Toward a better understanding of phase transformations in additive manufacturing of Alloy 718

(1)

ContentslistsavailableatScienceDirect

Materialia

journalhomepage:www.elsevier.com/locate/mtla

Full

Length

Article

Toward

a

better

understanding

of

phase

transformations

in

additive

manufacturing

of

Alloy

718

Chamara Kumara

a

_{, Arun Ramanathan Balachandramurthi}

a

_{, Sneha Goel}

a

_{, Fabian Hanning}

a

_,

Johan Moverare

a,b

a Division of Subtractive and Additive Manufacturing Processes, Department of Engineering Science, University West, 461 86 Trollhättan, Sweden b Division of Engineering Materials, Department of Management and Engineering, Linköping University, SE-58183 Linköping, Sweden

a

r

t

i

c

l

e

i

n

f

o

Keywords: Additive Manufacturing Alloy 718 Phase transformation Modelling

a

b

s

t

r

a

c

t

Thispaperpresentsadiscussiononthephase-transformationaspectsofadditivelymanufacturedAlloy718 dur-ingtheadditivemanufacturing(AM)processandsubsequentcommonlyusedpost-heattreatments.Tothisend, fundamentaltheoreticalprinciples,thermodynamicandkineticsmodeling,andexistingliteraturedataare em-ployed.TwodifferentAMprocesses,namely,laser-directedenergydepositionandelectron-beampowder-bed fusionareconsidered.ThegeneralaspectsofphaseformationduringsolidificationandsolidstateinAlloy718 arefirstexamined,followedbyadetaileddiscussiononphasetransformationsduringthetwoprocessesand subsequentstandardpostheat-treatments.Theeffectofcoolingrates,thermalgradients,andthermalcycling onthephasetransformationinAlloy718duringtheAMprocessesareconsidered.Specialattentionisgiven toillustratehowthesegregatedcompositionduringthesolidificationcouldaffectthephasetransformationsin theAlloy718.Theinformationprovidedinthisstudywillcontributetoabetterunderstandingoftheoverall process–structure–propertyrelationshipintheAMofAlloy718718.

1. Introduction

Inrecentyears,comparedwithconventionalmanufacturing meth-ods,additivemanufacturing(AM)ofnickel-basedsuperAlloy718shas attractedconsiderableattentioninaerospaceandpowergeneration ap-plications[1].InAM,acertainpartismanufacturedusinga layer-by-layerapproachusinga3D-digitalmodelofthepart.Animportant fea-tureofthisprocessisthatitallowstheproductionofcomplex, near-net-shapeobjects(eventhroughgenerativedesign),withminimalmaterial waste.Furthermore,itenablespartintegration,thusreducingthe num-berofassembliesandsub-assembliesrequiredinthefabricationprocess. Inaddition,itallowsmanufacturingondemand,therebyreducingthe needforalargeinventoryofspareparts.Forthesereasons,AMis consid-eredasuitablemanufacturingmethodfornickel-basedsuperAlloy718 componentsforaerospaceandpower-generationapplications[2,3].

However,certainchallengesshouldbeovercomebeforeAMcouldbe usedmoreeﬀectivelyinaerospaceandpower-generationapplications. Onesuchchallengeistoobtaintheappropriatemicrostructurethat pro-videsthedesiredmechanicalperformancetotheAMcomponent.During thelayer-by-layerdepositionprocess,thematerialundergoesmelting, solidiﬁcation,andthermalcycling.Thisinducesaliquid-to-solidphase transformation,aswellassolid-statetransformations.Therefore,the as-builtmicrostructureofthecomponentoftenleadstoaheterogeneous microstructurewithheterogeneousmechanicalpropertiesona macro-scopiclengthscale[1].Toovercomethis,thecommonpracticeistouse suitablepost-heattreatment(HT)protocols,withorwithout hot

iso-staticpressing(HIP),fortheas-builtpart.TheseHTsfurtherchangethe microstructureaccordingtoitscompositionsegregationlevel,phases, andgrainstructure(morphologyandtexture).Therefore,toachievethe desiredproperties,itisnecessarytounderstandthephase transforma-tionsduringtheformationoftheas-builtmicrostructureandunder dif-ferentHTsbeforetheappropriateoptimizationisperformed.

Thisstudyfocusesonunderstandingthephasetransformationof ad-ditivelymanufacturedAlloy718.Specifically,weusefundamental the-oreticalprinciples,thermodynamicandkineticsmodeling,andexisting literaturedatatoexplaintheobservedphasechangesduringtheAM andpost-HTofAlloy718.TwodifferentAMprocesses,namely, laser-directedenergydeposition(L-DED)andelectron-beampowder-bed fu-sion(EB-PBF)areconsidered.Generalaspectsofthephaseformation inAlloy718arefirstexamined,followedbyadiscussiononthephase transformationsduringthetwoprocessesandsubsequentHTs. 2. Alloy 718

Alloy718 isanickel–iron-based superAlloy718.It hasbeen ex-tensivelyusedinrocketmotors,aircraftengines,nuclearreactors,and pumps[4].Themainreasonsforitssuccessareitsrelativelylowcost andgoodmechanicalaswellascorrosionpropertiesatlowand inter-mediatetemperatures[5,6].However,its useinload-bearing compo-nents for elevated-temperatureapplicationsis limitedto650 °C ow-ingtothestrengthlossbeyondthistemperature[7].Thecomposition ofAlloy718iscomplexandinvolvesseveralAlloy718ingelements, https://doi.org/10.1016/j.mtla.2020.100862

Received8June2020;Accepted8August2020 Availableonline11August2020

(2)

Fig.1. Phaseprecipitationwindows([6,9,12]) inAlloy718duringnon-equilibrium solidiﬁca-tionandinsolid-statetransformation.

whichareaddedtoobtainthedesiredmicrostructureandproperties. Thenominalcompositionranges fortheAlloy718(accordingtothe ASTMstandard),theireﬀectonthemicrostructureandcrystalstructure informationofthecommonlyobservedphasesinAlloy718canbefound inreference[8].TheAlloy718microstructureisprimarilydominated byaface-centeredcubic(FCC)𝛾 matrix,whereinprecipitatessuchas

𝛾’/𝛾″(strengtheningphases),𝛿,Laves,MCcarbides,andnitridescanbe found[9].The𝛿 phaseistheequilibriumphaseofthemetastable𝛾″. Theexactmicrostructure(phasecomposition,phasedistribution, mor-phology,andvolumefraction)ofthisAlloy718ismainlygovernedby theprimary manufacturingtechnologyandsuccessivepost-HT condi-tions.Fig.1 showsthereportedphaseprecipitationwindowsforAlloy 718.Itcanbeseenthatthereisanoverlapbetweenthe𝛿 and𝛾’/𝛾″ pre-cipitationwindows.Although𝛿 isthermodynamicallymorestablethan

𝛾″,𝛿 precipitationupto~900°Cisalwaysprecededby𝛾″precipitation [10,11].

2.1. Non-equilibriumsolidiﬁcation

Duringsolidificationprocesses,suchascasting,welding,andAM, Alloy718tendstoformadendritic/cellularmicrostructure[9,13–15], thelengthscaleofwhichvariesaccordingtothesolidificationconditions [15,16].Knorovskyetal.[6] andAntonssonetal.[12] experimentally investigatedthesolidificationsequenceforAlloy718.Astemperature dropsintheliquid(L)melt,theL→ TiNtransformationfirstoccursabove theliquidustemperature.TheresultingTiNparticlesoccasionallyactas nucleationsitesforcarbideprecipitationatalaterstageinthe solidifi-cation[12].HoweverinthecontextofAM,TiNcanalsocomeviathe feedstockmaterialandcouldremainunmeltduetotheirhighermelting point(2930°C)[17].

Whenthetemperaturedropsbelowliquidus,thesolidiﬁcationofthe primary𝛾 phasetakesplace.Duringthesolidiﬁcationofthe𝛾 matrix,

segregationofAlloy718ingelementsistypicallyobserved[9].Thisis becausethesolubilityoftheAlloy718ingelementsinthematrixphase isdiﬀerentfromthatin theliquid.ElementssuchasNb,Mo,andTi, whichhavealowsolubilitylimitinthematrix,tendtosegregatetothe liquid.ElementssuchasCr,Fe,andAl,whichhaveahighsolubility limitinthesolid,tendtobetrappedinthesolid.Thissegregationalters thelocalthermodynamicsofAlloy718and,hence, thedrivingforce forphaseformation.Therefore,phasessuchasLavesandNbCbeginto formintheinterdendriticliquidregionduringsolidiﬁcation.Owingto theirlowsolubilityinthematrix,NbandCarecontinuouslyrejected intotheliquid.Asthematrixgrows,theNbandCcompositionsinthe liquidreachalevelthatenablesNbCformation.Thisreactionconsumes NbandthemajorityofCintheremainingliquid,shiftingtheremaining liquidcompositionbacktoalowerlevel.As𝛾 growsfurther,the segrega-tionofNbintheremainingliquidpromptsanothereutecticreactionL→

𝛾 +Laves,whichterminatesthesolidificationprocess.Thisisreferredto asnon-equilibriumsolidification(Fig.1).InthecontextofAMofAlloy 718,theformedprimarycarbidesandnitridesduringnon-equilibrium solidificationdonotchange(intermsofsizeanddistribution) notice-ablyatlowtemperaturesowingtotheirstabilityatthesetemperatures. Therefore,inthisstudy,thefocusisonunderstandingthe transforma-tionofthe𝛾’/𝛾″,𝛿,andLavesphases.

2.2. Effectofsolidificationconditionsonnon-equilibriumsolidification

Thepartitioningoftheelementsandthelevelofsegregationdepend onthesolidificationconditionsoftheprocess,anditaffectsthe forma-tionofthesecondaryphases[12]duringnon-equilibriumsolidification. Antonssonet al.[12] studiedtheeffectof thecoolingrate(from 2.5×10−1_to₂_×₁₀4_°C/s)_on_the_{solidification}_of_Alloy_718._During

therapidsolidiﬁcationthatoccursathighercoolingrates(>104_°C/s),_it

wasobservedthattheinterdendriticregioncontainslessNb,andLaves phasesarenot present.Owingtotherapidsolidification,thesolutes becometrappedinthemovingsolid/liquidinterfacebecauseofthe in-completesolutepartitioning[18].ThisresultsinlessNbsegregationto theinterdendriticliquidandhenceinsufficientconcentrationfor Laves-phaseformation.Incontrast,duringslowersolidification(similartothe conditionsintraditionalcasting),theinterdendriticregionwasobserved tohavemoredominantNbandLavesphases.

However,inlaser-beampowder-bedfusion(LB-PBF)ofAlloy718, Laves-phaseformationisobservedundernon-equilibriumsolidiﬁcation conditions[19,20],eventhoughthereportedcoolingratesduringthe solidiﬁcationofthemeltpoolareintheorderof105_°C/s_[₂₁_,₂₂_]._This

contradictstheobservationmadebyAntonssonetal.,highlightingthe factthatonecannotsimplyusethecoolingratealoneasacriterionto describewhetherthesolidificationprocesswouldresultinLaves-phase formation;rather,oneshouldalsoconsiderthesolid–liquidinterface velocity, thermalgradients,andundercoolingconditions[18].Inthe literatureonthemicrostructureofadditivelymanufacturedAlloy718, ithasbeendemonstratedthatLavesphasesarepresent,implyingthat non-equilibriumsolidificationconditionsoccur[23–26].Thisindicates thatthesolidificationconditionsduringtheseprocessesdonottrapa significantfractionofAlloy718ingelementsinthemovingsolid/liquid interface,allowingtheoccurrenceofelementpartitionandhence Laves-phaseformation(seeFig.2).

ThepresenceofLavesphasesinthemicrostructurehasanegative eﬀectonthemechanicalpropertiesofAlloy718(tensilestrength, duc-tility,fatiguelife,andfracturetoughness)[29,30],owingtoitsbrittle nature.Usually,liquationcrackingisoftenobservedinthe microstruc-turewhenLavesphasesarepresentinalong-chainmorphology[30]. ThelowmeltingpointoftheLavesphaseandthelong-chainmorphology willpromoteliquationcracking.AstheLavesphaseformstowardthe

(3)

Fig.2. Microstructuresobtainedfromdiﬀerentprocess,exhibitingtheformedLaves+NbC(whiteprecipitates)attheendofnon-equilibriumsolidiﬁcation.In L-DED,EB-PBF,andLB-PBF,thebuilddirectionisfrombottomtotop.

Source: Sources:cast[12],welding[27],L-DED[15],EB-PBF[14]andLB-BPF[28].

endofnon-equilibriumsolidification,itsmorphologyisdirectlyrelated tothemorphologyofthedendriticstructureatthattime.Therefore,by changingthedendritemorphologyduringthesolidificationprocess,the morphologyoftheLavesphasecanbechangedandpotentiallyaffect thepropensityforliquationcracking.

Itiswellknowninthecastingandweldingcommunitythatthe den-driticstructureduringthesolidiﬁcationprocesscanbealteredby chang-ingthesolidiﬁcationconditions,thatis,thermalgradient(G),cooling rate(̇𝑇), andliquid–solidinterfacevelocity(𝑅= ̇𝑇

𝐺)[31,32].This is trueinthecaseofAMofAlloy718aswell,whereitcanbeachievedby changingtheprocessparametersthatdirectlyaﬀectthethermal condi-tionsinthemeltpool.Dependingontheseconditions,theresulting den-driticstructurecouldhaveacolumnar,equiaxed,ormixedmorphology [33,34].

Theeffectofthethermalgradient(frombottomtotopinthedomain) andcoolingrateon thedendriticstructure ofAlloy718 isshown in Fig.3.Theseresults wereobtainedfrommultiphase-fieldsimulations ofAlloy718underdifferentthermalconditions.Informationaboutthis modellingispresentedinAppendixA.

ItcanbeseeninFig.3 (a)thatthedendriticstructureduring so-lidificationchangeswiththecoolingrateandthermal gradient.This ultimatelyaffects thesizeandmorphology of theLaves-phase parti-cles,asshowninFig.3 (b).Atlowcoolingrates,thedendritic struc-ture iscoarser. Therefore,theresultantLaves-phase structureis also coarse,andits morphologytends tothelong-chainform,which has beendemonstratedtohaveanegativeeffectonthemicrostructure,asit promotesliquationcracking[30].AsLavesphasesformduringthefinal solidificationstage,theybegintomeltaroundtheeutectic-forming tem-peraturewhenthetemperatureofthematerialisraised.IfaLavesphase hasalong-chain morphology,liquid can formalong withthis chain morphologyatthereheating stageduringthermal cyclingabove the eutectic-formingtemperature.Whencombinedwiththetensilestresses generatedatthereheatingstage,thisliquationcaneasilyleadto crack-ing.

Asthecoolingrateincreases,theamountof undercooling experi-encedbytheliquidbecomeshigherforagiventime.Higher undercool-ingresultsinhigherexcessfreeenergyintheliquid,whichisconsumed bytheliquid-solidinterfacethatiscreated(thoughnucleationand/or

growth).Whentheamountofexcessfreeenergybecomeslarge,more andmoreliquid–solidinterfacesarecreatedperunitarea.Consequently, theresultantdendriticstructureisfineronthelengthscale.Infine den-driticmicrostructures,thethicknessandspacingoftheprimaryand sec-ondarydendritearmsbecomesmaller.Therefore,towardtheendofthe solidification,theremainingliquidareasaretrappedbetweenthesefine dendricstructures.ThisresultsinfineanddiscreteLaves-phaseparticles attheendofthesolidification.WhentheLavesphasedistributesinthe microstructureasdiscreteandfineparticles,thepropensityforhot-crack formationisreduced,asacontinuousliquidfilmisdifficulttogenerate. Inaddition,asthecoolingrateincreases,theresultantLaves-phasearea fractionisreduced,asshowninFig.3 (c).

2.3. Solid-statephasetransformationafternon-equilibriumsolidiﬁcation

After non-equilibriumsolidiﬁcation, solid-state phase transforma-tionstakeplaceasthetemperaturecontinuestofall.Themainphases thatcanprecipitateinthesolidstateare𝛾’/𝛾″and𝛿.Immediatelyafter non-equilibriumsolidiﬁcation,thereexistsacompositiongradientfrom thedendritecoretotheinterdendriticregioninthe𝛾 matrix.Therefore, thelocalequilibriumconditionsanddrivingforcesforsolid-statephase transformationchangefromthedendritecoretotheinterdendritic re-gion[24,35,36].

Fig.4 showsasimulatedAlloy718microstructureattheendofa non-equilibriumsolidiﬁcation[23]intheL-DEDprocessandthe segre-gated1-Dcompositionproﬁlesalongalinefromthedendritecoreto theinterdendriticregion.Readerisreferedto[23] fordetailsaboutthe modellingwork.Inthissimulation,Alloy718ismodeled asa seven-element systemwithNi-Fe-Cr-Mo-Nb-Ti-Al.Themostsegregated ele-mentsareNbandFe,whereastheleastsegregatedareTiandAl.The generallyacceptedchemicalcompositionformulaforboth𝛾″and𝛿 is

Ni3Nb[4].Therefore,thedistributionofNbhasaprofoundeﬀectonthe

distributionof𝛾″and𝛿.Thegenerallyacceptedchemicalformulafor𝛾’

isNi3(Al,Ti)[4].Thus,thedistributionofAlandTiaﬀectsthe

distri-butionof𝛾’.Thesemodelingresultsagreewellwiththeexperimental observationsmadebySuietal.[4].Fromthe1-Dsegregationproﬁles, theAlsegregationlevelisquitelow,indicatingarelativelyhomogenous distributionofAlinthemicrostructure.However,thisisnotthecasefor

(4)

Fig.3. (a)Variationinthedendritestructureunder diﬀerentsolidiﬁcationconditions.

(b)VariationintheLaves-phasemorphologyunder dif-ferentsolidiﬁcationconditions.

(c)VariationofLaves-phaseareafractionunder diﬀer-entsolidiﬁcationconditions.

Both (a) and (b) were taken at the end of non-equilibriumsolidiﬁcationsimulationdescribedin Ap-pendixA.Domainsizeis80μm×80μm.

(5)

Fig.4. (a)As-builtmicrostructureattheendofnon-equilibriumsolidiﬁcation simulationperformedusingMICRESS,domainsize25μm×20μm.Scaleshows theNbdistributioninwt%.

(b)1Dcompositionproﬁlesextractedfromamultiphaseﬁeldsimulationatthe endofthenon-equilibriumsimulation.

Note:Simulationresultsweretakenfromaprevioussimulationworkpublished in[23].

Ti;preferentialsegregationofTiisclear,andthusitsdistributioncould aﬀectthedistributionof 𝛾’.Nevertheless,asthesegregationofNb is higherthanthatofAlandTi,itcanbeexpectedthatthevariationin thedistributionof𝛾″ismorepronouncedthanthatof𝛾’inthematrix.

Togaininsightintotheeﬀectofsegregationonthelocalequilibrium conditionsof Alloy 718,equilibrium volume fractiondiagrams were generatedusingtheJMatPro(JMatProisatrademarkofSun Microsys-tems,Inc-ver10.2)softwarepackageconsideringthecompositionin thedendritecoreandtheinterdendriticregioninFig.4.For compari-son,theequilibriumvolumefractionrelatedtothenominalcomposition ofAlloy718wasalsogenerated,andtheresultsareshowninFig.5.It canbeclearlyseenthatsegregationhasadirecteﬀectonthe equilib-riumconditionsfor𝛾’/𝛾″and𝛿.Awayfromthedendritecoreandclose totheinterdendriticregion,thereisanincreaseinthevolumefraction forthe𝛾’/𝛾″and𝛿 phases;moreover,theequilibriumtransformation temperatureforthephaseschanges.

Fig.5. Equilibriumvolumefractionsof𝛾’,𝛾″,and𝛿 predictedusingJMatPro fornominal,interdendritic,anddendritecorecomposition.Compositionforthe interdendriticanddendritecorewastakenfromFig.4(b).Inthecalculationof

𝛾″,𝛿 wassuspended.

2.4. Eﬀectofcoolingrateonsolid-statephasetransformationafter non-equilibriumsolidiﬁcation

The solid-state phase transformationin Alloy 718 is a diffusion-controlledprocess.Thatis,thephasetransformationisaffectedby el-ementdiffusion.Astheelementdiffusioncoefficientinthe𝛾 phaseis lowerthanthatintheliquidphase,theextentofsolid-statephase trans-formationisinfluencedbythecoolingrateoftheprocess.Inthecaseof slowcooling(asintraditionalcasting)showninFig.6,theas-solidified materialspendsarelativelylongertimeintheprecipitationwindowsof the𝛾’/𝛾″and𝛿 phases.Therefore,thereissufficienttimeforphase

(6)

nu-Fig.6. Schematicrepresentationofslowand fastcooling.

Fig.7. (a)SEMimageoftheinterdendriticareaofanas-castAlloy718.𝛾’/𝛾″and𝛿 phaseshaveprecipitatedaroundtheLaves.Takenfrom[9]. (b)Nano-SEMimageoftheinterdendriticregionofLMDEDAlloy718.𝛾’/𝛾″hasprecipitatedaroundtheLavesphase.Takenfrom[38].

cleationandgrowth.Attheendoftheslowcoolingprocess,𝛾’/𝛾″and

𝛿 phasescan beobservedinthemicrostructure.Thesetendto nucle-ateandgrowinsizeprimarilyinthesegregatedinterdendriticregions (closetoLavesphasesthatarealsopresent),asinthecaseofcastAlloy 718(Fig.7(a))[37].Thereasonforthisissimilartotheexplanationof thesegregationaﬀectingtheequilibriumconditionsandinducingphase transformationinAlloy718.Thesizegrowthofthesephasesisalso sup-portedbythehighelementconcentrationintheinterdendriticregion. However,asthesegregationofNbiscomparativelyhigherthanthatof AlandTi,thegrowthofthe𝛾″and𝛿 phasesintheinterdendriticregion isexpectedtobehighercomparedtothatofthe𝛾’phase.

Inthecaseofrapidcooling(asinAM),asshowninFig.6,the as-solidiﬁedmaterialspendsasmallamounttimeintheprecipitations win-dowsof𝛾’/𝛾″and𝛿.Therefore,thesephasesonlyhaveashortperiod timefornucleationandgrowthin themicrostructure.The precipita-tion,inthiscase,islikelytooccurininterdendriticregionsowingtothe compositionalsegregation.However,thesizeofthe𝛾’/𝛾″and𝛿 phasesis quitesmallcomparedwiththatofthecastmaterialuponinitialcooling totemperaturesbelow500°C.InanexperimentalstudybySegerstark etal.[38],wheretransmissionelectronmicroscopywasused,nano-size precipitationof𝛾’/𝛾″wasobservedintheinterdendriticregionofthe ﬁnallayerofmaterialthathadbeendepositedthroughlasermetal pow-derdirectedenergydeposition.However,no𝛿 phasewasobservedin

theas-builtmicrostructure.Itisinterestingtonotethatthe𝛿 phase pre-cipitatesintheinterdendriticregionwhentheprocesshasslowcooling (asincasting),butnotwhentheprocesshasfastercooling(asinAM). Thetimethatthemicrostructurespentinthe𝛿 precipitationwindowin AMisshorterthanthatduringcasting(Fig.5).Thus,thereismoretime forthe𝛿 phasetogrowincasting.

ThisimpliesthatthefinalmicrostructureofAlloy718isinfluenced bythecoolingrateoftheprocess,andthereforedifferent microstruc-turesresultingfromdifferentprocessconditionswillresponddifferently whensubjectedtofurtherHTs.

3. Laser-directed energy deposition

Laser-directedenergydepositionisadirectedenergydepositionAM process,wherelaserenergyisdirectedandfocusedintoanarrow re-gioninthesubstrate/previouslydepositedlayer,meltingboththe sub-strate/previouslydepositedlayerandthefeedstockmaterialthatis be-ingdeposited.Thelattercanbeintheformofeitherpowderorwire. This methodis widelyusedtorepaircorrodedandworngas-turbine componentsbecauseitinvolvesminimaldistortionanddilution[36]. Moreover,itisusedtoconstructsmallcomponentsoraddfeaturesto, forinstance,castcomponents,thusaddingcomplexitytoproducts[2]. Thereby,thecostofcomponentswithcomplexfeaturescanbereduced.

(7)

Fig.8. Schematicrepresentationofdifferent thethermalcyclesinL-DEDprocess. Notethatthestartofthethermalcyclesshows thefinalsolidification.Thatmeansthermal cy-clesthatcause the firstmelting andfurther remeltingarenotshowninthisschematic.

3.1. PhaseformationduringL-DEDofAlloy718

OwingtotheinherentnatureofL-DED,thematerialundergoes mul-tiplethermalcycleswhenadjacenttracks/beadsaredeposited. Depend-ingonthenatureofthesethermalcycles,phasetransformationsoccur, resultingindiﬀerentmicrostructuresthatcontaindiﬀerentphasesand phasefractions.

InL-DED,duringthesolidiﬁcationofAlloy718,phase transforma-tionsoccur as described in Section2. Afterthesolidiﬁcation of the material,theunderlyingmicrostructurehasasegregateddendritic mi-crostructurewithinterdendriticLavesandNbCphases[23].Thereafter, thismicrostructureundergoeschangesduringtheensuingthermal cy-clesowingtothesubsequentmaterialdeposition.Forabetter under-standingofthesolid-statephasetransformationduringthisthermal cy-cling,weconsiderthethermal cyclesshowninFig.8.These thermal cyclesresemblethethermalconditionsthataretypicallyencountered duringmaterialdeposition.

InthermalcycleA,thereissuﬃcienttimebetweenthedepositionof successivelayerssothatthematerialadditiondoesnotcausearisein theglobaltemperatureofthedepositedmaterial.Thistypeofthermal cyclingresemblesthethermalcyclesreportedin[23,38,39].Inthiscase, eventhoughthetemperaturegoesthroughtheprecipitationwindowsof

𝛾’/𝛾″and𝛿 duringthethermalcycling,nosignificantphase transforma-tionis observed.Thisisbecausethetimethatthematerialspendsin theseprecipitationwindowsisquiteshort.Thus,diffusionandhence phasegrowthcannotoccur.Inaddition,thistypeof thermalcycling doesnotsignificantlyaffecttheLavesphasesformedduringthefirst cy-cle.Thishasbeenexperimentallydemonstratedin[23].However,the situationmaybedifferent undersuchthermal conditionsasthosein thermalcycleB.

Thermalcycle Billustratesa casewherethe depositionof multi-ple layersis performed without sufficient inter-passwaiting timeto cooldownthepreviouslydepositedmaterialtoatemperaturecloseto roomtemperature.Intheseconditions,theglobaltemperatureofthe depositedmaterialrisesowingtoheataccumulation.The experimen-tallymeasuredtemperatureprofileinTianetal.[36] resemblessuch acondition.Thisriseintheglobaltemperaturehasaneffectsimilarto thatofanHT,bothtothedepositedandtothebasematerial,ifthe tem-peratureisabove~600°C.Ifthethermalconditionsduringprocessing resemblethoseinthermalcycleB,thenthismaybethoughtofasinsitu

ageingHTforthepreviouslydepositedmaterial.Therefore,withtime,

𝛾’/𝛾″beginstoprecipitateinthematerial,andthisprecipitationis in-ﬂuencedbythelocalcompositionalsegregationinthemicrostructure. Itshouldbenotedthattheriseintheglobaltemperature(>~600°C) alsoaﬀectsthesubstratematerial.Theamountofthe𝛾’/𝛾″precipitate intheinterdendriticregionislargercomparedwiththatinthedendrite

coreowingtothehigherequilibriumvolumefractionexpectedforlocal variationinthechemicalcomposition.Inaddition,these precipitates growinsizewiththetimethatthematerialspendsintheinsituageing. Thisgeneratesagradientof𝛾’/𝛾″precipitatesfromtheinterdendritic regiontothedendritecore,aswellasfromthebottomtothetopofthe depositedsample.Therefore,there willbe ahardnessgradientin the materialfromtoptobottom,whichwasexperimentallyconﬁrmedby Tianetal.[36] (Figs.4and7 intheirpaper).

ThermalcycleCissimilartothermalcycleB,buttheheat accumu-lationisgreater;therefore,theriseintheglobaltemperatureishigher thanthatin thermalcycleB.Such athermalconditionwasreported byZ.Lietal.[40] duringhigh-deposition-rateL-DEDofAlloy718.In thisthermalcycle,theglobaltemperatureofthematerialﬁrstrisesto ahighervalue(>1000°C)andthendropsgraduallyduringthe depo-sitionprocessowingtoheatconduction.Whenthetemperaturedrops, itpassesboththe𝛾’/𝛾″and𝛿 precipitationwindows.Thus,inthiscase,

𝛾’/𝛾″and𝛿 precipitateinthemicrostructure.However,their distribu-tionis notuniform owingtothelocal compositionalsegregation. As describedpreviously,𝛾’/𝛾″and𝛿 predominantlyprecipitatearoundthe Lavesphaseintheinterdendriticregion.Asthebottomofthesample spendsmoretimeatahighertemperaturethanthetop,therewillbe adiﬀerenceintheamountof𝛾’/𝛾″and𝛿 betweenthebottomandtop. Suchaheterogenousdistributionalongtheheightaswellasfromthe dendrite coretotheinterdendriticregionhasbeenreportedbyZ.Li etal.[34].

Itshouldbenotedthatforallthesethermalcycles,theLavesphase formedduringthenon-equilibriumsolidificationofagivenlayerdoes notsignificantlychangeowingtothethermalcyclingthatoccurs dur-ingsubsequentlayerdepositions.Thisisprimarilyduetothefactthatin thesethermalcycles,thetime-temperatureconditionsarenotsufficient todissolvetheLavesphasesignificantly.Therefore,tomodifytheLaves phase,theprocessparametersoftheL-DEDshouldbechangedsothat the thermal conditionsduringnon-equilibrium solidificationchange; otherwise, post-HT is necessaryto alter theformed Laves phase. In shouldalsobenotedthatthenon-equilibriumsolidificationconditions (thermalgradientsandcoolingrates)can,toacertaindegree,change duringthedepositionofthematerialowingtothegeometryofthe de-positedpart.ThiscouldinfluencetheLavesphaseformationlocally dur-ingthenon-equilibriumsolidificationfrombottomtotopofthebuild.

3.2. PhasetransformationduringheattreatmentofL-DEDedAlloy718

Heattreatments(homogenization,solutiontreatment,andageing) areoftenusedforAlloy718toremovecompositionalsegregationand obtaintheappropriatephase distributionsothat desiredmechanical propertiesfortherequiredapplicationmaybeobtained.Dependingon

(8)

Table1

StandardheattreatmentasperAMS5383forcastAlloy718

Homogenization Solution treatment Ageing

1093 ± 14°C for 1~2 h, followed by air cooling or faster cooling

(954~982) ± 14°C for more than 1 h, followed by air cooling or faster cooling

718 ± 8°C for 8 h, furnace cool to 621 ± 8°C at 55 ± 8°C/h, hold at 621 ± 8°C for 8 h, followed by air cooling

Fig.9.TTTdiagramobtainedusingtheJMatProsoftwarepackageby consider-ingcompositionsinthedendritecoreandtheinterdendriticregioninFig.4(b). Dottedlinesrepresent0.5%precipitationclosetoLavesphase,andsolidlines represent0.5%precipitationinthedendritecore.

theinitialmicrostructureandtheﬁnalapplication,diﬀerentHT combi-nationscanbeused.

InL-DEDAlloy718,theas-deposited(AD)microstructurevaries ac-cordingtothethermalconditionsoftheprocess,asdescribedpreviously. Therefore,theHTscouldbepreparedaccordingtotheAD microstruc-ture.Inaddition,theselectionofHTisbeinfluencedbythefinal appli-cation.Forexample,ifL-DEDwasusedtorepairdamagedorworn-out components,thentheHTshouldbeselectedsothatthebasematerialof thecomponentthatisinfullyheat-treatedconditionisnotaffected.

Intheliteratureonpost-HTforL-DEDAlloy718,thecommonlyused HTsareaspertheAMS5383standard,whichwasoriginallydeveloped forcastAlloy718[4,27,41–47].Thedetailsofthisstandardareshown inTable1.Directageing(DA),solutiontreatmentandageing(STA), homogenizationandSTA(HSTA),andhomogenizationandageing(HA) arethecommonHTsusedforL-DEDAlloy718.

3.2.1. Phasetransformationduringheattreatmentofamicrostructure resultingfromthermalcycleA

Asmentionedin theprevioussection,thermalcycleAproducesa microstructurewithsegregated𝛾 matrix, Laves,andMCphases. The presenceofthecompositionalsegregationinthematrixhasadirect ef-fectonthephasetransformationduringdiﬀerentpost-HTs.Toexplain thistransformation,time-temperaturetransformation(TTT)diagrams, asshowninFig.9,weregeneratedusingtheJMatProsoftwarepackage byconsideringthecompositioninthedendritecoreandthe interden-driticregionfromFig.4(b).

The JMatPropredictionsindicate thatthere is a cleardiﬀerence (morethananorderof magnitudediﬀerenceintime)in the precipi-tationkineticsof𝛾’/𝛾″and𝛿 owingtothecompositionalsegregation. Intheinterdendriticregion,precipitatesformmorerapidlythaninthe dendritecore.Moreover,thenoseoftheTTTcurvesrelatedtothe in-terdendriticregion(dottedlines)correspondstoahighertemperature thanthatofthecurvesrelatedtothedendritecore(solidlines).Further, ageingtreatmentisperformedtoprecipitatethestrengtheningphases

𝛾′/𝛾′′.AsshowninFig.5,theequilibriumvolumefractionof𝛾′/𝛾′′is higherintheinterdendriticregionthaninthedendritecore.

Accord-ingly,ahighvolumefractionof𝛾′/𝛾′′formstowardtheinterdendritic regionduringDAtreatment.Theacceleratedphaseprecipitation kinet-icscauses𝛾′/𝛾′′toprecipitateatanearlierstageoftheHTinthe inter-dendriticregionthaninthedendritecore.Hence,the𝛾′/𝛾′′precipitates intheinterdendriticregionexperiencegreatersizegrowththanthosein thedendritecore.Fromthesegregationleveloftheelementsforming𝛾′

and𝛾′′,asshowninFig.5,itisreasonabletoassumethatthegrowthof

𝛾′′issupportedmorethanthatof𝛾′.Therefore,eveniftheequilibrium predictionsindicateahighvolumefractionfor𝛾′intheinterdendritic region,thisfractionmaynotbehighcomparedwiththatof𝛾′′.

Theeffectofelementsegregationontheprecipitationof𝛾″during DAMICRESSsimulationsperformedusingtheADmicrostructurerelated tothermalcycleAisshowninFig.10(seeAppendixBformodeling de-tails).Theresultsqualitativelyagreewiththeexperimentalobservations [36,38].Incontrastwiththedendritecore,towardtheinterdendritic re-gion,thenumberdensityandsizeoftheprecipitatesincrease.Itcanbe assumedthatowingtothelackof𝛾′′precipitates,thedendritecoreis softerthantheinterdendriticarea.Consequently,thereisamechanical propertygradientfromthedendritecoretotheinterdendriticregion. However,atthemacrolevel,thisDAAlloy718microstructurepossesses arelativelyhigherstrengththanintheas-builtcondition,asshownin Fig.11.ItshouldbenotedthatthetemperatureinDAtreatmentisnot sufficientlyhightodissolve/changeaLavesphaseorgrain structure. Thus,theincreaseinstrengthinthisconditioncomparedwiththeAD conditionisduetotheprecipitationof𝛾′/𝛾′′.Thisraisesthefundamental questionofexplaininghowDAL-DEDAlloy718canresultinrelatively higherstrength(incomparisonwithAMS5662wroughtstandard)even thoughthedistributionofstrengtheningphasesisnon-uniforminthe microstructure.Apossiblereasoncouldbethatthemicrostructurehas apropertygradientfromthedendritecoretotheinterdendriticregion. This microstructurecouldbe treated asacompositematerialhaving asofterphase(dendritecorearea)embeddedinahardphase (inter-dendriticarea).Thesetwo(softandhard)areas mimicthedendritic structureinsidethegrain.Therefore,ourhypothesisisthata hierarchi-calstructurewiththesetwophasesanditsspatialstructureinsidethe grainsgiverisetothehigherstrengthonthemacroscale.Thiscouldbe verifiedbyusingcrystalplasticity/finiteelementmodelingtostudythe effectofsuchacompositemicrostructureonmacroscaleproperties;this isleftforfuturework.

SolutiontreatmentandageingisanothertypeofHTthatisusedin L-DEDAlloy718.Here,theADmicrostructureisfirstsubjectedto solu-tiontreatmentfollowedbyageing.Thesolutiontreatmentaccordingto AMS5383isusedtoprecipitatethe𝛿 phaseatthegrainboundaries,as itisperformedbelowthe𝛿 solvustemperature.Thepresenceofthe𝛿 phaseatthegrainboundarieshasbeendemonstratedtohavea benefi-cialeffectonstressruptureductility,andtoinhibitthegrowthtendency ofthematrixgrainsduringtheforgingprocess[48–50].However,the presence ofacompositional segregationinAD L-DEDmicrostructure causes𝛿 toprecipitateinsidethegrainsaswell,owingtothepresenceof multipledendrite/cellsinsideasinglegrain.Intheinterdendriticarea, the𝛿 phaseprecipitatesandgrowsinlargequantitiesowingtothelarge equilibriumvolumefraction(Fig.5)andincreasedprecipitation kinet-ics(Fig.9).Inaddition,theLavesphaseintheADstructureisalso dis-solvedtosomeextent.However,theamountofdissolutiondependson thesolutiontreatmenttemperature.Atlowtemperatures,the dissolu-tionoftheLavesphaseisslower.Thisisevidentfromthesimulations performedbyC.Kumaraetal.[23] andtheexperimentalobservations [38,42,51](Fig.12).Similarobservationshavebeen madefor

(9)

laser-Fig.10. MICRESSsimulatedADmicrostructureattheendofthermalcyclessimilartothermalcycle-A,andthesamemicrostructureafterDAsimulation. Domainsize25μm×20μm.ScaleshowstheNbdistributioninwt%.

Note:SeeAppendixBfordetailsaboutthemodellingwork.

Fig.11. Ultimatetensilestrengthvselongationof L-DEDAlloy718.Datatakenfrom[4,27,41–47].

(10)

Fig.12. Precipitationof𝛿 phaseintheinterdendriticregioninL-DEDmicrostructuressubjectedtosolutiontreatment.

(a)MICRESSsimulationshowingthemicrostructurenormaltothebuilddirectionbeforeandaftersolutiontreatmentat954°C/1h,domainsize25μm×20μm: Fig.showstheNbdistributioninwt%.

Note:MICRESSsimulationresultsweretakenfromaprevioussimulationworkpublishedin[23]. (b)Microstructurenormaltothebuilddirectionafter954°C/1h[23].

(c)Microstructureparalleltothebuilddirectionafter954°C/1h[38]. (d)Microstructureparalleltothebuilddirectionafter900°C/20min[51].

weldedAlloy718subjectedtosolutiontreatment[27].Theamountof

𝛿 thatprecipitatesintheinterdendriticregionisalsoaﬀectedbythe so-lutiontreatmenttemperature.Thevolumefractionof𝛿 ishigherfora solutiontreatmentat954°Cthanat980°C(seeFig.5).Noevidenceof precipitationof𝛾′/𝛾′′hasbeenreportedintheliteratureatthesolution treatmenttemperatureseventhoughtheJMatPropredictionsindicate thattherecouldbeapossibilityof𝛾′/𝛾′′precipitation(seeFig.9).As

𝛿 isrichin Nb,itsprecipitationconsumesacertainlevelof Nbfrom thematrixphase.Therefore,duringageingHT(afterthesolution treat-ment),alowervolumefractionof𝛾′′couldbeexpectedthanduringDA. Asthe𝛿 phaseisincoherentwiththe𝛾 matrix,itdoesnotcontributeto thestrengthoftheAlloy718asmuchasthecoherentandsemicoherent

𝛾′and𝛾′′phases,respectively.

HomogenizationHT isperformed todissolvethenon-equilibrium Lavesphase andhomogenizethecompositional segregationresulting from the non-equilibrium solidiﬁcation. L-DED Alloy 718 produces a microstructure with a ﬁner length scale than that in cast Alloy

718. Therefore, the Laves phase in L-DED Alloy 718 has a smaller size. In addition,L-DEDproducesa smallerdendritic structure than in thecastmaterial. Thus,its compositionalsegregationlength scale is smaller. Consequently, the homogenization kinetics of L-DED Al-loy 718 is faster than thatof cast Alloy 718;this is evident in the study byShang Sui etal. [52]. Asthe homogenization temperature increases, the dissolution of the Laves phase is faster, and eventu-allyhomogenizationofthecompositionwilloccur.However,whether this treatment results in a fully homogenized matrix without any Laves phases depends ontheLaves particle size,dendrite arm spac-ing, HT temperature, and time. During the treatment, 𝛾’/𝛾″ and 𝛿

do not form in the material, as thetemperatureis above their pre-cipitation temperature window. In addition, MC present in the mi-crostructure is not aﬀected signiﬁcantly, and grain growth is com-monlyobserved[4,25,43,45].Similarobservationscanbemade regard-ingthehomogenizationofmicrostructuresformedbythermalcyclesB andC.

(11)

IfSTA orDAweretobeperformedafterhomogenizationHT,the phasetransformationkineticswouldbediﬀerent,asthechemical com-positioninthe𝛾 matrixisdiﬀerentfromthatintheADcondition.Owing totherelativelyhomogenizedcompositioninthe𝛾 matrixafter homoge-nizationHT,the𝛿 phasetendstoprecipitatemainlyatgrainboundaries duringthe1hsolutiontreatmentofSTA[25,43,53].Atagrain bound-ary,thenucleationenergybarrierforsolid-stateprecipitationofaphase islessthanthatofthedefect-freematrix[54].Inaddition,𝛿 couldalso precipitateattwinboundarieswithinthegrainsbecauseatwin bound-aryalsohasalowerenergybarrierfornucleation[55].Theamountof

𝛿 thatprecipitatesduringthe1hsolutiontreatmentisbelowthe equi-libriumvolumefractionof𝛿 atthattemperature;thisisevidentfrom thestudybyS.Azadianetal.[11].Therefore,theprecipitated𝛿

dur-ingthesolutiontreatmentdoesnotconsumetheNbentirely,andthere remainsasufficientamountofNbtoprecipitate𝛾″duringageingHT. However,theNbdepletionatthevicinityofthe𝛿 precipitatehinders the𝛾″precipitation,resultingina𝛾″-freezonearoundthe𝛿 phase[10]. DuringageingHT,thestrengtheningphasesmainlystartto precip-itateuniformlyowingtothehomogenizedelementaldistribution[4]. However,duringtheprecipitationandgrowthof𝛾’and𝛾″,thelocal chemicalcompositionnearthevicinityoftheprecipitatecouldchange duetodifferenceinsolubilityofelementsinthesephases.Thiscould affectthenucleationandgrowthofsecondaryprecipitatesof𝛾’and𝛾″ [56].AsseeninFig.11,these𝛾’/𝛾″thatprecipitateandgrowduring agingHTwillincreasethestrengthofthematerial

3.2.2. Phasetransformationduringheattreatmentofamicrostructure resultingfromthermalcycleB

Asmentionedabove,thermalcycleBresultsinamicrostructurethat containsLaves,MC,and𝛾’/𝛾″phases.Therefore,thephase transforma-tionduringtheHTisrathercomplex.PerformingSTAmaynotbe suit-ableforthismicrostructure,astheexisting𝛾″canactasnucleationsites (stackingfaultsinthe𝛾″particles)thatprecipitateandgrow𝛿 phases

atthesolutiontreatmenttemperatures[10,11].Laves-phasedissolution canbeexpectedtoacertaindegree,asmentionedinSection3.2.1.

DuringDA,𝛾’/𝛾″phasesintheADconditioncanbeexpectedtogrow, asthetemperatureconditionsfavortheirgrowth.Inaddition,the pre-cipitationgradientobservedinthebuilddirectionofthemicrostructure isreduced.NochangestotheLavesphasecanbeexpected,asthe tem-peratureissolowthatcannotdissolvethem.

3.2.3. Phasetransformationduringheattreatmentofamicrostructure resultingfromthermalcycleC

ThermalcycleCresultsinamicrostructurethatcontainsLaves,𝛿,

MC,and𝛾’/𝛾″.PerformingSTAmaynotbesuitableforthis microstruc-ture,astheexisting𝛾″canactasnucleationsitefortheprecipitation andgrowthof𝛿.Inaddition,theexisting𝛿 canalsogrowlargerinsize. TheLavesphasewilldissolveonlytoacertaindegreeinthiscaseas well.

DuringDA,𝛾’/𝛾″phasesintheADconditioncanbeexpectedtogrow, asthetemperatureconditionsfavortheirgrowth.Theexisting𝛿 phase

isexpectedtobeunaﬀectedduringageingHT[55].Nochangestothe Lavesphasecanbeexpected.

Itshouldbenotedthatnodataareavailableinliteraturerelatingto Sections3.2.2 and3.2.3.Thus,thediscussionprovidedisbasedonthe understandingofphasetransformationspresentedinprevioussections andreferencestherein.

4. Electron-beam powder-bed fusion

Electron-beammeltingisatypeofpowder-bedAMtechniquethat wasﬁrstcommercializedin1997byArcamCorporationinSweden.In theEB-PBFprocess, anelectronbeamis usedtoselectivelymeltthe materialin alayer-by-layermanner.This processhasunique advan-tagesinrelationtothemanufacturingofbiomedicalimplantsand high-performancecomponentsusedinaerospaceandhigh-temperature

appli-cations:higherbuildratesowingtohighbeamenergyandspeed,lower residual stresses(duetoelevatedpowderbed temperature),the pos-sibilityoftailoringthemicrostructure[34,57,58],andreduced oxida-tionissues[59].However,owingtotheinherentnatureofthisprocess, thesemi-sinteredpowderisdiﬃculttoremovefromcomplex geome-tries.ComponentsmanufacturedbyEB-PBFhavehighersurface rough-nessthanthosemanufacturedbyLB-PBFandmaythusrequire post-treatment.

4.1. PhasetransformationduringtheEB-PBFprocessingofAlloy718

Owingtotheinherent natureofthisprocess,theprintedmaterial undergoesmultiplethermalcycleswhensuccessivelayersareprinted. Duringthissequentialprinting,theuppermostlayersundergoremelting. Dependingonthebeamparametersandmovement,thenumberof lay-ersthatareremoltenvaries.Whenlayermelting/remeltingoccurs, non-equilibriumsolidificationtakesplace,asdiscussedinSection2.The re-sultingas-solidifiedmicrostructurehasasegregateddendriticstructure withinterdendriticLavesandNbCphases[14,60].Uponbuilding fur-therlayers,themicrostructureexperiencesthermalcycling.However, itdoesnotchange significantlybythethermal cyclesowingtotheir shorterexposuretime.Themostsignificanteffectonthemicrostructure isexertedbytheelevatedbuildtemperaturethatoccursduringthe EB-PBFprocessingofAlloy718.ThistemperatureactsasaninsituHTand causesmicrostructuralchanges.Aseachlayerthatisbuiltisexposed totheelevatedtemperatureforadifferentamountof time,there ex-istsamicrostructuregradientfromthetoptothebottomlayer.Once theprintingiscompleted,theobtainedpartiscooleddownby inject-ingheliumintothebuildchamber.Duringthiscooling,furtherphase transformationstakeplace.

D.Dengetal.[14] andKumaraetal.[24] investigatedtheeffect of theelevated buildtemperatureon theas-solidifiedmicrostructure throughexperiments andmodeling.Thethermocoupledatafromthe bottomofthebuildplatefromthesestudiesareshowninFig.13.Itis evidentthatthroughoutthebuildingprocess,theglobaltemperatureof thepowderbedwasgreaterthan1020°C.Itwasdemonstratedthatthis temperaturecausedthematerialtohomogenize,graduallytowardthe bottom,bydissolvingtheLavesphaseandhomogenizingtheelement distributioninthematrix.Furthermore,thisdiffuseregionextended be-tween150μmto1800μmfromthetoplayer.Beyond1800μm,afully homogenizedregionwasobserved.Thishomogenizationprocesswas ac-celeratedbythesmallerdendritespacingandhencesmallersegregation lengthscaleandsmallerLaves-phaseparticlesize.TheMCphasewasnot affectedbythisinsitutreatment,asitwasmorestable.Noprecipitation of 𝛿 and𝛾’/𝛾″wasexpectedduringprinting,asthetemperaturewas abovetheprecipitationwindowsofthesephases.Precipitationof𝛿 and 𝛾’/𝛾″tookplaceduringthecoolingstageoftheprocess,asthe tempera-turedroppedthroughtheprecipitationwindowsofthesephases.Inthe as-solidifiedregion,𝛿 and𝛾’/𝛾″precipitatedprimarilyinthe interden-driticregionowingtothehighNbsegregation.Inthetransitionregion (150–1800μm)theprecipitationlevelwasinfluenced bythechange in thelocalchemical composition.Inthehomogenizedregion,few𝛿

particleswereobservedathigh-anglegrainboundaries.Thiscanbe at-tributedtothecoolingstage,asthemicrostructurespent~15mininthe

𝛿 precipitationwindow.The𝛾’/𝛾″phasealsoprecipitateduniformly ow-ingtotherelativelyhomogenizedcompositioninthisregion,resulting inahardnessof~420HV.Thisvalueindicatesthatasigniﬁcantvolume fractionof𝛾’/𝛾″precipitatedduringthecooling.Thetimethatthe mi-crostructurespentinthe𝛾’/𝛾″precipitationwindowwas~60min.This indicatesthattheaveragecoolingratethatthematerialexperiencedin the𝛾’/𝛾″processwindowwasapproximately5°C/min.Similarhardness observationshavebeenreportedintheas-builtconditionofEB-PBF Al-loy718[61,62].Theseobservationsagreewellwithcontinuouscooling experimentsconductedbyL.Gengetal.forwroughtAlloy718[63].In thefollowingsection,ithasbeendemonstratedthatthiscoolingdown

(12)

Fig.13. Thermocouplemeasurementfromthebottom ofthe buildplateof theEB-PBFprocess. Rawdata takenfromD.Dengetal.[14]andKumaraetal.[24].

Table2

StandardheattreatmentasperASTMF3055forcastAlloy718

HIP Solution treatment Ageing

1120 to 1185 ± 15°C in an inert atmosphere for 240 ± 60 min at ≥ 100 MPa, followed by furnace cooling to < 425°C

1066 ± 14°C (except not below 1038°C) for 1~2 h, followed by air cooling or faster cooling

760 ± 8°C for 10 ± 0.5 h, furnace cool to 649 ± 8°C, hold at 649 ± 8°C for total precipitation time of 20 h, followed by air cooling or faster cooling 927~1010 ± 14°C for 1 h (except not exceeding

1016°C), followed by air cooling or faster cooling

718~760 ± 8°C for 8 h, furnace cool to 621 ± 8°C, hold at 621 ± 8°C for total precipitation time of 18 h, followed by air cooling or faster cooling

stagecanbeusedasanalternativemethodforageingHTinsome ap-plications.

Kirkaetal.[60] havealsoinvestigatedthemicrostructure develop-mentinEB-PBFAlloy718.However,unlikeinD.Dengetal.[12]and Kumaraet al.[21], thepowder-bedelevatedtemperaturewas main-tainedat~975°C.Asthisiswithinthe𝛿 precipitationwindow,the𝛿 phaseprecipitatedatgrainboundariesaswellaswithinthematrix dur-ingthebuildingprocess.Thetransitionregion,inthiscase,waslonger, asthebedtemperaturewaslower.Theamountof 𝛿 phaseincreased towardthebottomofthesample,asthebottomlayersofthesample remainedforalongertimeat~975°Cthanthetoplayers.Therefore, thevolumefraction𝛾″thatprecipitatedduringthecoolingstageofthe processwashighertowardthetopofthesamplethanthebottom.The consequenceofsuchheterogeneitywithheightisevidentinthereported room-temperaturemechanicalproperties.Asthedistancefromthe bot-tomofthebuildincreased,theyieldandtensilestrengthaswellas elon-gationgraduallyincreased[60].

4.2. PhasetransformationduringheattreatmentofEB-PBFAlloy718

ForEB-PBFAlloy718,post-HTsaretypicallyusedtoremovethe het-erogeneousphasedistributionandreducedefects(porosityandlackof fusion)sothattherequiredmechanicalperformancemaybeachieved [64].TheHIPandpost-HTprotocolsforEB-PBFAlloy718,as recom-mendedintheASTMF3055standard[65],areshowninTable2.

HIPismainlycarriedouttohealdefects(porosityandlackoffusion) intheasbuiltmaterial.However,Oxidespresentatthedefectsurface preventcompletehealingofdefectsafterHIP[66].Asthisisperformed athightemperature(>1100°C)foralongerperiod(~4h),diﬀerent phases(Laves,𝛿, and𝛾’/𝛾″)in theas-builtcondition completely dis-solvebackintothematrixphase[67,68].However,owingtothe high-temperaturestability,carbidesdonotdissolveduringthisprocess[67]. AttheendoftheHIPcycle,thechemicalcompositionofthematrixcan beconsideredhomogenized.However,topreventphasetransformation duringthecooling-downpartoftheHIPcycle,arapidcooling (quench-ing)mustbeperformed,attheendofwhich,themicrostructureis

re-portedtohavea𝛾 matrixthathashomogenizedchemicalcomposition andisprecipitationfree,exceptforcarbides[69].

AftertheHIPcycle,theEB-PBFAlloy718issubjectedtoSTAor DA,dependingontherequirementsandapplication.Ifrapidquenching canbeperformed,thenthesolutiontreatmentat1066°Cisredundant, asthisdoesnotcauseanyphaseprecipitation;otherwise,solution treat-mentat1066°Cisnecessarytodissolvetheundesiredphasedistribution. However,comparedwithHIP,solutiontreatmentinvolvesrelativelylow temperatureandholdingtime,andthereforecompletephasedissolution mayormaynotbeachieved.Balachandramurthietal.[61] reported thatafter1066°C/1h,acertain𝛿-phaselevelwasobservedatsome grainboundaries.However,theirsizewasreducedcomparedwiththat intheas-builtconditionowingtothedissolutionprocessduringthe so-lutiontreatment.D.Dengetal.[70]reportedcompletedissolutionofthe

𝛿 phaseaftertreatingEBMAlloy718at1080°C/1h.Itisworth mention-ingthatthesizeofthe𝛿 phaseintheas-builtconditionobservedbyD. Dengetal.[70] wassmallerthanthatobservedbyBalachandramurthi etal.[61].

AftertheHIPcyclewithrapidquenching,tocontroltheprecipitation ofthe𝛿 phaseinthematerial,solutionHTshouldbeperformedwithin the𝛿 precipitationwindow.Insuchsolutiontreatments,grain-boundary

𝛿 precipitationisprimarilyobserved[70].AsdiscussedinSection3.2.1, solutiontreatmentatthelowerend ofthe𝛿 windowwillresultina highervolumefractionowingtothehigherequilibriumvolumefraction. InEB-PBFAlloy718,S.Goel[71]observedgreaterprecipitationofthe𝛿

phaseaftersolutiontreatmentat954°Cthanat980°C,bothperformed for1h.Furthermore,asdiscussedinSection3.2.1,theamountof𝛿 that

precipitatesduringthe1hsolutiontreatmentdoesnotgiverisetothe equilibriumvolumefractionof𝛿 atthesolutionHTtemperature.

S.Goel[71] demonstratedthepossibilityofperformingSTAinside theHIPvesselcombinedwiththeHIPcycle,therebyretainingahigh pressureduringSTA.Inthiscombinedcycle,theargongaspressure in-sidetheHIPvesselwasretainedabove100MPa.However,thesolution treatmentperformedat980°Cwithpressure(~160MPa)didnotyield any𝛿 precipitationinthemicrostructure,comparedwiththesame so-lutiontreatmentperformedwithoutpressure.Thiscouldbeexplained bytheClausius–Clapeyronrelation[54](Equation(1)),whichdescribes

(13)

Table3

TheoreticalcalculationsofClausius-ClapeyronrelationvaluesusingvaluesfromTTNI8database

Phase ΔH 𝛾 → precipitate (kJ/mol) T eq (K) Δ𝑉 ∗ = 𝑉 𝛾− 𝑉 𝑝𝑟𝑒𝑐𝑖𝑝𝑖𝑡𝑎𝑡𝑒 (10 −6 m 3 /mol) 𝑑𝑑𝑃𝑇𝑒𝑞 (K/100MPa)

𝛿 27.8 1283 7.3 – 7.9 = − 0.6 − 2.8

𝛾″ 27.2 1203 7.3 – 7.8 = − 0.5 − 2.8

𝛾’ 34.9 1140 7.3 – 7.6 = − 0.3 − 0.98

∗_calculated_using_the_data_from_TCNI8_database_since_volume_data_is_not_available_in_TTNI8

Fig.14. Microhardnessoftheas-builtandpost-treatedspecimensofEB-PBFAlloy718[71].

theeﬀectofpressureonthephase-transformationequilibrium temper-atureinagivensystem.

𝑑𝑇𝑒𝑞

𝑑𝑃 = 𝑇𝑒𝑞Δ𝑉

Δ𝐻 (1)

Here,Pisthepressure,Teqistheequilibriumphase-transformation temperature,ΔHistheenthalpychange,andΔVisthemolarvolume dif-ferencebetweenthetwophasesduringphasetransformation.If𝑑_𝑑𝑃𝑇𝑒𝑞 <0, thenanincreaseinpressurewillsuppressthephaseequilibrium tem-perature,whichisthecaseforthe𝛿 phase,asseenfromthetheoretical calculationoftheClausius–Clapeyronrelation(Table3).Onecanargue thatthevaluesshowninTable3 mayhaveanegligibleeﬀectinreality. However,thepurposeofthistheoreticalcalculationistodemonstrate thatanincreaseinpressurewilllowertheequilibriumtemperatureof the𝛾’/𝛾″and𝛿 phases.Experimentsarerequiredtoconﬁrmtheexact levelofdropintheequilibriumtemperatureowingtotheelevated pres-sureonAlloy718.

Finally,ageingHTis performedtoincreasethematerialstrength byprecipitatingthestrengtheningphases.Owingtothehomogenized compositiondistributionstateofthe𝛾 matrixafterthepreviousHTs, strengthening precipitation occurs uniformly in the matrix, thereby resulting in increased hardness of the material [71]. S. Goel et al. [67] demonstratedthatregardlessoftheinitialvariationinthe hard-nessofEB-PBFAlloy718,afterHIPandHTinvolvingageing,higher anduniformhardnesscanbeobtained.ItisworthmentioningthatHT operationsusedforconventionalcastandwroughtAlloy718are typi-callyemployedforEB-PBFAlloy718[58,72,73].Thisisalsoreflectedin theASTMF3055standardforPBFAlloy718,wheretheHTparameters appeartobetakenfromtheexistingAMS5363standardforcast[74], andtheAMS2774standardforwroughtAlloy718[75].However,as statedpreviously,itisevidentfromtheliteraturethatthe microstruc-tureof conventionalcast–wroughtAlloy718is highlydifferent from thatofEB-PBForotherAMAlloy718[20,76,77].Thus,usingstandard cast–wroughtAlloy718HTproceduresmaynotbetheidealsolutionfor EB-PBFmaterials.Therefore,thereisaneedforexploringHTprotocols tailoredtosuchmaterials.Previousworkbytheauthors’researchgroup hasdemonstratedthat,forprecipitationofstrengtheningphasesin EB-PBFAlloy718,theHTtime(particularlyforageing)canbesignificantly

reducedcomparedwiththeschedulerecommendedintheASTMF3055 standard(seeFig.14)[71,78].

InFig.14,itisseenthatthehardnessoftheAlloy718materialwas significantlyreducedaftertheHIPtreatment(1120°C,4h,100MPa). Thisisexpected,asHIPdissolvesthestrengtheningphasescompletely, andtheapplicationoffastcoolingaftertheholdingtimeinhibitsany significantre-precipitation[63].Inaddition,asexplainedearlier,HIP treatmentresultsinamatrixthathasahomogenizedcomposition dis-tribution,withcompositionalvaluesclosetothoseofthenominal com-positionoftheAlloy718.Itwasobservedthatinthefirstageingstep (Age1),hardnessincreasedastheageingtimeincreasedfrom1hto4 h,andwithprolongedholdingattheAge1temperaturefor4h,no fur-therhardnesschangewasnoticed.Similarly,duringthesecondageing step(Age2,followingan8htreatmentinthefirststep),hardness in-creasedafter1hofholdingtime.However,nofurthereffectonhardness wasobservedforlongerholdingtime.Therefore,hardnessappearedto flattenafter4hforAge1,andafter1hforAge2.Theseresultsare indicativeofthepossibilitytoreducethedurationoftraditional dou-bleageingtreatment.Detailedmicrostructurecharacterizationrelated toevaluatingtheshorteningoftheageingtimeandthecorrelationwith hardnessispartofanotherongoingstudyintheauthors’researchgroup. Asimilar hardnessincreasewithin5hduringthefirststep ofaging at 760 °Cis alsoreported (seeFig.15) in Fisket al.[79] for Alloy 718thatwashot-rolledandsolution-treatedat954°Cfor1hpriorto ageing.

An alternative approach to precipitatingstrengthening phases in HIPedEB-PBFAlloy718couldbecontinuouscooling,asmentionedin Section4.1.Experimentswereconductedtoevaluatethefeasibilityof thisapproachforperformingSTAthroughcontrolledcoolingafter hold-ingatHIPtemperature(seeAppendixCfordetails).Fig.16showsthe hardnessresultsfromthecontinuouscoolingexperiments,with pres-sure(STAwasperformedinsidetheHIPvesseldirectly aftertheHIP treatment)andwithoutpressure(STAinaheattreatmentfurnace).The resultsareingoodagreementwiththosebyL.Gengetal.[63].A de-tailedmicrostructurecharacterizationrelatedtothecontinuouscooling experimentsandthecorrelationwithhardnessispartofanother ongo-ingstudy.

(14)

Fig.15. (a)Contourplotofageingtime,temperature,andmicrohardness;(b)variationinmicrohardnesswithageingtimeatfourdiﬀerenttemperaturesforAlloy 718.Takenfrom[79].

Table4

TimethatAlloy718microstructurespentin𝛿 and𝛾’/𝛾″precipitationwindow.

Cooling rate (°C/min) 0.5 2 5 20

Time (min) spent in direct 𝛿 precipitation window (1020– 900°C) 240 60 24 6 Time (min) spent in 𝛾’/ 𝛾″ precipitation window (900–600°C) 600 150 60 15 Time (min) required to cool down to 500°C from 1120°C 1240 310 124 31

Fig.16. EﬀectofcoolingrateonhardnessofEB-PBFAlloy718afterholdingat 1120°C.

InthecontinuouscoolingexperimentsbyL.Gengetal.[63],the hardnessvariationwasexplainedbythevariationof𝛿 and𝛾″observed inthemicrostructureunderdiﬀerentcoolingrates.At5°C/min,apeak inhardnesswasobserved.Below5°C/min,asthecoolingratedecreased, hardnessdroppedgradually.Thiscanbeexplainedbytheincreaseinthe

𝛿 phasevolumefractionthatprecipitateswiththedecreaseinthe cool-ingrate.No𝛿 wasobservedforcoolingratesabove5°C/min[63].As seeninTable4,at0.5°C/min,themicrostructurehadspent approxi-mately4hinthedirect𝛿 precipitationwindow;thisallowsthe𝛿 phase

tonucleateandgrow,thusreachingahighervolumefractionand con-sumingmoreNbfromthematrix.Thislowerstheequilibriumvolume fractionfor𝛾″thatcanprecipitateandtherebyresultsinlower hard-ness.Thehigherhardnessat5°C/mincorrespondstoastructurewitha highdensityofﬁneanddiscrete𝛾″precipitatesformedduringthetime

(~1h)thatthemicrostructurespentinthe𝛾’/𝛾″precipitationwindow atacoolingrateof5°C/min.InS.Goel’sstudy[71],itwasreportedthat HIPedAlloy718withoutanyothersubsequentHT(thereforeno𝛾’/𝛾″) hadahardnessof~200HV.Comparingthiswiththehardnessresulting froma20°C/mincoolingimpliesthatduringthe15min strengthening-phaseprecipitationwindow,precipitationdidoccur.

AsseeninFig.16,thecontinuouscoolingexperimentsconducted in-sidetheHIPvesselunderpressureresultedinlowerhardness.Thiscould alsoberelatedtotheeﬀectofpressureonthephasetransformation,as discussedabove.

5. Summary and conclusions

Inthispaper,wepresentedanddiscussedaspectsofphase transfor-mationinAlloy718duringtheL-DEDandEB-PBFAMprocessesandthe commonlyusedsubsequentpostHIPandHTs.Datagatheredfromthe literature,theoreticalprinciples,andadditionalmodelingresultswere usedinthisstudy.Thepresentdiscussiondemonstratedthatthephase transformationinAlloy718iscomplexandinfluencedbyfactorssuch assolidificationthermalconditions,thethermalhistoryaswellasthe compositionsegregationobservedduringtheprocess.Therefore, con-trollingthethermalconditionduringtheAMprocessisoftheutmost importance. Inadditiontothat,thechangein localthermodynamics andkineticsoftheAlloy718due tothelocalchange incomposition influencethephasetransformationduringthestandardpostHTforthe Alloy718.

DiscussionpresentedinthisarticlerevealthatthecommonHTs orig-inallydesignedforcast-and-wroughtAlloy718maynotbetheoptimal solutionforadditivelymanufacturedAlloy718parts.Therefore,focus shouldbeplacedondesigningnewHTsspeciﬁcallyforadditively man-ufacturedAlloy718.Inthisstudy,thegrainstructurechangesduring HIPandHTswerenotdiscussedindetail.However,inthedesignofnew HTsforAMAlloy718,theeﬀectofgrainstructure(morphology, distri-bution,andtexture)changesshouldalsobeconsidered,particularlyif

(15)

theHTtemperaturesarehigh(>1000°C),wheregraingrowthcan oc-cur.Additivemanufacturingallowstheproductionofawiderangeof grainmicrostructures,fromcolumnartoequiaxedgrains,inasinglepart [1,2,33].Consequently,designingHIPandHTsapplicabletoallthese differentgrainstructureswilldefinitelybeachallengingtask,asthese structurestendtobehavedifferentlyunderpost-HIPandHTconditions. EventhoughLB-PBFwasnotconsidered,itisconceivablethatthe in-formationanddiscussionprovidedhereincouldbeusedtounderstand thephase transformationduringLB-PBFandthesubsequentHIPand HTs.AsLB-PBFisacolderprocess,thediscussiongiveninSection3may beastartingpointtounderstandthephasetransformationforthis pro-cessaswell.However,owingtotheincreasedcoolingrate,thelength scaleofLB-PBFwillbesmallerthanthatoftheL-DEDprocess.

Inconclusion,webelievethattheinformationanddiscussion pre-sented in this paper will promote the understanding of the overall process–structure–propertyrelationshipinAMofAlloy718.

Acknowledgements

The fundingfrom theEuropean Regional Development Fund for project3Dprint,andfromKKFoundation(StiftelsenförKunskaps-och Kompetensutveckling)forprojectSUMAN-Nextisalsoacknowledged. TheauthorsareverygratefultoMrMatsHögström,UniversityWestand MrJohannesGårdstam,QuintusTechnologiesAB,Västerås,Swedenfor carryingoutthecontinuouscoolingexperiments.

Declaration of Competing Interest None.

Appendix: A

ModelingtheeﬀectofsolidiﬁcationonAlloy718conditionsusingMICRESS

Herein, the MICRESS software package, which is based on the multiphase-fieldmethod[80–83],isusedtoinvestigatetheeffectof so-lidificationconditionofLaves-phaseformationinAlloy718.Themodels weresetupasfollows.

Modelingwasperformedin2D.Thesizeofthemodellingdomain was80μm×80μmwith0.1μmgridspacing.Aconstantthermal gra-dientwasappliedfromtoptobottomofthedomain.Aconstant cool-ingratewasappliedtothewholedomain.Thismimicsthemovingof theliquidusisothermfrombottomtotopinthesimulationdomain.At thebeginningofthesimulation,agrainwhichhasalmostﬂatliquid/𝛾 interface withrandomnoisewassetatthebottomof thesimulation domain. Theorientationof thisgrain wassetso thatits fast-growth directionisparalleltotheappliedthermalgradientdirectioninthe do-main.Thismimicsthegrowthofepitaxialdendritesfromasubstrate orthere-meltedlayers.Duringthesimulation,thenucleationofnew𝛾

grainsweresetrandomlyintheliquid.Forsimplicity,onlytheliquid/𝛾

interface wasmodeled asananisotropicinterface havingcubic crys-talanisotropy.Therestofthesimulationparameters,including simpli-fiednominalAlloy718composition,nucleationofLavesphase, interfa-cialenergyvalues,interfacethickness,andinterfacialstiffness/mobility coefficientforanisotropyweretakenfrom[23].Thermodynamicand mobilitydataweretakenfromtheTCNI8andMOBNI4databasesfrom Thermo-Calc[84].

TheMICRESSmodelsetupweusedhereissimilartothemodelsetup reportedinNieetal.[30].However,inNieetalhasusedastochastic modellingapproachandmodelledtheAlloy718systemasNi-Nb

bi-Fig.A1. (a),(c)PredictedmorphologyoftheLavesphaseattwodiﬀerentthermalconditionsand(b),(d)respectiveexperimentallyobservedLavesphasemorphology inthemicrostructure(size80μmx80μm).

(16)

narysystemwithoutconsideringtheLavesphaseformationexplicitly. WhereasinourcasewehavemodelledtheAlloy718as6element sys-tem(Ni-Cr-Fe-Mo-Nb-Ti)andexplicitlymodelledtheformationofLaves phaseinthemicrostructure.Inadditiontothat,Nieetal.haveusedﬁxed orientationforthenew𝛾 grainsthatformintheliquid.Themodel pre-dictionsinourcurrentstudywerealsoinabetteragreementwiththe experimentalobservations(seeFig.A1)reportedinNieetal.[30]. Appendix: B

Directageingheat-treatmentsimulationsusingMICRESS

Herein,DA,asper AMS5383,wassimulatedusing MICRESS.The microstructure resulted from theL-DED thermal cycle simulationin [23] wasusedastheinitialstartingmicrostructure.Duringthe simu-lation,onlythenucleation(randomly)of𝛾″inthematrixwassetfor simplicity. Thegrid resolution of the L-DEDsimulation was 50 μm. Therefore,thecorrectsizeandmorphologyof𝛾″couldnotberesolved correctlyunderthisgrid resolution.Toovercome this,theanalytical curvaturemodel[82,85,86]implementedin MICRESSwasused.The anisotropyof𝛾 /𝛾″wasneglected,andanisotropicinteractionmodel wasadopted.Aninterfacialenergyof1×10−5 _J/cm2_[87]_was_used.

NophaseinteractionbetweentheLavesphaseand𝛾″wasmodeled.The restoftheparameters,includingsimpliﬁedAlloy718composition,were takenfrom[23].

Appendix: C

Continuouscoolingexperimentswithoutpressure

HIPedEB-PBFAlloy718sampleswereusedforthecontinuous cool-ingexperiments.Samples(17mm×15mm×55mm)wereprintedwith anArcamA2XmachineusingArcamStandardparameters.Beforethe continuouscooling,thesampleswereagainhomogenizedat1120°Cfor 1h.Theexperimentsrelatedto0.5°C/minand2°C/mincoolingrates wereconductedusinganaluminatubefurnace(modelR120/500/13, NaberthermGmbH,Germany)inaninertargonatmosphere.Therestof thecoolingrateexperimentswereconductedusingaGleeble3800D sys-tem(DynamicSystemsInc,Poestenkill,NY,USA).Allthecoolingrates weremaintainedupto500°C,andthereafterthesampleswerequenched toroomtemperature.Thesamplesweremetallographicallypreparedfor hardnessevaluation.Hardnessmeasurementswereperformedusinga ShimadzuHMV-2microhardnesstesterwith1kgloadand15sdwell time.

Continuouscoolingexperimentswithpressure

ThecontinuouscoolingratestudiesinsidetheHIPfurnace(Model QIH21,QuintusTechnologies,Sweden)wereperformedunderpressure. EB-PBFAlloy718sampleswereﬁrstHIPed at1120°C/100MPa for 4hbeforebeingcontinuouslycooledat5°C/minand20°C/min. How-ever,duringcooling,thepressuregraduallydroppedwithtemperature toavalueof40 MPa.Thesamplesweremetallographicallyprepared forhardnessevaluation.Hardnessmeasurementswereperformedusing aShimadzuHMV-2microhardnesstesterwith1kgloadappliedfor15 s.

References

[1] S.S. Babu , et al. , Additive Manufacturing of Nickel SuperAlloy 718 s: Opportunities for Innovation and Challenges Related to Qualiﬁcation, Metall. Mater. Trans. A 49 (9) (Sep. 2018) 3764–3780 .

[2] T. DebRoy , et al. , Additive manufacturing of metallic components – Process, structure and properties, Prog. Mater. Sci. 92 (Mar. 2018) 112–224 .

[3] M.M. Attallah , R. Jennings , X. Wang , L.N. Carter , Additive manufacturing of Ni-based superAlloy 718 s: The outstanding issues, MRS Bull 41 (10) (2016) 758–764 .

[4] S. Sui , et al. , The inﬂuence of Laves phases on the room temperature tensile properties of Inconel 718 fabricated by powder feeding laser additive manufacturing, Acta Mater 164 (2019) 413–427 February .

[5] M. Anderson , A.-L. Thielin , F. Bridier , P. Bocher , J. Savoie , 𝛿 Phase precipitation in Inconel 718 and associated mechanical properties, Mater. Sci. Eng. A 679 (Jan. 2017) 48–55 .

[6] G.A. Knorovsky , M.J. Cieslak , T.J. Headley , A.D. Romig , W.F. Hammetter , INCONEL 718: a solidiﬁcation diagram, Metall. Trans. A 20 (10) (Oct. 1989) 2149–2158 . [7] R.W. Hayes , Creep Deformation of Inconel Alloy 718 in the 650°C To 760°C, in:

SuperAlloy 718 s 718, 625 and Various Derivatives, 1991, 1991, pp. 549–562 . [8] C. Kumara , Microstructure Modelling of Additive Manufacturing of Alloy 718, 2018 . [9] J.F. Radavich , The physical metallurgy of cast and wrought alloy 718, SuperAlloy

718s 718 Metall. Appl. (1989) 229–240 .

[10] M. Sundararaman , P. Mukhopadhyay , S. Banerjee , Precipitation of the Delta-Ni3Nb phase in two nickel base superAlloy 718s, Metall. Trans. A 19 (3) (1988) 453–465 . [11] S. Azadian , L.Y. Wei , R. Warren , Delta phase precipitation in inconel 718, Mater.

Charact. 53 (1) (2004) 7–16 .

[12] T. Antonsson , H. Fredriksson , The eﬀect of cooling rate on the solidiﬁcation of IN- CONEL 718, Metall. Mater. Trans. B 36 (1) (Feb. 2005) 85–96 .

[13] J.T. Tharappel , J. Babu , Welding processes for Inconel 718- A brief review welding processes for Inconel 718- a brief review, IOP Conf. Series: Materials Science and Engineering, 330, 2018 .

[14] D. Deng , R.L. Peng , H. Söderberg , J. Moverare , On the formation of microstructural gradients in a nickel-base superAlloy 718 during electron beam melting, Mater. Des. 160 (2018) 251–261 .

[15] Y. Zhang , Z. Li , P. Nie , Y. Wu , Eﬀect of cooling rate on the microstructure of laser-remelted INCONEL 718 coating, Metall. Mater. Trans. A 44 (12) (Dec. 2013) 5513–5521 .

[16] J.K. Tien , T. Caulﬁeld , SuperAlloy 718s, supercomposites and superceramics, Aca- demic Press, 1989 .

[17] W.M. Haynes , CRC Handbook of Chemistry and Physics - Google Books, 97th ed., CRC Press, 2016 .

[18] R. Trivedi , W. Kurz , Solidiﬁcation microstructures: a conceptual approach, Acta Met- all. Mater. 42 (1) (1994) 15–23 .

[19] Z. Wang , K. Guan , M. Gao , X. Li , X. Chen , X. Zeng , The microstructure and mechanical properties of deposited-IN718 by selective laser melting, J. Alloy 718s Compd. 513 (2012) 518–523 .

[20] D. Deng , R.L. Peng , H. Brodin , J. Moverare , Microstructure and mechanical properties of Inconel 718 produced by selective laser melting: sample orientation depen- dence and eﬀects of post heat treatments, Mater. Sci. Eng. A 713 (October 2017) 294–306 2018 .

[21] S. Ghosh , et al. , Single-track melt-pool measurements and microstructures in Inconel 625, Jom (2018) February .

[22] S. Ghosh , M.R. Stoudt , L.E. Levine , J.E. Guyer , Formation of Nb-rich droplets in laser deposited Ni-matrix microstructures, Scr. Mater. 146 (2017) 36–40 .

[23] C. Kumara , et al. , Microstructure modelling of laser metal powder directed energy deposition of Alloy 718, Addit. Manuf. 25 (2019) 357–364 .

[24] C. Kumara , D. Deng , F. Hanning , M. Raanes , J. Moverare , P. Nylén , Predicting the Microstructural Evolution of Electron Beam Melting of Alloy 718with Phase-Field Modeling, Metall. Mater. Trans. A Phys. Metall. Mater. Sci. 50 (5) (2019) 2527–2537 . [25] H. Qi , M. Azer , A. Ritter , Studies of standard heat treatment eﬀects on microstructure and mechanical properties of laser net shape manufactured INCONEL 718, Metall. Mater. Trans. A 40 (10) (Oct. 2009) 2410–2422 .

[26] W. Huang , J. Yang , H. Yang , G. Jing , Z. Wang , X. Zeng , Heat treatment of Inconel 718 produced by selective laser melting: Microstructure and mechanical properties, Mater. Sci. Eng. A 750 (2019) 98–107 February .

[27] G.D. Janaki Ram , A. Venugopal Reddy , K. Prasad Rao , G.M. Reddy , J.K. Sarin Sun- dar , Microstructure and tensile properties of Inconel 718 pulsed Nd-YAG laser welds, J. Mater. Process. Technol. 167 (1) (Aug. 2005) 73–82 .

[28] X. Li , et al. , Eﬀect of heat treatment on microstructure evolution of Inconel 718 Alloy 718 fabricated by selective laser melting, J. Alloy 718s Compd. 764 (2018) 639–649 .

[29] J.J. Schirra , R.H. Caless , R.W. Hatala , The eﬀect of laves phase on the mechanical properties of wrought and cast + HIP Inconel 718, in: SuperAlloy 718s 718, 625 and Various Derivatives, 1991, 1991, pp. 375–388 .

[30] P. Nie , O.A. Ojo , Z. Li , Numerical modeling of microstructure evolution during laser additive manufacturing of a nickel-based superAlloy 718, Acta Mater 77 (Sep. 2014) 85–95 .

[31] F.D.J. Kurz W. , Fundamentals of solidiﬁcation, 1988, CRC Press, 1998 June 1 . [32] Ø. Grong , Metallurgical modelling of welding, Institute of Materials, 1997 . [33] A.R. Balachandramurthi , J. Olsson , J. Ålgårdh , A. Snis , J. Moverare , R. Pederson , Mi-

crostructure tailoring in Electron Beam Powder Bed Fusion additive manufacturing and its potential consequences, Results Mater 1 (2019) 100017 August .

[34] N. Raghavan , et al. , Numerical modeling of heat-transfer and the inﬂuence of process parameters on tailoring the grain morphology of IN718 in electron beam additive manufacturing, Acta Mater 112 (2016) 303–314 .

[35] G. Lindwall , et al. , Simulation of TTT curves for additively manufactured Inconel 625, Metall. Mater. Trans. A 50 (1) (Jan. 2019) 457–467 .

[36] Y. Tian , et al. , Rationalization of microstructure heterogeneity in INCONEL 718 builds made by the direct laser additive manufacturing process, Metall. Mater. Trans. A 45 (10) (Sep. 2014) 4470–4483 .

[37] R.G. Carlson , J.F. Radavich , Microstructural characterization of cast 718, SuperAlloy 718s 718 Metall. Appl. (1989) 79–95 .

[38] A. Segerstark , J. Andersson , L. Svensson , O. Ojo , Microstructural characterization of laser metal powder deposited Alloy 718, Mater. Charact. 142 (Aug. 2018) 550–559 June .