甲醇氧化详细动力学机理构建与多工况验证（AComprehensveMechanismforMethanolOxidation）摘要：本文构建了覆盖633–2050K、0.26–20atm、当量比0.05–2.6的甲醇氧化详细动力学机理，包含完整H₂/O₂、CO、CH₂O与甲醇子机理。通过静态反应器、流动反应器、激波管、层流火焰多工况验证，明确HO₂/H₂O₂体系在低温链分支、高温链终止的关键作用，修正CH₃OH+OH分支比与HCO分解等核心反应速率。机理精准复现点火延迟、火焰速度与组分分布，阐明甲醇经CH₂OH/CH₃O→CH₂O→CO→CO₂的反应路径，为宽工况甲醇燃烧模拟提供可靠详细机理。ABSTRACT:Acomprehensivedetailedchemicalkineticmechanismformethanoloxidationhasbeendevelopedandvalidatedagainstmultipleexperimentaldatasets.Thedataarefromstatic-reactor,flow-reactor,shock-tube,andlaminar-flameexperiments,andcoverconditionsoftemperaturefrom633–2050K,pressurefrom0.26–20atm,andequivalenceratiofrom0.05–2.6.Methanoloxidationisfoundtobehighlysensitivetothekineticsofthehydroperoxylradicalthroughachain-branchingreactionsequenceinvolvinghydrogenperoxideatlowtem-peratures,andachain-terminatingpathathightemperatures.Thesensitivitypersistsatun-usuallyhightemperaturesduetothefastreactionofCH2OH+O2=CH2O+HO2comparedtoCH2OH+M=CH2O+H+M.ThebranchingratioofCH3OH+OH=CH2OH/CH3O+H2Owasfoundtobeamoreimportantparameterunderthehighertemperatureconditions,duetotherate-controllingnatureofthebranchingreactionoftheH-atomformedthroughCH3Othermaldecomposition.◎1998JohnWiley&Sons,Inc.IntJChemKinet30:805–830,1998INTRODUCTIONDetailedchemicalkineticmechanismscanbeusefulengineeringtools,whichpermitexplorationofthemi-croscopicchemicalprocessesthatunderlieandsome-timescontrolthemacroscopicphysicalprocesses,suchasflamespeedorautoignitiontime.Themech-anismsaresystemsofmanyelementaryreactions,withrateconstantsdetermined,wherepossible,byfunda-mentalkineticexperimentsortheoreticaltreatment.Theexaminationofthecontributionoftheelementaryreactionswithinthecontextofthelargersystemaidsinidentificationofrateconstantsorreactionchannelsthatwarrantfurtherinvestigationbyfundamentalki-neticists.Detailedmechanismsareoftendevelopedinre-sponsetoandvalidatedagainstasinglesetofexper-imentalmeasurements.Asaresult,therangeofap-plicabilityofthemechanism,asdefinedprimarilybytemperature,pressure,andequivalenceratio,islimitedtothatcoveredbythedataset.Theterm“comprehen-sive”impliesthattherangeofvalidityofthemecha-nismhasbeenextendedtothemaximumpracticalex-tentbycomparisontomultipleexperimentaldatasets.Thecreationofsuchmechanismshasadditionalutil-ity.Manycombustionprocessesoccuroverawiderangeofconditions(Anexampleistheautoignitionprocessinspark-ignitionengines.Thefuel/airchargeentersthecylinderatnear-ambientconditions,andinatransientprocess,iscompressedto10–40atm,andanunburnedgastemperatureof1000–1200K).Acomprehensivemechanismmaybeusedtonotonlyidentifyimportantreactions,butalsothecontrollingchemicalregimesduringthetransientprocess.Inthepresentwork,adetailedkineticmechanism,initiallydevelopedincomparisontolowandinter-mediatetemperature(<1100K)flowreactordata,isextendedbycomparisontoshock-tube,flame-speed,andstatic-reactordata.Reactionpathandsensitivityanalysisisusedtoidentifycontrollingreactionchan-nelsandrateconstants,andtoindicateseveralreac-tionsthatrequireadditionalstudy.PREVIOUSMODELINGSTUDIESWestbrookandDryerThefirstcomprehensivedetailedkineticmodelofmethanoloxidationwasdevelopedbyWestbrookandDryer[1].Inclusionofreactionpathsthatwereim-portantatbothhighandintermediatetemperaturesproducedsuccessfulreproductionofflow-reactorandshock-tubedata.Usingasimplifieddiffusionmodel,thestoichiometriclaminarflamespeedofapremixedmethanol/airmixturewascalculatedtobe44±2cm/satoneatmosphere,298Kinitialconditions.Theworkwashamperedbyalackofelementaryrateconstantandreactionpathinformation.Manyim-portantrateconstants,includingthoseformethanolandhydroxymethyl(CH2OH)thermaldecomposition,andHandOHabstractionreactions,wereestimatedinthecontextofthedetailedmodeling.Themethoxyradical(CH3O)wasneglectedinthemechanism.Con-siderableimportancewasplaceduponthedehydrationreactionCH3OH+OH=CH3+H2OasasourceofmethaneandC2hydrocarbons.Laterwork[2]hasindicatedthatthisreactionrouteisneg-ligible,andothersourcesaresufficienttoaccountforexperimentalmethaneandC2speciesmeasurements.Theeffectsofpressure-dependentreactionswerenotincludedinthisearlymodelingeffort.NortonandDryerOxidationMechanism:InanefforttoresolveseveraldifficultieswiththeWestbrook/Dryer[1]mechanism,NortonandDryerupdatedthemodelusingmorecur-rentrateconstantsandaconsistentsetofthermochem-icalparameters[3].Inafewinstances,differentprod-uctpathswereproposedbasedonmorerecentwork.Theupdatedmechanismwascomparedtoanewsetofatmospheric-pressureflow-reactordata[4],encom-passingthetemperaturerangeof1025–1090Kandequivalenceratiosfrom0.6–1.6.Improvedagreementwiththeflow-reactordatawasattained,andtheim-portanceofthehydroperoxylradical(HO2)tometh-anoloxidationkineticswasidentified.Noattemptwasmadetorecalculatethefullexperimentalbasissetfromtheoriginalreference[1].ComprehensivePyrolysisMechanism:Thesameau-thorsexpandedupontheirpreviousworktodevelopacomprehensivemodelformethanolpyrolysis[2].Al-thoughnotanoxidationmechanism,manyoftheim-portantreactionpathsinhigh-temperatureoxidationareidenticaltothosefoundinpyrolysisstudiesaswell.Themechanismwascomparedsuccessfullytostatic-reactor,flow-reactor,andshock-tubedata.Egolfopoulos,Du,andLawAmorerecentattempt[5]atcreatingacomprehensivemethanoloxidationmechanismwasbasedprimarilyuponpremixedlaminarflame-speedmeasurementsoverarangeofinitialtemperaturesandpressures.Ex-cellentagreementwasattainedforboththelaminarflamespeedandatmospheric-pressureflow-reactordataset[4].TheagreementwithBowman’sshock-tubeignitiondelaymeasurements[6]waslesssatis-factory,andonlyasmallsubsetofthedatawerere-ported.Laminar-flame-speciesprofileswerecomparedtodataofVandoorenandVanTiggelen[7,8],Pauwelsetal.[9],andBradley[10].Althoughthecalculationtechniquewasnotspecified,itisapparentthatthespe-ciesprofileswereforcedtomatchthedatapointclos-esttotheburnerface.Itisthereforedifficulttojudgetheaccuracyofthepremixedflameprofilecalcula-tions.Unfortunately,theproductsofaCH3+OHreac-tionwereerroneouslyassignedasCH2OH+HratherthanCH3O+H[11].Theresultingreversereactionratecoefficientishigherthancollisional,andsignifi-cantlyaffectsthecalculatedresults.Withtheproductchannelcorrectlyspecified,themechanismcalculatesflamespeedsinsubstantialdisagreementwiththeau-thors’experimentalmeasurements.TheerroralsooveremphasizestherelativeimportanceofC2chem-istryduetothehigherCH3productionrateandaltersthesensitivityofthecalculationstotheCH3OH+OH=CH2OH/CH3O+H2Obranchingratio,asdis-cussedfurtherbelow.GrotheerAlthoughnotdescribedasacomprehensivemecha-nism,themodelofGrotheeretal.,[12,13]hasbeenappliedtobothpremixedlaminarflamespeedcalcu-lationsandtoautoignitioninaspark-ignitionengine[14].Thepublishedcomparisonofflamespeedtoex-perimentaldataisexcellent.Gradientsensitivityco-efficientsforflamespeedidentifyanumberofimpor-tantreactions,includingHO2+H=productsandhydroxymethyldecomposition.Theauthorsalsoidentifiedthebranchingratiobe-tweenCH3OH+OH=CH2OH/CH3O+H2O(de-finedaskCH2OH/ktotal)asanimportantparameterincal-culatingflamespeeds.Thevaluechosentogiveoptimalagreementwiththemeasurementsis0.85.SeveralindependentexperimentalandtheoreticalstudiesoftheOHabstractionreaction,indicateanin-creasingcontributionofthemethoxyradicalpathwithincreasingtemperature,approachingavalueofthebranchingratioof0.5above865K.Substitutionofthelowerbranchingratiointothebaselinemechanismwouldresultinasignificantincreaseinthecalculatedlaminarflamespeeds.DETAILEDMECHANISMDEVELOPMENTMechanismDevelopmentThedevelopmentofadetailedkineticmodelisahi-erarchicalprocedure.Thebasisforanyhydrocarbonoxidationisthesubsetofreactionsinvolvinghydro-gen,oxygen,andtheirassociatedintermediatesandproducts.TheseincludeH-andO-atoms,hydroxyl(OH)andhydroperoxyl(HO2)radicals,andhydrogenperoxide(H2O2).andwater.Thissubmechanismde-terminestoalargeextentthecharacteristicsoftherad-icalpoolresponsibleforchainpropagation,termina-tion,andbranching.Theoxidationofcarbon-containingspeciesfollowsabasicseriesofsteps,beginningwithinitiationreac-tions,followedbyradicalattackonthefuel,produc-tionof(generally)smallerintermediates,andfinallyachainofaldehyde→CO→CO2steps[15].Takenininverseorder,thesestepsformthebasicreactionhi-erarchyinadetailedkineticmechanism.Thissectiondescribestheproceduresandsourcesusedincompilingthedetailedmodelsofthisstudy.Althoughattentionispreferentiallyfocuseduponre-actionsofthemostsignificancetothemechanism,dis-cussionoftherelativeimportanceofindividualreac-tionsappearsinthefollowingsectionswhererelevant.CO/H2/O2Thecarbonmonoxide/hydrogen/oxygenreactionsys-temusedinthisstudyistakenprimarilyfromthemechanismofYetteretal.[16],whichhasbeenre-centlymodifiedtoreflecthigh-pressurestudiesofKimetal.[17,18].Thissubmechanismwasverifiedagainstaseriesofflow-reactor,static-reactor,andshock-tubeexperiments,andiswellestablishedasprovidinganaccuratedepictionofhydrogenandcarbonmonoxideoxidationoverawiderangeofconditions.CH2OFormaldehydeoxidationkineticsareofgreatimpor-tancetotheoxidationoflargerhydrocarbonandox-ygenatedhydrocarbonspecies.Undermostcircum-stances,nearlyallofthecarboninthesespeciesisoxidizedthrougharouteinvolvingformaldehyde.Methanoloxidationisnotanexceptiontothisgener-alization.Thus,accuratemodelingofmethanoloxi-dationrequiressignificantattentiontothedetailsoftheformaldehydeoxidationmechanismaswell.Thedifficultiesinvolvedingeneratingformalde-hydehaveresultedinarelativelysparseexperimentaldatasetformechanismvalidation[19].Thecompre-hensivemodelingeffortofHochgrebandDryer[20]encompassedthefullrangeofavailabledata,includingshock-tube,flow-reactor,andstatic-reactordata.Forthepresentmodelingstudy,theCH2Osubmechanismwasmodifiedasdescribedinanearlierarticle[21].TherateconstantsofCH2O+HandCH2O+OHwerereducedfromtheirhighvaluesintheearlierwork.Thesechangeswerenecessarytoreproducepeakformaldehydeyieldsintheflowreactorexperi-ments[22].Thenewrateconstantsforthesereactionsallowedaccuratecalculationoftherelativeformalde-hydeandmethanoldestructionratesinflow-reactorexperimentsontheoxidationandpyrolysisofmixturesofthesetwospecies.Recalculationofthedatasetusedintheoriginalmechanismdevelopment[20]showedlittlechangeinthecalculatedspeciespro-files,withtheexceptionoftheoxidativepyrolysisflow-reactorexperiments.However,thepresenceofanuncontrolledtracecontaminationofoxygenintheseexperimentscastsuncertaintyupontheiraccuracy.Consideringthisuncertainty,thealterationofthecal-culatedformaldehydeconsumptionrateiswithintheerrorlimitoftheexperiments.CH3OHThemethanolsubmechanismrequiresadditionofthefollowingspecies:methanol(CH3OH),hydroxy-methyl(CH2OH),andmethoxy(CH3O).Thehigh-pressureflow-reactorexperimentsindicatedthepres-enceofformicacid(HCOOH)and1,2-ethanediol(ethyleneglycol,HOC2H4OH)asminorintermediates.Forthepresentwork,tentativeformationmechanismsforthesespeciesaredescribedbelow.Thetwometh-anolmechanismsofNortonandDryerprovidedtheinitialbasisforthepresentmethanolsubmechanism.Anumberofrateconstantswerechangedfromtheoriginalmechanismstoreflectmorerecentratecon-stantmeasurements.Thedetailsofthemostsignificantportionsofthemechanismarediscussedbelow.Initiation/Decomposition:Althoughtherateconstantsofinitiationreactionsareseldomimportantunderflow-reactor,static-reactor,orlaminar-flamecondi-tions,theymayplayamoresignificantroleinshock-tubestudies.Atleastfourdifferentdecompositionre-actionsarepossible:CH3OH+M=CH3+OHΔHr=92.2kcal/mol=CH2OH+H98.0kcal/mol=CH3O+H104.1kcal/mol=1CH2+H2O91.8kcal/molThefirstreactionpredominates,accountingfor75%to90%ofthetotaldecompositionrateinvariousstud-ies[23–25].Themethoxyradicalchannelisthermo-dynamicallyunfavorable.Thehydroxymethylchannelisusuallyassumedtoaccountfortheremainderoftheinitiationrate,althoughthesingletmethylenechannelisaninterestingalternativeproposedbyDombrowskyetal.[23].Asyetthereisnodirectevidencetosupportthischannel,butitcouldbeconsideredinfuturestud-ies.Forthepresentmodel,aTroefitfortheCH3+OHchannelwasgeneratedbasedonthefalloffparam-etersofTsang[26].Forsimplicity,thehydroxymethylchannelisassignedarateof10%oftheprimarychan-nel.Amoreaccurateexpressionshouldtakeintoac-countthehigheractivationenergyexpectedwiththemoreendothermicreactionpath;however,thecalcu-lationsareinsensitivetotherateconstantofthisre-action.OHAbstraction:TheabstractionreactionsofOHarethepredominantfuelconsumptionroutesinthemeth-anolmechanism.Theabstractionmayoccurateitherthemethylorhydroxylgroup,formingCH2OH+H2O(83)orCH3O+H2O(84),respectively.Exceptatthehighesttemperatures,thetwospeciesreactbyconsiderablydifferentmechanismsCH2OH+O2=CH2O+HO2(70)CH3O+M=CH2O+H+M(42)Itisthereforeimportanttomaintainadistinctionbe-tweenthetwoCH3Oisomers.TworecentstudiesreportedoverallratesforthereactionofmethanolwithOH.HessandTully[27]obtainedanexpressionfork83+k84=3.54×104T2.6exp(883/RT)overthetemperaturerange293–803K*.Isotopicsubstitutionallowedanesti-mateofthe“branchingratio,”definedask83/(k83+k84),thatincreasedfromasmallvalueto0.5attheirhighesttemperature.Morerecently,BottandCohen[28]obtainedavaluefork83+k84of5.2×1012at1200K,inexcellentagreementwiththevalueof5.1×1012obtainedfromthepreviousexpression.Theircalculatedsitespecificexpressionsyieldabranchingratiothatincreasesfrom0.39at1000Kto0.51at2000K.Thepresentmechanismusestheex-pressionofBottandCohen.ReactionwithH:Methanolmayreactwithhydrogenatomsbyabstractionfromeithersite,orbydehydra-tion,formingmethylandwater.Thelatterreactionwassuggestedasasignificantsourceofmethylradicals[1],althoughsubsequentstudies[2]havefailedtode-tectevidenceofthischannel.Thepresentmodeldoesnotincludethisreaction.Itsinclusionattheratecon-stantsuggestedbyNorton[3]hasnodiscernibleeffectontheoverallkinetics.Theabstractionreactionconsumesasignificantfractionofthemethanol,particularlyunderfuelrichconditions.Theproductchannelyieldinghydroxy-methylis6.1kcal/molexothermic,whilethemethoxychannelisalmostthermoneutral.ConsistentwiththepyrolysisstudyofNortonandDryer[2],theratecon-stantofWarnatz[29]wasappliedwitha20%contri-butionbythemethoxyradicalpath.CH3O/CH2OHIsomerization:Apossibleisomeriza-tionreactionbetweenCH3OandCH2OHwasfirstsug-gestedasalossmechanismforCH3Oinafundamentalkineticsexperiment[30].Becausethermodynamicequilibriumstronglyfavorshydroxymethyl,theisom-erizationprimarilyconvertsmethoxytohydroxy-methyl.BecauseiteliminatestheH-atomproducedbymethoxydecomposition,theisomerizationreactioncouldbeveryimportantifitoccursataratecompa-rabletothedecompositionrate.Theoreticalandther-mochemicalestimatesoftheisomerizationratecon-stant[31–33]consistentlyplaceitsvalueatorbelowabout10%ofthatformethoxydecomposition.Ex-perimentalwork[34]alsosuggeststhattheupperlimitfortheisomerizationreactionis10%ofthedecom-positionrate.Atthisupperlimit,theisomerizationre-actiondoesnotsignificantlyaffecttheresultsofthedetailedmodel.Sincenodirectevidenceexiststosup-portthisreactionpath,thepresentmechanismdoesnotincludeit.MinorSpeciesFormation:InfraredspectracollectedduringtheVPFRexperimentsindicatedthepresenceofdetectableamountsofformicacidasaninterme-diatespecies(approximately50ppmwith4000ppminitialmethanol).Apostulatedformationmechanismisacombinationreactionbetweenhydroxymethylandhydroperoxylradicals,followedbydecompositionorrearrangementanddecomposition:CH2OH+HO2=CH2(OOH)OHCH2(OOH)OH=CH2(O)OH+OH-or-CH2(OOH)OH=HCOOH+H2OCH2(O)OH=HCOOH+HSpangenbergetal.[35]originallypostulatedthede-hydrationroute,whichinvolvesafairlyhighlystrainedtransitionstate.Thedecompositionrouteseemsmorelikely,butthereisnodirectevidencetosupportonepathovertheother.Thepresentmecha-nismincludesbothreactionswithestimatedrateco-efficientsof3.0×1013cm3/mol-s.Thevariationofformicacidmolefractionwithequivalenceratioisnotwellreproduced,andthusthismechanismcanonlybeconsideredtentative,atbest.Inthemostfuel-richhigh-pressureflow-reactorex-periments,aspectralfeatureidentifiedas1,2-ethane-diol(ethyleneglycol)wasdetectedaftertheoxygenhadbeencompletelyconsumed[21,22].Clearly,thisspeciesisaproductofhydroxymethyl(CH2OH)di-merization.Althoughthequalityofthespectrumwasinsufficienttopermititsquantification,thepossibilitythat1,2-ethanediolformationisanimportantradicalterminationpathforfuel-richconditionsledtothein-clusionofthisreactioninthemechanism.However,itsimpactontheoverallpredictionsofthemechanismisnegligible.C2Species:Becausesmallamountsofmethylradicalsarecreatedduringtheoxidationandpyrolysisofmeth-anol,C2orlargerhydrocarbonspeciesmaybeformedbytheirrecombination.Egolfopoulosetal.[5]re-portedsignificanteffectsoftheinclusionofadetailedC2submechanismontheircalculatedlaminarflamespeeds.Becauseoftheincorrectlyhigh-rateconstantforCH2OH+H=CH3+OHusedintheirmecha-nism[11,13],itislikelythatthisconclusionisinflu-encedbyanerroneouslyhighCH3productionrate.Inthepresentwork,asimpleC2mechanismwasassem-bledprimarilyfromthecompilationofTsangandHampson[36]forpurposesoftestingtheinfluenceofhighercarbonnumberkineticsonmethanoloxidation.Underallcasessimulated,noinfluenceofC2chem-istrycouldbedetectedinanyofthecalculatedspeciesprofiles,overallreactionrates,ignitiondelaysorpre-mixedflamespeeds.Summary:Thecumulativemechanismusedforcom-parisontothemethanolexperimentsappearsinTableI,withtheforwardratecoefficientsandreferences.Thereversereactionratesarecalculatedbydetailedbalanceandthermodynamicparameters,listedinTa-bleII.MostofthesedataarefromtheSandiather-modynamicdatabase[37].TheenthalpyofformationforCH2OHhasbeenchangedtoreflecttherecentmea-surementsofSeetulaandGutman[38].SolutionTechniqueSixdifferenttypesofexperimentsweresimulatedinthisstudy,staticreactors,flowreactors,shocktubes,premixedflamesextrapolatedtothefreely-propagat-ing,unstretchedcondition,andburner-stabilizedflatpremixedflames.TheChemkin-IIpackage[39]wasusedforthesimulations.Thefundamentalmodelingassumptionsaresummarizedasfollows.StaticReactor:Constantvolume,spatiallyhomoge-neous.Theassumptionofaspatiallyhomogeneousmixturerequiresthatthereactiontimeismuchlongerthanthecharacteristicthermalandmassdiffusiontimestothereactorwalls.Theimplicationsofthesecharacteristicsofstaticreactorexperimentsaredis-cussedbelow.FlowReactor:Constantpressure,adiabatic,zero-di-mensional.Theconstantpressureassumptionisessen-tiallyalowMachnumberassumption.Adiabaticityisapproximatedintheexperimentsthroughtheuseofpreheatedreactortubewallsandashortlengthtodi-ameterratiointhereactortube.Zero-dimensionalityisvalidinthecaseofnegligibleaxialandradialdif-fusion.Radialdiffusionisheldtoaminimum,againbylimitingtheexperimenttoL/Dvaluessuchthattheflowisessentiallyanentry-regionflow,wherethede-velopingboundarylayersdonotinteractstronglywitharadiallyuniformcoreflow.Finally,axialdiffusionisnegligiblewherethecharacteristicdiffusionlengthismuchgreaterthantheconvectivelength.Althoughthisassumptionisreasonablyvalidovermuchofthereactionzone,itbreaksdowninregionsofhighcon-centrationgradients,suchasintherapidtransitionintheoxidationrateofCOaccompanyingthedepletionofhydrocarbonspeciesinFigure12.Thefinite-ratemixingoffuelandoxidizer,recir-culationzonesnearthemixingregion,andresidualeffectsofaxialdiffusioninthisregionallresultinuncertaintyinspecificationofanabsolute“zerotime”forflowreactorexperiments.Modelingsimulationoftheseeffectsusingstirredreactor-plugflowcoupledmodelshasshownthattheyallservetotranslatethecalculatedspeciesandtemperatureprofilesalongthetimeaxistowardtheorigin.Furthermore,theinitialperturbationsofthesystemarequicklyrelaxedandresultinnohistoricaleffectsdownstreamofthisre-gion,otherthanshiftingtheentirereactionprofiles(withoutperturbations)withrespecttothe“zerotime.”Whenthecalculatedprofilesareartificiallytemporallyalignedatanarbitraryreferencepointwithinthedownstreamreactionzonewheremethanoldisappear-anceisobserved,thecalculationsandexperimentalprofilesoverlayoneanothernearlyperfectly.Thus,incomparingcalculationswithexperimentaldata,thetimeaxisofthedataiseffectively“translated”toachieveaminimumRMSerrorwiththecalculatedfueldecayprofile.Themagnitudeoftherequiredshiftisnotedinthefigurecaptions.Timeshiftingandthere-sultsachievedbytheaboveapproachareentirelycon-sistentwiththeassumption(notedabove)thatwithintherangeofextentsofreactiontobecomparedwiththecalculation,axialdiffusiontimescalesaremuchlongerthankineticandconvectivetimescales.Math-ematically,thesolutionoftheconservationequationsthenbecomesaninitialvalueproblem,andanysinglematchingpointbetweentheexperimentandcompu-tationisequivalent(Computationally,calculationscanbemarchedupstreamordownstreamofthematchingpointwithoutconcern).ShockTubes:Thethermalenvironmentinthepost-shockregioncanbesafelyassumedtobeadiabatic.Also,theshortreactiontimescalesrelativetodiffusivetimespermitsthezero-dimensionalapproximation.Thetreatmentofthefreeboundaryofthereactionzoneisopentosomedebate.Alimitingcase,frequentlyapplied,assumesaconstant-volume(density)bound-ary,whichimpliesthatthebulkexpansionofthefluidduetotemperatureriseandaveragemolecularweightchangeoverwhelmstheinertialeffectsofthesur-roundingfluid.However,thesituationisrarelyasclear-cutasthesimplifiedmodelwouldindicate.Shortofsolvingtheone-dimensionalmomentumequation,thebestthatcanbeassumedisthatrealityliesbetweenthelimitingcasesofconstantdensityandconstantpressure.Bothcaseswerecalculatedinthisworkandrepresentativepointsareindicatedonthefiguresastheaverageparameterwitherrorbarsindicatingthelim-itingcases.TheSENKINprogram[40]wasusedtocalculatealltheprecedingthreecases,i.e.,casesinvolvingstatic-reactor,flow-reactor,andshock-tubecompari-sons.PremixedLaminarFlameSpeeds:TheChemkinpro-gramPREMIXwasusedtosimulateafreely-propa-gating,one-dimensional,constant-pressureadiabaticflame.Themulticomponentdiffusionmodelwasused,andthermaldiffusionofHandH2wasincludedinthecalculations.Thewindwarddifferencingnumericalschemewasusedformostofthecalculations,duetoitssuperiorconvergenceproperties.Atthehighgridresolutionused(approximately150nodeswithintheflame,and50inthepreheatandpost-flameregions),lessthan0.5cm/sdifferenceincalculatedflamespeedresultswhenusingthemoreaccuratebutlessstablecentraldifferencingscheme.Burner-StabilizedPremixedFlatFlames:Speciespro-filesthroughseverallow-pressure,burner-stabilizedflameswerecalculatedusingPREMIX.Themeasuredtemperatureprofileswereusedasinputstothemodelduetotheunquantifiedheatlossestotheburner.Thesamediffusionmodelandnumericalparameterswereusedasforthefreely-propagatingflamecalculations.RESULTSANDDISCUSSIONStaticReactorsStaticreactorsaretypicallyusedtostudylowtemper-atureoxidationchemistry,wherethereactiontimescalesaremeasuredinminutes.Theprincipaladvan-tagesoftheseexperimentsaretheirsimplicity,andessentiallyunlimitedtimeavailableforobservationofslowreactions.However,theinfluenceofsurfaceshaslongremaineddifficulttohandlefornumericalmod-elers.Theexperimentsareoftendifficulttocontrolwithouttreatmentofsurfacesbyrigorouscleaningprocedures,coatingwithvarioussubstances,and/or“aging”or“seasoning”ofthevesselsbynumerousrepetitions(oftenhundreds)ofexperiments.Thesurfacesinteractwiththeexperimentsboththermallyandchemically.Thermalinteractionin-volvestransferofthereactionenthalpythroughthewallsofthevessel.Forthepresentwork,thiseffectwastreatedbyassumingalumpedheatcapacitymodel,withanoverallcharacteristicthermaltransferrate.Thenumericalvaluesusedinthecalculationswereinitiallyestimatedbasedonthethermaldiffusiv-ityofthemajorspeciesandthereporteddimensionsofthereactionvessels.Thedetailsofthechemical/surfaceinteractionsarediscussedbelow.Thefivestaticreactorexperimentsselectedforsim-ulationaresummarizedinTableIII.Thetemperaturesstudiedrangefrom633to873K,andtheinitialpres-suresareatmosphericorbelow.ThereactorsurfacesintheexperimentswereuncoatedPyrexorsilica.FortandHinshelwood[41]followedtheextentofreactionbymonitoringthepressureriseduetothedecreasingaveragemolecularweightofthereactingmixture.Thecharacteristicthermaltimeoftheirreactorvesselwasmuchshorterthanthereactiontimescale,andnearlyisothermalconditionsweremaintained.ThiswasalsothecaseintheexperimentsofBoneandGardner[42],andBellandTipper[43].TheprimarydiagnosticinBoneandGardner’sexperimentwasalsopressurerise,althoughlimitedspeciesanalysiswasperformed.Car-bonmonoxidewastheprimaryproductdetected,withsmalleramountsofformaldehyde,formicacid,carbondioxide,andanunidentifiedperoxide.Themorede-tailedspeciesmeasurementsofBellandTipper[43]wereinagreementwithBoneandGardner’sresults,althoughinaddition,hydrogenandwaterwerealsomeasured,andtheperoxideidentifiedashydrogenper-oxide.Areactionschemeforlow-temperaturemeth-anoloxidation,involvingHO2astheprimarychaincarrier,wasalsoproposed.Cathonnetetal.[44]stud-iedmethanoloxidationathighertemperaturesasafunctionofequivalenceratio.Temporalspeciespro-filesweremeasuredbygaschromatography,andade-tailedreactionmechanismwasproposed,whichrepro-ducedtheirmeasurementswithareasonabledegreeofaccuracy.Finally,inthecontinuously-stirredstaticreactorstudyofAniolekandWilk[45],overallreac-tionrateasafunctionoftemperature,pressure,andequivalenceratiowasmeasuredbypressurerise.Asinglesetofspeciesmeasurementswasalsore-ported.Ignitioneventswereobservedatfuel-richcon-ditions.Initialmodelingattemptsusingthebaselinemech-anismresultedinpooragreementwithalltheexperi-mentaldata,withthecalculatedreactiontimescalesbeingmuchshorterthanmeasured.Becausethemech-anismsuccessfullysimulatedhigh-pressureflow-re-actorexperimentsattemperaturesnearingthoseinthestaticreactorstudies[21],threepossibilitieswerecon-sidered.First,thelow-pressureconditionsofthestatic-reactorexperimentsmayopennewreactionpathwaysnotconsideredinthemechanism,orapressure-depen-dentratecoefficientmaynotbeadequatelyspecified.However,thelowerpressuresofthestaticreactorex-perimentsdonotfavorgas-phasechainterminationreactions,asrequiredtocauseanoveralldecreaseinreactionrate.Also,detailedfalloffexpressionshavebeenincorporatedforallrelevantreactions.Second,theoverallreactionratemeasuredintheflowreactorexperimentsmaybesystematicallytoohigh.However,flowreactorexperimentsaregenerallylessaffectedbysurfacereactionsthanarestaticreactorexperiments,particularlyiffluidelementresidencetimesintheflowreactoraremuchshorterthandiffusiontimesto/fromthereactorwalls.Thus,heterogeneousreactionsmayplayanimportantroleinmodifyingthechemicalrateobservedinstaticreactorsfromthatcharacteristicofhomogeneousgas-phaseconditions.Thetwomainreactionsgenerallyassumedtoberesponsibleforthechemicaleffectofsurfacesarelossofhydrogenperoxideandhydroperoxylradicalthroughheterogeneoustermination[43,46].Thesere-actionsweremodeledbytheoverallprocess:HO2(+wall)一H2+O2H2O2(+wall)一→H2OO2详细内容请见附件免责声明：本页面/内容部分素材来源于互联网公开信息，旨在传递更多信息，不代表本平台立场。版权归原作者或机构所有，如涉及侵权，请通过平台联系我们，我们将在核实后第一时间处理。本平台对转载内容的真实性、准确性不作任何保证，用户需自行判断并承担使用风险。
