TribologyInternational100(2016)388–399ContentslistsavailableatScienceDirectTribologyInternationaljournalhomepage:www.elsevier.com/locate/tribointTribologicalassessmentoftheinterfaceinjectionmold/plasticpartN.Crisana,S.Descartesa,n,Y.Berthiera,J.Cavoreta,D.Baudb,F.MontalbanocaUnivLyon,CNRS,INSALyon,LaMCoS,UMR5259,F‐69621Villeurbanne,FranceCentreTechniquedelaPlasturgieetdesComposites,2RuePierre&MarieCurie,Bellignat01100,FrancecChristianDallozSunoptics(CDS),RoutedeGenèveBP155,Saint-Claude39206,FrancebarticleinfoArticlehistory:Received3October2015Receivedinrevisedform13April2016Accepted13April2016Availableonline25April2016Keywords:InjectionprocessMoldsurfaceThermoplasticpolymerRoughnessabstractOneofthecurrentchallengesoftheplasticinjectionprocessislinkedtotheimportancegiventoproductdesignthatenablesastrongdifferentiation.Thekeyindevelopingoptimalsurfacemoldsthatcanovercomethepresentdisadvantagesliesinthecomprehensionoftheinteractionsthatoccuratthemold/injectedpieceinterface.Thispaperfocusonidentificationandevaluationofthecontactconditionsatthatinterface,takingintoaccounttheeffectofthepolishing,ofthemoldgeometryandoftheinjectedmaterial.Acriticalcharacterizationofthesurfacetopographywasperformedtostudythecorrosion-mechanicalattackandthemechanical-physico-chemicaloneontwomolds.&2016ElsevierLtd.Allrightsreserved.1.IntroductionThesectorofplasticsprocessingisrelativelyyoungcomparedtocastiron,steelorglassindustry.Soitstillhasaverystrongdevelopmentpotential.Oneofthecurrentchallengesoftheplasticinjectionprocessislinkedtotheimportancegiventoproductdesignthatenablesastrongdifferentiation[1].Plasticpartswithanincreasedtechnicallevelofsurfaceaccuracyarerequiredintheareasofluxury,packaging,automotive,includingthemedicalandoptics.Theirdevelopmentinvolvestheimprovementofthefab-ricationprocess,andoneofthekeysliesinthemasteryofthesurfaceconditionsofthemolds.Injectionmoldingisacyclicprocess,characterizedby5phases:dosing,injection,packing,coolingandejection(Fig.1).Therawmaterialthatisdosedinthemachinemustbepureandconservedbeforeuseatanadequatetemperatureinordertobeasdriedaspossible.Thisisnecessarytoavoidcondensationinsidethemold.Theinjectionphaseischaracterizedbyhighflowratesandhencehighshearrate(tangentialeffect).Asthemoltenmaterialentersthemold,twoheattransfermechanismsoccur:convection(betweenthemeltedmaterialandmoldsurface)andviscousdissipation(duetotheeffectofinjectionspeedontheinjectedmaterialviscosity).Asthefillingiscomplete,themoldisupholdatapredeterminedpressureandsothepackingphaseisinitiated.Duringthisphase,themoltenpolymercontinuestobeintroducedintothemoldtocompensatefortheshrinkageofthealreadyninjectedmaterialasitcoolsdown.Afteraspecifictime,thecoolingphase(contrarytothecoolingstatewhichbeginsduringinjectionandpackingphase)oftheentireassemblystartsandsoalsothesolidificationprocessoftheplasticpart.Asthematerialsolidifiesandshrinksinthemold,thedominantheattransfermechanismisconduction.Whenthepartissufficientsolidified,itisejectedfromthemold.Duringthislastphase,anormaleffectcanbeattributedtotheejectionforceandadhesionphenomenacanoccurbetweenthemoldsurfaceandtheplasticpart[2].Despitetheundeniablediversityofconfigurationsavailable(intermsofcombination:moldmaterial,surfacefinishandprocessedmaterials),theproducersarefacedwithsimilardifficulties.Thus,thekeyshortcomingsthatstandout,moreorlesscombined,canbesummarizeasfollows:●thefoulingphenomenawhichrequirefrequentstopsforcleaning;●corrosionphenomenathatcangreatlylimitthelifetimeofthemoldcavityasafunctionofthetypeofinjectedpolymer;●problemsofstickingandreleasinginfunctionoftheinjectedmaterialsandsurfacequality;●problemsinkeepingthepolishingquality;●scratchesorshocksduringuseorstorage.Theavailableliteratureapproachesexperimentaland/ornumerical,variousaspectsregardingtheplasticinjectionmoldingprocess.Oneofthemisthefillingandflowbehaviorofmoltenpolymers.BociagaandJaruga[3]studiedtheformationofflow,weldandmeldlinesbydevelopinganewmethodofflowvisua-lization,whichcanprovehelpfulintheidentificationofweakCorrespondingauthor.E-mailaddress:sylvie.descartes@insa-lyon.fr(S.Descartes).http://dx.doi.org/10.1016/j.triboint.2016.04.0150301-679X/&2016ElsevierLtd.Allrightsreserved.N.Crisanetal./TribologyInternational100(2016)388–3993Fig.1.Plasticinjectionprocessphases[1].areasoninjectedparts.Alsotheeffectofpressureandcavitythicknesswereassessed.SametopicwastreatedbyOzdemiretal.[4],comparingthebehaviorofmoltenHDPE(highdensitypoly-ethylene)andPEexperimentallyandnumerically.Duringmolding,frictionforcesactfirstbetweenthemoldsurfaceandthemoltenpolymerandsecondwhentheplasticpartisejectedfromthemold.Bulletal.[5]adaptedtheASTMrubberwheelabrasiontesttosimulatetheconditionsofwearproducedbytheglassfilledpolymersonthebarrelsurfaceofaninjectionmoldingmachine.Variouscoatingsweretested,butunfortunatelytheytendedtohaveaweakperformanceonaccountofthetestconditions.Pouzadaetal.[6]developedaprototypeapparatustostudythefrictionpropertiesofmoldingthermoplasticsduringejectionphase.Themeasuredfrictioncoefficienthadatendencytoincreasewiththeroughness.Butwhentheroughnesswasreducedthefrictioncoefficientincreaseduetotherisingadhesionforceseffect.Thescanningelectronmicroscopyimagesofthemoldsur-faceandtheonesforthepolycarbonate(PC)andpolypropylene(PP)plasticparts,revealedaclearreplicationofthemoldsurfaceontheparts.Transientinnature,injectionmoldingprocessinvolvesnotonlyseveralheattransfermechanisms,phasechangeandtimevaryingboundaryconditions,butgoesfurtherinaddingtheeffectofmaterialpropertiesandgeometryoftheinjectedpart.Bendadaetal.[7]performedastudytoevaluatethenatureofthermalcontactbetweenpolymerandmoldthroughthedifferentphasesofatypicalinjectioncycle.Theirfindingsconcludedthatthethermalcontactresistancewasnotnegligible,notconstantwithtimeandwasstronglylinkedwiththeprocessconditions.Theexistingnumberofstudiesconcerningthephenomenapresentattheinterfacemoldsurface/polymerisrelativelylowtootherrelatedtopics.Besides,theydon’tfocusonstudyingthecurrentlimitationsoftheplasticinjectionprocessatamicroscopicscale,takingintoaccountvariousmacroscopicinfluences.Toovercomeplasticinjectionmoldingshortcomings,thecontactconditionsattheinterfacebetweenmoldsurfaceandplasticparthavetobeidentified.Thisworkfocusontheeffectofthepolishingquality,themoldgeometryandtheinjectedmaterialonthatinterface,bystudyingthecorrosion-mechanicalattackandthemechanical-physico-chemicalone.2.Methodandmaterials2.1.MaterialsFourpolymerswerechosentobeinjected:ionomerresin(E-MMASurlynsPC2000),styrene-acrylonitrileresin(SANTyril790),polyamidewith25%glassfibers(PA66GF25)andpoly-carbonate(PCMakrolonLQ27).Surlynsisacopolymerofethyleneandmethacrylicacidwheresomeoftheacidgroupsareneutralizedtoformthesodiumsalt.Theacidinthepolymergivespolarityandreducescrystallinity.Theionicbondingbetweenthepolymerchainsgivesoutstandingmeltstrength,toughnessandclarity.ThereasonofchoosingSurlynswasbasedontheexperi-enceofourindustrialpartners,whichfinditparticularlycorrosivedespiteitsgoodproperties.SANisalsoacopolymer,opticallytransparentandbrittleinmechanicalbehavior.It'sconsideredinthisstudyareferencematerial,usuallyusedincosmetics,luxuryandautomobiledomains.PA66GF25isanaliphaticpolyamide,reinforcedwith25%glassfibers.PA66hasanexcellentbalanceofstrength,ductilityandheatresistance.Theglassfibersexertanabrasiveeffectandthusaffectthemechanicalprotectionofthepolishing.PCiscomposedbycarbonategroups.Ithasahighimpact-resistance,lowscratch-resistanceandishighlytransparenttovisiblelight.Itisusuallyusedfortheproductionofeyewearlensesandexteriorautomotivecomponents.2.2.MoldsTwomolds,madeofhardenedsteel(52HRC)containing13%to15%ofChromium,withdifferentgeometrieswereused,onewithamirrorpolishedsurface(complexgeometry)(Fig.2a,b)andanotherwithanopticalpolishedsurface(simplegeometry)(Fig.2c,d)(Table1).Themoldhastwoparts:thestampandthematrix.Forthemoldwithcomplexgeometrythestampisof149Â119Â80mminsizeandthematrixof149Â119Â50mm.Incaseoftheonewithasimplegeometry,thestampisof50Â70mminsizeandthematrixhasacylinderformwithadiameterof70mm.Thesurfacefinishofthemirrorandopticalpolishedmoldsinvolvedapolishingclothanddiamondpaste.Furtherdetailsonthepolishingprocessareconfidential.ThemirrorpolishedmoldwasspeciallydesignedforthisstudybyTechnimold(amoldmaker)tohighlighttheroleofanglesandobstaclesintheformationofdefaults.Alsothemolddesigndidnotincludeaspecialfeaturethatcanevacuatetheair.Thiswasdoneintentionallyinordertosubmitthepolishedsurfacestoaggressiveconditions.Themoldingprocesswasperformedat“CenterTech-niquedelaPlasturgieetdesComposites”(IPC,France)ona50TEngelmachine.Theinjectionparameters,listedinTable2,werechoseninaccordancewithstandardspecificationsfortheinjectedpolymers.Basedonanumericalsimulationtheywereadaptedtorespondinconformitywiththemolddesign.Twoinjectioncam-paignswereconductedonthistypeofmold.Afterthefirstcam-paign,ontheplanepartofthemoldstamp,aninsertwithadia-meterof12mmandaheightof8mm,wasmountedtofacilitatethemorphologyassessment.FortheSurlynsinjection,3000partswereinjectedinthefirstcampaign.Aftersurfaceanalysis,themoldwassubmittedtotheindustrialcleaningoperation.Thesecondcampaignconsistedintheinjectionof3700moreparts.SANandPA66GF25wereinjectedonthesamemold.Duringthefirstcampaign,only8000SANpartswereinjected.Beforestartingasecondcampaign,themoldwaspolishedentirely.Thesecondcampaignconsistedintheinjection390N.Crisanetal./TribologyInternational100(2016)388–399Fig.2.Mirrorpolishedmold(complexgeometry):a)thestamp;b)thematrix.Opticalpolishedmold(simplegeometry):c)thestamp;andd)thematrix.Table1Configurationsmold-materials.InjectedmaterialMirrorpolishedmoldSANSurlynsPA66GF25PCxxxOpticalpolishedmoldxWhy?Table3InjectionParametersfortheopticalpolishedmolds.InjectionparametersInjectedpolymersSANReferencematerialCorrosiveAbrasiveOpticalapplicationsMetaltemperaturewhichheatsthematerial(°C)Injectionspeed(mm/s)Quantityofinjectedmaterial(cm3)Pressuretoavoidshrinkage(bar)Timetomaintainthepressuretoavoidshrinkage(s)Timetocooldown(s)Clampingfore(kN)280303055-308151300PC280303060-308151300xTable2InjectionParametersforthemirrorpolishedmolds.InjectionparametersInjectedpolymersSANSurlynsPA66GF25Metaltemperaturewhichheatsthepolymer(°C)Temperatureinsidemold(°C)Injectiondebit(cm3/s)Quantityofinjectedmaterial(cm3)Injectiontime(s)Pressuretoavoidshrinkage(bar)Timetomaintainthepressuretoavoidshrinkage(s)Timetocooldown(s)Clampingfore(T)Lengthofejectionpin(mm)Ejectionforce(kN)Ejectionspeed(mm/s)2403026.62153030423008020231.6141070302040processwasalsoperformedbyCDS,ona145TEngelmachine.TheinjectionparametersarelistedinTable3.2.3.MethodThesurfaceexpertizeconsistedintwomainsteps:themicro-scopyanalysisandtheinterferometrymeasurementsbeforeandafterinjectionprocess.Duetotheirlargedimensionsandelevatedmass,thesurfaceanalysisofthemirrorpolishedmoldswasper-formedusingaclassicopticalmicroscope.Fortheopticalpolishedone,thankstosmallerdimensions,themicroscopeanalysiscouldbecarriedoutusinganumericalopticalmicroscope(Keyence)andahighresolutionenvironmentalscanningelectronmicroscope(FEIXL30ESEM).Althoughtwoinjectioncampaignshavebeenperformed,theresultspresentedinthispaper,referonlytothesurfaceexpertizeperformedattheendofthesecondcampaign.Fortheinjectedplasticparts,onlytheinterfacebetweenmoldmatrixplanepartandplasticpartisdiscussedinthispaper.Inordertoidentifythechemicalcompositionofdifferentdepositsfoundonthemirrorpolishedmoldsurfaces,aFourierTransformInfrared(FTIR)spectrometerwasusedfortheanalysis.0.763.352075025.23130327032.5550of300partsofSAN.Theinsertwaschangedbeforestartingtheinjectionof12200PA66GF25parts.TheopticalpolishedmoldwasprovidedbyChristianDallozSunoptics(CDS)andwasusedtoevaluatethepolishingmechan-icalprotectionandfoulingphenomenonconsideredascomposingelementsofthemechanical-physico-chemicalattack.TheinjectionN.Crisanetal./TribologyInternational100(2016)388–399391Fig.3.Interferometrymeasurementslocalizationforthemirrorpolishedmold.(Forinterpretationofthereferencestocolorinthisfigure,thereaderisreferredtothewebversionofthisarticle.)Theinfraredtechniquecouldnotbeappliedfortheopticalpolishedmold,becausethepossiblepresentlayersanddepositaretoothintobedetectedwiththismethod.Theinterferometrymeasurementsweremadeusinga3Dnon-contactopticalprofiler(Sensofar)whichcombinestheconfocalandinterferometrytechniques.Twoobjectiveswereused,oneÂ10withanacquisitionareaof1200Â800mm²andÂ50withanacquisitionareaof250Â150mm².TheacquireddatawasprocessedusingMoun-tainsMapUniversalsoftware.Thefollowingheightparametersfor3Darealsurface(ISO25178)wereconsideredinthisstudy:thearithme-ticalmeanheightofthesurface(Sa),therootmeansquareheightofthesurface(Sq),themaximumpeak-to-valleydistance(Sz),theheightdistribution(Ssk)andtheflatnessoftheheightdistribution(Sku)[8].Bigerelleetal.[9]developedanoriginalmethodologyappliedtoagrainedmoldsurface,toanalyzetheinfluenceofinjectionparametersontheroughnessofinjectedplasticparts,byfindingtheevaluationlengthonwhichclassicalparameterscanbeesti-mated.Thisworkraisedthequestionoffindingtherightcut-offfiltertoensurethattheheightparametersvaluesareduetotheroughnessofthesurfaceandnotofthewaviness.Inthisstudyafivedegreepolynomialwasappliedtotheacquireddatatoremovetheform.Bycomparingthevaluesoftheheightparameters:Sa,Sq,Szatdifferentscalesofthesamearea,itwasobservedthatvaluesareidenticalorveryclose.Soifistakenintoaccountthatthelevelofpolishingisveryhigh(verysmoothsurface)andthevariationofthevaluesisinsignificant,itbecanconcludedthatthevaluesoftheheightparameterspresentedinthispaperaretrulyrelatedtotheroughnessofthesurfaceandnotofthewaviness.AstudytakingintoaccountthemethodologydevelopedbyBigerelle[9]andVanGorp[10]couldbeenvisagedforthefutureinordertoapplyitonaverysmoothsurfaces.TheinterferometrymeasurementsonthemoldusedfortheSurlynsinjectionwereperformedasshowninFig.3(redcolor).23measurementswereperformedincaseofthestampand17measurementsforthematrix.ThelocationofthemeasuredareasforthemoldusedforSANandPA66GF25injectionisalsoshowninFig.3(bluecolor).Inthiscase,31measurementswereconsideredforeachmoldcomponent(Fig.4).Thedifferentdefaultsanddepositspresentonthesurfaceswerenotconsiderforthecalculation,theyweremasked.Thevalueslis-tedinTables4and5representtheaverageofthemeasuredvalues.3.Resultsanddiscussions3.1.Mirrorpolishedmold3.1.1.InjectionSurlynsAllalongthestampplanepart,depositsdifferentintextureandconsistencecanbeobserved(Fig.5).TheirlocationandmorphologyFig.4.Interferometrymeasurementslocalizationfortheopticalpolishedmold.seemtoindicatetheflowdirectionofthemoltenpolymer.Alsoitcanbenoticed,towardstheendoftheflow,thedepositsgrowintermsofthicknessandoccupiedsurface.ThetypeofdepositobservedinFig.5eandfisalsoobservedafterthefirstinjectioncampaign(3000injectedparts),andappearedthatthecleaningoperationhasbeenabletoremoveit,butformedagainduringthesecondinjectioncampaign(3700injectedparts).Thisparticulardepositislocatedbetweentheextremityoftheovalbumpandtheholewhereoneoftheejectionpinsacts.Alsointhislocationtheflowchangesdirection,morepreciselymakesaleftturn;factalsorevealedbythedepositmorphology.Itsexistencecanbeexplainedstartingwiththeeffectoftheinjectionspeedonthemoltenpolymerviscosity,whichisconsideredtobeaheattransfermechanismthatoccursduringtheinjectionprocess.Duetothegeometryfactor,theviscousdis-sipationcreatesatemperaturegradientwhichsensitizesthisarea.Duringthepackingphase,asthemoldcontinuestobefilled,thelocationidentifiedisoneofthelasttobereachedbythemoltenpolymer.Astheholdingphasebeginsandwithitthesolidification,thetemperaturegradientthatappearsintheinjectionphasecontinuestoactandbydoingsoitdelaysthesolidificationinthisarea.Whentheestablishedtimefortheholdingphaseexpires,themechanismofejectionissetinmotion.Theejectionpinisclosetotheidentifiedlocationandasitwasaffectedbythetemperaturegradientandhasnotyetbeenentirelysolidifies,itwillalsobethefirstareatobeseparatedfromthemoldsurface.Allthesecanexplaintheappearanceoftheadhesionphenomenon.InFig.5d,thedepositappearslikeathinfilmandisalsolocatedinanareawheretheflowchangesdirection.Itcouldalsobejus-tifiedbythetemperaturegradient,butitsaspectandcompositionsuggestthatmayanotherphenomenacanoccur.Theinfraredanalysisperformedonthisarea(Fig.6c)suggestthatonlysomeofthewavenumbersmatchwiththeonesfromthespectrumregis-teredfortheinjectedpart(Fig.6a).Itispossiblethatthegasesreleasedfromthecontactofthemoltenpolymerwiththemoldsurfacereactedwiththeadditivesfromtherawmaterialcompo-sitionandfacilitatedtheseparationofthethinlayerthatstickon392N.Crisanetal./TribologyInternational100(2016)388–399Table4ThevaluesofroughnessparametersforthemoldusedfortheSurlynsinjection.Parameters(ObjectiveÂ10)MoldbeforeinjectionStampSa(nm)Sz(nm)Sq(nm)SskSku5.4470.8663.96713.736.8671.15À0.070.183.1470.36Matrix5.5271.3966.6971.796.9971.79À0.15670.123,2970.38SurlynsinjectionStamp8.71.23113.8875.7410.8871.56À0.01770.143.2170.55Matrix6.8871.1580.66721.58.6671.44À0.04170.143.2370.38Table5ThevaluesofroughnessparametersforthemoldusedfortheSANandPA66GF25injections.Paramètres(ObjectiveÂ10)MoldbeforeinjectionStampSa(nm)Sz(nm)Sq(nm)SskSku5.4470.8663.96713.736.8671.15À0.070.183.1470.36Matrix5.5271.3966.6971.796.9971.79À0.15670.123,2970.38SANandPA66GF25injectionStamp5.9371,765.52718.77.4772.150.000770.143.1370.32Matrix11.0273.18124.99742.7713.8373.99À0.0570.243.2470.44Fig.5.MoldstampusedfortheSurlynsinjection–varioustypeofdepositsfoundontheplanepart.themoldsurface.Alsothe“scraped”aspectofthisdepositindi-catesthatismorelikelythatthistypeofdeposithasformedduringtheinjectionphase.Themoldmatrixsurfacepresentssmall“islands”depositsthatfollowtheflowdirection(Fig.7a).ThisaspectisinconcordancewithwhatitcanbeobservedontheplasticpartsurfaceintheN.Crisanetal./TribologyInternational100(2016)388–399393Fig.6.MoldstampinjectedwithSurlyns–registeredinfraredspectrumfor:a)theinjectedpart–finalproduct;b)depositidentifiedonthemoldsurface(Fig.5e);c)depositidentifiedonthemoldsurface(Fig.5d).samearea,shownatamagnifiedscaleinFig.8a.Thistypeofdepositduetothemoltenpolymerviscosityhasledtothefor-mationofburrs(excessmaterialinthinlayer)possibleduringtheejectionphase.Holes(from14.6nmto404nmdeep)areobservedbeforeinjectionprobablyduetopolishing.Theirmorphologyevolvesduringinjectionprocess:theholesexpandinoccupationareaanddepth(39.7nmto877nm).InFig.7bandc,thepointingredarrowsindicatethepresenceoftheevolvedholes.Theyexhibittwotypesofmorphology.ThefirsttypeillustratedinFig.7bshowsverysmallholesfocusedaltogetherinsmallerorlargerspotsandthesecondtypeillustratedinFig.7cpresentsaholesurroundedbya“cloud”ofsmallholes.Duetotheinclusionsinthebulkmaterial,grainsdislocationcouldoccurcausingtheformationofholesduringpolishingpro-cess.Thoseholesaremodifiedintermofdepthandareaduringinjectionprocess.Asreportedin[11],stresscorrosioncrackingcanaffectthemolds,startingatamicroscopiclevelandrevealingitselfascrack.Theprimarycausalelementsarethemetallurgyofsteel,thepresenceofchlorineinthewaterusedinthecoolinglinesofthemoldandthestressesonthetoolduringmolding.ItisknownthatChromiumgivesthesteelcorrosionresistance,byprovidingaprotectiveoxidelayer.Thusitispossiblethatduetothepolishingdefects(holes),thethicknessofthislayeriscompromisedandthuswhenahighviscouscorrosivepolymer,likeSurlyns,isinjected,theareasaffectedbyholes,aresubmittedtocorrosionattack.ThefactthatthefeaturetoevacuatetheairwasexcludedfromthemolddesigninconjunctionwiththecorrosionnatureofSurlyns(basedontheexperienceofindustrialprojectpartners),cancreateanaggressiveenvironmentatthemold/moltenpolymerinterfaceduetothegasesrelease.ThehighviscosityofSurlynsanditscapabilitytostickontothemoldsurfacealsoplaysaroleintermsofexertingamechanical-physico-chemicalattackontheareawherethedefaultsarelocated.Allthesestatementsallowtocatalogthisdefaultascorrosionpit.AscanbeseeninFig.8bandc,thesecorrosionpitshavealsoaneffectontheplasticpartsurface.Theirinfluenceismanifestedbytheformationofaccentuatedburrs(Fig.8b)oraccumulationofdebrisorfragmentspresentedinsameformthatcorrosionpitsexhibitonthemoldsurface.Althoughthedamagemechanismexhibitedbythestampisdifferentthantheoneforthematrix,somedepositscanbefoundonthesurface.Neartheendofinjection,wheretheflowchangedirection,smalldepositscanalsobeobserved(Fig.7d).Theirlayingoutindicatetheflowdirection.Finescratchescanbeobservedintheareawheretheflowchangesdirection,butatacloserlook,infacttheyaretracesofdeposits.Asignificantdepositcanbeobservedallalongtheareawherethetwomeltingpolymerflowsencounter(Fig.7e).Theinfraredanalysisrevealthatthisdeposithasadifferentcompositionthantheinjectedpolymer.Theplasticpartsurfacepresentsaweldline(Fig.7e),onthecorrespondingarea.Depositsarealsofoundontopofcorrosionpits,inachangingflowdirectionarea(Fig.7f).Duetothelimitationscausedbythedimen-sionsofthematrix,theoptimalconditionsforanappropriateinfraredanalysiswerenotpossible.Butoptically,thetextureofthesedepositsareverysimilarwiththeonefoundonthestamp(Fig.5e).TheinitialroughnessparametersforthetwopartsofthemoldarelistedinfirsttwocolumnsofTable4.TakingintoaccounttheinitialvaluesofSkewnessparameter(Ssk)andKurtosis(Sku),itcanbeconcludedthatthemoldsurfacehasaslighttendencytohavemorevalleysthanpeaks(Ssko0)andtheflatnessofthehighdis-tributioniswide(Sku43),sothesurfaceisratherplane.ItcanbeobservedalsofromTable4,thatthevaluesofSa,SzandSqarehigherafterthemoldhasbeensubmittedtotheinjectionprocess.Althoughinitially,thereisnosignificantdifferencebetweenthevaluesofSa($0.08nm)andSq($0.13nm)forthestampandthematrix,aslightmoresignificantoneisobservedafterinjection($1.76nmforSaand$2.2nm).Thisdifferenceconsistsinhighervaluesoftheroughnessparametersafterinjectionforthestamp.Thepeak-to-valleyheightrepresentedbySzhasahighervalueafterinjectionincaseofthestamp.Thissupportsthestatementthatthestampandthematrixpresentdifferentdamagemechanism.394N.Crisanetal./TribologyInternational100(2016)388–399Fig.7.MoldmatrixusedfortheSurlynsinjection–identifieddamagedareasontheplanepart.Fig.8.InjectedSurlynspart–identifieddamagedareasontheplanepart.N.Crisanetal./TribologyInternational100(2016)388–399395Fig.9.Surlynsinjection–arithmeticmeanroughness(Sa)variationalongtheflowdirectionforthe:a)stamp;andb)matrix.Althoughduringdataprocessingthedefaultsanddepositsaremasked,itappearsthatthesticking-releasingphenomena(foundonthestamp)affectsmoretheintegrityofthesurfacethanthecorrosionpits(foundonthematrix).Alltheroughnessparametersevolvedalongtheflowdirection.ThevalueofSaparameterwaschosentohighlightthissurfaceevolutionalongtheflowdirection.TheSavalueconsideredtorepresentthisvariationistheaverageofthemeasuredvalueslocatedonalineperpendicularontheflowdirection.Forbothstampandmatrix,thelowestSavaluecanbefoundatthebeginningofflowontheplanepart(C1onthegraphsinFig.9).Forthematrix,thisvalueisnotmuchhigherthantheonemeasuredbeforeinjection.IftheSavariationisconsideredonthelinesofobservations,itcanbeobservedforthestampthattheroughnesshasthetendencytoincreaseontheflowdirection(Fig.9a).Alsointhecaseofthestamp,theSavaluealsovariesinfunctionofobservationcolumn.Ithasthetendencytoincreaseanddecreasefromonemeasuredareatothenext.Ifthemicroscopyanalysisperformedonthestampiscorrelatedwiththeroughnessvaria-tions,thepresenceofalotofdepositswithvariousthicknessjustifiestheincreaseofroughnesstowardstheendofflow.Incaseofthematrix,theSavalueincreasesalongthecenteredobservationlines(L3,L4,L5,L6–Fig.9b)towardstheendofflow.Butitdecreaseontheobservationlinessituatedontheedge.ThevariationofSavalueontheobservationcolumnsC3andC4(Fig.9b)isverypronounced,asSavaluedecreasesorincreasesdramaticallyfromameasuredareatothenext.Apparentlythematrixsurfaceflatnessisseverelycompromised,butitisnotaninjectionconsequence,itisduetothepolishingprocessthathasproventobechallengingwhenitcomestohighsurfacefinishingoncomplexgeometries.3.1.2.InjectionSANandPA66GF25SANandPA66GF25polymerswereinjectedsuccessivelyonthesamemold.ThefirstpolymertobeinjectedwasSAN.Whentheinjectionfinished,theinsertmountedonthestampwaschangedwithanewoneandtheinjectionofPA66GF25began.Nosup-plementarycleaningoperationswereperformed.ThemicroscopyanalysispresentedbelowwasperformedafterthePA66GF25injection,andsoonlytheplasticpartmadeofthispolymerwasconsideredfortheresultspresentedinthispaper.Finescratchescanbeobservedonthestampsurfaceatbeginningoftheplanepart(Fig.10a).Thesescratchesbecomedenserandpronouncedintheareastowardstheendofinjection,wheretheflowchangedirection.OvalholesliketheoneinFig.10carescatteredalloverthemoldsurface.Initiallyformedduringpolishing,itseemsthattheywereenlargedonaperpendiculardirectiontotheinjectionflow,bytheabrasiveactionoftheglassfibers.Theinsertandtherestofthestampsurfacepresentdifferentmorphologyintermsthatontheinsertthescratchesaremorehighlighted(Fig.10b).TheinterferometrymeasurementsconfirmaslightdifferencefortheSavalue($0.76nm).ItispossiblethataverythinlayerofpolymerremainedaftertheSANinjection,andsothesurfacewassomehowprotectedagainsttheglassfibersaction.Thearea,wherethetwoinjectionflowsencounter,presentsburnmarksandalotofsmalldepositslike\"drops\"(Fig.10d).Theinfraredanalysisrevealthatthesedepositsarenotsimilarincompositionoftheinjectedpolymers.Thematrixexhibitsalsoelongatedholesproducedbytheglassfibers(Fig.11a).Onlyasmallpolymerdeposit,locatedonahole,wasfoundonmatrixsurface(Fig.11b).Thetypeofdefaultpre-sentedinFig.11c,isalsoobservedbeforeinjectionsandsotheirpresencecanonlybecatalogedasapolishingdefault.Unfortu-nately,thisdefaultcanaffectthequalityoftheinjectedpartintermsofestheticalaspect.Fortheexample,thedefaultinFig.11ccanbeobservedontheplasticpartinFig.12b.Theeffectofmoldgeometrycombinedwiththeinjectedmaterialeffectmanifeststhroughthedamageeffectofglassfibers,andismorevisibleintheareaneartheovalbump.Burnmarksanddepositslike\"drops\"arealsofoundinthesameareathaninthecaseofthestamp(Fig.11d).Alsotheplasticpartpresentsburnsmarksinthesamearea(Fig.12c).Thiscanbeduetothelackinthemolddesignofthefeaturethatwouldhavepermittedairevacuation.Butitcanalsobecausedbytheabrasiveeffectexertedbytheglassfibersinconjunctionwithphysico-chemicalreactionsatthemoldsurfaceasperitieslevel.ThePA66GF25partsurfacepresentsblackareas,thatseemtobeburned(Fig.12a).Smallvoidpocketscanalsobeidentifiedontheplasticpartsurface(Fig.12d),probablyduetodifferentsoli-dificationtimesand/orthewatervaporsthathavemanagedtoinfiltratethemold.InTable5,theaveragevaluesoftheconsideredroughnessparametersarelistedandthesameinitialroughnessparametersareconsidered.Thevaluesofroughnessparameters(Sa,SzandSq)measuredafterinjectionarehigherinthecaseofthematrix.Thedifferencebetweenis$5.5nmforSa,$6.36nmforSqand$5947nmforSz.Forthestamp,thevaluesofSa,SqandSzareveryclosetotheonesbeforeinjection.ThiscanconfirmthesuspicionofathinlayerremainingonthesurfaceaftertheinjectionofSAN.InthiscasealsotheSaparameterwaschosentoanalyzesurfacemodificationalongflowdirection.Forthestamp,theSavalueincreasesalongthecenteredobservationlines(L4,L5,L7,L8–Fig.13a)towardstheendofflow.Butalongtheobservationlines396N.Crisanetal./TribologyInternational100(2016)388–399Fig.10.MoldstampusedfortheSANandPA66GF25injections–identifieddamagedareasontheplanepart.Fig.11.MoldmatrixusedfortheSANandPA66GF25injections–identifieddamagedareasontheplanepart.locatedattheedge,theSavaluedecreasesintheflowdirection.ThevariationofSavaluealongtheobservationcolumnsC2andC3,isconstantforthemajorityofmeasuredareas.Buttowardsthetopborder(thetopedgeasviewedinFig.3),theSavalueincreasedconsiderably(forexample,forC3increasesfromavalueof3.96nmto11.1nm).ThevariationofSavalueforthematrix,hasthesamebehaviorastheoneforthestamp.ButalongtheobservationcolumnsC4andC2(Fig.13b),thevalueofSaincreasesanddecreasessig-nificantlyfromonemeasuredlocationtothenext.ThehighestSavalue(about14nm)canbefoundintheareaaroundthepinejectionholes.LikeinthecaseofthemoldusedforSurlynsinjection,thesurfaceflatnessofthetwomoldcomponentsisalsoaffected.3.1.3.ComparisonbetweenSurlynsinjectionandSAN/PA66GF25injectionThemoldsurfaceinjectedwithSurlynsisgovernedbyadhe-sionandcorrosionpits.InsteadforthemoldinjectedwithSANandPA66GF25,althoughscarcedepositsmaybefound,polishingdefaultsstandoutthemost.ThetypeofpolymerdepositsfoundontheplanepartofthestampincaseoftheSurlynsinjectionisobservedneitherinthecaseoftheSANinjection,norPA66GF25injection.Anexplanationisthatthechangeofflowdirectioninducesaneffectontheinjectionspeed.Buttheinjectionspeedcanbelinkedwiththeinjectedmaterialeffect,morepreciselytheviscosityofthemoltenpolymerexertsathermictransferbetweenthemoltenpolymerandthemoldsurface.Theholesfoundonthemoldsurface,differintermsofmor-phologyasafunctionofinjectedmaterial.FortheSurlyns,theN.Crisanetal./TribologyInternational100(2016)388–399397Fig.12.InjectedPA66GF25part–identifieddamagedareasontheplanepart.Fig.13.SANandPA66GF25injections–arithmeticmeanroughness(Sa)variationalongflowdirectionforthe:a)stamp;andb)matrix.holesaremoreorlessround,surroundedbyothersmallones.Meanwhile,theonesforthePA66GF25injection,areelongatedonaperpendiculardirectiontotheflowone.Theinterferometrymeasurementshasbeenusefultodeter-mineaclassificationforthetwomoldpartsinfunctionoftheSavalue.Forthestamp,theaverageSavaluebeforeinjectionissmallertheonesusedfortheSAN/PA66GF25andSurlynsinjec-tions.ButthemoldstampinjectedwithSurlynshasahigheraverageSavaluethantheoneusedfortheinjectionofSANandPA66GF25.Alsoincaseofthemoldmatrix,theaveragemeasuredSavaluebeforeinjectionremainssmallerthantheonesobtainedfortheinjectedmolds.ButthemoldmatrixinjectedwithSurlynshasahigherSavaluethantheoneusedfortheSANandPA66GF25injections.3.2.OpticalpolishedmoldTheopticalpolishedmoldsurfacesandtheinjectedplasticpartwereanalyzedbyinterferometryandopticalmicroscopy.Theplasticpartswereanalyzedonlyonthesurfacethatcomesincontactwiththemoldstampforbothinjectedpolymers(SANandPC).Duetoitssmallersize,thestampcouldbeanalyzedwiththehighresolutionenvironmentalscanningelectronmicroscope(ESEM).AsrevealedbyESEManalysis,themoldstampsurfacefortheSANinjectionpresentverynarrowgrooves(Fig.14a).Ontheotherhand,frictiontrackscanbeobservedforthePCinjection(Fig.14b).Thesecanoccurduringinjectionandcanbecorrelatedwiththepresenceoftheflowlinesontheinjectedplasticpart.SAN“followed”verywellthemorphologyofthesurfaceofthemold.Inconsequencethemicroindentationspresentonthemoldstamp,canbeidentifiedasdistinguishedpeaksontheplasticpartsurface(Fig.15).TheSANpartssurfacepresentsnovisibleflowlines.OnthecontrarythesurfaceoftheinjectedpartsinPCexhibitshighlyvisibleflowlines(Fig.16a).Buttheiramplitudeandperiodicitydecreasefromtheregionneartheinjectionsitetowardstheareaneartheendoftheinjection(Fig.16b).InTables6and7thevaluesforthechosenheightparametersarelisted.Forthemoldstamp,thevaluesarehigherincaseoftheoneusedforthePCinjection.ThevaluesforthemoldmatrixaresimilarforthePCandSANinjections.IncaseoftheplasticpartitseemsthattheoneinjectedwithPChashighervaluesthantheoneinjectedwithSAN,understandableconsideringthepresenceofflowlineonthePCpart.398N.Crisanetal./TribologyInternational100(2016)388–399Fig.14.ESEManalysisfortheopticalmoldstampinjectedwith:a)SAN;andb)PC.Fig.15.Interferometrymeasurements:a)thesurfaceontheupperpartofthemold;andb)surfaceoftheinjectedpiece.Fig.16.Interferometrymeasurementssamelineofobservation:a)surfaceneartheinjectionpoint;andb)surfaceattheendoftheinjectionflow.N.Crisanetal./TribologyInternational100(2016)388–399399Table6RoughnessparametersforthetwomoldpartsusedfortheSANinjectionandfortheSANpart.Parameters(ObjectiveÂ10)MoldbeforeinjectionStampSa(nm)Sz(nm)Sq(nm)SskSku1.2670.6729.38714.281.6670.97À0.09870.0224.5870.83Matrix1.2970.1633.46710.011.6570.2À0.170.10.9372.14MoldafterinjectionStamp1.7570.2424.5674.212.2470.3À0.00470.1663.5170.3Matrix1.4370.1622.9376.871.8470.240.15670.3163.970.91.8170.1140.173.032.3570.12À0.01970.3614.5870.5SANpart(stampside)Table7RoughnessparametersforthetwomoldpartsusedforthePCinjectionandforthePCpart.Parameters(ObjectiveÂ10)MoldbeforeinjectionStampSa(nm)Sz(nm)Sq(nm)SskSku2.7970.4748.76.983.5670.60.39570.1213.6670.6Matrix1.770.2134.5777.272.1870.28À0.02570.1253.8270.37MoldafterinjectionStamp2.3770.4933.57710.672.9970.620.40270.1413.4170.35Matrix1.2670.1121.6276.361.6170.15À0.08770.1313.5770.622.9470.6347.47713.783.7370.79À0.47770.1063.7470.43PCpart(stampside)4.ConclusionsThisstudyhasallowedtheidentificationandevaluationofdefaultsthatoccurduringplasticinjectionprocess,atmicroscopicscale.Theresultsobtainedhighlightthedifferentdamagemechanismssustainedbythemoldsurface,asafunctionofpol-ishing,geometryandinjectedmaterial.Itcanbealsoobservedthatforeachmaterialinjectedthereisadifferenceoflevelofwearanddamagemechanismbetweenthestampandthematrix.Surlynsinjectionexhibitedconsiderableamountofdepositsonthemoldstamp.Itseemsthatthephysico-chemicalconditions,createdduringtheinjectionbythistypepolymer,favoredtheadhesion.Alsointhiscase,thecouplingeffectsofpolishingqual-ity,theinjectedmaterial,adhesionandthelackofthemoldfea-turethatevacuatesair,tendtoformcorrosionpitsonamirrorpolishedsurface.SANandPA66GF25polymerswereinjectedsuccessivelyonthesamemold.Themoldsurfacepresentedpolishingdefaults(holes)beforeinjection.Theholeswereenlargedinthedirectionper-pendiculartotheinjectionflowduetoabrasiveeffectofglassfibers.Acriticalcharacterizationofthemoldsurfacetopographywasperformedinordertoidentifythelocationandthetypeofdefaultsthatoccurwhenmoreorlessaggressivematerialswereinjectedinmoldswithdifferentgeometries.Alltheresultsprovidedcanbetakenintoconsiderationforthedesignofa“chameleon”coatingthatcanovercomepresentdrawbacks.P.Jehan(Technimold,France)forfruitfuldiscussions.AlsotheauthorswillliketothankPh.Sainsot(LaMCoS,INSALyon,France)forhisadvicesoninterferometrymeasurements.Theauthorsgratefullyacknowledgethefinancialsupportfromcompetitive-nessclustersArveIndustriesandPlastipolis,alsofrom“leconseilrégionaldel’Ain”,“leconseilrégionalFranche-comté”and“larégionRhone-Alpes”.References[1]ChaillyM.Influencedestraitementsdesurfacedemouledansleprocédéd’injectionmoulage.Applicationauxdéfautsd’aspects,Ph.D.thesisINSAdeLyon,Infrench;2006.[2]KennedyPK.Practicalandscientificaspectsofinjectionmoldingsimulation.Eindhoven,TheNetherlands:TechnischeUniversiteitEindhoven;2008Proefschrift.-ISBN978-90-386-1275-1.[3]BociagaE,JarugaT.Experimentalinvestigationofpolymerflowininjectionmould.ArchMaterSciEng2007;28(3):165–72.[4]OzdemirA,UluerO,GuldasA.Flowfrontadvancementofmoltenthermo-plasticmaterialsduringfillingstageofamoldcavity.PolymTest2004;23:957–966.[5]BullSJ,DavidsonRI,FisherEH,McCabeAR,JonesAM.Asimulationtestfortheselectionofcoatingsandsurfacetreatmentsforplasticsinjectionmouldingmachines.SurfCoatTechnol2000;130:257–65.[6]PouzadaAS,FerreiraEC,PontesAJ.Frictionpropertiesofmouldingthermo-plastics.PolymTest2006;25:1017–23.[7]BendadaA,DerdouriA,LamontagneM,SimardY.Analysisofthermalcontactresistancebetweenpolymerandmoldininjectionmolding.ApplThermEng2004;24:2029–40.[8]PessolesX,TournierC.Automaticpolishingprocessofplasticinjectionmoldsona5-axismillingcenter.JMaterProcessTechnol2009;209:3665–73.[9]BigerelleM,VanGorpA,IostA.Multiscaleroughnessanalysisininjection-moldingprocess.PolymEngSci2008:1725–36.[10]VanGorpA,BigerelleM,NajjarD.Relationshipbetweenbrightnessandroughnessofpolypropyleneabradedsurfaces.PolymEngSci2016:103–17.[11]EdSeverson.Fighthiddencorrosioninstainless-steelmolds.PlasticsTech-nology.September2014issue.Onlineedition,〈http://www.ptonline.com/articles/fight-hidden-corrosion-in-stainless-steel-molds〉.AcknowledgmentsThisstudywasperformedaspartofthecollaborativeresearchprojectPOLIREV(FUIN°F1310027V).Thanksareexpressedtotheothersprojectpartners,C.Chouquet(DMXsociety,France)and
