Tong Liu,Tiefeng Wang,Yunlong Huang,Chao Wang,Jinfu Wang*

Beijing Key Laboratory of Green Reaction Engineering and Technology,Department of Chemical Engineering,Tsinghua University,Beijing 100084,China

Keywords:Detailed kinetics Methylphenyldichlorosilane Chloromethylsilylene Gas phase condensation

ABSTRACT Methylphenyldichlorosilane(MPDS,CH3C6H5SiCl2)is an important silicone monomer for the synthesis of highperformance polymethylphenylsiloxane polymers.In this work,the mechanism of the synthesis of MPDS from methyldichlorosilane and chlorobenzene by gas phase condensation was studied,and a kinetic model with 35 species and 58 elementary reactions was established.Experiments were carried out in a tubular reactor under a wide range of reaction conditions.The calculated mole fractions of the reactants and products were in a good agreement with the experimental results.A mechanism of the insertion of chloromethylsilylene into the C-Cl bond of chlorobenzene was proposed,which was proved to be the main pathway of MPDS production.The established kinetic model can be used in design and optimization of the industrial reactor for MPDS synthesis.

1.Introduction

Methylphenyldichlorosilane(MPDS,CH3C6H5SiCl2)is an important silicone monomer for the synthesis of high-performance polymethylphenylsiloxane polymers,which can be used in the fields of coating,electronics,aeronautics and astronautics[1–3].Compared with polymethylsiloxane,polymethylphenylsiloxane has phenyl groups,which enhance its thermal stability and radiation resistance[4].The main methods to prepare MPDS include the gas phase condensation[5],liquid phase condensation[6],Grignard method[7],catalytic cracking[8]and disproportionation[9].Among these methods,the gas phase condensation method is very attractive because the process is short and efficient.However,this process is hard to control and the reactor design is very difficult.The reaction rate is fast and the product distribution is sensitive to the reaction temperature.At present,only a few companies are able to use this method to produce MPDS.

The overall reaction of MPDS production from methyldichlorosilane(MH,CH3SiHCl2)and chlorobenzene(PhCl,C6H5Cl)by gas phase condensation is shown in Eq.(1),which is accompanied by the generation ofmethyltrichlorosilane(M,Cl3SiCH3)and benzene(C6H6),as shown in Eq.(2).Actually,this reaction system follows a free radical reaction mechanism,and is far more complicated than Reactions(1)and(2).For the reactor design,the reaction kinetics is needed.Due to the complexity of the reaction system,the traditional macro kinetics are unable to give reliable predictions in a wide range of operating condition,therefore a detailed chemical kinetics should be developed.

The studies on the kinetics of synthesis of MPDS by gas phase condensation have not been reported in literature.Nevertheless,there are some kinetic studies on the decomposition of chlorobenzene and methyldichlorosilane.Ritter proposed a detailed kinetic model for the thermal decomposition of C6H5Cl diluted in hydrogen[10],which included 39 species and 39 reactions.Ring reported that the initial step of CH3SiHCl2decomposition was the elimination of methane to generate dichlorosilylene(SiCl2),while the Si-C bond break and HCl elimination might also contribute to the decomposition but were not dominant[11].According to Ring's viewpoint,when C6H5Cl reacts with CH3SiHCl2,the SiCl2molecule generated from CH3SiHCl2willinsert into the C-Cl bond of C6H5Cl and form C6H5SiCl3,rather than MPDS[12].This obviously conflicts with the experimental results.Therefore,the mechanism of CH3SiHCl2decomposition needs further study.

In our previous publication[12],the mechanism and kinetics of synthesis of C6H5SiCl3from C6H5Cl and SiCl3H were reported.Because the methyl group is introduced in the MPDS molecule,the detailed kinetics of C6H5SiCl3synthesis is inapplicable to describe the kinetics ofMPDS synthesis.Therefore,this work aims to establish a new detailed kinetic model for the synthesis of MPDS from C6H5Cl and CH3SiHCl2by gas phase condensation,based on existing kinetics and mechanisms of decomposition of the reactants.

2.Kinetic Model

2.1.Reaction network

The development of a new detailed kinetic model was very complicated,therefore the reaction network for synthesis of MPDS from methyldichlorosilane and chlorobenzene was developed starting from the reported kinetic models for the thermal decomposition of C6H5Cl[10],CH3SiHCl2[11]and Cl3SiCH3[13].Then the unique reactions are added to this reaction network.Finally,the kinetic parameters are modified according to the experimental results.

The kinetic model established for the synthesis of MPDS contained three sub-models:the decomposition of CH3SiHCl2,the decomposition of C6H5Cl,and the interaction between CH3SiHCl2and C6H5Cl.These sub-models are listed in Table 1.

2.1.1.Thermal decomposition of CH3SiHCl2

The sub-model of CH3SiHCl2decomposition was mainly from the kinetic model of Ring[11].Considering that there was no experimental evidence showing the existence of wall capture species RSiRwalland Ring did notgive the kinetic data ofRSiRwallgeneration,the wallcapture reactions of RSiRwallwere not included in the present work.

Four reactions from the Cl3SiCH3decomposition model by Ge[13]were adopted,which are R3,R14,R21 and R22 in Table 1.Among these reactions,R3 and R14 described the generation and consumption of CH3SiCl,which were as important as the reactions of SiCl2generation and consumption.Because CH3SiCl and SiCl2are similar in molecule structures and chemical properties,the contribution of CH3SiCl to the synthesis of MPDS cannot be ignored.The reactions of R21 and R22 were also included in the present work to describe the formation of Cl3SiCH3.

In this work,the main decomposition reactions of CH3SiHCl2are the Si-C bond break reaction(R1),CH4elimination reaction(R2),and HCl elimination reaction(R3).Because the reactant are the mixture of CH3SiHCl2and C6H5Cl,the unimolecular decomposition reactions R1,R2 and R3 were rewritten as CH3SiHCl2+M=CH3+SiHCl2+M,CH3SiHCl2+M=CH4+SiCl2+M and CH3SiHCl2+M=CH3SiCl+HCl+M,where M represented a third body gas[14],and its concentration was equal to the total molar concentrations of the components in the mixture.In addition,silicon deposition was found during the experiments.To account this phenomenon,R23 and R24 was included in the reaction kinetics by analogy with reactions reported in our previous publication[12].

2.1.2.Thermal decomposition of C6H5Cl

The sub-model of C6H5Cl decomposition was fully taken from the work of Ritter[10].Since the reaction temperature of MPDS synthesis is not high enough for the ring-opening cracking of aromatics,the ring-opening reactions in Ritter's model were not included in the present work.In addition,the third body M was also included in the C6H5Cl decomposition reactions,as shown in R25 and R26.

2.1.3.Interaction of C6H5Cl and CH3SiHCl2

The interaction reactions between the intermediates from the decomposition of C6H5Cl and CH3SiHCl2have not been reported before.However,they are the key reactions for the synthesis of MPDS,so the establishment of this sub-model is of great significance in this work.

The CH3SiClinsertion reaction and the free radicalchain termination reaction(shown as R52 and R53)were the main formation reactions of MPDS.In order to prove the importance of R52 to MPDS formation,the following experiment was conducted.

The CH3SiCl molecule can be trapped by a conjugated diene molecule,such as isoprene,to form a cyclic molecule,and this is commonly used to verify the existence of silylenes[15–17].In this work,C6H5Cl,CH3SiHCl2and isoprene were mixed to react at 813 K,0.1 MPa and C6H5Cl:CH3SiHCl2:isoprene ratio of 1:1:1.The 1,4-cycloaddition product of CH3SiCl and isoprene was obtained by GC–MS with a mole fraction of 5%,as shown in Eq.(3).Meanwhile,MPDS was not produced at the above condition,indicating that the R52 pathway was blocked and the insertion of CH3SiCl into the C-Cl bond of C6H5Cl was the main pathway of MPDS formation.Besides,the addition product of SiCl2and isoprene was not detected,indicating the absence of SiCl2.

The insertion reactions of silylenes into the C-H,O-H,O-C and C-X bonds have been reported by many researchers[18–22],where X represents halogen.By analogy with the insertion mechanism of dichlorosilylene into the C-Cl bond of C6H5Cl[12],a mechanism of the insertion of CH3SiCl into the C-Cl bond for R52 is proposed,as shown in Fig.1.Firstly,the empty p orbital of CH3SiCl accepts the lone pair electrons of the chlorine atom by electrophilic attraction to form transition state TS1.Secondly,the lone pair electrons in the σ orbital of CH3SiCl attack the carbon atom of the C-Cl bond to form transition state TS2 by nucleophilic insertion.And finally,MPDS is formed.

According to our previous research[12],the product of SiCl2into C-Cl of chlorobenzene was phenyltrichlorosilane,not MPDS.Because phenyltrichlorosilane was not detected in the product by gas chromatography,the insertion of SiCl2into C-Cl of chlorobenzene could be ignored.Besides the MPDS formation reactions,this sub-model also included the important free radical chain-transfer reactions,shown as R55 and R57.In addition,the reactions of Cl radical with CH3SiHCl2and Cl2SiCH3were also included,shown as R54 and R58.

2.2.Determination of the kinetic parameters

The parameters of the reaction kinetics were determined with the following method.Firstly,the kinetic parameters from the literature were used as the initial value.For those reactions not mentioned in the literature,the kinetic parameters of similar reactions were used.Using these parameters,the predicted product distribution was signi ficantly different from the experimental results,indicating that the parameters needed to be regressed according to the experimental data of MPDS synthesis.

To study the main reactions that have a major effect on the product distribution,the sensitivity analysis was performed.The sensitivity coefficient sijwas evaluated by the following expression:

where ciwas the concentration of species i and kjwas the reaction rate constant of reaction j.For the species i with a mole fraction larger than 1%,the kinetic parameters of the reactions having sensitivity coef ficients largerthan 0.01 were regressed based on the selected experimental data.Generally,only the pre-exponential factors A were fitted.However,the activation energy Eawere also fitted for the interaction reactions,such as R52,R55 and R57,because no kinetic parameters were reported in the literature for these reactions.Overall,12 parameters,which were labeled by*in Table 1,were redetermined by the experimental data of MPDS synthesis.

The mole fractions of CH3SiHCl2,C6H5Cl,Cl3SiCH3,C6H6and MPDS measured at three reaction conditions(15 experimental data,listed in Table 2)were used to determine these 12 parameters by least squaresregression.The regression results were very similar when choosing experimentaldata fromdifferenttemperatures and pressures.The thermochemical properties of the species needed in the calculations were listed in Table 3.The details of the fitting method have been reported in our previous work[12].For each species,the formation or consumption rates can be obtained from the rate equations.The quasi-steady state assumption and the element mass balance equations were used to simplify the calculation.The Runge–Kutta Method was used to solve the differential rate equations in MATLAB.

Table 1 Rate constants of the reactions in synthesis of methylphenyldichlorosilane①

Fig.1.Mechanism ofthe insertion ofchloromethylsilylene into C-Clbond ofchlorobenzene.

3.Experimental

3.1.Experimental apparatus

The schematic of the experimental apparatus is shown in Fig.2.The reactants,CH3SiHCl2(Aladdin Chemistry Co.,Ltd.,99.5%)and C6H5Cl(Sinopharm Chemical Reagent Co.,Ltd.,99.5%),were separately pumped into the vaporizers by two single-cylinder plunger pumps.The vaporization temperature was 503 K,at which the decomposition of reactants could be neglected.The reactants CH3SiHCl2and C6H5Cl were mixedand fed into a stainless steel tubular reactor with a length of 800 mm and inner diameter of20 mm.The reaction temperature were measured and controlled accurately by the thermocouples and digital temperature controller.The reaction products were condensed by a water-cooling condenser and a cold trap in series.Both the condensed liquid and noncondensable gas were collected for GC analysis.

Table 2 Experimental data used for the parameter fitting

Table 3 Standard gas phase thermochemical properties

Fig.2.Schematic of the experimental apparatus.1—high purity nitrogen；2,4—feed tank；3,5—single cylinder plunger pump；6—pressure reducing valve；7,8,10,11—one way valve；9,15—mass flowcontroller；12,13,14—vaporizer；16—tubularoven and tube reactor；17—water cooling condenser；18—productcollection tank；19,21—corrosion resistance valve；20—cold trap；22—buffer tank；23—electromagnetic valve；24—pressure transducer.

3.2.Analytical method

A Techcomp 7890II gas chromatograph(GC)equipped with a thermal conductivity detector(TCD)was used to analyze the condensed samples from the water-cooling condenser and the cold trap.The column temperature increased from 90 to 240°C,with a programmed heating rate of 10 °C·min-1.The injection temperature and the TCD temperature were also 240°C.The TCD bridge current was 120 mA.High purity hydrogen(99.999%)was employed as the carrier gas for the packed column.The column pressure was 0.08 MPa.The species in the products were identified by the retention times ofstandard samples.The experimental results confirmed that the loss of silicon and carbon elements in the product was less than 5%.The typical analysis results of gas chromatography are shown in Fig.3.

Fig.3.Typical analysis results of gas chromatography.

4.Results and Discussion

4.1.Discussion of the CH3SiHCl2 decomposition

There are several experimental and theoretical studies of the competing pathways of CH3SiHCl2decomposition,and some contradictory results have been reported.The low-pressure pyrolysis of CH3SiHCl2was studied by Davidson[23],who concluded that R1 was the main channel,and the rate constant of CH3SiHCl2decomposition was about 4 times larger than that of CH4formation.Ring et al.[11]reported that the Si-C bond break and the CH4elimination were competing reactions,and the rate constants for R1 and R2 were k1=1017.2exp(-95,100/RT)s-1and k2=1013.0exp(-67,400/RT)s-1,respectively.Ge[24]obtained the rate constants for R2 and R3 using the secondorder perturbation theory(MP2)and aug-cc-pVDZ basis set.However,the rate constant for R2 reported by Ring et al.[11]was 1.8×106times larger than that by Ge et al.[24]at 1000 K.

In this work,because the third body gas Mwas used in R1,R2 and R3,the pre-exponential factors of these three reactions were also regressed by the experimental data.The pre-exponential factor A2was 100 times larger than that reported by Ring et al.[11],and A3was 8.3×105times larger than that reported by Ge et al.[24],which was mainly caused by including the third body gas M in the present work.

4.2.Model validation

The predicted mole fractionsofthe reactants and products(CH3SiHCl2,C6H5Cl,Cl3SiCH3,C6H6,MPDS and C6H5C6H5)were compared with the experimental data in Figs.4–7.All of the experimental data used in the parameter regression are different from those used for model validation.Only the reaction products with a mole fraction over 1%are plotted.In general,the model gave good predictions at conditions of 773–903 K,0.1–0.8 MPa,space time of 20–120 s,and n(C6H5Cl)/n(CH3SiHCl2)of 0.5–2.5.

In orderto evaluate the errors,allofthe predicted mole fractions versus experimental data were plotted in Fig.8.The two lines representing±10%errors were also shown in Fig.8.The calculated results agreed well with the experimental data.In this work,only three groups of experimental data were used to determine the kinetic parameters.The resulted reaction kinetics give good predictions of the experimental results in Fig.4–7,showing that the detailed kinetics developed in this work have a good prediction extrapolation ability because the reasonable complex reaction network was considered.

Fig.4.Experimental(symbols)and predicted(lines)mole fractions of reactants and products at different temperatures(p=0.5 MPa,τ=60 s,n C6H5Cl/n CH3SiHCl2=1).

Fig.5.Experimental(symbols)and predicted(lines)mole fractions of reactants and products at different pressures(T=873 K,τ=60 s,n C6H5Cl/n CH3SiHCl2=1).

Fig.6.Experimental(symbols)and predicted(lines)mole fractions of reactants and products at different space times(T=873 K,p=0.5 MPa,n C6H5Cl/n CH3SiHCl2=1).

4.3.Discussion of experimental results

Fig.7.Experimental(symbols)and predicted(lines)mole fractions of reactants and products at different n C6H5Cl/n CH3SiHCl2 ratios(T=873 K,p=0.5 MPa,τ=60 s).

Fig.8.The model prediction accuracy analysis.

Fig.4 compared the experimental and predicted mole fractions of the reactants and products at different temperatures.The decomposition of CH3SiHCl2was sensitive to the reaction temperature.As the reaction temperature increased from 760 K to 840 K,the mole fraction of CH3SiHCl2decreased from 50%to 0%,while the decomposition of C6H5Cl was less sensitive than that of CH3SiHCl2.Because the mechanism of soot formation in this reaction system has not been well understood,the kinetics of soot formation was not directly considered in the present reaction kinetics.However,the biphenyl(C6H5C6H5)formed in the reaction was considered to be the precursorofsoot[25,26],therefore the mole fraction of biphenyl was used to estimate the intensity of soot formation.When the temperature was higher than 810 K,the biphenyl began to form,which indicated that the sensitive temperature of biphenyl formation was 810 K.

Fig.5 shows the effect of pressure on the reactions.As the pressure increased from 0.1 to 0.8 MPa,the mole fraction of MPDS increased from 11%to 17%.The yield of MPDS in an industrial reactor can be increased by increasing the operating pressure.At 873 K and 0.5 MPa,the reactions reached equilibrium at a space time of 60 s,as shown in Fig.6.Therefore,the space time in an industrial reactor should be less than 60 s to avoid overmuch soot formation.As the nC6H5Cl/nCH3SiHCl2ratio increased from 0.5 to 2.5,the mole fraction of MPDS decreased from16.1%to 10.2%,as shown in Fig.7.However,the yield and selectivity of MPDS based on CH3SiHCl2increased with increasing nC6H5Cl/nCH3SiHCl2ratio.Considering that CH3SiHCl2is more expensive than C6H5Cl in the industrial process,a higher nC6H5Cl/nCH3SiHCl2ratio should be used to enhance the utilization of CH3SiHCl2.

4.4.Path flux analysis

The path flux analysis[27]was used to analyze the formation and consumption fluxes of each species and identify the important reaction pathways for these species.For each reaction containing species A,the net reaction rate was calculated as a function of the space time.Then the relative contribution of one reaction to the production or consumption of the selected species was calculated by time integration of the corresponding reaction rate.

Based on the reaction flux analysis at typical reaction conditions of 873 K,0.5 MPa,60 s and n(C6H5Cl)/n(CH3SiHCl2)=1,the production and consumption reactions for the selected species are listed in Table 4,respectively.The results show that 99.8%of MPDS is formed by the insertion of CH3SiCl into chlorobenzene(R52),and the contribution of the chain termination reaction offree radical(R53)is only 0.2%.Forconsumption,the percentages of CH3SiHCl2decomposing into CH3SiCl(R3)and CH3SiCl inserting into C6H5Cl to form MPDS(R52)are 28.5%and 63.3%,respectively.Therefore,the main pathways of the synthesis of MPDS are CH3SiHCl2→CH3SiCl+HCl and C6H5Cl+CH3SiCl→MPDS.

5.Conclusions

A detailed kinetic modelconsisting of 35 species and 58 reactions was developed for the synthesis ofMPDS frommethyldichlorosilane and chlorobenzene by gas phase condensation.The mole fractions ofthe reactants and products were measured in a tubular reactor at temperature of 793–953 K,pressure of 0.1–0.7 MPa,space time of 30–120 s and n(C6H5Cl)/n(CH3SiHCl2)of 0.5–2.5.The kinetic parameters were regressed from some of the experimental data,and the resulted reaction kinetic model gave good predictions of the other experimental data.A mechanism of the insertion of chloromethylsilylene into the C-Cl bond of chlorobenzene to form MPDS was proposed,which was validated by trapping the CH3SiCl with isoprene.The established kinetic model in this work can be used in design and optimization of an industrial reactor for MPDS synthesis.

Table 4 Consumption or production pathways for selected species at 873 K,0.5 MPa,60 s and n C6H5Cl/n CH3SiHCl2=1

Nomenclature

A pre-exponential factor,s-1or L·mol-1·s-1

Cpconstant pressure heat capacity,J·mol-1·K-1

ciconcentration of species,mol·L-1

Eaactivation energy,J·mol-1

ΔHf,298standard enthalpy of formation at 298 K,kJ·mol-1

kjreaction rate constant,s-1or L·mol-1·s-1

m temperature correction term of Arrhenius equation

p reaction pressure,MPa

S298standard entropy at 298 K,J·mol-1·K-1

sijsensitivity coefficient

T reaction temperature,K