A study of fiber laser welding of galvanized steel using a suctionmethod

JournalofMaterialsProcessingTechnology214(2014)1456–1465

ContentslistsavailableatScienceDirect

JournalofMaterialsProcessingTechnology

journalhomepage:http://wendang.chazidian.com/locate/jmatprote

c

Astudyof berlaserweldingofgalvanizedsteelusingasuctionmethod

ZhichunChena,ShangluYangb,ChunmingWanga, ,XiyuanHua,XinyuShaoa,JunWanga

ab

SchoolofMaterialsScienceandEngineering,HuazhongUniversityofScienceandTechnology,Wuhan430074,ChinaChinaScienceLab,GeneralMotors,56JinwanRoad,Shanghai201206,PRChina

article

info

abstract

Articlehistory:

Received2July2013

Receivedinrevisedform3January2014Accepted7February2014

Availableonline18February2014

Keywords:LaserweldingGalvanizedsteel

Gap-freecon gurationZincvapor

Suctionmethod

Inthispaper,wedescribeamethodemployingasuctiondeviceadaptedtoprovideanegativepressurezoneonthesurfaceofthekeyholetoallowthehighly-pressurizedzincvaportoescape.Themicrostruc-tureandpropertiesofthelapjointwerestudied,andthedistributionofzincelementinthejointswasanalyzedusingsynchrotronradiationX-ray.Ahigh-speedvideocamerawasusedtorecordthedynamicbehaviorofthelaser-inducedplasmaplumeandthezincvapor.Experimentalresultsdemonstratedthatthissuctionmethodcannotonlyfacilitateventingofthehighpressurezincvaporfromthemoltenpoolandthekeyhole,butalsostabilizesthelaser-inducedplasmaplumeduringtheweldingprocess.Alapjointwithagoodsurfacequalityandexcellentmechanicalstrengthwasobtainedusingthismethod.

©2014ElsevierB.V.Allrightsreserved.

1.Introduction

Laserweldingofgalvanizedsteelsheetsinanoverlapjointcon- gurationiscommonlyusedintheautomotiveindustry.Technicalproblemsarisewhenlaserweldinggalvanizedsteelinazero-gapoverlapjointcon gurationduetotheboilingpointofzinc(about1200K),beingwellbelowthemeltingpointofsteel(1800K).Highlypressurizedzincvaporiseasilygeneratedonthefayingsurfaceoftwometalsheetsduringtheweldingprocess.Thezincvaporventsoutthroughtheweldpoolandresultsintheformationofblowholes,spatterandporeswhichdramaticallyreducestheweldstrength,degradesthecorrosionresistanceandaffectstheaesthet-icsoftheweldseam.

Inordertoresolvetheseissues,severaltechniqueshavebeenproposedduringthepastdecade.OneproposedsolutionistoremovethezinccoatinglayeraroundtheweldingregionusingalaserandthenplatewithaNicoatedlayer.Akhteretal.(1991)demonstratedthismethodexperimentally.Whilethisproceduredoessolvetheimmediateproblemitissomewhatimpracticalinaproductionenvironmentduetotheadditionalprocessesandresult-ingreductionincorrosionresistance.Amethodintroducedandstillinusetodaytoovercomethisproblemistosetagapbetweenthe

Correspondingauthor.Tel.:+8613871541964;fax:+862787543894.

E-mailaddresses:cmwang@http://wendang.chazidian.com,dachun1230@http://wendang.chazidian.com(C.Wang).

sheetspriortothewelding,whichprovidesachannelfortheescapeofthezincvapor.Grahametal.(1994)proposedcreatingagapusingastampingprocessormechanicalmethodpriortoinitiatingtheweld.Thegapthuscreatedestablishedachannelthroughwhichthezincvaporcanescape.Themajorchallengeforthismethodisthedif cultyinaccuratelycontrollingthegapdimension.Grahametal.(1996)demonstratedthataspeciallyredesignedcon gurationofthelapjointcouldalsoformagapforventingthegas.Whilethiswaseasilyachievedinproduction,additionalequipmentisneces-sarytoproducethespecialcon guration.Gu(2011)http://wendang.chazidian.cominga lletjointweldingcon gurationcanalsomitigatetheproblemofZnexpansion,butitrequiresedgetrackingdevices.AnUSpatenthasproposedanapproachtolapweldingzinccoatedsteelinazerogapwhichistomodifythezincevaporationbyaddingathinlayerofanalloyingagentonthefayinginterfaceoftheoverlappedsheetstoformachemicalreactionduringtheweldingprocess(Mazumderetal.,2007).ThesematerialscouldbethinsheetsofAlorCuorpowderstoformZn–AlorZn–Cualloy.Theresultisahigherboilingpoint,eliminatingthezincvaporpres-sureandthusavoidingformationofthewelddiscrepancies.Zhangetal.(2012)useda“sandwich”methodtoweldzinccoatedsteel.DasguptaandMazumder(2008)usedcopperastheaddedmaterialwhenLaserlapweldingzinccoatedsteel.Butsomenewproblemswillbegeneratedusingthisapproachsuchasthesoftnessandcor-rosionresistanceofthejoints.Someresearchershaveemployedan

http://wendang.chazidian.com/10.1016/j.jmatprotec.2014.02.0130924-0136/©2014ElsevierB.V.Allrightsreserved.

Z.Chenetal./JournalofMaterialsProcessingTechnology214(2014)1456–1465

1457

Table1

Theprocessvariables.

Parameter

Unit

Value

3.2,3.4,3.6,2.40

Vacuumsuction0.3

Lapjoint

LaserpowerWeldingspeedDefocus

WeldingconditionDiameteroflaserspotJoint

kWm/minmm–mm–

steel,Kielwasseretal.(1998)usedapulsedCO2lasertoweldgalva-nizedsteelresultinginimprovedweldqualitybutthereremainedsomewelddiscrepancies.Inparticular,astablekeyholecouldnotbeproducedastheweldingspeedwasincreased.

Amethodusingavacuumsuctionnozzletoreplacethecon-ventionalshieldinggasnozzlewasappliedduringlaserweldingzinccoatedsteel.Oneofourteam,Yangetal.(2013)presentedthisprocessandshowedabetterresultsbutitwasjustprimaryachieve-ments.Inthispaper,deepandcomprehensiveworkshavebeendonebyourteamandmoreresultsweregiveninthisstudy.Theresultsobtainedfromahigh-speedvideocameraandthezincele-mentmappingfurthercon rmedthatwhenusingavacuumsuctionnozzle,http://wendang.chazidian.comingthevacuumsuctionnozzle,bothlaser-inducedplasmaandspatterscanbeeffectivelycontrolledandsoundweldscanbeachieved.

2.Experimental

Fig.1.Experimentalsetupforgalvanizedsteelswelding:(a)schematicdiagramofweldingcon guration;(b)photooflaserweldinghead.

assistantlaseroragastungstenarcweldingprocessasasecondheatsourceinthelaserweldinggalvanizedsteels.Salmanetal.(2010)documentedtheuseofanassistantlasertocutaslotwhiletheotherlaserwasusedforthe nishwelding.Maetal.(2013)alsousedatwo-passlaserweldingzinccoatedinazerogapsteelandadefect-freejointwasobtainedusingthismethod.Insteadofadualbeamlaser,YangandKovacevic(2009a,b)conductedatrialusingarc-laserhybridweldingofgalvanizedsteel.Theyconcludedthataslowersolidi cationtimeandanenlargedmoltenpoolfacilitatestheremovalofthezincvapor.However,thismethodwasdif culttouseinproductionduetotherequirementfortwoheatsources.Asimplersolution,“pulsedlasers”hasbeenusedtoweldgalvanized

Thematerialusedwasdual-phasesteel(DP590),bothsidesofwhicharegalvanizedwithazinccoatinglayerapproximately25 mthick.Thespecimenswerecuttothedimensionsof100mm×100mm×1.2mm.TheexperimentswerecarriedoutwithanYLR-4000 berlaserwithapeakpowerof4.0kWandanIRB4400robot.Thelaserheadandthede nitionsoftheweld-ingparametersareshowninFig.1.ProcessvariablesusedintheexperimentareshowedinTable1andtheexperimentalsetupisillustratedinFig.1.Thevacuumsuctionnozzlewaslocatedonthelaserheadusingafour-dimensionalcontroltable.Theworkpieceswerepressedtightlytogetherusingajig.Beforewelding,thegapbetweentwosheetswascheckedusingafeelergauge.Thegapbelow0.02mmwasdeemedazero-gap.Differentweldingcondi-tionswereestablished.Inthe rstmethod,galvanizedsteelsamples

Fig.2.Welddefectsofautogenouslylaserlapweldingofgalvanizedsteel:(a)topsurface;(b)bottomsurface(c)transversecross-section.

1458

Z.Chenetal./JournalofMaterialsProcessingTechnology214(2014)1456–1465

Table2

Thesurfacesofweldsindifferentlevelsofsuction.

No.

Weldingconditions

Typicalsurfacesofthelapjoint

#1

P=3.4kWV=2.4m/minf=0

weaksuction

#2

P=3.4kWV=2.4m/minf=0

strongsuction

#3

P=3.4kWV=2.4m/minf=0

without

suction

wereweldedwithoutanyotherprotection.Inthesecond,avacuumsuctiondevicewasemployedlocatedforwardofthelaserbeam.Ahigh-speedvideocameraoperating1000framespersecondwasusedtorecordthelaserinducedplasmaplumeduringtheweldingprocess

Theweldqualitywas rstevaluatedbydirectobservationstodetermineifdiscrepanciessuchasspattersorblowholeswereproducedinthewelds.Thereafter,metallographicexamination,micro-hardnessandshear-tensiletestswereconductedtoexam-inemicrostructurefeaturesandstudythemechanicalperformance.Allsampleswerepolishedwithdifferenttypesofsandpapersandetchedwithsolutionsincluding5%picricacidmixedwithnital.Themicro-hardnessdistributionandphaseconstitutionswerethenanalyzed.Themicro-hardnessequipmentmodelandparametersusedinthemicro-hardnessanalysiswere:theDHA-1000microhardnesstestersetfor100gfloadingforceanda15sdwelltime.

ThezincelementwasmappedintheselectedareaoftheweldsbysynchrotronradiationX-ray.Samplesselectedforthetwodifferentweldingconditionswerechosenandpolishedtoa2.0mmthickness.Themaximumenergyselectedwas20keVtoensurea30%transmissionrate.Theminimumspotsizewas1.6 m×1.8 minsidethesamplewithafocusphoton uxden-sityof1.8×1011phs/(s m2)at10keV.The uorescentintensitiesweremeasuredusingaliquid-nitrogen-cooled-elementenergy-dispersivehigh-puritySi(Li)detector(Wuetal.,2013).Thesampleswerescannedinthe‘step-by-step’modeatBL15U.Thespotsizeweusedwas100mm×100mmandthetotalpixelswere720.

Defectsobservedwerespatters,throughholes,blowholesandporosities.Fig.2showsthetop,bottomandcross-sectionalviewsoftypicalweldsobtainedbyasinglelaserbeamwithoutsettingagapatthefayingsurfaceoftwoworkpieces.Whenthesuctiondevicewasaddedtothelaserweldingprocessandotherparametersheldconstant,thesurfacequalityoftheweldsimprovedsigni cantly.WecanalsoobservethevariationoftheweldsurfacequalityinTable2whenchangingthesuctionlevel.Ithasbeenprovedthatastrongsuctionduringtheweldingprocesscanmakeabigcon-tributiontoobtainingabettersurfacequalitybycomparingabovethreeconditions.Thesurfacequalityofweldswaskeptsmoothbyvaryingincidentpowerfrom3.2kWto3.4kWand3.6kW.Thebestresultingsurfacequalitywasobtainedat3.4kW.Table3andFig.3showthemacro-morphologyandtransversecross-sectionviewsofthelapjointsusingthesuctionmethodduringthelaserweldingprocess.Anearlydefect-freelapjointwasobtained.Noporosi-tiesorblowholeswereproducedintheweld,butsmallundercutwasobservedonthetopandbottomoftheweld.However,unlikeincentpower,weldingvelocityhasanenormousin uenceontheweldsurfacequality.WecanseeinTable4thatthesmoothsurfacecanbeobtainedatalowerweldingvelocity(lessthan3m/min),andwhentheweldingvelocityincreasesto4m/min,thesurfaceofweldsbecomesporousandrough.Theresultscon rmthatthesuctionmethodiseffectiveinsigni cantlyimprovinglaserweldingofgalvanizedsteelinagap-freelapjointcon guration,particularlywhentheweldingvelocityisbelow3m/min.

3.2.Microstructureandhardness

3.Resultsanddiscussion

3.1.Macro-morphologyoftheweldedjoints

Foragivensetofincidentpowerandtheweldingvelocity,with-outanyotherprecautionsduringtheweldingprocess,variouswelddefectsappearwhichsigni cantlydecreasedtheweldstrength.

Toinvestigatethemicrostructureandhardnessofweldedjoint,theexperimentswereperformedwithlaserpowerof3.4kW,weld-ingvelocityof2.4m/minandsuction.Themicro-hardnessoftheweldedjointwasmeasuredalongthelineshowninFig.3.Fig.4showsthemicro-hardnessdistributionoftheweldedjoint.Hard-nessdistributionwasnotuniformalongtheweld.Themaximum

Z.Chenetal./JournalofMaterialsProcessingTechnology214(2014)1456–1465

1459

Table3

Thesurfacesofweldsindifferentincentpower.

No.

Weldingconditions

Typicalsurfacesofthelapjoint

#1

P=3.2kWV=2.4m/minf=0

vacuumsuction

#2

P=3.4kWV=2.4m/minf=0

vacuumsuction

#3

P=3.6kWV=2.4m/minf=0

vacuumsuction

andthemartensitephaseincreasedintheHAZ;Inaddition,thegrainsizeneartheweldcenterwaslargerthanthatfurtherremovedfromthecenter.Duetothehightemperatureandrapidcoolingrate,themicro-structureoftheweldinthefusionzoneconsistsmainlyofthehardmartensite.Alsoaffectedbythehightemperatureofthelaserweldingprocess,somelargebainitewasobservedinthecenteroftheweld,asshowninFig.5fandg.Theresearchesonthemicrostructureandhardnessofweldedjointindicatedthatusingsuctionduringtheweldingprocesswouldreducetheinsideporesofthejointswithoutsofteningthejoint.

Fig.3.Thecross-sectionoftheweldseamwithsuction.

3.3.AnalysisofzincelementdistributionbysynchrotronradiationX-ray

hardnessvalueof410.7HVwaslocatedatthecenteroftheweldwhiletheminimumhardnessvalueof184.7HVwasfoundinthebasematerial.Thisanalysisclearlyindicatesthattheweldedzoneandtheaffectedzone(HAZ)areharderthanthebasematerial

Microstructurefeaturesinthedifferentweldzonesoftheweldedjointwereobservedbymeansofopticalmetallographicmicroscope.Fig.5showstheexperimentalresults.Fig.5aandbdemonstratedthefeaturesinthebasematerial,theHAZandthecenteroftheweldatlowmagni cation.Highmagni cationwasalsousedtoobservethefeaturesindifferentzonesoftheweldedjoint,asshowninFig.5c–g.Typicalmicrostructureconstituentsofdualphase(DP)steelswerefound:softferriteandtheislandsofthehardmartensitedispersedthroughouttheferritematrix.Thehardmartensiteprovidessubstantialstrengthwhiletheferritematrixcontributestogoodductility.AsshowninFig.5,thesegregationofmicrostructureinthejointwasformedbecauseoflaserweldinghavingafastcoolingspeed,butweldingdefects,suchasporosities,werenotfoundinsidetheweld.Theferritephasecontentdecreased

Weldedjointswithlaserpowerof3.4kW,weldingvelocityof2.4m/minanddifferentsuctionlevelswereselected,sectioned,andpolishedinordertoanalyzethezincelementdistributionintheweldedjointwithasynchrotronradiationlightsource(synchrotronradiationX-ray).Fig.6showsthespatialdistributionofthemajorZnelements.Theelementalmappingsaregradedusingacontinuouscolorbar.Eachbarisassignedtoanelementallevel,forwhichredisthehighest,greenismediumandblueisthelowest.Fig.6aandbindicatestheexperimentalresultsfortheweldedjointsobtainedwithandwithoutuseofthesuctiondevice.InFig.6,wenotethatthezinccontentisnotuniformlydistributedindifferentlocationsoftheweldedjoint.Theareasinthetopandbottomsurfacesandthefayinginterfaceofthemetalsheetshavearichercontentofzincthanintheotherareas.WhenwecompareFig.6awithb,weseethatthezinccontentintheweldzoneofFig.6awasricherthanthatinFig.6b.Thisisexplainedbythefactthatoneemployedvacuumsuctionwhiletheotherdidnot.Duringlaserwelding,thezinccoatingevaporatedduetoitslowerboilingpoint.Portionsoftheevaporatedzinccoatingescapedthroughthekeyholeandthe

1460

Z.Chenetal./JournalofMaterialsProcessingTechnology214(2014)1456–1465

Table4

Thesurfacesofweldsindifferentweldingvelocity.

No.

Weldingconditions

Typicalsurfacesofthelapjoint

#1

P=3.0kWV=2.0m/minf=0

vacuumsuction

#2

P=3.0kWV=3.0m/minf=0

vacuumsuction

#3

P=3.0kWV=4.0m/minf=0

vacuumsuction

moltenweldpoolwhilesomequantityofthezinccoatingvaporremainedintheweld.Whenthevacuumsuctionisused,itaccel-eratestheescapeofzincvaporfromthemoltenweldpoolandthekeyhole.Asaresult,lesszincremainedinthejointtocontaminatetheweld.

WeexaminedthetransverseandthedepthdirectionsalongthemeasuredlineshowninFig.6toobservethevariationinzinccon-tent.TheresultsareillustratedinFig.7a–d.InFig.7awenotethatthezincwasalittlericherintheweldzonethanitisintheHAZ;whileFig.7cpresentsthattheweldzonehadlesszincthanotherareaswhenusingvacuumsuction.However,inthedepthdirection,bothweldswerecharacterizedbyaconsistenttrendofricherzincdistributioninthetopofthewelds.Thisresultcanbeexplainedastheinabilityofthehighpressurezincvaportoventrapidlyenoughtoovercometherapidcoolingrateofthelaserweldingprocessand

thezincvapordesublimatedatthelowertemperatureasitrosetothesurface.

3.4.Resultofthehigh-speedvideocamera

http://wendang.chazidian.com-parisonofFigs.8and9,showsthatwhenusingvacuumsuction,theplasmaplumebecomesmorestableandcanbeshapedinacertaindirectionandacertainsize.TheunstableplasmaplumeshowninFig.8resultsinnumerouslargesizespatters.Inaddition,thecou-plingspotsizevarieddramaticallyastheplasmaplumefrequentlychanged.Behavioroftheplasmaplumehasasigni cantin uenceonthestabilityoftheweldingprocess.Hamadouetal.(2004)have

Fig.4.Vickersmicro-hardnesspro le.