at 450°C for 3 h in air.The final catalysts were defined asxVMn-MS, wherexrepresents the molar ratio of V/Mn.

For better comparison,mesoporous silica modified by Mn and V were prepared by the co-impregnation method.Pure HMS was prepared by the reported method [22], and the obtained HMS was then added to ethanol solutions of VOSO4·4H2O and MnCl2·4H2O.All remaining steps are the same as above.The prepared catalyst was marked as VMn-MS-P.Subsequently, the V-MS-P sample was also prepared by the same method excluding the addition of MnCl2·4H2O.

2.3.Catalyst characterizations

Powder X-ray diffraction (XRD) patterns were recorded by Smartlab TM 9 kW.

Fourier Transform Infrared (FT-IR) spectra were recorded on Bruker VECTOR22 resolution (Bruker, Germany).

The N2adsorption-desorption measurements were conducted on BELSORP-MINI volumetric adsorption analyzer.

SEM images were collected on a Hitachi S4800 field emission scanning electron microscopy (Hitachi, Tokyo, Japan).

High-resolution transmission electron microscopy (HRTEM)equipped with elemental mapping analysis was recorded on a JEM-2010 EX microscope (JEOL, Tokyo, Japan).

Diffused reflection UV-Vis spectra were recorded by Lambda 950 spectrophotometer (PerkinElmer, Waltham, MA, USA).

The X-ray photoelectron spectra (XPS) were acquired on a PHI 5000 Versa Probe X-ray photoelectron spectrometer (ULVAC-PHI,Kanagawa, Japan).The C1s peak at 284.8 eV was used as the reference.

The metal content was measured by inductively coupled plasma optical emission spectrometer(ICP-OES,PerkinElmer,Waltham, MA, USA).

Hydrogen temperature-programmed reduction (H2-TPR) measurements were performed using a fixed-bed reactor under a flow of 10% H2/Ar gas mixture and a heating rate of 10 °C·min-1from room temperature to 800 °C.Before the TPR analysis, the catalyst was pretreated by 30 ml·min-1flowing argon at 300 °C for 1 h.The consumption of H2during the reduction was measured continuously using a thermal conductivity detector (TCD).

2.4.The assessment of catalytic activity

The model oil was prepared by dissolving DBT in a solution ofnoctane with a sulfur content of 1000 mg·L-1.ODS experiments were carried out in a 50 ml three-necked round bottom flask fitted with a reflux condenser.Typical, 0.04 g catalyst, 10 ml model oil was added to the reactor.Afterward, 0.1 ml of TBHP was injected drop-wise into the above mixture under vigorous stirring.Then,the reacted oil was collected at 15 min time interval and was analyzedviaan internal standard (n-tetradecane) method by an SP-6890 gas chromatograph.The sulfur removal (%) was calculated by the following equation:

whereCtrepresents the concentration of sulfides at any timet,andC0represents the initial concentration of model oil.

Finally, the stability of the catalyst was assessed through conducting multiple reaction cycles,where the spent catalyst was separated from the completed reaction solution and reused for the next reaction cycles after a simple wash using acetone.

3.Results and Discussion

Fig.1 describes the path of V and Mn inserting HMS nanocomposites synthesized by a two-step strategy.First of all, Metal cations (Mn2+) and neutral surfactant dodecyl amine (DDA) were assembled to form metalized micelles similar to cationic templates through the coordination effect[21,23].The negatively charged silicate oligomers accumulate and deposit on the surface of the metallomicelle through electrostatic forces.A thin layer of manganese oxide (Mn3O4) is then anchored to the pore wall by removing the pore-forming agent.Secondly, the vanadium precursors were implanted into the Mn3O4-modified mesoporous channel by the post-impregnation approach, so that vanadium species were deposited around the pre-covered Mn3O4on the mesoporous surface.Finally, the nanocomposite with bimetal VOx-MnOxspecies adjacent contact is obtained by the secondary quenching treatment.Note that the post-introduced V species would insert into the internal structure of MnOxto form solid-solution like active species by the reconstructed process.

3.1.Characterizations of structure and morphology

Low-angle XRD patterns of obtained pure Mn-MS and seriesx-VMn-MS sample are shown in Fig.2(a), an intense diffraction(1 0 0) peak at 2θ = 2.4° was observed, which was the typical(1 1 0) characteristic peak of HMS molecular sieve [24], proving that the addition of Mn did not destroy the hexagonal ordered structure of HMS.It is worth noting that the characteristic peak at 2θ=2.4°is slightly shifted to a smaller angle after the introduction of foreign vanadium.When the nanopore of silica was occupied by the metal oxide, the pore wall thickness would be enlarged and the pore volumes and pore size would be compressed,resulting in the increase of unit cell parameter(a0).Meanwhile, the thermal treatment process after wet impregnating V precursor would cause the distortion of mesopore thereby resulting in the slight enlargement ofa0, enables the diffraction peak to shift to a smaller angle slightly.Besides, the diffraction peak intensity gradually weakens with ascending of vanadium loading in the series samples, until the V content is increased to V/Mn = 1, the diffraction peak almost disappears, which indicates that the mesostructure in Mn-MS is deteriorated to great extent,owing to post-implanted vanadium species in the mesopores.

The wide-angle XRD patterns of the prepared samples are collected and shown in Fig.2(b), an obvious broad diffraction peak in the range of 20°-30° can be observed in all samples, which is ascribed to the characteristic peak of amorphous silicon oxide[25].For initial Mn-MS, several weak diffraction peaks at around 32°, 36° and 59° were observed, corresponding to the lowcrystalline Mn3O4phase(PDF#24-0734)pre-covered on the mesoporous surface, indicating that manganese oxide has trace crystal particles in the mesoporous.Interestingly, with the increase in the amount of vanadium introduced into the catalyst,the intensity of the diffraction peak corresponding to Mn3O4showed a gradual downward trend.This could be attributed to the fact that the vanadium implanted later was inserted and mixed in manganese oxide nanocrystals under a high-temperature driven effect on the silicon oxide surface-induced the reconstruction of active V-Mn oxides,which formed the composite metal oxidation to some extent destroyed the crystallinity of manganese oxide.Further increasing the amount of vanadium to V/Mn ratio greater than 0.8, it can be seen that all Mn3O4characteristic peaks were not observed owing to the crystallinity of the Mn3O4microcrystal was completely destroyed by the post-impregnated vanadium.The above phenomena indicate the strong interaction and reconstruction between vanadium and manganese in the prepared bimetallic catalyst.

Fig.1. Depiction for the synthetic procedure of targeted VMn-MS catalyst.

Fig.2. (a) Low-angle and (b) wide-angle XRD patterns of pure Mn-MS and series samples of xVMn-MS.

Fig.3. N2 adsorption/desorption isotherms (a) and pore size distribution (b) of Mn-MS and series V-bearing xVMn-MS.

The textural information of the obtained sample was further collected by N2adsorption-desorption measurement and shown in Fig.3(a), For the Mn-MS sample, the typical type-IV isotherms ascribed to the presence of mesoporous are observed [26,27],which agrees with the low-angle XRD result.Note that the type IV-like hysteresis gradually evolved into a type I-like isotherm and accompanied with a weakening of capillary condensation with the increase of extraneous vanadium, this evidence suggests that the pore size of the target sample evolved from mesoporous to microporous level owing to the implantation of foreign vanadium onto the surface of Mn3O4within the confined channels.Besides,the corresponding size distribution of the obtained sample is shown in Fig.3(b).Note that the pore size of the Mn-MS sample is within the mesopore range, and with the gradual introduction of vanadium,the pore size of the seriesxVMn-MS samples showed decreased trend and was tailored into almost micropore size.Thereinto,the pore size decreased obviously as the vanadium content increased from 0.8 to 1 for the ratio of V/Mn,attributed to the residual vanadium species being coated on the Mn3O4surface resulting in pore size reduction and wall increase after partial vanadium were completely mixed with manganese oxide due to the introduction of excess vanadium.Moreover,their detailed textural structure information is further displayed in Table 1.As expected that the specific surface area, gas adsorption capacity,and pore size of the series VMn-MS samples show a significant decrease with the introduction of foreign vanadium, compared with pure Mn-MS,and the pore wall thickness increases.The above comprehensive phenomena suggest that vanadium particles were successfully implanted around Mn3O4in the mesoporous silicon channel,and all the obtained samples almost retain superior structural characters, which can provide good structural space for the catalytic reaction process.

Table 1Textural properties and detailed parameters of series catalyst samples

The microscale morphological characteristics of a series of prepared samples were characterized by SEM and representative SEM images are presented in Fig.4.As shown in Fig.4(a), it is obvious that the original Mn-MS sample with uniform spherical morphology with the particle size diameter between 200 and 300 nm can be observed.Besides, the variation in uniform spherical particle morphology is not remarkable before and after the introduction of trace vanadium oxide into the restricted mesopore, which suggested that vanadium oxides were uniformly distributed in the mesoporous channel (Fig.4(b)-(f)) rather than the outside region.Also, the element mapping analysis of the selected region in the 0.6VMn-MS sample can be presented in Fig.4, where V and Mn were coexisted and uniformly dispersed in the target sample,confirming the successful progressive insertion of two metal species.

Fig.4. Representative SEM images of series catalyst samples(a)Mn-MS,(b)0.2VMn-MS,(c)0.4VMn-MS,(d)0.6VMn-MS,and(e)0.8VMn-MS,(f)1VMn-MS,and(g)elemental mapping of 0.6VMn-MS.

To further get insight into the distribution of metal species and microscale structure of the target catalyst,the representative TEM images of 0.6VMn-MS samples were collected.Fig.5(a),(b)exhibit a typical micron-sized spherical particle morphology with reserved sponge-like frame and wormhole-like pores corresponding to HMS mesoporous material even undergoing successive implantation of the metal species[28].This observation is in good agreement with the previous N2adsorption-desorption result.What’s more,among all the collected TEM results, the aggregated vanadium and manganese particles are invisible in the presented HRTEM images(Fig.5(c), (d)), indicating that the metal oxide species were well dispersed in the mesoporous matrix with ultrafine sizes.And the corresponding fast Fourier transform (FFT) images (Fig.S1) of the catalyst in the selected region exhibit only diffuse Debye rings rather than distinct dots, further verifying the dominant amorphous nature of the 0.6VMn-MS.Apart from that, EDS results(Fig.5(e))and elemental mapping images(Fig.5(f))further proved that vanadium and manganese particles were successfully introduced into the confined channel and highly dispersed into the target sample.We speculate that the effect of Mn3O4covered in advance on the mesoporous surface of silica making them strongerintermetallic interaction with vanadium that facilitates vanadium dispersion.

Fig.5. TEM images (a-d) and EDS analysis (e) and corresponding elemental mapping images (f) of 0.6VMn-MS.

3.2.State of active metal species in the catalyst

FT-IR was used as a convenient and effective technique to indirectly reflect the dispersion and chemical environment of metal species on mesoporous silica surfaces.The FT-IR spectra of pure MS, Mn-MS, and seriesxVMn-MS are shown in Fig.6(a).All the samples exhibited the absorption peaks atca3000-3400 cm-1attributed to the surface adsorbed water.Meanwhile, the absorption peaks of all samples at 1090, 802, and 462 cm-1were observed, assigned to the symmetric and asymmetric stretching vibrations of Si-O-Si [29,30].Additionally, the absorption band at 967 cm-1is recognized as the characteristic of Si-OH groups[31].However, it is interesting that compared with pure MS, the characteristic peak is significantly reduced after thein-situintroduction of manganese species,which is attributed to the depletion of Si-OH due to the formation of Si-O-Mn by replacing H on the silicon wall with Mn, indicating that Mn3O4was successfully covered on the inner surface of mesoporous silicon.What’s more, we observed that the characteristic peak changes attributed to Si—OH were feeble after the further injection of the second vanadium metal species, which associates with the fact that vanadium oxide was surrounded by manganese oxides rather than silicon walls.This present result also favors the proposed synthetic route and further implies close contact and possible strong interaction between the two metal oxides.

Fig.6. FT-IR (a) and DRUV-vis (b) spectra of pure MS, Mn-MS, and series xVMn-MS.

To further explore the electronic structure and coordination environment of the introduced two metal species, UV-vis diffuse reflectance spectra are collected and shown in Fig.6(b).For 0.6 V-MS, two absorption peaks were identified at 250 nm and 340 nm, and the absorption band at 250 nm was assigned to the low-energy charge transfer transitions tetrahedral oxygen ligand to the V (V) in the isolated VO43-species [32,33].Another band at 340 was attributed to the V5+species bonding to the pore surface with V=O and V-O bonds [34].Besides, the band at about 520 nm could be assigned to polymeric octahedral bulk V2O5microcrystals on the surface [35].On the other hand, the typical characteristic bands at around 279 nm in Mn-MS were associated with the charge-transfer transition of O2-→Mn3+in octahedral coordination [36].It was noteworthy that the characteristic peak at 279 nm was shifted to 258 nm after the post-impregnation of vanadium, and the two bands corresponding to 340 nm and 520 nm in 0.6VMn-MS are invisible, indicates the strong interaction between adjacent VOxand Mn3O4and possible solidsolution existence in the confined channel thereby impeding the further aggregation of vanadium oxide in the channel.Besides,the broad absorption band at about 410 nm was obviously observed as the amount of vanadium introduced exceeded V/Mn = 0.6, which was attributed to the oligomer pseudotetrahedral VOxspecies by excessive vanadium species[33],which also implied excessive vanadium species might induce the reduction in catalytic performance.

H2-TPR is considered to be a convenient and efficient detection technique for identifying the reduction process of bimetallic V and Mn oxide-supported nanoporous molecular sieve, the H2-TPR profiles of series comparable samples are depicted in Fig.7.Clearly,the one-pot post-impregnated 0.6 V-MS-P samples were identified with a single reduction peak of around 516 °C, attributed to the reduction of low oligomeric V-O-V groups or dispersed tetrahedral vanadium species [35,37].Meanwhile, two visual peaks at 320°C and 500°C,ascribed to the reduction of surface Mn species(Mn3O4→ MnO) of strongly-interacted and weakly-interacted with the silica in initial monometal Mn-MS [38].Besides, we observed that the reduction temperature of vanadium inxVMn-MS samples was significantly shifted to higher temperature compared with 0.6 V-MS-P samples,the progressive shift of the H2consumption peak to high temperature with the V loading suggest that vanadium species affording close contact with Mn3O4in channel trigger the strong intermetallic interplay (Fig.7(b)), which leads to the delayed reduction of vanadium species.For the 1VMn-MS sample, the peak was shifted to a higher temperature of 580 °C,which could be caused by the formation of polymeric vanadium species at a higher vanadium loading which will further retard the reduction of vanadium species [35], which is consistent with the results of N2adsorption/desorption isotherms.Besides, the reduction peak at 320 °C in the Mn-MS sample was observed to shift towards lower temperatures as the implantation of foreign vanadium.This could be attributed to the weakening of the interaction between Mn and surface silica, which further proves the neighboring contact and strong interaction between VOxand Mn3O4.

Fig.7. H2-TPR profiles of pure Mn-MS, 0.6 V-MS, and series samples of xVMn-MS.

The surface composition state and chemical environment of 0.6VMn-MS along with Mn-MS or 0.6 V-MS were further studied by the high-resolution XPS technique.Fig.8(a) suggests the targeted catalyst composed of the main elements of V, Mn, O, and Si.Fig.8(b) shows the high-resolution V2p XPS spectra of 0.6VMn-MS and Mn-MS samples.The two main peaks located at around 524.0 eV and 516.7 eV correspond to V2p1/2and V2p3/2core level, which confirms that the oligomeric vanadium(V) species[39].However, it is worth noting that the binding energy of V2p in the 0.6VMn-MS sample is slightly lower than that of the 0.6 V-MS sample, which is caused by the influence of its adjacent manganese oxide in the strong interplay state.In addition, the high-resolution Mn2p XPS spectra are presented in Fig.8(c).For Mn-MS samples,the peaks located at 654.3 and 642.2 eV represent the Mn2p1/2and Mn2p3/2transition splits,respectively,ascribed to the existence of Mn3O4in the monometal material [40].For the 0.6VMn-MS sample,Mn2p1/2and Mn2p3/2correspond to the binding energy of 654.1 eV and 641.9 eV respectively.Also,the satellite peaks attributed to Mn(II)in the oxide state were also observed at 645.2 eV.Compared with single metal Mn-MS samples,the binding energy of Mn2p1/2in bimetal samples showed significant deviation, which was related to the electron cloud density transformation of Mn atoms in the confined channel, suggesting that the surrounding vanadium species post-introduced in Mn-MS exert a certain influence on Mn3O4in the interaction state.Specifically,three specific Mn states (Mn2+, Mn3+, Mn4+) are observed in the deconvoluted Mn2p XPS spectra (Fig.8(c)).Interestingly, the ratio of Mn3+species decreased from 33.3%(atom fraction) to 27.8%(atom fraction), and the ratio of Mn2+and Mn4+species shows the corresponding variation from 40.1% (atom fraction) to 42.9%(atom fraction),26.6%(atom fraction)to 29.3%(atom fraction)after introducing VOxspecies (Table S1) on the surface of Mn3O4, originating from the strong interaction between Mn3O4and VOx.On the other hand, the spectrum of O1s core level is also shown in Fig.8(d),the main peak of 532.8 eV is assigned to the Si-O-Si in the silicon framework [41].The faint peak at 531.2 eV represents the oxygen atoms bonded to the metal atoms (—O—metal—O—, Mn or V species) [42].The peak at 533.8 eV is classified as surface adsorbed water.After introducing the VOxspecies,the peak representing the oxygen atoms bonded to metal atoms shifted to higher binding energy,which should be attributed to the oxygen bonding to V in the vanadium oxide cluster near Mn3O4species.

Fig.8. XPS spectra of 0.6VMn-MS and Mn-MS: (a) survey scan, (b) V2p, (c) Mn2p, and (d) O1s.

3.3.Evaluation and analysis of catalytic oxidative desulfuration performance

In the oxidative desulfurization process, the dibenzothiophene as a model sulfur compound was used to assess the catalytic activity of different synthesized catalysts.The catalytic results obtained using TBHP as oxidant was shown in Table 2.Obviously, the Mn-MS catalyst exhibited very poor catalytic activity in the absence of V species, in the contrast, the sulfur removal also only reaches up to 68.6%for the 0.6 V-MS-P sample despite affording the higher TOF value,indicating that vanadium acted as the more active reactive center in the oxidative desulfurization process compared to Mn active center.While the catalytic activity was obviously increased in prepared bimetalxVMn-MS by the two-step progressive insertion method.It is satisfactory that the sulfur removal gradually increases basically as the increased vanadium content,in which 0.6VMn-MS catalyst showed the highest sulfur removal(99.4%) and lowest residual S level (6 mg·L-1) among allxVMn-MS catalysts.Note that the sulfur removal efficiency decreased slightly when further increasing vanadium content in the Mn-MS,which might be caused by the aggregation of vanadium coated the surface of Mn3O4.Besides,in contrast,the sulfur removal of the synthesized comparative 0.6VMn-MS-P sample by the coimpregnation only reached 73.3%,which was much lower than that of thexVMn-MS series samples.To visually observe the sulfur removal variation of different catalysts, the time-dependent catalytic reaction courses of oxidative desulfurization of sulfide compounds are showed in Fig.9(a).As observed,compared to 0.6VMn-MS-P,allxVMn-MS series catalysts exhibited faster desulfurization rates at the initial stage of the reaction, and the sulfur removal rates tended to be slow over time and reached equilibrium near 60 min.Moreover,judging from the time course of desulfurization,the desulfurization process of 0.6VMn-MS affords an ultrafast reaction rate in the 40 min and ultrafast reaction rates constants(Fig.S2, 0.0766 min-1), whose sulfurs removal reaches above 90%.Furthermore, we also evaluated the desulfurization performance of mesoporous silicon modified with different content of monometal vanadium synthesized by post-impregnation.As shown in Fig.9(b),the reaction course result reveal that,compared to 0.6VMn-MS, the increasing amounts of vanadium in monometallicxV-MS-P samples only improve the sulfur removal degree to a certain extent.When the content of vanadium is increased to a higher proportion (1VMn-MS) compared to 0.6VMn-MS, the maximum desulfurization efficiency does not increase significantly, and at this point, the desulphurization rate only reaches 86.6%.These desulfurization results show thatx-VMn-MS samples prepared by the two-step strategy can not only improve the desulfurization efficiency but also reduce the used dosage of vanadium, which further proves the superiority of our proposed strategy.

In addition, to further verify the feasibility of the target catalysts for removing different sulfur compounds.Different types of aromatic sulfides such as 4-MDBT and 4, 6-DMDBT were selected for oxidative desulphurization reaction under the same reaction condition and the results are shown in Fig.10(a).It is obvious that the order of sulfur removal for different substrates is depicted as follows:DBT ＞4-MDBT ＞4,6-DMDBT.This result could be caused by differences in electrostatic resistance effects.Compared with DBT, the electrostatic resistance of 4-MDBT and 4, 6-DMDBT is stronger, which has a negative effect on the oxidation reaction,and consistent with the reported literature [11,43,44].Also, the nucleophilic character of the sulphur atom of the S-containing compound should be another key factor affecting the desulfurization efficiency of the catalyst.As we all know,due to the low nucleophilic character of the sulphur atom in these molecules, the electron pair of thiophenic compounds could participate in the aromatic delocalization which is less available for electron donation.As expected,the desulphurization efficiency of different substrates follows the trend:DBT ＞4-MDBT ＞4,6-DMDBT,which agrees well with the order of their nucleophilic character.

With 0.6VMn-MS as the target catalyst,the desulfurization efficiency at different reaction temperatures was investigated to identify the reaction temperature function.As shown in Fig.10(b), the desulfurization efficiency showed an increasing trend with the raised reaction temperature.Specifically, the sulfur removal was only 78.4% at the reaction temperature of 40 °C, while the sulfur removal rate can reach 99.4% in 60 min at 80 °C.Moreover, it is noteworthy that the initial reaction is significantly accelerated with the raised reaction temperature, which is attributed to the increment of effective collision frequency of reactant in higher reaction temperature,which is conducive to the oxidation reaction.At the same time,we have compared the desulphurization performance of DBT between our target samples and different reported vanadium-based catalysts under optimal conditions and are shown in Table 3.The statistical results suggest that the 0.6VMn-MS catalyst studied in this work exhibits more excellent desulfurization efficiency under the conditions of comprehensive comparison.Typically, as for the high concentration DBT contaminant(1000 mg·L-1), 0.6VMn-MS exhibits the best desulfurization efficiency (99.4%).Even if the research showed the higher DBT removal efficiency (Entry 5＞99.6%), this catalytic reaction should be conducted with a lower DBT initial concentration(500 mg·L-1).Also, some researches found the DBT contaminant could be removed under low temperature(30°C)with a great removal efficiency (98.7%), but the reaction needs a longer time of 240 min compared to that of catalyst 0.6VMn-MS.

The recycling performance and reusability of catalysts is a critical factor in practical applications.After the first experiment, the catalyst was separated by simple filtration.The recovered catalyst was washed several times with acetone to desorb the sulfur compound and dried at 80 °C for 4 h.The treated catalyst was then reused for the next reaction cycle.As shown in Fig.11, a slight reduction in sulfur removal can be observed and the catalytic removal efficiency remains above 95% after eight cycles, showing good recyclability.More importantly, as shown in Table S2, the ICP results further indicate the only trace leaching (0.06% (mass)for manganese and 0.1% (mass) for vanadium)of two metal active species after eight reaction cycles compared to the original freshcatalyst,verifying the good stability of the resulted catalyst in ODS system.

Table 2Catalytic properties in sulfur removal of DBT using comparative catalysts

Table 3The oxidative desulfurization performances of DBT for reported vanadium-based catalysts

Fig.9. The catalytic activity at different reaction courses of (a) different various samples and (b) series post-impregnated xV-MS-P.

Fig.10. Catalytic conversion of (a) using different reaction substrates at different time courses and (b) catalytic conversion of DBT at different reaction temperatures using 0.6VMn-MS.

3.4.Possible catalytic reaction mechanism

Gas chromatography-mass spectrometry(GC-MS)analysis was used to detect the product after the ODS process.In general, the catalyst was collected through simple filtration, centrifugation,and extraction by carbon tetrachloride.As shown in Fig.12(a), a strong retention time peak at 19.6 min was detected.Through mass spectrometry analysis, the peak of the charge mass ratio(m/z) = 216.0 was observed, which proves the existence of DBTO2,and other identified lowm/zpeaks are assigned to the fragment ion peaks of DBTO2.In addition,the FT-IR spectra of 0.6VMn-MS catalyst before and after the ODS reaction is also shown in Fig.12(b).Compared with the catalyst before the ODS reaction, several newly-emerged peaks at 1290, 1165, 1155, and 1047 cm-1were observed, respectively, and originated from the used 0.6VMn-MS after the reaction.Thereinto, the peaks at about 1290, 1165, and 1155 cm-1are identified and assigned to O=S=O symmetric and asymmetric stretching vibration modes in DBTO2[45,47].Besides,a weak characteristic peak was observed at 1047 cm-1,which corresponded to the characteristic peak of S=O in DBTO, indicating that trace DBTO was generated in the ODS process [48].These results suggest that DBT was oxidized to DBTO/DBTO2and simultaneously adsorbed on the surface of the catalyst.

Fig.11. Stability assessment of 0.6VMn-MS in the recycling experiment.Reaction conditions for each cycle: 10 ml of model oil, 0.04 g of catalyst, 0.1 ml of TBHP,80 °C, and 60 min.

To our best knowledge,TBHP is used as an oxidant in oxidative desulphurization, which could be decomposed to produce superoxide radicals (·O2-) and hydroxyl radicals (·OH) during the ODS reaction.To explore the real role of the generated free radicals in the ODS process, excessivep-benzoquinone (p-BQ) andisopropanol (IPA) are used as the scavengers to trap radicals of·O2-and·OH, respectively.As shown in Fig.13(a), it is obvious that almost no change in sulfur removal through the addition ofpbenzoquinone, indicating no radical quenching occurs usingpbenzoquinone.On the contrary, the experiment with isopropanol shows an obvious drop in the desulfurization efficiency, which is proportional to the amount of isopropanol.This suggests that·OH is generated and plays a decisive role in the ODS process.However, it is interesting that there is still certain desulfurization performance even with the addition of excessive amounts of isopropanol.To further verify the reaction mechanism, a series of verification experiments were designed and depicted in Fig.13(b).For V-MS-P samples, the desulfurization performance was almost unchanged regardless of the addition ofp-benzoquinone or isopropanol, while the ODS test of Mn-MS catalyst after the addition of isopropanol hardly showed catalytic activity.The above results indicate that the mechanism of the vanadium active metal participated process possibly excludes free radical participation but is contrary to the ODS process using the manganese as reaction active site.The ODS process of the target material may be the combined result of free radical reaction and cyclization reaction.Besides, we also recognized that the catalytic activity of 0.6VMn-MS was reduced by 78% owing to the addition of IPAs, hinting the quenching activity derived from Mn active sites occupied 78%of the original activity of 0.6VMn-MS, obviously higher than that(43.9%)of monometal Mn-MS.This phenomenon suggests the possible reconstruction of Mn-V active species after introducing the VOxand makes the Mn reactive center more active.Also, the V atoms surrounding the Mn sites changed the outer electron structure of Mn thereby improving the activation of TBHP to produce more oxidative active free radicals.Therefore, the bimetalbearing V-Mn catalyst receives the generated catalytic synergy of two close-contact metals, which further favors the viewpoint for the presence of synergetic sites.

Based on the above analysis and reported comprehensive researches [8,49,50], we proposed a plausible dual reaction cycles mechanism(Fig.14).As shown in Fig.14,this oxidative desulfuration reaction mainly contains two typical processes:Procedure I(free radical reactions) andProcedure II(ring-forming reactions).

Firstly, TBHP is susceptible to attack by V-O and Mn-O active species to form two different peroxometallic intermediates.As for the manganese peroxometallic intermediate inProcedure I, the O—O bond in the five-membered ring is easy to break to produce hydroxyl radical with incredible oxidation ability.As an active species, it oxidizes dibenzothiophene to the corresponding sulfoxide(DBTO)in the ODS process,and DBTO is further oxidized to sulfone(DBTO2) by ·OH radicals.On the other hand, inProcedure II, the TBHP would be adsorbed on the V site and form a bridged fivering structure, and the formed vanadium peroxometallic intermediates with high electrophilic activity attack adsorbed S atoms of DBT to form DBTO and regenerated to the VMn-MS species.After that, the reduced catalyst is coordinated with TBHP to form a five-membered ring intermediate again.Subsequently, DBTO is further oxidized to form DBTO2through the peroxometallic intermediates.DBTO2with high polarity produced in the above two processes is easy to be adsorbed on the surface of the catalyst and then remove the catalyst by centrifugation to achieve deep desulfurization.

Fig.12. (a) The GC-MS profile of products of the catalyst-involved reaction phase after the reaction, and (b) FT-IR spectra of 0.6VMn-MS catalyst before and after ODS reaction.

Fig.13. (a) Influence of different scavengers and scavenger’s dosage on DBT conversion over 0.6VMn-MS catalyst (x IPA, x refers to IPA/TBHP ratio), and (b) different quenching experiments for different samples.

Fig.14. The proposed possible reaction mechanism for the ODS reaction process.

4.Conclusions

In summary, we successfully developed an excellent oxidative desulfuration heterogeneous catalyst that V-Mn bearing mesoporous molecular sieve through a two-step insertion synthesis approach.Impressively, the post-insertion of VOxspecies into mesopore containing confined Mn3O4induced the reconstruction of dual metal oxides as the improved catalytic active centers,which was demonstrated by various convincing characterizations for the existed strong interplay of intermetallic species.In addition,the catalytic results further demonstrated the synergetic catalytic effect of bimetals in the confined mesopore for the enhanced catalytic oxidative desulfuration efficiency in only 60 min.More importantly, the reaction mechanism study concerning the free radicals quenching experiments indicates the provided dual metal sites afforded two different reaction pathways over respective two active sites involving free radicals reactions and ring-forming reactions,where the Mn sites acted as the activator triggers the generation of active free radicals,while the V center was regarded as the specific active site for the ring-forming reaction in the TBHPparticipation process, and synergetically promoted the activated ability of Mn for active free radicals.Finally, the obtained catalyst was still demonstrated for the long-life utilization in the catalytic oxidative desulfuration.This work also paves the way for the catalytic oxidative desulfuration by exploiting the synergetic catalytic effect of dual metals by right of two different reaction mechanism processes.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this work.

Acknowledgements

The authors acknowledge the National Natural Science Foundation of China (No.21908085, 21776129, and 21706121), Natural Science Foundation of Jiangsu Province (No.BK20170995 and BK20190961), General Program for University Natural Science Research of Jiangsu Province (No.16KJB530003) and the Project of Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2021.05.015.