Bonding-in-Cementitious-Composites--How-Important-

3BONDINGINCEMENTITIOUSCOMPOSITES:HOWIMPORTANTISIT?SIDNEYMINDESSDepartmentofCivilEngineering,UniversityofBritishColumbia,2324MainMall,Vancouver,BritishColumbia,V6TIW5,CanadaABSTRACTAnumberofdifferenttypesofinterfacialbondingoccurinconcrete,including:i)bondingbetweenvariousphases(includinganhydrouscement)inthehydratedcementpaste(hcp)system;ii)bondingbetweencementandaggregates;iii)bondingbetweencementormortarandthefibresinfibre-reinforcedconcrete;andiv)bondingbetweenconcreteandsteelreinforcingbarsorprestressingcables.Theimportanceofthesetypesofbondswithrespecttothemechanicalbehaviourofconcreteisdiscussed.Itisconcludedthatinsomesystems,themechanicalpropertiesaregovernedprimarilybytheinterfacialbond;inothersystems,however,thisbondhasonlyasecondaryeffect.INTRODUCTIONFreshconcretemaycontain,inadditiontoportlandcement,waterandchemicaladmixtures,anumberofothermaterials:i)flyashorotherpozzolanicmaterials;ii)aggregates,withabroadsizedistributionandvariablemineralogicalcomposition;iii)steelreinforcingbarsorprestressingcables;andiv)discontinuousfibres,ofsteel,glass,ornaturalandsyntheticorganicmaterials.Therefore,inhardenedconcrete,agreatmanydifferenttypesofinterfacecanexist;i)interfacesbetweenthevariousphasesthatmakeuphydratedportlandcement(hcp),includingthatbetweenthehcpandtheunhydratedcement;ii)interfacesbetweenthehcpandtheunreactedpozzolanicmaterial;iii)interfacesbetweenthehcpandtheaggregate;iv)interfacesbetweentheconcrete(orhcp)andthesteelreinforcement;andv)interfacesbetweenthemortar(orhcp),andthediscontinuousfibresthatmaybepresent.Theinfluenceoftheseinterfacesonconcretepropertieswasnotedaslongagoas1905,whenSabin[1]observedthatsincecementmortarsareusuallyemployedtobindothermaterialstogether,itfollowsthattheadhesivestrengthisofthegreatestimportance.Whilethereisalmostuniversalagreementwiththisstatement,quantitativeinformationaboutthecontribu-tionoftheinterfacialbondtothepropertiesofconcreteisdifficulttofind,forthesamereasongivenbySabin[i]:OnaccountofthedifficultyMat.Res.Soc.Symp.Proc.Vol.114.1988MaterialsResearchSocietyofmakingtestsofadhesivestrength,however,thedataconcerningitareverymeager.Thus,whilethemorphologiesofthevarioustypesofinter-facesarereasonablywellunderstood,theircontributionstoconcretepropertiesarestilllargelyamatterofconjecture.Theelasticpropertiesofcementitiouscompositescanbepredictedreasonablywellfromsimplecompositemodels,suchasthoseshowninFig.l(a).Forplainconcrete,forinstance,thesefourmodelscanbeusedtorelatetheelasticmodulustothevolumefractionofaggregates,asshowninFig.l(b).Theupperboundrepresentsthecaseofperfectbondbetweenthecementandtheaggregate;thelowerboundrepresentsthecaseofnobond;forordinaryaggregatesandpastes,thesetwoextremesarenotverydifferent.However,itisvastlymoredifficulttodevelopsimilarequa-tionsforstrength,becausestrengthdependsnotonlyonthevolumefractionsofthecomponents,butalsoonthemechanismoffailure.Sincethereisgeneralagreementthatincementitiouscompositesitistheinter-facialregionsthataretheweakest,thenthewayinwhichthevariousphasesarebondedtogethermusthavealargeeffectonstrength.Hence,anymodelforstrengthmusttakeintoacountthewayinwhichmicrocracksdevelopintheinterfacialregion,andhowtheysubsequentlypropagatethroughthematerialtocausefailure.Unfortunately,westilldonotknowenoughaboutthedetailsofthesefailuremechanismstopermitsensiblestrengthmodelstobedeveloped.THENATUREOFTHEINTERFACIALREGIONSAllofthedifferenttypesofbondreferredtoabovemustresultfromsomecombinationofmechanicalinterlock,physicalbondinginvolvingvanderWaals'forcesand,insomecases,chemicalreactionsbetweenthephasesinquestionaswell.Therelativeimportanceofthesethreemechanismsisstillnotclear;itcertainlyvaries,dependingonwhichphasesarebeingconsidered.Thephysicalandchemicalnaturesofmostoftheinterfacesmentionedwillnotbedescribedhereinanydetail,sincetheyhavebeenreviewedextensivelyinrecentyears[2-6].Itisclearthatthemicrostructureoftheinterfacialzoneisusuallyprofoundlydifferentfromthatofthebulkcementpaste.Notonlyisthereadditional,well-orientedcalciumhydroxideattheinterface,buttheinterfacialregionoftenisalsomoreporousthanthebulkpaste,duetobleedingandothereffects.And,whileitisintuitivelyappealingtoassumethattheinterfacialregioninvolvesonlyarelativelysmallvolumeofthehcp,thisisnotthecase.Extensivestudies,usingbothSEMobservationsandx-raydiffraction,haveshownthattheinterfacialregionisabout40-50pmthick(thoughonlyaboutthefirst20pmhavesignificantlydifferentmechanicalpropertiesfromthoseofthebulkpaste).However,anexaminationofpolishedconcretesurfaces[7]revealsthatthemeanspacingbetweenaggregateparticlesisonly75-100pm,thoughofcoursethevariabilityislarge.Ifdiscontinuousfibresareintroduced,themeanspacingbetweeninclusionsisreducedevenfurther.Thus,witha40-50pmthickinterfacialzone,mostofthehcplieswithintheinterfacialzone,andonlyarelativelysmallvolumeofbulkhcpexists.Thereareatleastthreequestionsthatarise:1.Whateffectdothesevariousinterfaceshaveonthemechanicalproper-tiesofconcrete?2.Isitpossibletoaltertheinterfacesinsomecontrolledwaytoimprovethepropertiesofconcrete?