Haoyu Yao,Jiangcheng Li,Jiangyan Li,Xiangfeng Liang,Gang Wang4,,Haiyan Luo

1 Environmental Resources, Green Chemical Separation Group, Qingdao Institute of Bioenergy, Bioprocess Technology, Chinese Academy of Sciences, Qingdao 266101, China

2 Beijing Insight Chemistry Co., Ltd, Beijing 101121, China

3 Merck Sharp & Dohme Research and Development (China) Co., Ltd, Beijing 100101, China

4 Institute for Catalysis, Hokkaido University, N-21, W-10, Sapporo 001-0021, Japan

5 CAS Key Laboratory of Green Process and Engineering, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China

Keywords:Polyoxymethylene dimethyl ethers/ZrO2 -TiO2 Chain propagation Kinetics Deactivation behavior

ABSTRACT Polyoxymethylene dimethyl ethers (OMEs) with physical properties similar to those of diesel has received significant attention as green additives for soot emission suppression.Herein,series of/ZrO2—TiO2 catalysts were developed for OMEs production from dimethoxymethane (DMM) and 1,3,5-trioxane through sequential formaldehyde monomer insertion into C-O bond of DMM.Not Lewis but Br?nsted acid sites were identified to be active for the decomposition of 1,3,5-trioxane into formaldehyde unit,however,both of them are effective for the chain propagation of DMM via formaldehyde unit insertion into C-O bond.Kinetic studies indicated each chain growth step exhibited the same parameters and activation barrier on corresponding Br?nsted and Lewis acid sites due to the same reaction mechanism and very similar chemical structure of OMEs.Also,the catalytic stability investigation suggested the deactivation behavior was derived from the carbon deposition,and the decay factor could be exponentially correlated with the amount of coke accumulation.

1.Introduction

Despite the high thermal efficiency,the utilization of diesel suffers from the significant emissions of NOxand soot [1-3].For the purpose of these emissions mitigation,polyoxymethylene dimethyl ethers (OMEs),especially the OME3-5,having physical properties similar to those of diesel can be used as fuel additives to suppress the formation and emission of soot,attributing to their chemical structure consists strictly of carbon and oxygen backbones without a single carbon-carbon bond [4-8].In addition,OMEs have lower vapor pressure,higher viscosity,and higher cetane number compared with other additives,such as dimethyl ether and dimethoxymethane (DMM) [9,10].Therefore,they are receiving increasing attention as ideal diesel fuel additives in recent years.

OMEs can be produced starting from the commercial chemicals such as bio-derived methanol under mild conditions and also by gas-to-liquid technology[11-20].There are multiple routes to synthesize OMEs and one of the crucial pathways starts with DMM,which is the simplest OME (namely OME1),and 1,3,5-trioxane as feedstocks [21].As described in Fig.1,this reaction process,typically catalyzed by solid acid catalyst,involves 1,3,5-trioxane decomposition and followed sequential formaldehyde insertion steps,each inserts a formaldehyde unit into the OME backbone,thus elongating the carbon-oxygen chain.DMM and 1,3,5-trioxane are found to be excellent terminal group and methoxyl group suppliers,respectively,because there is no water produced during reaction,as it will cause catalyst poison and significant side reactions.Homogeneous catalysts,such as mineral acid [21-23]and ionic liquids [24-29],show excellent performance,but they are very corrosive,environmentally malign and difficult to separate.In contrast,heterogeneous solid acid catalysts are noncorrosive,easy to separate and recover,however,the main challenging problem is to enhance their activity and stability.

Fig.1.Transformative pathways of DMM and 1,3,5-trioxane to OMEs.

Among the solid catalysts,ion-exchange resins[23,30-38],carbon materials[23,39,40],zeolites and their modifiers[41-45],solid superacid[46-50],metal chlorides[51]and other aluminosilicatebased materials [52-55] have been investigated for OMEs production.Xue and Baranowski found that Br?nsted is more effective than Lewis acid sites for this reaction on Al-SBA-15 [55],Sn-Beta[44] and Tin-Montmorillonite catalysts [56].Other researchers observed the identical rate and equilibrium constants of these CH2O insertion steps on ion-exchange resin [30,31,33,57],/TiO2[46,48]and ZrO2/Al2O3[49,54],irrespective of the ether chain length.Zhaoet al.[23] and Liuet al.[31] revealed that the distributions of OMEs followed Schulz-Flory law.Wanget al.[24],Liuet al.[31] and Goncalveset al.[58] elucidated 1,3,5-trioxane decomposition was the kinetic-relevant step of this reaction.As for the catalytic mechanism,two processes were proposed in the reported literatures for OME production,one is initial,growth,and termination pathway,and the other is sequential addition route[21,24,38,58].However,these research were based on a wide range of reactants,reaction conditions and catalysts in liquidphase system,which still lack sufficiently comprehensive understanding on the active sites,catalytic mechanism,kinetic requirements and catalyst deactivation behavior for enabling this relatively simple and repetitive C-O bond formation chemistry.

Here,we explore the nature of active sites for this transformative process,catalytic mechanism and kinetics of these sequential CH2O insertion steps,and deactivation behavior on/ZrO2-TiO2catalyst series with different Br?nsted and Lewis acid site densities.To achieve these purposes,the characterization technologies including X-ray diffraction (XRD),FT-IR,pyridine IR,chemical pyridine titration,and thermogravimetry (TG) were employed in combination with kinetic studies.

2.Experimental

2.1.Main materials

DMM (purity,99%),1,3,5-trioxane (purity,99%),titanium chloride (TiCl4,purity,99.8%),zirconyl chloride octahydrate (ZrOCl2-·8H2O,purity,98%),ammonium sulfate (purity,99.0%),pyridine(purity,99.0%),and concentrated ammonium hydroxide solution(concentration,28% in water) are supplied by Sigma-Aldrich.The commercial NKC-9 resin with Br?nsted acid site density of 4100 μmol·g-1is provided by Chemical Plant of Nankai University(Tianjin,China).Prior to use,the NKC-9 resin should be rinsed with saturate solution of sodium chloride to remove the impurities,followed by being washed with hydrochloric acid (10% (mass)) and distilled water sequentially,and dried in oven under vacuum at 60 °C.

2.2.Catalyst preparation

2.3.Catalyst characterization

The physical N2adsorption and desorption isotherms for specific surface area and pore structure measurements were conducted using AUTOSORB 6AG (Yuasa Ionics Co.,Ltd.,Japan).The X-ray diffraction (XRD) patterns were recorded using Cu-K radiation on a Miniflex benchtop X-ray diffractometer (Rigaku Corp.,Japan)under ambient conditions.

Acid site density titration with pyridine and its sequential temperature-programmed desorption (TPD) were conducted for/ZrO2-TiO2catalysts prepared with different impregnated(NH4)2SO4concentration.For the pyridine titration experiment,about 50 mg of sample was supported on a quartz frit in a cylindrical quartz reactor.The sample was first treated under flowing He(10 ml·min-1) at a constant heating rate of 5 °C·min-1to 200 °C and maintained isothermally for 2 h.Then the pyridine was introduced at the flow rate of 2.05×10-6mol·min-1by a gas tight syringe mounted on a syringe infusion pump into a vaporization zone maintained at 120°C and located upstream from the reactor.In the vaporization zone,pyridine was evaporated and mixed with the flowing He stream.The mixture of pyridine and He was then introduced to the catalyst sample and the amount of pyridine in the effluent stream became identical to that of the feed stream,at which point the difference in pyridine molar flow rate between the feed and effluent stream was less than 5%.After the titration,the sample was purged with flowing He stream (10 ml·min-1) at 200°C for 30 min and then the temperature-programmed desorption of pyridine (pyridine-TPD) was performed.During the TPD experiments,the temperature was increased to 800 °C at 5 °C·min-1in flowing He stream (10 ml·min-1).During the entire period of the pyridine-TPD experiment,the amount of pyridine in the effluent stream was quantified with a flame ionization detector (FID) without chromatographic separation.

The pyridine FT-IR was detected on a FT/IR-4200 spectrometer equipped with anin situcell.A catalyst pellet (13 mm diameter)prepared with about 25 mg of powder samples was charged into a self-supporting plate and placed in the IR cell.The catalyst pellet was first pretreated under vacuum at 300 °C for about 30 min.After the cell was cooled down at room temperature,the pyridine vapor was introduced into the cell and adsorbed for 10-15 min.Then the pyridine chemically bounded on catalyst sample was desorbed for 30 min at room temperature,driving the condition to higher vacuum.After that,the cell was scanned at 150 °C for 30 min.

The total amount of carbon deposition on spent catalyst at different temperatures was determined using thermogravimetry(TG)analyzer (NETZSCH TG 209 F1 Libra?).Typically,about 5 mg of catalyst samples were placed and heated to 900 °C at a rate of 10 °C·min-1in air flow of 30 ml·min-1,during which the mass change of sample was recorded.

2.4.OME synthesis

The OME synthesis reaction on/ZrO2-TiO2catalyst series having different Lewis and Br?nsted acid site densities were conducted on a fixed-bed reaction system,as shown in Fig.2.The catalyst layer was charged and fixed at the middle position of tubular quartz reactor(8 mm inner diameter),the temperature at the position of catalyst layer was controlled by a thermocouple monitored PID controller during reaction.Prior to the synthesis reaction,the catalyst layer was heated to the required reaction temperature in the flow of He.Then the DMM-1,3,5-trioxane solution mixture with required molar fraction,which was provided together with the experimental results in the main text,was introduced by a syringe mounted on an infusion pump into a vaporization zone maintained at 150°C and located upstream from the reactor.To change the molar ratio of DMM to 1,3,5-trioxane,the DMM was introduced by another syringe mounted on the infusion pump.In the vaporization zone,the reactants were evaporated and mixed with the flowing He stream to pass through the catalyst layer.After reaction,the outlet stream was charged into an on-line GC for quantitative analysis.It should be noted here that the reaction line in the downstream of reactor was kept warm at 250 °C.As for the assessments of transformative rates of 1,3,5-trioxane (or DMM),the conversion of 1,3,5-trioxane (or DMM) should be controlled below 10%,and then the change of partial pressure of 1,3,5-trioxane(or DMM)after passing the catalyst layer under each residence time (the contact time between reactants and catalyst layer) could be obtained.By plotting the conversion amount(calculated from partial pressure) with residence time,the slope is the transformative rate.

Fig.2.Experimental apparatus for OMFs synthesis from DMM and 1,3,5-trioxane.

2.5.Quantitative analysis method

The reaction product mixture was quantitatively analyzed on an online GC (GC-2014,Shimadzu) equipped with a methanizer(MTN-1,Shimadzu) and flame ionization detector (FID),which was connected to a DB-1 column (30 m × 320 μm × 1 μm).The temperature of both injector and detector was set as 250 °C.The temperature of column was initially kept at 40 °C for 5 min,then it was increased to 250 °C at the heating rate of 10 °C·min-1and held for 5 min.The octane was used as internal standard substrate for quantitative analysis.The relative mass factors of DMM and 1,3,5-trioxane to octane were obtained from the prepared standard mixture with different concentration,and that of formaldehyde to octane was considered to be equal to the correction factor of 1,3,5-trioxane.While those of OMEn(n≥2)to octane were estimated by the effective carbon number method [59].

3.Results and Discussion

3.1.Catalyst characterization

Fig.3.FI-IR spectra of /ZrO2-TiO2 catalyst series.

Table 1 summarizes the physical N2adsorption and desorption isotherms analysis results.The BET surface area of/ZrO2-TiO2catalyst increases first to a maximum value and then decrease with the continuous increase of impregnation concentration of(NH4)2SO4.As illustrated in the Ref.[48],the agglomeration degree of metal oxide precursor in process of hightemperature calculation would be significantly influenced by the amount of free hydroxide on the particles surface.During the impregnation process,theions would strongly interact with surface hydroxides.With the increasing concentration ofions,more surface hydroxides would be covered,leading to the decrease in free hydroxides and increase in specific surface area [65].However,the further increased concentration ofions would cause pore blockage [66],as verified by the decrease of both pore volume and average pore size.These two mentioned effects resulted in the reduction of specific surface area with the further increased impregnation concentration of (NH4)2SO4,which was higher than 0.50 mol·L-1.Similar experimental phenomenon was also observed for the solid superacid catalysts of/TiO2and/ZrO2-TiO2[48,67].

Table 1 Physical N2 adsorption and desorption isotherms analysis results

Table 2 Quantitative analysis results of acid properties of /ZrO2 -TiO2 catalyst series and NKC-9 resin

Table 2 Quantitative analysis results of acid properties of /ZrO2 -TiO2 catalyst series and NKC-9 resin

Fig.4.XRD patterns of /ZrO2-TiO2 catalyst series.

Fig.5.(a) Pyridine IR spectra and (b) pyridine-TPD profiles of /ZrO2-TiO2 catalyst series.

3.2.Identification of catalytic sites

Fig.6 shows the transformative rates of DMM and 1,3,5-trioxane during OMEs synthesis on/ZrO2-TiO2catalyst series(with both Br?nsted and Lewis acid sites),pristine ZrO2-TiO2composite oxides (with only Lewis acid sites) and NKC-9 resin (with only Br?nsted acid sites).The rates of both DMM and 1,3,5-trioxane increase first to a peak value with the concentration of impregnated (NH4)2SO4rising to 0.75 mol·L-1and then decrease with further increase in the concentration,which is consistent with the tendency of each acid site density.To reveal the nature of active sites for 1,3,5-trioxane decomposition during OMEs synthesis,the rates calculated on per different acid site are compared,as shown in Fig.6(a).It can be observed that the measured rates of 1,3,5-trioxane conversion on per Br?nsted acid site over/ZrO2-TiO2catalyst series are almost the same,which are also close to that observed on NKC-9 resin.In addition,the rate of 1,3,5-trioxane decomposition measured on pristine ZrO2-TiO2composites is much slower than that on NKC-9 and/ZrO2-TiO2catalysts.Therefore,it can be concluded that the Br?nsted other than Lewis acid sites are the main active sites for 1,3,5-trioxane decomposition.The proton affinity of 1,3,5-trioxane is not known,while the basicity correlated with proton affinity have been found to be 7.3,compared with 10 of ethylene oxide which has proton affinity of 183 kcal·mol-1(1 kcal=4.18 kJ) [70-72].So it can be considered that the 1,3,5-trioxane shows relatively strong interaction with Br?nsted acid sites,during which it can be readily protonated and activated,which is in accordance with experimental phenomenon identified on Tin-montmorillonite catalyst[56].There is another similar case reported by Baranowski that the Br?nsted acid sites on Sn-Beta zeolites are effective for 1,3,5-trioxane decomposition to formaldehyde monomers,however,the Lewis acid sites are inactive in this process [44],which can also be considered as a strong support for our results.

Fig.6.Rates of (a) 1,3,5-trioxane and DMM transformation and (b) OMEs formation during DMM-1,3,5-trioxane reactions on /ZrO2 -TiO2 catalyst series,pristine ZrO2-TiO2 composites and NKC-9 resin at the reaction conditions of 5% 1,3,5-trioxane and 10% DMM in He stream (50 ml·min-1),1 g (1 ml) catalyst,60 °C and 0.1 MPa.

With regard to the rate of DMM,it also increases first and then decreases with the continuously increasing concentration of impregnated (NH4)2SO4,as presented in Fig.6(a).In order to demonstrate the effect of acid sites on the transformation of DMM,the rates obtained on per acid site are also compared.It is found that the Br?nsted acid sites-basedrDMMis higher than corresponding Lewis acid sites-basedrDMMon each/ZrO2-TiO2catalyst and Br?nsted acid sites-basedrDMMon NKC-9 resin,which has only Br?nsted acid sites.While the total acid sites-based rates measured on/ZrO2-TiO2catalysts are very close to each other,which are much slower than that obtained on NKC-9 resin.These experimental results suggest both Br?nsted and Lewis acid sites are effective for DMM conversion,whereas,the former is more active than the later.Fig.6(b) shows the total acid sitesbased formation rates of OMEs measured on pristine ZrO2-TiO2,/ZrO2-TiO2and NKC-9 resin catalysts.The rate of each OME observed on/ZrO2-TiO2catalyst series is very similar,which are slower than that measured on NCK-9 resin.Thus,it can be considered that not only Br?nsted but also Lewis acid sites are active for chain propagation of OMEn(n≥1)viaformaldehyde insertion into C-O bond due to the same reaction mechanism,similar results are also identified on/TiO2and Sn-Beta zeolites[44,62].The rate of DMM conversion on ZrO2-TiO2is significantly slow,which is limited by the 1,3,5-trioxane decomposition on Lewis acid sites.By comparing the rates of DMM with 1,3,5-trioxane on each catalyst,it can be seen that the 1,3,5-trioxane decomposition is the kinetic-relevant step.The density of Br?nsted acid sites increases positively with that of total acid sites in/ZrO2-TiO2catalyst,as shown in Table 2,which favours the 1,3,5-trioxane decomposition into formaldehyde and consequential DMM transformation.Besides,the proton affinity of DMM(204.9 kcal·mol-1)is higher than that of 1,3,5-trioxane(lower than 183 kcal·mol-1using ethylene oxide as references based on basicity),so the DMM molecules will be competitively adsorbed on Br?nsted acid sites during the 1,3,5-trioxane decomposition on Br?nsted acid sites.Thus,the DMM conversion is still positively related to the total acid sites though the 1,3,5-trioxane decomposition is the kinetic-relevant step.

Table 3 summarizes the OMEs distribution and OME3-5selectivity over pristine ZrO2-TiO2,/ZrO2-TiO2catalyst series and NCK-9,which reveals the effect of acid properties on the product distribution and selectivity.For all the catalysts,the percentage of each OME in reaction mixture decreases as the chain length increases,indicating the gradual chain growth by formaldehyde insertion steps.As it can be seen that the selectivity of OME3-5increases with the rising ratio between Br?nsted and Lewis acid site density in/ZrO2-TiO2catalyst.According to the experimental results shown in Fig.6(a) and (b),the decomposition of 1,3,5-trioxane to formaldehyde monomers proceeds on Br?nsted instead of Lewis acid sites,while the chain growth of DMM to OMEs occurs on both Br?nsted and Lewis acid sites.In addition,the comparison of catalytic activity in DMM transformation to OMEs between NKC-9 and ZrO2-TiO2confirms the Br?nsted acid sites are much more efficient than Lewis acid sites for not only 1,3,5-trioxane decomposition but also DMM chain propagation process.Therefore,it can be considered that the higher ratio of Br?nsted to Lewis acid site density in/ZrO2—TiO2catalyst will favor the higher selectivity of OME3-5during DMM-1,3,5-trioxane reactions.

Table 3 OME distribution and OME3-5 selectivity on /ZrO2-TiO2 catalyst series,pristine ZrO2-TiO2 composites and NKC-9 resin at the reaction conditions of 5% 1,3,5-trioxane and 10% DMM in He stream (15 ml·min-1),1 g (1 ml) catalyst,60 °C and 0.1 MPa (unit: %)

Table 3 OME distribution and OME3-5 selectivity on /ZrO2-TiO2 catalyst series,pristine ZrO2-TiO2 composites and NKC-9 resin at the reaction conditions of 5% 1,3,5-trioxane and 10% DMM in He stream (15 ml·min-1),1 g (1 ml) catalyst,60 °C and 0.1 MPa (unit: %)

3.3.Reaction pathways

According to the detailed analysis of Fig.6 mentioned in Section 3.2,it can be confirmed that the 1,3,5-trioxane decomposition into formaldehyde monomers occurs on Br?nsted instead of Lewis acid sites,while both of them are effective for chain propagation of DMM though sequential formaldehyde unit insertion step.In addition,the Br?nsted acid sites are more efficient than Lewis acid sites for OMEs productionviaDMM chain elongation.According to the reported mechanism of DMM activation on Br?nsted acid sites over/TiO2,which is also a typical solid superacid shows similar acid properties and characteristics to/ZrO2-TiO2reported herein,the terminal oxygen atom will be protonated,followed by insertion of one formaldehyde unit for the chain growth.As for the formaldehyde activation mechanism,it was proposed by Bara-nowski with experimental support that the interaction between the formaldehyde molecules and Lewis acid sites resulted in an activated carbonyl bond,which was likely the initial step for insertion of formaldehyde into OME backbone as a consequence of ratelimiting 1,3,5-trioxane dissociation step on Br?nsted acid sites.By combining the reported mechanism and our experimental results,the catalytic mechanism for DMM-1,3,5-trioxane reactions on/ZrO2-TiO2is proposed,as described in Fig.7.The C-O bond activation and further ring opening of 1,3,5-trioxane molecule to a linear trioxymethylene intermediate is triggered by the protonation of oxygen atom on Br?nsted acid sites.Then,this active linear-structured substance consecutively decomposed into three units of formaldehyde on Br?nsted acid sites.This two-step decomposition mechanism of 1,3,5-trioxane on Br?nsted acid sites is also observed by other researchers through theoretical DFT calculation [24,58].Unlike 1,3,5-trioxane,both Br?nsted and Lewis acid sites are active for OMEn(n≥1) chain growth,which is expressed as path I and II,respectively.In path I,the terminal oxygen atom of OMEn-1(n≥2)is protonated on the Br?nsted acid site,followed by 2-C-O bond activation and one formaldehyde unit insertion to produce OMEn(n≥2).With regard to path II,the formaldehyde molecule is activated on the Lewis acid site,then the activated formaldehyde unit inserts into the 2-C-O bond of OMEn-1to produce OMEn(n≥2).These two pathways of OME production on Br?nsted and Lewis acid sites,namely the so-called sequential addition routes,are also proposed by the other scholars[44,48,56].It should be noted here that the hemiacetals,which are reported as intermediate products during OMEs production,are not detected in our experiments due to its high reactivity on the catalyst.Although the Langmuir-Hinshelwood-Hougen-Watson(LHHW) mechanism on catalyst with only Br?nsted acid sites is also reported by the literature [30],namely both monomeric formaldehyde and OME should be chemisorbed and activated on acid sites,the Eley-Rideal(E-R)mechanism[44]seems to be more feasible for our reaction process as a consequence of the ratecontrolling 1,3,5-trioxane decomposition step occurs on Br?nsted acid sites,DMM conversion proceeds on both Br?nsted and Lewis acid sites,and activation of DMM and formaldehyde on different kinds of acid sites.

Fig.7.Proposed mechanism for OMEs synthesis from DMM and 1,3,5-trioxane on /ZrO2-TiO2 catalyst.

3.4.Kinetic studies

With the understanding of reaction pathway,the kinetic studies on this transformative process on the optimal 0.75S-Cat were conducted.Prior to the mechanism-based kinetic modelling and regression,the internal and external diffusion limitation was eliminated by regulating the catalyst particle size and changing the feed rate(F,ml·min-1)and catalyst mass(W,g)with the same multiple,respectively.As depicted in Fig.S1,when the catalyst particle size is smaller than 0.30-0.42 mm,the influence of internal diffusion can be excluded,while the reaction will not suffer from the external diffusion when the feed rate and catalyst weight are both increased to 2 folds.According to the proposed reaction mechanism for transformation of DMM and 1,3,5-trioxane to OMEs,the detailed elemental reaction steps for kinetic modelling are proposed as Fig.8,with the assumption of surface reaction controlling theory.In addition,due to the same reaction mechanism of C-O bond insertion of OME with formaldehyde unit and the very similar structural property,the reaction rate constant of each chain propagation step is considered the same on Lewis and Br?nsted acid sites,respectively,as well as the adsorption and desorption equilibrium constant of each OME product.

Fig.8.Proposed elementary reaction steps for OME synthesis from DMM and 1,3,5-trioxane on SO42-/ZrO2-TiO2 catalyst.

Due to the highly reactivity of those active intermediate species produced during 1,3,5-trioxane decomposition process,their transient concentrations can be obtained according to the steadystate approximation theory.

whererB,rCandrDare the rates of intermediate species produced during 1,3,5-trioxane decomposition process as shown in Fig.7;k1,k2,k3,k4,k5andk6are the rate constants;?A,?B,?Cand ?Dare respect to the adsorbed concentration of 1,3,5-trioxane and intermediate species produced during 1,3,5-trioxane decomposition.

Therefore,the ?B,?Cand ?Dcan be expressed by ?Aas follows:

The whole reaction is assumed as surface reaction controlling process,therefore,the ?A,?E,?F,?G,?H,?Iand ?Jcan be described based on the adsorption and desorption equilibrium.

where ?A,?E,?F,?G,?H,?Iand ?Jare respect to the adsorbed concentration of DMM,OME2,OME3,OME4,OME5and formaldehyde;?Vis the site without adsorption;K1,K2andK3are adsorption equilibrium constants of 1,3,5-trioxane,DMM and formaldehyde;K4is the desorption equilibrium constant of OMEn(n=2-5);PTRI,PDMM,,,,andPFAare the partial pressure of 1,3,5-trioxane,DMM,OME2,OME3,OME4,OME5and FA.

Thus,Eqs.(4)-(6) can be transformed to Eqs.(14)-(16).

So the kinetic equation of each chemical component in reaction mixture can be described by Eqs.(18)-(24).

The kinetic parameters in the equation group can be obtained by regression of Eqs.(18)-(26) using the experimental data collected at 50-80 °C,without considering the temperature distribution.The regression curves of kinetic modelling compared with the experimental results are presented in Fig.9 and the obtained parameters are summarized in Table 4.The comparison between experimental results and calculated data under different DMM/1,3,5-trioxane ratios are provided in Figs.S2 and S3.It can be believed that the proposed kinetic model can well reflect the transformative process of DMM and 1,3,5-trioxane to OMEs,considering the acceptable deviation between experimental and calculated results,and the obtained parameters are credible.Together with the product distribution,the data of carbon balance of each run are also provided in the related figures,suggesting the formaldehyde polymerization and carbon deposition during one-run reaction is not very heavy.In addition,the product distribution seems to be different from the results in liquid-phase reaction system reported by Burger using batch reactor[30],it may be due to the different reaction system,which is gas-phase herein,and short residence time (the contact time between reactants and catalyst layer in fixed-bed reactor).

Table 4 Kinetic parameters obtained from model regression

Fig.9.Kinetic regression results compared with the experimental data collected at(a)50°C,(b)60°C,(c)70°C,80°C with other reaction conditions of 5%1,3,5-trioxane and 10% DMM in He stream,1 g (1 ml) 0.75S-Cat and 0.1 MPa.

Fig.10 shows the Arrhenius and van’t hoff plots,the calculated activation barrier of each reaction step and enthalpy change of each equilibrium reaction are summarized in Table 5.It can be concluded that the ring opening of 1,3,5-trioxane is the ratedetermining step as the highest activation energy among the whole steps,and the formaldehyde generation from the decomposition of linear-structured intermediates will become more kinetically favored as the decreasing chain length.In addition,the OME formation on Br?nsted acid sites (path I) showslower activation energy than that on Lewis acid sites,suggesting Br?nsted acid sites are more efficient for this sequential C-O bond insertion process.

Table 5 Kinetic parameters obtained from model regression

Fig.10.(a) Arrhenius and (b) van’t hoff plots for activation barrier and enthalpy change calculation.

3.5.Catalyst deactivation

Then the catalytic stability of 0.75S-Cat was tested,as shown in Fig.11(a).The rates of DMM transformation measured at 50-80°C gradually decrease with the extension of duration time to 60 h.The spent catalyst was collected and characterized with IR spectroscopy,which was presented in Fig.11(b).Compared with the spectrum of fresh catalyst,additional three absorption bands can be observed for the spent sample.The band appearing at 1613 cm-1is attributed to the deposited coke,while the band at 1541 cm-1is assigned to the alkylnaphthalenes or polyphenylene structures,and the one at 1486 cm-1is indicative of δs(CH2),δas(CH2),=C+H,-C+H2(deformations of primary or secondary carbocations)or CCC stretching of allylic carbocations [73].After regeneration at 500 °C in furnace under static air atmosphere,these IR bands related to the coke substances disappear (Fig.11(b)),and the spectrum is highly similar to that of fresh specimen.In addition,the rate of DMM conversion measured on the regenerated catalyst is relatively close to that on the fresh sample.Therefore,it can be referred that the deposition of carbonaceous compounds is the main reason for catalytic deactivation of 0.75S-Cat.

Fig.11.Results of (a) catalytic stability examination and (b) IR spectra of fresh,spent and regenerated 0.75S-Cat.

As illustrated in literatures,the catalytic decay factor described by the rates on fresh and deactivated catalyst(Eq.(27))can be correlated with the amount of coke accumulated on the catalyst with different functionalities,including linear (Eq.(28)),inverse (Eq.(29)) and exponential relationship (Eq.(30)),when the catalyst deactivation is caused by the deposition or fouling of carbonaceous compounds [74,75].

where α is catalytic decay factor;rdandrFare respect to the rates measured on fresh and deactivated catalyst,respectively;λ is constant;andCCis the amount of carbon deposition.

Fig.12(a)shows the time-dependent evolution of carbon deposition on 0.75S-Cat,the longer time-on-stream and higher reaction temperature will lead to heavier coke formation.While Fig.12(b)presents the evolution of catalytic decay factor over duration time,which is calculated from the data in Fig.11(a).With the extension of duration time and enhancement of reaction temperature,the catalytic decay factor decreases consequently.By correlation of catalytic decay factor with the amount of carbon deposition,an exponential relationship is observed,as shown in Fig.12(c),and the related constant λ is 1.1.Therefore,the rate of DMM transformation on deactivated catalyst deriving from carbon deposition can be described by the rate on fresh sample(rF)and coke accumulation amount in the form of exponential functionality.

Fig.12.Deactivation examination results of (a) carbon deposition (g coke/g cat.)and (b) catalytic decay factor evolution over duration time and (c) relationship between the amount of carbon deposition and catalytic decay factor.

4.Conclusions

Data Availability

No data was used for the research described in the article.

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

For this part of work,we would like to thank the fund from the National Natural Science Foundation of China (22208349),the Innovation Academy for Green Manufacture (Chinese Academy of Sciences,IAGM2020C20),Shandong Provincial Natural Science Youth Fund (ZR2022QB244),Japan Society for the Promotion of Science (P20345).

Supplementary Material

Supplementary material to this article can be found online at https://doi.org/10.1016/j.cjche.2023.03.018.