fsn_ 1127-wolfgang kunther -2. hydration of portland cement - fp

1、ThermodynamicmodelingofPtlcement–metakaolin–limestoneblendsWolfgangKunther一ZhuoDaiJrgenSkibstedAarhusUniversityDepartmentofChemistryInterdisciplinaryNanoscienceCenter(iNANO)Langelsgade140DK8000AarhusCDenmarkAbstractThepa

2、rtialreplacementofPtlcementbydifferentsupplementarycementitiousmaterials(SCM’s)hasbeeninvestigatedextensivelyinrecentyearswiththeaimofreducingenergyconsumptiontheCO2footprintofconcretebasedonPtlcementbinders.Metakaolin(M

3、K)isanidealcompoundtomodelaluminosilicaterichSCM’sincementblendsduetoitshighreactivitycomparedtootherSCM’sthegeneralavailabilityofkaoliniteasapurestructurallywelldefinedclay.Ontheotherhlimestone(LS)isacommonlyusedconstit

4、uentincommercialcementswhereitsbeneficialeffectsarewellknownfafewpercentlimestoneadditions.InthiswkwehaveemployedafixedPtlcementsubstitutionlevel(35wt.%)crespondingtothemaximumallowedamountofSCMaccdingtotheEuropeancement

5、stardEN1971:2000investigatedtheeffectofvariationsoftheMKLSratioonthephasesfmedexperimentallybymodeling.PowderXRD27Al29SiMASNMRspectroscopyareusedtoacterizethehydrationofthePtlcement–metakaolin–limestoneblends.29SiNMRprov

6、idesdatafboththeamphouscrystallinephasesallowingdeterminationofthedegreeofhydrationfalitebelitetheamphousMKphase.Theacquireddataisusedasinputfthermodynamicmodelingstudiesassumingthatthedissolutionoftheindividualphasesist

7、heratelimitingstepfphasefmation.ThehydrationofthesilicatecontainingclinkerphasesMKismodeledbyanexponentialrelationshipwhichemploysvariablesthatdescribethequantityofreactedmaterialanempiricaldissolutionrateparameterthehyd

8、rationtime.Theseparametersareoptimizedbasedontheexperimentaldata.ThephaseassemblagesofthehydratedblendschangefveryhighMKcontentsfromaCASHptlitecalcitemonocarbonateettringitesystemtoaphaseassemblagethatinadditioncontainss

9、trtlingiteotherAFmphases.TheposityispredictedtodecreasewithincreasingMKadditionsinceettringiteisstabilizedthemonocarbonatecontentincreases.MeoverthequantityofCSHispredictedtoincreasetheCaSiratiooftheCSHphasestodecreasewi

10、thincreasingMKcontentwhichissupptedexperimentallybytheobservationoflongeraveragealuminosilicatechainlengthsftheCASHphaseby29SiMASNMR.TheimplementationofhydrationkiicsfapurewPc–MKsubstitutionserieswithoutanylimestoneaddit

11、ion(notapartofthepresentwk)hasrelativelysmalleffectsonthepredictedhydratephaseassemblage.Incomparisonthelimestoneadditionhasaverystronginfluenceonthehydratephaseassemblagethemaindifferencebeingthatthefmationofstrtlingite

12、islargelysuppressedsinceincreasingamountsofmonocarbonatearepredictedfthewPc–MK–LSblends.Thisisconfirmedbyexperimentaldatawhereasitcouldnotbepredictedbythermodynamicequilibriumcalculationsalone.Overalltheimplementationoft

13、hehydrationkiicsreproducestheexperimentalresultsinaconvincingmanner.iginalitySolidstateNMRspectroscopyhasgainedincreasingattentioninthecementcommunityduringthelasttwodecadesmainlybecauseNMRtechniquesallowacterizationofXr

14、ayamphousconstituentssuchasCSHCASHphases.InthiswkNMRisutilizedtodeterminethedegreeofhydrationfthecalciumsilicateclinkerphasesmetakaolin.Thesedataareusedfthefirsttimeasinputtothermodynamiccalculationsallowingmodelingofhyd

15、rationkiicsinadditiontothefinalequilibratedstate.Althoughtheexponentialequationusedfthehydrationkiicsisempiricalinnatureithastheadvantagethateachparametercanbedirectlyrelatedtothedegreeofhydrationthereactionrates.Thusitm

16、ayalsobeusedtodescribethehydrationofcementitiousbindersbasedonhydrationdegreesdeterminedbyothermethods.Ptlcement–calcinedclaybindersareofhighimptancetofuturecementproductionwithareducedCO2footprintwheretheapplicationofca

17、lcinedclayshastheadvantagethatlargeclaydepositsarespreadacrosstheglobetheiractivationrequireslessthermalenergycomparedtoPtlclinkerfmation.Keywds:calcinedclaylimestonethermodynamicmodelinghydrationkiicsNMR.Crespondingauth

18、:Tel:4587155916Fax:4586196199usingahomebuiltCPMASNMRprobef7mmouterdiameterzirconia(PSZ)rotsaspinningspeedof?R=6.0kHzsinglepulseexcitationwithapulsewidthof3.0?sfanrffieldstrengthof?B12?=40kHzarelaxationdelayof30stypically

19、2048scans.The29Sichemicalshiftsarereferencedtoneattetramethylsilane(TMS)usingamineralsampleof?–Ca2SiO3(larniteScawtHillN.Irel?(29Si)=–71.33ppm)asasecondaryreference.Thedeconvolutionsofthe29SiMASspectraareperfmedutilizing

20、theVarianVnmrsoftwaredeconvolutionapproachesdescribedelsewhere(ersenM.D.etal.2003PoulsenS.L.etal.2009DaiZ.etal.2014).The27AlMASNMRspectrawereacquiredonaVarianDirectDriveVNMR–600spectrometerusingahomebuiltCPMASprobef4mmou

21、terdiameterPSZrotsaspinningspeedof?R=13.0kHzapulsewidthof0.5?sfanrffieldstrengthof?B12?=60kHzarelaxationdelayof2stypically6.200scans.Thespectra27Alchemicalshiftsarerelativetoa1.0MaqueoussolutionofAlCl36H2O.2.3XRDexperime

22、ntsPowderXRDwasusedtoidentifythephaseassemblageswithspecialfocusonthedistinctionbetweendifferentAFmphases(e.g.monosulfatehemicarbonatemonocarbonate).ThemeasurementswereacquiredonaBrukerD8advanceddiffractometerequippedwit

23、haLynxEyepositionsensitivedetect(PSD)employingCuKαradiation(λ=1.5418?).2.4ThermodynamicmodelingThermodynamiccalculationswerecarriedoutusingthegeochemicalcodeGEMS3(KulikD.etal.2013WagnerT.etal.2012).Thedefaultdatabasesare

24、expedwiththeCEMDATA07database(LothenbachB.etal.2008)whichcontainssolubilityproductsofsolidsrelevantfcementitioussystems.ThedatasetincludesthermodynamicdataofcommoncementmineralssuchasAFt(ettringite)AFmphaseshydrotalciteh

25、ydrogarsCSHphases.Onlythefmationofsiliceoushydrogar(C3AS0.8H4.4)thaumasitearesuppressedwhichfmextremelyslowlyatambienttemperature.Aferriteclinkerphasewerenotconsideredinthemodelingbecauseoftheverylowironoxidecontentofthe

26、wPc(0.24wt.%Fe2O3)whichimpliesthatnearlyallironareincpatedasguestionsinthealitebelitetricalciumaluminate(C3A)phasesofthewPcclinker(TaylH.F.W.1997).Thedegreeofhydrationfthetricalciumaluminateclinkerphaseisnotquantifiedsin

27、ce27AlMASNMRrevealsthatthisphasehascompletelyreactedafterafewdaysofhydration.ThusfullreactionftheC3Aclinkerphasewasassumedinthemodelingatalltimes.Furthermodelingdetailsadditionaldiscussionsaredescribedinarelatedwk(Kunthe

28、rW.etal.inpreparation)thatprimarilyfocusesontheimplementationofhydrationkiicsinthemodelingapproach.3.ResultsDiscussion3.1.acterizationofthephaseassemblagesTable3summarizestheacteristicphasesthathavebeenobservedbyXRD27AlM

29、ASNMR.ThemostimptantinfmationfromtheseanalysesisthedistributionofAFmphasese.g.monosulfatehemicarbonatemonocarbonatestrtlingiteasarelatedphasewhicharefmedinsignificantquantitiesasaresultoftheconsiderableamountofaluminumre

30、leasedbythehydrationofMK.GenerallythedominatingphasesinthehydratedwPc–LS–MKblendsareCASHphasecalcitemonocarbonateettringite.PtliteisonlyexpectedftheblendswithsmallquantitiesofMKsincehighMKcontentswillconsumeallptlitebyth

31、epozzolanicreactionbetweenthesephases.HoweverthepresenceofsmallamountsofptliteisoftenobservedexperimentallyathighSCMsubstitutionlevelswhichmayreflecttheheterogeneityofcementitiousmaterials.ThequantitiesoftheAFmphasesincr