Guixian Li,Tao Tian,Hanxu Li,Jinlian Li,Tingna Shao,Qi Zhang,Peng Dong,2,

1 School of Petrochemical Engineering, Lanzhou University of Technology, Lanzhou 730050, China

2 Key Laboratory of Low Carbon Energy and Chemical Engineering of Gansu Province, Lanzhou 730050, China

Keywords:Twinned HZSM-5 Encapsulated metal Shape-selective catalysis Anti-carbon deposition

ABSTRACT Toluene methylation with methanol to produce para-xylene has been extensively and intensively studied.However,the methanol-to-hydrocarbons(MTH)side reaction in this reaction is difficult to be inhibited,which will cause a mass of carbon deposition and cover the catalyst surface,resulting in catalyst deactivation.Here,a dual-functional Ru@HZSM-5 catalyst with high para-selectivity and low carbon deposition was prepared by encapsulating Ru metal with HZSM-5.According to catalytic performance studies,the Ru@HZSM-5 catalyst produced xylene selectivity of 98% and para-xylene selectivity of 96%.Meanwhile,we find that carbon precursors (e.g.ethylene) were very little when Ru catalyst was used,but the results of HZSM-5 catalyst were completely opposite.Ru@HZSM-5 catalyst achieves a lower carbon deposition rate of only 6% of HZSM-5.The main possible reason for this is that the initial C-C bond between methanol and the olefin is difficult to form.

1.Introduction

Para-xylene(PX)is one of the products with higher added value in aromatics,which is mainly used in the production of terephthalic acid (PTA),ethylene terephthalate (PET) and dimethyl terephthalate (DMT),and used as raw material for the production of medicine,fragrance,ink and other industries[1].In the route of toluene alkylation with methanol topara-xylene (MTPX),the capacity of toluene is excessive,while methanol may be produced easily using methods involving coal and natural gas,meanwhile,the high selectivity of PX in the product can reduce the energy consumption for subsequent separation.Therefore,MTPX is considered to be the most promising PX process production route [2].

Nevertheless,the most difficult aspect of the MTPX process was the development of an efficient and stable catalyst.All kinds of zeolites such as MOR [3],TNU-9,SSZ-33 [4],MCM-22 [5] and SAOP-41[6]had been reported to be applied in the MTPX reaction,while HZSM-5 with a large number of acidic sites and unique MFI structure exhibited excellent catalytic performance in MTPX.Meanwhile,thep-xylene selectivity could exceed 95% when HZSM-5 was modified with B[7],Mg[8],Si[9],etc.And the result of that the reduction of pore size after modification was more conducive topara-selectivity,but inhibited the diffusion ability of zeolites,which made it more difficult for the carbon deposition precursor to diffuse out of zeolites and deposit in the pore mouth or even inside of zeolites.In this case,the active site was covered and the diffusion of molecules was hindered,which eventually led to the inactivation of zeolites [10-12].Carbon deposition species in MTPX were mainly derived from olefins produced by MTH,and vinylbenzene,methyl-ethylbenzene or other C9+[2,13]produced by the methylation of toluene.This was due to the fact that it was difficult to form macromolecular olefins and polycyclic aromatic hydrocarbons under the constraints of MFI space[14].In the study for the anti-carbon deposition of zeolites,recently,Chenget al.[15] treated ZSM-5 with mixed alkalis (tetrapropylammonium hydroxide and sodium hydroxide) to prepare an ordered mesoporous structure.This study showed that the introduction of mesopores can effectively increase the diffusion capacity of zeolites and the diffusion rate of carbon precursors.Wanget al.[16]synthesized zeolites with the sandwich structure (Si@HZSM-5@Si),which was designed to cover the acidic sites on the external surface of zeolites and effectively improve thepara-selectivity.The pore of zeolites was cut offviathe internal SiO2core,thus reducing the residence time of products in the pore,which could further inhibit the deep reaction of xylene to form C9+,etc.In addition,researchers found that HZSM-5 with certain hydrogenation ability was beneficial to the alkanation of olefins produced by side reactions,which could reduce the formation of carbon precursors[17].Huet al.[18] found that Pt-modified HZSM-5 had a strong inhibitory effect on the production of ethylene.Hanet al.[17]used Pt and Ni to modify HZSM-5,and the results showed that CO and CO2were generated by the dehydrogenation of methanol over Pt-Ni/HZSM-5 catalyst while carbon oxides were hydrogenated on the active center of Pt and Ni,effectively promoting the conversion of precursors prone to carbon deposition.In short,the current measures to suppress carbon deposition include the following:First,by improving the diffusion capacity of zeolites,the carbon precursors could faster discharge the internal of zeolites,which could be achieved by increasing the pore size (but not conducive to the high selectivity ofp-xylene or shortening the length of the channel.Secondly,metal modification could be used to inhibit the formation of carbon precursors and accelerate the transformation of carbon precursors.It should be noted that we have found that Ru/HZSM-5 catalyst had such a function,but it was easy to cause the loss of Ru and lose the function of effectively inhibiting MTH side reactions.At the same time,the carrier of catalyst was commercial HZSM-5 and the PX selectivity was low [19].

In this work,the twinned zeolites-encapsulated metal structure Ru@HZSM-5 was prepared by hydrothermal crystallization.The catalytic performance of twinned HZSM-5 and Ru@HZSM-5 in the alkylation of toluene with methanol was investigated.The influence of gas phase product distribution over twinned HZSM-5 and Ru@HZSM-5 catalysts on carbon deposition was analyzed,which was helpful for the high PX selectivity and the preparation of zeolites with high carbon deposition resistance.It would broaden the idea for the subsequent research of anti-carbon deposition in the alkylation of toluene with methanol.

2.Experimental

2.1.Preparation of catalysts

Twinned ZSM-5 was prepared by the conventional hydrothermal method [20].Zeolites precursor gel was obtained by mixing and dissolving silica sol,sodium metaaluminate (NaAlO2)aqueous solution,tetrapropyl ammonium bromide (TPABr) aqueous solution,n-butylamine and water,and the composition of gel molar was 1Na2O: 1Al2O3: 300SiO2: 135TPABr: 106C4H11N: 10160H2O.After the gel was dispersed ultrasonically for 30 min,it was transferred into a hydrothermal reactor with polytetrafluoroethylene and crystallized in a homogeneous reactor at 180 °C for 48 h.The crystallized solid precipitate was centrifuged,washed and dried,and then calcined at 550 °C for 5 h under flowing air to remove the residual templating agent.HZSM-5 were ion-exchanged with 1 mole ammonium nitrate (NH4NO4) solution at a volume of 40 ml·g-1at 80°C for 3 h,repeated three times.Finally,the samples were centrifuged and dried.HZSM-5 was obtained after air roasting at 550°C.The preparation of Ru-modified HZSM-5 was to add a certain amount of ruthenium chloride(III)hydrate to Zeolites precursor gels and repeated the above steps to obtain Ru@HZSM-5.

2.2.Characterization

X-ray diffraction(XRD)spectra were obtained with a Bruker D8 Advance diffractometer (with 40 kV,100 mA Cu K*radiation) at a rate of 5 (°)·min-1from 5° to 90°.X-ray photoelectron spectroscopy (XPS) is used to analyze metal valence distributions and is obtained by the following steps: an appropriate amount of catalyst is pressed onto a sample tray,the sample is placed in the sample chamber of the Thermo Scientific K-Alpha XPS instrument(excitation source Al Kα rays,hv=1486.6 eV),and the sample is fed into the sample chamber at a pressure of less than 2.0×10-5Pa,analysis chamber with a spot size of 400 μm,an operating voltage of 12 kV and a filament current of 6 mA;the full spectral scan fluence energy was 150 eV in 1 eV steps;the spectral data were recorded.Electron micrographs were used to analyse the morphological structure of the catalysts:trace amounts of catalyst samples were glued directly onto the conductive adhesive and subsequently sprayed with gold for 45 s using an Oxford Quorum SC7620 sputter coater,followed by morphological photography and mapping using a Gemini-300 scanning electron microscope(SEM)with an acceleration voltage of 3 kV for morphological photography and 15 kV for mapping,with a SE2 secondary electron detector.TEM photographs were recorded by FEI Talos F200X scanning transmission electron microscopy to analyze the distribution of metal sites.Ammonia temperature-programmed desorption(N2-TPD)was obtained by using a 3-station fully automated specific surface area analyzer,Quantachrome Autos orb IQ3 model,USA.The samples were subjected to nitrogen adsorption and desorption tests under -196 °C liquid nitrogen conditions,and when the instrument had finished its analysis isothermal adsorption and desorption curve was obtained,the total specific surface area of the material was obtained by the BET method,and the pore size distribution was obtained by the 2D-NLDFT method.On the one hand,the Micromeritics AutoChem 2920 II instrument was used to measure the adsorption and desorption of NH3during the temperature rise and fall to obtain the acidic characteristics of the samples,and on the other hand,the Thermo fisher Nicolet iS50 was used to measure the signal value of the samples at a certain temperature using pyridine as the probe molecule to investigate the acidic distribution of the samples.Finally,the STA 449 F3 Jupiter?simultaneous thermal analyzer was used to analyze the post-use catalyst carbon build-up by heating the post-use sample in flowing air from 50°C to 800°C at a rate of 10°C·min-1and recording the sample mass at various times.

2.3.Catalyst evaluation

The reaction was carried out in a fixed-bed reactor.Typical reaction conditions were: 0.5 g ZSM-5 powder,0.4 MPa,460 °C.And the feedstock withn(toluene)/n(methanol)=6 was fed into the reactor together with twice the water of the feedstock using two pumps.Hydrogen atmosphere was used as the carrier gas to assist the mixtures into the reactor at a molar ratio of H2/(toluene+methanol)=2.The reactor was equipped with a condensing tank at the end of the reactor to condense the exhaust gas to collect the liquid phase products.The liquid phase product was fed into the 8890GC equipped with an AT-WTX capillary column to detect toluene and aromatic product distribution.At the same time,exhaust gas condensed was introduced into Shimadzu GC-2018 equipped with a SE-30 capillary column for online analysis to analyze the hydrocarbon results.Before starting the feed,the reactor was warmed up to 500 °C and reducedin situwith hydrogen for 1 h.The conversion of toluene (CT),the selectivity ofp-xylene(Spx) and the selectivity of xylene (Sx) were calculated from the following equations:

3.Results and Discussion

3.1.Characterization of catalyst

Fig.1 shows the XRD results of Ru@HZSM-5 and ZSM-5.Notably,all samples showed well diffraction peak resolution.There were two representative diffraction peaks in the range of 7°-10°and three peaks in 22.5°-25°,which were the typical of MFI topology structure resonance [21].Meanwhile,Ru diffraction peaks were not seen in Ru@HZSM-5 sample,which was attributed to the low Ru loading and well-dispersed Ru nanoparticles.Fig.2 shows that the Ru 3d core level spectra were measured from the Ru@HZSM-5 samples.The presence of Ru 3d5/2and Ru 3d3/2peaks in the spectra was a result of spin-orbital splitting.The Ru 3d3/2peak overlapped with the C 1s peak at～285 eV resulting from CHximpurities.For this reason,so Ru 3d5/2was taken as the benchmark for discussion here [22,23].Two binding peaks of Ru 3d5/2were observed near～280 eV and～281 eV.The binding peak at～280 eV was the characteristic peak of Ru,while～281 eV was the characteristic peak of Ruδ+species,which indicated that Ru was passivated by air or not completely reduced [24,25].

Fig.1.XRD patterns of HZSM-5 and Ru@HZSM-5 catalysts.

Fig.2.XPS survey spectra of Ru@HZSM-5 catalysts after in-situ H2 reduction.

Fig.3 shows the SEM images of HZSM-5 and Ru@HZSM-5 catalysts,the crystals are the regular polyhedron with twin crystal structure,and the length ofc-axis with sinusoidal channels is about 50 μm.Longer sinusoidal channels are conducive to highp-xylene selectivity,which is caused by the fact that by taking advantage of the faster diffusion rate of PX within zeolite pores,MTPX is able to control the reaction kinetically and produce a higher yield ofpara-xylenes.As seen in Fig.3(b),(d),the front of theb-axis with straight apertures is covered by the intergrowth twin crystal.Consequently,xylene molecules diffused through the sinusoidal channels,which led to the formation of paraxylene molecules [26].Furthermore,amorphous zeolites appeared on the surface of both crystals,which might be the result of separate nucleation during the crystallization process.For encapsulating different metals,organic ligands are usually needed to prevent metal loss.However,as a special case,RuCl3or Ru(NH3)6-Cl3were directly used as precursors without any additional organic ligands,which Cl-and NH3could help stabilize the metal sites during the zeolite crystallization [27,28].The TEM imaging analysis could further investigate the distribution of Ru metals on zeolites(Fig.4).As seen in Fig.4,Ru nanoparticles were uniformly distributed on the surface or inside of the zeolites.Meanwhile,Ru particle size was less than 1 nm and no obvious agglomeration.

Fig.4.TEM images (a) and mapping images (b) of Ru@HZSM-5 catalysts.

The N2adsorption/desorption isotherms and the corresponding 2D-NLDFT pore size distribution (PSD) of as-calcined catalysts are illustrated in Fig.5 and Fig.6,respectively.The specific hole structure parameters are shown in Table 1.Here,the N2adsorption/desorption isotherms of catalysts are of a classic type-I isotherm(Fig.5).At a relative pressure of 0 ＜P/P0＜0.01,the amount of adsorption increases sharply due to the enhanced adsorbentadsorbent interaction in the narrow micropores,which leads to the filling of micropores at very low relative pressures and reaches adsorption saturation after a certain amount.No significant nitrogen desorption hysteresis was observed after the relative pressure reached 0.4,indicating that the pore channels of zeolites were predominantly microporous [29].The pore size distribution would further (Fig.6) support the conclusion that zeolites were predominantly microporous and most of the pores are 0.56 nm and 0.58 nm micro-pores.The pore size close to the kinetic diameter ofpara-xylene was certainly more conducive to the high selectivity of PX.This was ascribed to the fact that the high selectivity ofparaxylene might come from its faster diffusion rate,where small poresize micropores could hinder the diffusion of other xylenes within the pore channel,highlighting the proliferation advantage ofparaxylene [30].

Table 1 Pore structure parameters of HZSM-5 and Ru@HZSM-5 catalysts

Fig.5.N2 absorption/desorption isotherms of HZSM-5 and Ru@HZSM-5 catalysts.

Fig.6.Pore size distributions of HZSM-5 and Ru@HZSM-5 catalysts.

The NH3-uptake profiles of two catalysts are shown in Fig.7.As shown in Fig.7,HZSM-5 owns three NH3desorption peaks near～100 °C,～300 °C and～500 °C,while Ru@HZSM-5 has only two obvious NH3desorption peaks near～100°C and～300°C.The desorption peaks at different temperatures represented acid centers of different intensities.Three of the desorption peaks at～100 °C,～300 °C and～500 °C could be classified as weak,medium and strong acid centers,and the peak areas correspond to their acid amounts [31].It was obvious that the strong acid sites of HZSM-5 after Ru modification were much less than those of HZSM-5,which was attributed to the high metal dispersion covering most of the strong acid sites after Ru modification,thus leading to the reduction of the strong acid amount center.To study the Lewis acid site and Br?nsted acid site changes of the catalyst,Py-IR characterization was performed (Fig.8).Three absorption peaks exist for both HZSM-5 and Ru@HZSM-5 catalysts with the Lewis (L) acid center at 1445 cm-1,the Br?nsted (B) acid center at 1544 cm-1,and the absorption peak located between them at 1489 cm-1for the combined effect of B and L acids [32].In addition,the B acid/L acid of the catalysts at 250°C were both 0.21.The analysis by NH3-TPD and Py-IR characterization showed that the introduction of Ru could reduce the strong acid sites of the zeolites but not change the acid type.

Fig.8.FT-IR spectra of pyridine absorption at 250 °C.

3.2.Catalytic performance

3.2.1.Evaluation of catalyst performance

The toluene conversion (CT),p-xylene selectivity (Spx) and xylene selectivity (Sx) as a function of reaction time in the alkylation of methanol and toluene over HZSM-5 and Ru@HZSM-5 catalysts were listed in Fig.9 and Fig.10.As could be seen from Fig.9,HZSM-5 could maintain the conversion rate of～7%toluene and the selectivity of～80 %p-xylene at the beginning of the reaction,but the selectivity of xylene was decreased from～95% to～90% after the reaction for 1000 min.This might be because the MTH reaction would produce a large number of olefin products,such as ethylene and propylene,which could react with toluene to form C9+.It affected the proportion of xylene in aromatic products.At the same time,HZSM-5 had shown a short life with toluene conversion plummeting to 2.2% at 1350 min.As could be seen from Fig.10,Ru@HZSM-5 catalyst could show a high selectivity forp-xylene and a low conversion rate for toluene of～96% and～3.7%,respectively,while the selectivity for xylene was stable at a high of about 98%.It was not difficult to see from the comparison of Fig.9 and Fig.10,Ru@HZSM-5 catalyst exhibited the low conversion of toluene and high selectivity ofp-xylene.The low conversion of toluene could be attributed to the high dispersion of Ru metal,which could cover the strong acid sites and strong acid sites were more likely to make toluene alkylation reaction activated.HZSM-5,by contrast,had lowerp-xylene selectivity,the reason was that the outside surface of HZSM-5 could expose more acid sites,making some reactants react on the outer surface acid sites.The results show that the reaction would move toward thermodynamic equilibrium to produce morem-xylene rather thanp-xylene,because the toluene alkylation reaction with methanol would lose the channel shape selection function.This would result in a relatively lowp-xylene selectivity of HZSM-5.

Fig.9.Catalytic performance of HZSM-5 catalyst in MTPX reaction.

Fig.10.Catalytic performance of Ru@HZSM-5 catalyst in MTPX reaction.

3.2.2.Carbondepositionresistanceofcatalysts

In MTPX reaction,ethylene and propylene produced by MTH reaction was the main source of carbon deposition.The product distribution of alkane/alkene over HZSM-5 and Ru@HZSM-5 catalysts in MTPX reaction were seen in Fig.11 and Fig.12.As shown in Fig.11,with the progress of the reaction,the content of ethylene on HZSM-5 catalyst increased from 24% to 50%,while ethane,propylene,propane and C4+showed an increasing trend,while the content of methane decreased gradually to～5%.However,in the alkane olefin distribution of Ru@HZSM-5 catalyst (Fig.12),the content of ethylene and propylene as precursors of carbon deposition were much lower than that of HZSM-5 without modification,and the content of methane over Ru@HZSM-5 catalyst was much higher than that of HZSM-5.The thermogravimetric (TG)curves of HZSM-5-22 (22 stands for sample after 22 h of reaction)and Ru@HZSM-5-30 (30 h) catalysts were seen Fig.13.According to Fig.13,the first significant mass loss of the catalyst occurred between 25 °C and 300 °C,which was attributed to the loss of water adsorbed by HZSM-5 or the desorption of organic matter.The second mass loss between 300 °C and 800 °C was attributed to the combustion of carbon deposition[33].In the first mass loss,the mass loss rate of HZSM-5-22 was 0.0299%per hour,while that of Ru@HZSM-5-30 was 0.0337% per hour,which might be that Ru@HZSM-5-30 could absorb more water than HZSM-5-22 after longer feeding time and result in a higher mass loss rate in Ru@HZSM-5-30 during the first stage.In the second mass loss,the mass loss rates of HZSM-5-22 and Ru@HZSM-5-30 were 0.0297% per hour and 0.0190% per hour,respectively.Ru@HZSM-5-30 owned a lower carbon deposition rate,and the carbon deposition rate was only 64%of that of HZSM-5.Therefore,Ru@HZSM-5 catalyst had a longer life than HZSM-5,and there was no obvious deactivation after 30 h.Moreover,the amount of Ru was hardly lost because Ru was encapsulated,where the content of Ru metal changed from 0.0836% to 0.0815% after 30 h by ICP test.In addition,the TG curve of HZSM-5-22 appeared obvious concave at 550 °C,which indicated that the carbon deposition was burnt toappear weightless at 550 °C,while that of Ru@HZSM-5-30 did not appear obvious concave at 550 °C,so it’s very stable.

Fig.11.Alkane alkene distributions over HZSM-5 catalyst in MTPX reaction.

Fig.12.Alkanes alkene distributions over Ru@HZSM-5 catalyst in MTPX.

Fig.13.TG curves of HZSM-5-22 and Ru@HZSM-5-30: (a) TG differential curve of HZSM-5-22;(b) TG differential curve of Ru@HZSM-5-30.

Based on the experimental data and characterization analysis,the excellent anti-carbon deposition performance of Ru@HZSM-5 might be because Ru metal could alter the formation of the first C—C bond and could not carry out the MTH reaction process.It was generally believed that the MTH reaction could be simplified into the following stages: the initial reaction in the first stage included methanol dehydration to dimethyl ether (DME),which was further decomposed at the Lewis acid site to CH4,HCHO,formate,etc.[34,35],and CO was generated by methanol dehydrogenation [36].The second stage was similar to an autocatalytic induction phase to generate more reactive intermediate.The third stage was the core of reaction,where the first C—C bond was formed,and then the first C—C bond could react with methanol or itself to generate multi-carbon compounds.There’s been a lot of debate about how the first C—C bond was formed.Liuet al.[37] proposed a C—C bond formation mechanism involving CO.At the initial stage of the reaction,the composition of C was electrophilic materials (CH3OH and DME),nucleophilic material (CO)and inert material (CH4).Then the electrophilic methyl group was formed by the dehydration of methanol at the Lewis acid site and the first C—C bond was formed with nucleophilic CO.However,the introduction of Ru might have impeded the formation of C—C bonds on this basis.Because CO was catalyzed by Ru to form CH4at high temperature in the presence of hydrogen (Fig.14),and Ru-based catalysts had high selectivity in the methanation of carbon and oxygen compounds[38,39].The methanation of CO could shift the balance of C—C formation to the left,thus reducing the number of multi-carbon compounds in the precursors of carbon deposition.The experimental and characterization results were reflected in the higher selectivity of methane and lower mass loss rate.

Fig.14.Possible methanol auto recreation pathway catalysts by Ru@HZSM-5.

4.Conclusions

In the study we presented,the encapsulated Ru@HZSM-5 catalyst was prepared by hydrothermal synthesis method,and the catalytic performance of HZSM-5 and Ru@HZSM-5 catalysts in toluene alkylation with methanol were investigated.Compared with HZSM-5,the higherpara-selectivity (～96%) and the lower conversion of toluene(～3.7%)were obtained over Ru@HZSM-5 catalyst.In addition,compared with HZSM-5,a lower proportion of olefin and a higher proportion of methane were found over Ru@HZSM-5 catalyst,resulting in carbon deposition rate of Ru@HZSM-5-30 was only 64% of HZSM-5-22.The results showed that the high selectivity ofp-xylene was derived from the catalyst matching the product shape selection pore size.The introduction of Ru would cover the strong acid site of HZSM-5,resulting in a low conversion of toluene.Metal Ru might change the formation of the first C—C bond in the auto-reaction of methanol and reduce the formation of carbon deposition precursor in the reaction process to enhance the anti-carbon deposition performance of catalysts.

Data Availability

Data will be made available on request.

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 paper.

Acknowledgements

We acknowledge financial support from the Hongliu Outstanding Young Talents Funding Program of Lanzhou University of Technology (02/062214).