ContentslistsavailableatScienceDirect
Combustion and Flame
journalhomepage:www.elsevier.com/locate/combustflame
Detailed experimental and kinetic modeling study of toluene/C 2 pyrolysis in a single-pulse shock tube
Wenyu Sun
a,∗, Alaa Hamadi
a, Said Abid
a,b, Nabiha Chaumeix
a, Andrea Comandini
a,∗aCNRS-INSIS, I.C.A.R.E., 1C, Avenue de la recherche scientifique, 45071 Orléans cedex 2, France
bUniversité d’Orléans, 6 Avenue du Parc Floral, 45100 Orléans, France
a r t i c l e i n f o
Article history:
Received 17 September 2020 Revised 27 November 2020 Accepted 27 November 2020
Keywords:
Toluene Acetylene Ethylene
Single-pulse shock tube
Polycyclic aromatic hydrocarbons (PAHs)
a b s t r a c t
Acombinedexperimentalandkineticmodelingstudyiscarriedouttoexploretheinfluencesofacetylene andethyleneadditiononthespeciesformationfromtoluenepyrolysis.Experimentsareconductedsepa- ratelywithfourdifferentargon-dilutedbinarytoluene/C2mixturesinasinglepulseshocktubeatanom- inalpressureof20baroveratemperaturerangeof1150−1650K.Alltheexperimentalmixturescontain about100ppmtolueneanddifferentamountsofC2fuels(50,216,459ppmacetyleneand516ppmethy- lene).Speciesconcentrationsasafunctionofthetemperatureareprobedfromthepost-shockgasmix- turesandanalyzedthroughthegaschromatography/gaschromatography-massspectrometrytechniques.
Akineticmodelis developed,whichsuccessfullypredictsthe absolutespeciesconcentrationmeasure- mentsaswellasthechangesbroughtbythevariedfuelcompositions.Withtheneatfueldecomposition profilesasareference,bothfuelcomponentsexhibitincreasedreactivityinthepyrolysisofallstudied binarymixtures,indicatingtheexistenceofobvioussynergisticeffects.Inparticular,sucheffectsbecome moreremarkable whenincreasingthe initialacetyleneconcentration. Thisessentiallyresults fromthe addition-eliminationreactionofatoluenefuelradicalandaC2 fuelmoleculeleadingtoaC9 molecule and ahydrogenatom.Indeneisidentifiedas thepredominantC9 productinallstudiedcases, andits peakconcentrationssharplyincreasewiththe initialacetylenecontentsintoluene-acetylenepyrolysis.
Ontheotherhand,indane,producedfromtheaddition-eliminationreactionbetweenbenzylandethy- lene,isonlydetectedattracelevelsintoluene/ethylenepyrolysis.Thisindicatesarelativelyweakerinter- actionbetweenbenzylandethylene,comparedtothatofbenzylandacetylene.Apartfromtheincreased concentrationsofhydrogenatomsandC9 aromatics,interactionsbetweentolueneandC2fuelsalsodi- rectlyresultinareducedlevelofC7radicalsinthereactionsystem.Overall,PAHspeciescanbedivided intotwogroupsaccordingtotheway theirpeakconcentrationsvaryingwiththeinitialfuelcomposi- tions.Forthoselargelydependingonbenzylreactions,suchasbibenzyl,biphenylmethane,fluoreneand phenanthrene,thepeakconcentrationsdecreasewiththeaddedC2 fuels.Incontrast,increasingtrends areobserved inthe peakconcentrationsofthe PAHs whichrelyonindenyl as aprecursor,including naphthalene,methyl-indene,benzofulveneandacenaphthalene.
© 2020TheAuthor(s).PublishedbyElsevierInc.onbehalfofTheCombustionInstitute.
ThisisanopenaccessarticleundertheCCBY-NC-NDlicense (http://creativecommons.org/licenses/by-nc-nd/4.0/)
1. Introduction
Being found as the most predominant aromatic in gaso- line, toluene serves as an essential component of the ternary (n-heptane/iso-octane/toluene) surrogate fuel mixtures that are widely formulated andutilizedinthelast decade[1].Givensuch importance, a wealth of studies were carried out to establish
∗Corresponding authors.
E-mail addresses: wenyu.sun@cnrs-orleans.fr (W. Sun), andrea.comandini@cnrs- orleans.fr (A. Comandini).
thekinetic mechanismresponsiblefortolueneconsumptionunder combustionrelevantconditions.Availableexperimentaldatacover globalparameters suchasignitiondelaytime[2,3],laminarburn- ingvelocity[4]anddetailedspeciescompositionsinjet-stirredre- actoroxidation[5-7],premixedflames[8,9]andflow-reactor[7]or shock tube [10-12] pyrolysis. The decomposition of toluene pro- ducesvariousresonantly stabilized radicalssuch asbenzyl, fulve- nallenyl,cyclopentadienyl and propargyl. Reactionsof theseradi- calsamongthemselves[13-16]orwithotherabundantcombustion radicals/molecules [17-19] are identified as important pathways leadingto polycyclicaromatichydrocarbon(PAH)compounds.Ac- curate kinetic characterization of relevant channels need to be
https://doi.org/10.1016/j.combustflame.2020.11.044
0010-2180/© 2020 The Author(s). Published by Elsevier Inc. on behalf of The Combustion Institute. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ )
included in developing predictive combustion kinetic models of PAHformation,andsuchmodelsarefundamentalinnumericalde- signstowardsnewcleancombustionstrategies.
SimpleC2 hydrocarbons,acetylene(C2H2)andethylene (C2H4) are usually abundant intermediates in practical combustion sys- tems. The decomposition of C2 species and the build-up of larger hydrocarbonsunderpyrolyticconditionsis aclassical topic and has been long-studied in the combustion kinetics commu- nity [20-31]. The interactions between C2 and aromatic radicals, forinstance,thewell-known Hydrogen-Abstraction-C2H2-Addition (HACA) routes [32,33], play a vital role of PAH formation and growth.More specifically,theinteraction betweenC2H2 andben- zyl isdeemedasanimportantsource ofindene,andtherateco- efficientsofrelevantreactionshavebeenderivedthroughtheoreti- calcalculations[18,34].Toluenepyrolysis,especiallywiththepres- enceofC2H2,canprovideidealenvironmentstoinspectthesere- action channels. The pyrolysis of toluene-C2H2 was studied in a few previous works through experimental and kinetic modeling approaches. Shukla et al. [35] compared the PAH products from neat toluene and toluene/C2H2 pyrolysis in a low-pressure flow reactor by using in-situ directsampling massspectrometry tech- nique. Theysuggested relevantPAH formation schemes basedon the products massspectra,though no quantitative measurements were obtained[35].Later,a kinetic modelwasproposed by Mat- sugi and Miyoshi [36] to interpret two- and three-ring PAH for- mation fromthe pyrolysisof tolueneandtoluene-C2H2 mixtures, and the model could satisfactorily predict the qualitative experi- mental observations in [35]. Very recently, Li et al. [37] investi- gated the co-pyrolysis of toluene and C2H2 in a flow reactor at atmosphericpressure over1057–1340 K.Molefractionsofcrucial speciesconcentrationswere measuredviaa photoionizationmass spectrometer.Throughcomparisonswiththecasesofneattoluene andneatC2H2 pyrolysis,Lietal.[37]pointedoutthat intoluene and C2H2 co-pyrolysis, thedecomposition ofboth fuels isgreatly stimulated.Suchphenomenawereexplainedwithacirculatingre- action sequence convertingtoluene+C2H2 to indeneandH2 [37]. TheworkbyLietal.[37]focusedonthesynergisticeffectsoffuel decomposition,soonlylimitedPAHspeciationmeasurementswere reported.However,theinteractionsbetweentolueneandC2H2can directly or indirectly alter the PAH formation scheme. This sub- ject merits further investigations, in particular, at elevated pres- sures thatare morerelevant topractical conditions.Furthermore, byvaryingtheinitialfuelcomposition,i.e.theproportionbetween toluene andC2H2,thereliability ofkinetic parameters forimpor- tant channelscan beexaminedunderreactionenvironments con- tainingdifferentchemicalcompositions.ThoughC2H4isalsoacru- cial combustion intermediate, how it reacts with aromatic radi- calsreceivedrelativelylessattention.Ourrecentwork[38]showed that C2H4 andC2H2 havedifferentinfluences ontheformationof PAHs, specifically, naphthalene and acenaphthalene, frompheny- lacetylene pyrolysis. Definitely, valuable insights can be acquired byexploringtheinteractionsbetweenC2H4 andradicalsproduced from toluene pyrolysis andby comparing the differenteffects of C2H2andC2H4.
Tothisend,twomajorgoalsareconcernedinthecurrentstudy:
i)comparetheeffectsofaddingC2H2andC2H4totoluenepyroly- sis,regardingthefueldecompositionreactivityandthedirectlyor indirectly affected speciation behaviors; ii)track andaccount for the variation trends of crucial PAH concentrations when adding different amounts of C2H2 to toluene pyrolysis. Shock tube py- rolysis experiments are carried out with different toluene/C2H2 (C2H4) mixtures having the same amount of toluene but vary- ing C2 contents.Speciesconcentrationsthatevolvewiththepost- shock temperature are obtained via the gas chromatography/gas chromatography–mass spectrometry(GC/GC–MS)technique. A ki- netic model isdeveloped by incorporatingthe reactions between
Table 1
Compositions of the gas mixtures used for experiments.
Composition Gas mixture
Toluene (C 7H 8) Ethylene (C 2H 4) Acetylene (C 2H 2)
TE 106 ppm 516 ppm –
TA_50 107 ppm – 50 ppm
TA_216 106 ppm – 216 ppm
TA_459 105 ppm – 459 ppm
Fig. 1. A measured endwall pressure history and the definition of the reaction time.
C2Hxandtoluenerelatedspecies.Throughacombinationofexper- imentalmeasurementsandmodelinganalyses,theadditioneffects ofC2 fuelsontolueneconsumption,aswellasthefurtherimpacts onPAHspeciationwillbeillustrated.
2. Shocktubepyrolysisexperiments
The experimental work is carried out with the single-pulse shocktubefacilityatICARE,Orléans,France.Detaileddescriptions abouttheshocktubeapparatusandthecoupledgaschromatogra- phy/gaschromatography–massspectrometry(GC/GC–MS)diagnos- tictechniquecanbefoundinourpreviouspublications[12,38,39]. Briefly,thedriven(78mmininnerdiameterand6.0minlength) anddriversectionsareseparatedbyadoublediaphragm.Tooper- atetheshocktubeina single-pulsefashion,adump tank,witha volumefivetimesbiggerthanthedrivensection,isplacednearthe diaphragmonthedrivensectionside.Thedrivensectionisheated upat90°Ctopreventthecondensationorabsorptionoffuelsand reactionproducts ontheshocktubeinnersurface. Fourargondi- lutedbinaryfuelmixturesareusedinthecurrentexperiments,in- cludingonetoluene-ethylenemixtureandthreetoluene-acetylene mixtureswithdifferentacetylenecontents.The detailedcomposi- tionsoftheexperimentalmixturesarelistedinTable1.
Four pressure sensors (CHIMIEMETAL A25L05B) are mounted along the sidewall ofthe driven section at intervalsof 150 mm, withthelastone being82mmawayfromtheendwall.The time whentheshockwavepassesindividualsensorsisrecordedtode- rivetheincidentshockvelocity.Thepost-shockconditions(p5and T5) issubsequentlydetermined bysolving conservationequations together with the ideal gas lawand variable heat capacityratio.
The uncertainty of the calculated T5 is estimated to be ±30 K, mainly duetothe uncertaintyin thedetermination ofthe actual positions of the pressure sensor sensitive surfaces. Such spatial uncertaintiesare estimatedto be 2mm, equal to the sensor di- ameter.APCB Piezotronicspressure sensorshieldedbya layerof room-temperaturevulcanizing(RTV)siliconeisplacedontheend- wall. Oneof therecorded endwall pressurehistories is shownin Fig.1,andthecorrespondingreactiontimeisdefinedastheinter- valbetweenthearrivaloftheshockwaveandthetimewhenthe
pressuredropsto 80%p5.Thereactiontime isatanominalvalue of4.0mswiththecurrentexperimentalconfiguration[12].Anair- actuatedvalve,suppliedbytheHighPressureEquipmentCompany (HiP),ismountedinthecenteroftheendwalltosamplethepost- shock reaction mixtures. The operation of the sampling valve is triggered by the endwall pressure signal with a delayof 4.0 ms (see Fig.1), equaltothe nominalreactiontime.The openingand closingofthevalvetakeshundredsofmilliseconds,withdrawinga sample ofa relatively bigquantity (equivalent to about20ml at 2bar),toensuresufficientsensitivityfortracePAHmeasurements.
Duetothebigsamplevolume,theaverageshockvelocity,instead of the extrapolated value at the endwall, is used for the subse- quentdeterminationofT5 andp5.Inthisway,therealconditions encountered by thesamples can be better represented.For most experiments,theshockvelocityattenuation iswithin2.5%,sothe difference betweenthe T5scalculated basedon theaveraged and the extrapolated(local at the endwall) shockvelocities is below 20K,withintheuncertaintyspecifiedabove.
Thesampledgasistransferredtotheanalyticalsystemthrough SilcoTek tubes,whichareheatedat250°Ctoavoidcondensation.
The first GC (Agilent 7890) is equipped with a flame ionization detector(FID) connectedtoa DB-17-mscolumnforthemeasure- ments of heavy species.A thermalconductivitydetector (TCD)is coupled to a Molsieve 5A column to monitorthe absence of air.
An external valve boxwhich can regulate the temperatureup to 320 °C is used for this GC to minimize the loss of heavy com- poundsduetocondensation.AnFID coupledtoanHPPlotQcol- umnis installedin thesecond GC(Thermo TraceGCUltra),used for themeasurements of smallhydrocarbons, anda DSQTM (dual stagequadrupole)massspectrometerisalsoconnectedtohelpthe speciesidentification.Theconnectionsintheanalyticalsystemand theoperationparametershavebeenoptimizedinthiswork,bring- ing benefitsfromdifferentaspects, includingincreasedsensitivity oftracePAHspeciesandbetterseparationofPAHspeciesthathave similar retention time. Forinstance, theC14H10 isomers, phenan- threne and anthracene, whichappear at11.41 and11.50 min,re- spectively,canbebetterseparatedandquantifiednow.PAHspecies uptofourrings,withthepeakconcentrationabove10−2ppmlevel canbedetectedinthecurrentexperiments.
The identification ofPAHspecies mainlydepends ontheir re- tention time known through injections of standards. For species without available standards, the mass spectrometer provides molecular weights information, from which it is possible to in- fer the element composition, and it also suggests possible iso- mericcandidates.TheadditionofC2fuelsintroduces newspecies, such asa seriesof C9 species that were negligible in ourprevi- ous work on neat toluene pyrolysis [12]. Figure 2 shows the GC signals for the C9 species. Indene and indane are identified ac- cording to their retention time. For other small peaks, their for- mulas are known from mass spectrometry, but their structures cannot be unambiguously identified, because some possible iso- mershavesimilarfragmentationpatternsandstandardmassspec- tra for some candidates are not available in the library. As pre- sented in Fig. 2, these C9H8 species intoluene/acetylene pyroly- sisareassignedasethynyl-tolueneandpropynyl-benzeneisomers, andtheC9H10 compoundsintoluene/ethylenepyrolysisare taken as vinyl-toluene and propenyl-benzene isomers. In the later dis- cussion, summed concentrations of C9H8 (excluding indene) and C9H10(excludingindane)are providedtocomparewithmodeling results.
The quantification of species concentrations relies on calibra- tions for the FID response, and the procedures of the calibra- tion experiments are detailedin [12,40]. Smallhydrocarbons, ex- cept diacetylene(C4H2)andtriacetylene (C6H2) arecalibratedus- ing standard mixtures withknown compositions. The calibration of C4H2 and C6H2 is achieved in acetylene (C2H2) hightemper-
Fig. 2. GC signals for C 9species detected in the toluene/C 2pyrolysis experiments.
The molecules in gray are possible candidates according to the mass spectra.
ature pyrolysisexperiments based on thecarbon atom conserva- tion.MajorPAHspeciesuptothreerings,includingindene,indane, naphthalene,acenaphthalene,bibenzyl,biphenylmethane,fluorene, phenanthreneandanthracene,arecalibratedingasphasethrough thefollowingprocedures:i)knownamountsofspecific PAHcom- poundsaredissolvedinmethanoltopreparethecalibrationsolu- tions;ii)asmallquantity (1–5μl)ofthesolutionisinjected into aglass vessel(witha volumeof500 ml)usingaGCsyringe; the solutionvaporizesimmediatelyintheglassvessel,whichisprevi- ously vacuumed and heatedup at 150 °C; iii) the vessel is then filled with argon to a pressure of around 900 Torr, and the re- sulting gas mixture containing PAH standards stands for around 10–15 min to guarantee good homogeneity; iv) the gas mixture is injected into the GC system at least three times to check for consistency and the peak areas for the corresponding PAH stan- dard are normalized by the injection pressure; v) the procedure is repeated at several standards mole fractions which cover the neededrange for experiments,and thecorresponding calibration factorsarederived.Four-ringPAHs(pyreneandfluoranthene)can- notbesteadilyvaporizedwiththeaboveapproach.Therefore,their calibration factors are determined through an extrapolation from thoseofthetwo- andthree-ring PAHs,namely,naphthaleneand phenanthrene.Theuncertaintyintheconcentrationmeasurements is expectedto be within 5% fordirectly calibrated smallspecies, and10%−15% for PAHspecies calibrated in gas phase, while the errorinmeasuredconcentrationsoffour-ringPAHspeciesmayin- crease to 50%. Measurementspresented in our previous publica- tions [12,38,39]haveproven thereliabilityandthe goodrepeata- bilityoftheexperimentalset-up.
For the chemicals used in the experiments, toluene (with a purity of 99.8%) and PAH standards are purchased from Sigma- Aldrich, and the gases, including acetylene (>99.5%), ethylene (>99.5%),argon(>99.9999%)andhelium(>99.995%),aresupplied byAirLiquide.A450BMathesongaspurifier witha454cartridge is connected to the acetylene bottle to remove possible acetone traces,andacetone is foundbelow the detectionlimit of theGC systeminallexperiments.Theexperimentalmixturesareprepared in a 136 L electropolished stainless steel cylinder which is pre- viously evacuated to pressures below 10−5 mbar with a turbo- molecular pump. To preparean experimental mixturewith a re- quiredcomposition, the fuel componentsare introduced into the
cylinderandthepartial pressuresare measuredwitha 0–10Torr MKS Baratron pressure transducer (model 122BA). Argon is then added to a pressures of around 10 bar, monitored by a 0–
10,000TorrMKS Baratronpressure transducer (model627D). Be- foretheexperiments,thegasmixturerests overnighttohomoge- nizeandtheactualcompositionisanalyzedwiththeGC,aslisted in Table 1. The driven section is vacuumed with a turbomolecu- lar pump till the pressure is below 10−5 torr before beingfilled withthe experimental mixture.The carbonbalanceasa function ofT5 ineachexperimental setispresentedinFig.S1.The carbon balance keepsabove 95% asT5 is lessthan 1400 K,whiledeteri- orating to70−80% at highertemperatures. Thisis mainly dueto the formation of heavy PAH species beyond our detection capa- bilities andsoot particles. Besides, the carbon recovery improves when reducingthetoluenecontents intheexperimental mixture, suggestingthe importanteffects ofaromaticspecieson thesoot- ingtendency.Theinnersurfaceoftheshocktubeiscleanedevery day toremove thecarbon deposit.Eachofthe fourexperimental sets includesaround20experiments,forwhichthenominalp5is 20 bar andT5 ranges from 1100 K to 1650 K. The experimental results are provided inthe Supplementary Material, includingthe post-shockconditionsT5andp5,thespeciesconcentrationsaswell astheendwallpressurehistoriesupto10msofindividualexperi- ments.
3. Kineticmodeling
Continuous efforts are made, based on our recent works to- wards constructing a predictive kinetic model for PAHformation fromthepyrolysisofaromaticfuels.Previously,reactionsequences describing PAHspeciation fromtoluenedecompositionwasdevel- oped andvalidatedagainst thedetailedmeasurements oftoluene pyrolysis [12].Later, benzylrelatedpathways were further tested withtheexperimentsofC8−C10alkylbenzenepyrolysis[39],where the PAH formation is also controlled by benzyl chemistry. The C2 sub-mechanismwasupdated inthediscussion ontheinterac- tionsbetweenphenylacetyleneandacetylene/ethylene[38].Inthis work,particularattentionispaidtothereactions betweenC2 and toluene fuel radicals. Reactions updated in the current work are listed inTableS1intheSupplementaryMaterialtogether withthe adoptedratecoefficients.
The reaction system of benzyl (C7H7)+acetylene (C2H2) was studiedinprevioustheoreticalworks[17,18],sinceitisapotential indene source under combustion-relevant conditions. Vereecken and Peeters [18] performed RRKM/ME calculations based on a density functional potential energy surface (PES). The formation of indene (C9H8)+H was identified as the predominant chan- nel and a ratecoefficient at its high-pressurelimit was reported [18]. Later, Mebel et al. [17] revisited the PES of C7H7 + C2H2 reactions at G3(MP2, CC) level of theory and derived pressure- dependent rate coefficients of involved reactions over the tem- perature rangeof 500−2500 K. Only the formation of two pairs of bimolecular products, indene (C9H8)+H and 3-phenylpropyne (C6H5CH2C≡CH)+H, has kinetic significance under the currently concerned high-pressure conditions. Besides, C6H5CH2C≡CH can also convert to C9H8 through H-assisted isomerization. The re- actions between phenyl (C6H5) and C3H4 (allene and propyne) also result in the formation of C6H5CH2C≡CH, and these reac- tionsweretheoreticallyexploredinthesameworkbyMebeletal.
[17]. All mentioned relevant reactions with their theoretically- determined rate coefficients[17] are incorporated in the current model.Reactionsbetweenbenzyl(C7H7)andethylene(C2H4)have notbeenaddressedthroughtheoreticalapproachesinliteratureto the best of our knowledge. Herein, we consider that they pro- ceed in a similar manner with C7H7 + C2H2 reactions, lead- ing to the formation of the cyclic species indane (C9H10) and
thesingle-chain aromatic3-phenyl-1-propene(C6H5CH2CH=CH2).
The rate coefficient of C7H7+C2H2 = C9H8+H from [18] is as- signedtotheC9H10 formationchannel,andtheratecoefficientof C7H7+C2H4=C6H5CH2CH=CH2+Hisdeterminedthroughananal- ogy to the reaction CH3+C2H4 = C3H6+H. In particular, the re- action rateconstant ofthe C7H7+C2H2 reaction in[18](used for theC7H7+C2H4reactioninthecurrentmodel)isabout9–10times lowerinthetemperaturerange1200–1550Kcomparedtotheone calculated by Mebel et al.[17] at10 atm. The choice of a lower reaction rate constant for the C7H7+C2H4 reaction is backed up by experimental evidencesasdiscussed later inSection 4.2.Nev- ertheless, future theoretical works are highly demanded to map outthereactionsontheC7H7 +C2H4 surfaceandtoderiveaccu- ratekinetic parameters.Regarding theconsumption ofC9H10, be- sidestheH2 eliminationformingindene(C9H8),ring-openingpro- cessesleadingtoo-vinyl-toluene(CH3C6H4CH=CH2)and3-phenyl- 1-propene(C6H5CH2CH=CH2)arealsotakenintoaccount.Rateco- efficients for these ring-opening reactions are evaluated through analogies to the isomerization of cyclopentene to 1,3-pentadiene (CH3CH=CHCH=CH2) and 1,4-pentadiene (CH2=CHCH2CH=CH2) [41],respectively.Methyl-phenyl(CH3C6H4) isconsideredasami- norfuel radical oftoluene, incomparisonto its resonantly stabi- lizedcounterpartC7H7.ThedetectionofC9specieswiththestruc- ture of CH3C6H4C2Hx in the current experiments, however, sug- geststheparticipationofCH3C6H4 inrelevantformationreactions.
Addition-eliminationreactions betweenCH3C6H4 andC2Hx form- ingCH3C6H4C2Hx+Hareconsideredinthecurrentmodel,andthe ratecoefficientsoftheanalogousphenyl+C2Hxreactionsareused incorrespondingcases.TheconsumptionofCH3C6H4C2Hxiscon- sidered to proceed through a similar scheme to that of toluene.
Thecurrentmodelincludesreactionsbetweenindenyl(C9H7)and acetylene producing ethynyl-indene(C9H7C2H) andmethylnaph- thaleneradical(C10H7CH2),forwhichtheratecoefficientoriginate fromarecentkineticmodelforindenepyrolysis[42].
Asmentionedintheexperimentalsection, theC14H10 isomers, phenanthrene (PC14H10) and anthracene (AC14H10), can be well quantifiedinthecurrentexperiments.PC14H10wasfoundasama- jorthree-ringPAHspecies,andAC14H10wasalsomeasuredofnon- negligibleamounts.Theformationofthree-ringPAHsfrombenzyl recombination was theoretically explored by Sinha and Raj [16], andPC14H10wasidentifiedasthemajorproduct.Relevantstepwise isomerizationandring-closurereactionswere alreadyincludedin our kinetic model [12]. These pathways help in improving the modelpredictionforPC14H10concentrations,however,cannotform sufficient AC14H10 to matchthe experimental measurements. The isomerizationbetweenPC14H10 andAC14H10,withtheratecoeffi- cientsreported by Colket andSeery [43], is alsoincluded in the model, but this reaction results in a net production of PC14H10 overtheinvestigatedtemperaturerange.Discussionontheforma- tionofAC14H10isrelativelylimitedinliterature,comparedtothat of PC14H10. In the latest CRECK model [44] describing PAH for- mation, two lumped reactions, C7H7+CH3C6H4 = C14H10+2H+H2 and CH3C6H4+CH3C6H4 = C14H10+2H+H2, are included in the mechanism to interpret the formation of the lumped C14H10 species. As shown in Scheme 1, the recombination steps result intwo differentC14H14 molecules,2-methyldiphenylmethane and 2,2-dimethylbiphenyl, separately. Similar to the case of benzyl self-recombination, the subsequent dehydrogenation and isomer- ization processes may lead to both AC14H10 and PC14H10. In the current model, AC14H10 and PC14H10 are assigned as the prod- ucts of the lumped reactions starting from CH3C6H4+C7H7 and CH3C6H4+CH3C6H4,respectively.Such an assumptionis basedon the structuralfeatures ofthe corresponding C14H14 molecules,as canbenotedfromScheme1.
Since the kinetic modeldevelopment in our works [12,38,39]
is based on the latest version of CRECK model [44], the
Scheme 1. Potential pathways starting from toluene fuel radicals leading to C 14H 10
isomers.
thermochemical data used in our model are mostly from the CRECK model [44]. For speciesthat are missing fromthe CRECK model[44],suchasindane(C9H10),indanyl(C9H9)andanthracene (AC14H10), corresponding thermochemical parameters come from thedatabaseofBurcat[45].Thekineticmodel,includingthereac- tion mechanismandthespecie thermochemical data,isprovided in the Supplementary Material. Simulations in the present work were performedwiththehomogenous reactormodelofthe soft- ware COSILAB[46] using twodifferent methods: anominalreac- tion time of 4.0 ms and a constant pressure of 20 bar are used inthefirstmethod.Theconstantpressureassumptioninsimulat- ing the speciation sampled from single-pulse shock tube experi- mentswaswelljustifiedinpreviouspublications[47,48].Theother methodistoemploythemeasuredpressureprofilesupto10.0ms tosimulateeachshockoperationandthengatherthefinalspecies concentrations that no longer vary with the time. This takes ac- count of the reactions that may occur in the quenching period [41,49,50], particularlythose involving resonantly stabilized radi- calssuchasbenzyl.Simulatedresultsforspecificspecieswiththe twodifferentmodeswillbecomparedinthefollowingdiscussion.
4. Resultsanddiscussion
Predictive performances of the kinetic model are examined with species concentration measurements, in particular, for im- portantaromaticintermediatesandproducts.Forcomparisonpur- pose, the kinetic model proposed inthe recentwork [37] is also usedtosimulatethespeciationmeasurementsobtainedinthecur- rentwork. Experimentalandmodelingspeciesconcentration pro- filesforindividualspeciesinallfourinvestigatedsetsareshownin Figs. S2–S5 intheSupplementaryMaterial.The currentmodelcan better capturethefueldecompositionreactivity,andmoreimpor- tantly, the speciation ofthearomaticproducts that istargeted in thiswork.Overall,themajorPAHspecies,withthepeakconcentra- tions above 1 ppm,include indene,naphthalene,acenaphthalene, bibenzyl, fluorene andphenanthrene, andthe relative abundance ofthesePAHs arealteredby thefuelcomposition. Thespeciation of most PAHspecies occurs in the temperatureregime of1400–
1600 K,though forspecific species such asindene, thiswindow shiftstolowertemperatureswhenaddingC2fuels.Inthefollowing discussion, experimental observationsandanalyses withthe vali- dated kinetic modelare combinedtoreveal the influencesofthe C2H4 andC2H2 addition onthespeciationfromtoluenepyrolysis.
Attention is paid tocompare the effects broughtby the separate addition of C2H2 andC2H4, andto illustrate thevariation trends ofspecificspeciesconcentrationswhendifferentamountsofC2H2 arepresentinthefuelmixtures.
4.1. Fueldecompositionandmono-aromaticringintermediates
Figure 3displays the fuel concentrations asa function ofthe post-shock temperature T5 in separate cases of toluene/C2H2
(or C2H4) pyrolysis. The kinetic model satisfactorily predicts the measured fuel conversion profiles in individual studied cases. To
Fig. 3. Measured (symbols) and simulated (lines) fuel concentrations as a function of the post-shock temperature T 5in toluene/acetylene (ethylene) pyrolysis. The dark dashed lines are simulated fuel decomposition profiles of (a) neat toluene (106 ppm in argon), (b) neat acetylene (50, 216 and 459 ppm in argon) and (c) neat ethylene (518 ppm in argon). The inset of (a) shows toluene concentrations over the temper- ature range of 130 0 −140 0 K.
better illustratethe mutualeffects betweenthe binaryfuelcom- ponents, simulated fuel decomposition profiles of neat toluene, neatC2H2 andneat C2H4 arealsoshowninFig.3 asa reference.
Given the almost identical contents of toluene in separate cases andan estimateduncertainty of ±30 Kin the calculated T5, the effects of added C2 fuels on toluene decomposition reactivity are not noticeable on the largely overlapped toluene concentra- tion profiles. Thus, toluene concentrations over 1300−1400 K, where the consumption of toluene just begins, are zoomed in and shown in the inset of Fig. 3(a). Experimental and modeling results consistently suggest that the increased C2H2 contents in the binary mixtures promote toluene decomposition. The C2H4 additionalso enhancestoluenedecompositionreactivity, butto a lesserextentthanC2H2.Accordingtotherate-of-production(ROP) analyses for toluene consumption, in all studied cases, toluene decomposition is initiated via the loss of the benzylic hydro- gen (C7H8(+M) = C7H7+H(+M)), while at higher temperatures, tolueneisconsumedmainly throughthehydrogen abstractionby Hatomformingbenzyl(C7H7)(C7H8+H=C7H7+H2).Simulated Hatom concentrationsasafunction ofT5 inindividualcases are shown in Fig. 4(a). Among all the cases of toluene/C2 pyrolysis, thetoluenedecompositionreactivityfollowsthesameorderofthe Hatomconcentrationsbelow1450K.Theslightdecreaseandthe subsequent rise at elevated temperatures in C2H2 concentrations (seeFig.3(b))comefromanoveralleffectofthereactionsbetween
Fig. 4. Simulated H atom and benzyl radical (C 7H 7) concentrations in the pyrolysis of neat toluene (106 ppm in argon) and toluene/acetylene (ethylene) binary mixtures.
Fig. 5. Measured (symbols) and simulated (lines) concentration profiles of (a) ethy- lene in toluene/acetylene pyrolysis and (b) acetylene in toluene/ethylene pyroly- sis. Simulated concentrations in neat toluene (106 ppm in argon) and neat acety- lene/ethylene pyrolysis are also shown as a reference.
C2H2 and toluene related radicals, the C2H2 thermal decompo- sition and the C2H2 formation from toluene consumption. Such phenomenacan bewell reproduced bythe currentkineticmodel in all three cases of binary toluene/C2H2 mixtures pyrolysis.
Moreover,comparedtothecaseofneatC2H2pyrolysisatidentical initialconcentrations,C2H2startstodecomposeatlowertempera- turesinthebinarymixturepyrolysis,andsucheffectsaremoreob- viouswhenraisingtheinitialC2H2contents.Thesynergisticeffects between toluene andC2H2 is also presentedin the recent study by Lietal.[37],andtheauthorsattributedtheenhancedfuelde- compositionreactivitytothereactionC7H7+C2H2=C9H8+H.This process consumes C2H2 at relatively low temperatures, and pro- ducesHatomswhichcanfacilitatethehydrogenabstractionsfrom bothtolueneandC2H2.Thepresenceoftoluenealsopromotesthe decomposition ofC2H4,but toa smallerextent incomparisonto that ofC2H2.Thesynergistic effectsobservedintoluene/C2H4 py- rolysis arise froma similarmechanism, mainly through reactions between benzylandC2H4. Detailed discussion onthe interaction between the binary fuel components and the directly affected speciationbehaviorswillbepresentedinSection4.2.
C2H2 and C2H4 have distinct decomposition reactivity, and in reality, they also exhibit different speciation behaviors as inter- mediates in thecurrentlystudied cases.Concentration profiles of C2H4 intoluene/C2H2 pyrolysis and C2H2 in toluene/C2H4 pyrol- ysis are presented in Fig. 5. The simulated concentration distri- butionsinneat toluene, acetyleneandethylene pyrolysisarealso shown forcomparison.It isseen that thepyrolysis ofC2H2 does not contribute to the formation of C2H4, and instead, the addi- tionofC2H2 reducestheC2H4 productionfromtolueneconsump- tion (see Fig.5(a)). As will be detailedlater, the added C2H2 al-
Fig. 6. Experimental (symbols) and modeling (solid lines) benzene (C 6H 6) concen- trations as a function of the post-shock temperature T 5in (a) toluene/C 2H 2pyroly- sis and (b) toluene/C 2H 4pyrolysis. The inset in (a) plots the measured (dark open square) and simulated (gray cross) peak C 6H 6 concentrations against the initial C 2H 2contents; linear regressions (the dashed lines) of the data points are shown as a visual guide. The simulated C 6H 6concentration profile in neat toluene pyrolysis (the gray solid line) is shown as a reference.
terstheconsumption pathwayoftolueneandits radicals. Oneof the consequences is that the formation of C2H5, the major pre- cursor of C2H4, is inhibited. C2H5 mainly comes fromthe dehy- drogenationofC2H6 followingtheCH3 recombination,andCH3 in thereactionsystemoriginatesfromtheCH3side chainoftoluene molecule.Differently, intoluene/C2H4 pyrolysis,all C2H4 isessen- tiallyconvertedtoC2H2 (seeFig.5(b)). Thisobservationindicates that the toluene-C2H2 chemistry also plays an important role in toluene/C2H4pyrolysis.
Benzene(C6H6) isoneofthemajor productsoftoluene/C2 bi- narymixturespyrolysis,andtheexperimental andmodelingC6H6 concentrations, as a function of T5, are shown in Fig. 6. Simu- latedC6H6speciationprofileinneattoluene(106ppminargon)is alsoprovidedasa reference.TheadditionofbothC2H2andC2H4 lowers the C6H6 formation temperature. This is because the in- creasedHatomconcentrations(seeFig.4(a))facilitateC6H6forma- tionthroughtheipso-substitution reactionC7H8+H= C6H6+CH3. However,thetwoC2 fuelshaveoppositeeffectsonthepeakcon- centrations of C6H6: the added C2H2 slightly reduces C6H6 peak concentrations, and such effects are more pronounced when in- creasingtheinitialC2H2concentration.Differently,thepresenceof C2H4resultsinanincrementinC6H6peakconcentrations,incom- parisontoneattoluenepyrolysis.Thecurrentmodelcancorrectly reproducetheearlyC6H6 formationbroughtbyaddedC2,andac- curately capture the variation trendof C6H6 peak concentrations whenalteringtheinitialfuelcompositions.
TobetterunderstandthedifferentinfluencesonC6H6peakcon- centrations of C2H2 and C2H4 addition, sensitivity analyses for C6H6 are performed at 1500 K, where the C6H6 concentrations areapproachingtheirmaxima.Theanalyzedsensitivityspectraare
Fig. 7. Sensitivity analyses for C 6H 6at T 5= 1500 K in toluene/C 2binary mixtures pyrolysis.
presented in Fig. 7. In all studied cases of toluene/C2 pyrolysis, thepredominantC6H6formationreaction,H+C7H8 =CH3+C6H6, has the highest positive sensitivity coefficient promoting C6H6 production, while its competing channel, the hydrogen abstrac- tion reaction, H + C7H8 = H2+C7H7, has remarkable inhibit- ing effects. Apart from the reactions directly producing C6H6, reactions that generate H atoms also have positive coefficients on the sensitivity spectra. As mentioned above, the reaction C7H7+C2H2 = C9H8+Hisan importantsourceofHatoms, which improvesthefueldecompositionreactivityintoluene/C2H2 pyrol- ysis. However, thisreaction is not among the most sensitive re- actions forC6H6 formation.The reasonliesinthefact that atel- evated temperatures such as the analyzed 1500 K, H atoms are predominantly contributed by the C7H7 decomposition through C7H7→C7H6→C7H5,insteadoftheaddition-eliminationreaction C7H7+C2H2 =C9H8+H.However,thelevelofC7H7 issignificantly reduced(seeFig.4(b))duetoitsrecombinationreactionwithC2H2 formingindene+H.Thepropargyl(C3H3)formationfromC7H5de- composition (C7H5 = C3H3+C4H2) is alsoimpeded,andtherefore the sensitivity of the C3H3 self-recombination forming C6H6 de- creases when increasing the C2H2 contents in the fuel mixtures.
Concerningthetoluene/C2H4 pyrolysis,aremarkableamountofH atoms is produced at elevated temperatures (see Fig. 4(a)) from C2H4 consumption. Besides, the inhibition on C7H7 decomposi- tion is not as pronounced as in the cases of toluene/C2H2 py- rolysis, becauseof the lower reactivityof the benzyl+C2H4 reac- tion,aswillbe discussedinSection4.2.Furthermore,thestyrene (C6H5C2H3)decompositionreaction,C6H5C2H3 =C6H6+H2CC,ap- pears as an alternative C6H6 producing channel in toluene/C2H4 pyrolysis. Thesemultiplefactors explain theincreased C6H6 peak concentrationintoluene/C2H4pyrolysis.
Figure 8 displays the experimental and modeling concentra- tionprofiles ofC8mono-aromatichydrocarbon(MAH)species,in- cludingethylbenzene(C6H5C2H5),styrene(C6H5C2H3)andpheny- lacetylene (C6H5C2H), in the pyrolysis of toluene/C2 binary mix- tures. Forall three C8 MAHs, the currentkinetic model can well capture the variation trends of their peak concentrations when altering the initial fuel compositions. The added C2H2 andC2H4 reduces the formation of C6H5C2H5, and C2H2 exhibits more pronounced effects if the same amount of C2H2 and C2H4 are added. Such inhibiting effects arise from the fact that a part of C7H7 reacts with the added C2H2 or C2H4, so that the car-
Fig. 8. Experimental (symbols) and modeling (solid lines) concentrations of C 8
MAH species as a function of the post-shock temperature T 5. The left panel:
toluene/C 2H 2pyrolysis; the upper left insets plot the measured (dark open square) and simulated (gray cross) peak concentrations of separate species against the ini- tial C 2H 2 contents; linear regressions (the dashed lines) of the data points are shown as a visual guide. The right panel: toluene/C 2H 4pyrolysis. The simulated species concentration profiles in neat toluene pyrolysis (the gray solid lines) are shown as a reference.
bon fluxinto C6H5C2H5 through C7H7+CH3 recombinationis de- creased. C2H2 and C2H4 have opposite effects on the forma- tion of C6H5C2H3. A cleardecreasing trend is noted in the peak C6H5C2H3 concentrations when increasing the initial C2H2 con- tents, while the C6H5C2H3 production from toluene pyrolysis is doubled by the addition of C2H4. Such differences can be ex- plained by the varied dominant C6H5C2H3 formation pathways:
in neat toluene and toluene/C2H2 mixture pyrolysis, C6H5C2H3 is mainly produced from the consumption of C6H5C2H5, while the reaction C6H5+C2H4 = C6H5C2H3+H governs the C6H5C2H3 formation when C2H4 is introduced. The speciation temperature windows for the above-mentioned C6H5C2H5 and C6H5C2H3 are not significantly changed by the addition of C2 hydrocarbons.
However, the formation of C6H5C2H shifts towards lower tem- peratures when C2H2 or C2H4 is present in the fuel mixtures.
Meanwhile, the peak concentrations increase, particularly when C2H2 is added to toluene pyrolysis. The formation of C6H5C2H relies on the reactions between phenyl (C6H5) and C2H2, and C6H5 mainly comes from toluene unimolecular decomposition (C7H8(+M)=C6H5+CH3(+M)).Inneattoluenepyrolysis,thepro- duction of C6H5C2H is limited by the level of C2H2, which is largelyproducedonlyatelevatedtemperaturesabove1450K.This explains how C2H2, either already existing in fuel mixtures or produced from C2H4 decomposition, enhances the production of C6H5C2Hatrelativelylowtemperatures.
4.2. Intermediatesandproductsfromtolueneradicals-C2interactions As already mentioned above, the synergistic effects on fuel decomposition reactivity in toluene/C2 binary mixture pyrolysis
Fig. 9. Experimental (symbols) and modeling (solid lines) concentrations of C 9H 8
species as a function of the post-shock temperature T 5in toluene/acetylene pyrol- ysis. The upper right insets plot the measured (dark open square) and simulated (gray cross) peak concentrations of separate species against the initial C 2H 2con- tents; linear regressions (the dashed lines) of the data points are shown as a vi- sual guide. Simulated concentration profiles in neat toluene pyrolysis (the gray solid lines) are shown as a reference.
derive from the interactions between toluene radicals and the C2 fuels. A direct proof is offered by the observation of abun- dant C9 species.Intoluene/C2H2 pyrolysis experiments,indeneis identified as the dominant C9H8 species,and minor amounts of other C9H8 isomers, propynyl-benzene (C6H5C3H3) and ethynyl- toluene (CH3C6H4C2H) species, are also detected (see Fig. 2).
Figure 9 presents the experimental and modeling concentra- tion profiles for indene and the sum of other C9H8 isomers in toluene/C2H2 pyrolysis,andthevariation trendsof thepeak con- centrationsagainsttheinitialC2H2concentrationsarealsoshown.
Byadoptingtheratecoefficient reportedbyMebel etal.[17],the currentkinetic modelcan preciselypredictindeneconcentrations overtheentiretemperaturerangeinallthreetoluene/C2H2exper- imental setswithdifferentinitialC2H2 concentrations.The minor C9H8specieshaveasimilarspeciationtemperaturewindowtothat ofindene,andtheyformvia theadditionofC7H7 orCH3C6H4 to thetriplebondofC2H2.ForbothindeneandotherminorC9H8iso- mers, their peak concentrationsalmost linearlyincrease withthe quantity ofC2H2 addedtotoluenepyrolysis,andsuchatrendcan bewellpredictedbythekineticmodel.Indenespeciationprofilein neattoluene(106ppminargon)pyrolysisisshownasareference:
it starts to formslightly above 1400 K, andreaches a peak con- centrationbelow1 ppmataround 1500 K.The presenceofC2H2 lowerstheonsettemperaturetoabout1200K,andremarkablyfa- cilitatesindeneproduction:Evenbyaddingonly50ppmC2H2,the peak concentration of indene reaches over 2 ppm. This suggests theroleofthereactionC7H7+C2H2=C9H8+Hasanefficientpath- wayproducingindene.
The indene concentration profile in toluene/C2H4 pyrolysis is shown in Fig. 10, which also displays the concentration profiles of C9H10 species, including indane and other isomers, namely, propenyl-benzene (C6H5C3H5) and vinyl-toluene (CH3C6H4C2H3) species (see Fig. 2). C2H4 inthe fuel mixture can alsolower the
indene speciation temperature and enhance the indene produc- tion, but to a lesser extent in comparison to C2H2 if the same amounts of C2H2 and C2H4 are concerned. With over 500 ppm C2H4 added to toluene pyrolysis, the resulting peak indene con- centration is about 4 ppm, falling between those in TA_50 and TA_216pyrolysis(seeFig.9).Indaneisexpectedtobeabundantin toluene/C2H4 pyrolysis,asits productionfromtheinteraction be- tweenC7H7 andC2H4throughC7H7+C2H4 =C9H10+Histhought to be efficient. However, only a trace quantity (10−1 ppm level) ofindaneismeasured intoluene/C2H4 pyrolysis.The limitedpro- ductionandthequickconsumptionofindanetogetheraccountfor thelowconcentrationsofindaneobservedintoluene/C2H4pyrol- ysis.Morespecifically,theindaneformationfromC7H7+C2H4pro- ceeds at a slower rate than the C7H7+C2H2 reaction producing indene; and indane rapidly decomposes to indene or isomerizes throughring-openingprocessesundertheinvestigatedconditions.
At relativelylow temperatures,the decompositionof indanegov- ernstheindeneformation;whileatelevatedtemperatures,theef- ficientindeneformationchannel,C7H7+C2H2=C9H8+H,becomes dominant,asconsiderableC2H2 isformed fromC2H4 decomposi- tion.Asmentionedinthemodelingsection,theratecoefficientof thereactionC7H7+C2H4=indane+Hisassessedthroughananal- ogytothereactionC7H7+C2H2 =indene+Hreportedin[18].This practicepotentially introducesuncertaintiesinthe indanespecia- tionkinetics,thoughthe currentmodelsatisfactorily capturesthe measurementsofC9speciesintoluene/C2H4pyrolysis.Futurethe- oretical works on the C7H7+C2H4 reaction system and how in- danefurtherdehydrogenatestoformindenearehighlynecessary.
Compared to indene and indane, the monocyclic C9H10 isomers (see Fig. 10(c)) havesignificant contents in toluene/C2H4 pyroly- sis.Allylbenzene(C6H5C3H5–1),1-propenyl-benzene(C6H5C3H5–2) andlumped o-, m- andp- vinyltoluene isomers (CH3C6H4C2H3) are included in the current kinetic model. Simulated concentra- tionprofilesofindividualabove-mentionedC9H10isomersarepre- sentedin Fig.S6 in the SupplementaryMaterial.According to the modelpredictions,bothC6H5C3H5–1andCH3C6H5C2H3bothhave peakconcentrationsofabout1ppm;the formerpeaksataround 1400 K while the later reaches its maximum ataround 1500 K.
The consumption of these C9H10 isomers partly contributes the formation of indene through dehydrogenation and ring-closure steps.
Ithasbeen discussedinour previouswork [12]that theben- zyl(C7H7) chemistry controlsthe speciation,particularly forPAH species,intoluenepyrolysis.Toexploretheinfluencesoftheadded C2 fuels on C7H7 consumption, ROP analyses of C7H7 are per- formedat 1400K,whereabundantC7H7 existsin thetoluene/C2 pyrolysisreactionsystems(see Fig.4(b)).According totheresults inFig.11,thefateofC7H7 isverysensitivetotheinitialfuelcom- position.C7H7 eitherdecomposes,orreactswithotherspecies,in- cludingtheC2 fuels,CH3 andC7H7 itself.Thecontributionsofre- actionsinvolvingC2fuelsincreasenotably withtheinitialC2con- centrations.Thus,othercompetingpathwaysareinhibited,among
Fig. 10. Experimental (symbols) and modeling (solid lines) concentrations of C 9species as a function of the post-shock temperature T 5in toluene/ethylene pyrolysis. Simu- lated concentration profiles in neat toluene pyrolysis (the gray solid lines) are shown as a reference.
