Quantitative analysis of the effect of inoculation and magnesium content on compact graphite irons — Experimental approach

HAL Id: hal-03185484

https://hal.archives-ouvertes.fr/hal-03185484

Submitted on 30 Mar 2021

HAL is a multi-disciplinary open access

archive for the deposit and dissemination of

sci-entific research documents, whether they are

pub-lished or not. The documents may come from

teaching and research institutions in France or

abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est

destinée au dépôt et à la diffusion de documents

scientifiques de niveau recherche, publiés ou non,

émanant des établissements d’enseignement et de

recherche français ou étrangers, des laboratoires

publics ou privés.

magnesium content on compact graphite irons

-Experimental approach

Anna Regordosa, Urko de la Torre, Jon Sertucha, Jacques Lacaze

To cite this version:

Anna Regordosa, Urko de la Torre, Jon Sertucha, Jacques Lacaze.

Quantitative analysis of

the effect of inoculation and magnesium content on compact graphite irons - Experimental

ap-proach.

Journal of Materials Research and Technology, Elsevier, 2020, 9 (5), pp.11332-11343.

OATAO is an open access repository that collects the work of Toulouse

researchers and makes it freely available over the web where possible

Any correspondence concerning this service should be sent

to the repository administrator:

tech-oatao@listes-diff.inp-toulouse.fr

This is a Publisher’s version published in:

http://oatao.univ-toulouse.fr/27592

To cite this version:

Regordosa, Anna and de la Torre, Urko and Sertucha, Jon and Lacaze,

Jacques

Quantitative analysis of the effect of inoculation and magnesium

content on compact graphite irons — Experimental approach. (2020) Journal

of Materials Research and Technology, 9 (5). 11332-11343. ISSN 2238-7854

https://www.journals.elsevier.com/journal-of-materials-research-and-technology Availableonlineatwww.sciencedirect.com

Original

Article

Quantitative

analysis

of

the

effect

of

inoculation

and

magnesium

content

on

compact

graphite

irons

—

Experimental

approach

Anna

Regordosa

_,

_Urko

_de

_la

_Torre

_,

_Jon

_Sertucha

_,

_Jacques

_Lacaze

a_{InvestigaciónyDesarrollodeProcesosMetalúrgicos,Azterlan,BasqueResearchTechnologicalAlliance,Durango,Bizkaia,Spain} b_{CIRIMAT,ENSIACET,UniversitédeToulouse,Toulouse,France}

a

r

t

i

c

l

e

i

n

f

o

Articlehistory:

Received5February2020 Accepted2August2020 Availableonline23August2020

Keywords:

Compactgraphiteiron Magnesiumfading Inoculation Solidification Thermalanalysis

a

b

s

t

r

a

c

t

Inmanyindustrialdomains,compact graphitecastirons aredevelopingrapidly atthe expenseoflamellargraphiteirons.Theformationofthemicrostructure,especiallygraphite shape,duringthesolidificationstageofthesealloysishoweverstillnotclearlyunderstood, showingcharacteristicssimilartobothlamellarandspheroidalgraphiteirons.Theaimof thisworkwastoprovidequantitativeinformationonthesolidificationofcompactgraphite castirons,whetherinoculatedornot,bythermalanalysisinthefoundryshop.Thisincludes theeffectoftheamountofnodularizerontheundercoolingbeforesolidificationstartsand theamountofrecalescencewhenbulkeutecticsolidificationsetsup.Comparing quanti-tativeanalysisoftheas-castmicrostructurewiththecharacteristicsofthecoolingcurves giveshintstoabetterunderstandingofthemicrostructureformationincompactgraphite irons.Thisworkthusprovidesasetofquantitativedatanecessarytoverifytherelevanceof anysolidificationmodellingapproachforcompactgraphitecastiron.Onapracticalpointof view,itsuggeststhatthermalanalysiscouldcertainlybeausefulmeansforcontrolofmelt preparationforCGIcastingbyaddingverylowlevelofinoculantinthestandardthermal analysiscups.

©2020TheAuthor(s).PublishedbyElsevierB.V.Thisisanopenaccessarticleunderthe CCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc-nd/4.0/).

1.

Introduction

Compared to lamellar graphite cast irons (LGI), compact graphitecastirons (CGI)have improvedpropertiesand are replacing them formoreand morecomponents [1]. As for other cast irons, alloying allows fine tuning of the final mechanical properties [2] mainly bymodifying the matrix microstructure.Bazdaretal.[3]showedthatthese mechan-ical properties depend also on the so-called compactness

∗ _{Correspondingauthor.}

E-mail:jacques.lacaze@ensiacet.fr(J.Lacaze).

ofgraphiteparticleswhichtheychangedbysulfuraddition. Nevertheless,thesatisfactoryproductionofCGIissomewhat difficultbecausethedesiredgraphitedistribution,i.e.a mix-tureofcompactandspheroidalparticles,isverysensitiveto variousprocessparameters.Thecoolingratebecomes preva-lentinthinwalledcastingasinvestigatedbyCharoenvilaisiri et al.[4]. Also,chemicalcomposition, andinparticularthe presenceofminorelementssuchasAl,SandTi[5–7],affects the compactnessasdoesalsotheinoculationlevel[8].The needforadefinitionofasimpletestingprocedureensuring lownodularityandmaximumcompactnessishamperedby thefactthataclearunderstandingofcompactgraphite for-mationisnotyetavailable[9].

https://doi.org/10.1016/j.jmrt.2020.08.008

2238-7854/©2020The Author(s). PublishedbyElsevier B.V.Thisis anopen accessarticleunderthe CC BY-NC-NDlicense (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

ThecheapestandmostcommonmethodforproducingCGI consistsinusinglowadditionofanodularizer,whichismost oftenanFeSiMgalloycontainingrareearthadditions. Check-ingthattheamountofnodularizerinthemeltiscorrectcan bedonepriortopouringbycontrollingtheoxygenlevel[10]

althoughthisisnotacommonprocedure.Infact,therecent developmentofCGIreliesonthermalanalysis[11].Thermal analysis(TA)usestheparticularitythatbulkeutectic solidi-ficationofnon-inoculatedCGIstartsatahighundercooling, whichissometimeslarger[12,13]andsometimessmaller[14]

thanthatencounteredwithspheroidalgraphiteirons(SGI). Oncesolidificationhasstarted,recalescenceishighasforLGI unlesscementite precipitates concurrently [15]. Becauseof thesecomplicatedfeatures,thecharacteristicsofthethermal recordsarehighlyscattered[16].Toovercomethisproblemin theuseofTAformeltcontrol,acomplexprocedureusinga specialcruciblewithtwothermocoupleshasbeendeveloped bySintercast,asindicatedinDawson[11].Analternativewas soughtbySunetal.[17]whosuggestedapatternrecognition basedmethodusingadatabasewherepreviousrecordsand analysesarestored.

Duetothedeepundercoolingbeforebulkeutectic solid-ification,CGIisinoculated.Thisincreasesnodularityinthin sectioncastings[9].Itwouldthusbeofinterestinbeingable to correlate characteristic features of thermal analysis of non-inoculated CGI with final microstructure of the same alloysbutinoculated.Suchacorrelationshouldbeestablished foravariablelevelofnodularizer,whichcanbeachievedby holdingthemeltforanincreasingamountoftime,aswas donebyHernando etal.[18]andJinhaietal.[13],amongst others.Thepresentstudywasdedicatedtothestudyofthe effectofholdingtimeandinoculationrateontheformation ofcompact graphite using an alloy held liquid forseveral hoursinalargepressurizedpouringunit.Atregularintervals, twoTAcupswerefilledwithliquidandtheircoolingcurves recorded.Oneofthecupscontainedaninoculant,theother didnot. Thechange in coolingrecords and the associated evolution of the microstructure during melt holding are presented. This provides quantitative information on the effectsofinoculationandnodularizercontentthatcouldlater beusedtovalidateamodellingapproachorbeenteredintoa databaseformeltcontrol.

2.

Experimental

procedure

Attheendofanormalproductionshift,4tofcastironwere leftinthe8tpressurizedpouringunitoftheBetsaideS.A.L. foundry(Basque Country,Spain)where thetests were car-riedout.Thiscastironhadbeenpreparedfortheproduction ofspheroidalgraphitecastingsusingaFeSiMgspheroidizer containingsomeceriumandlanthanum[19].Justbeforeeach pour,thepouringbasinwasfilledtwicetoensurechemical homogeneity,especially of the magnesium. Approximately every25min,asetofanalyseswasmade,andamelt sam-pletakenoutfromthepouringbasinwasusedtofilltwoTA cups and toobtain a medalsample,which was thenused todeterminethechemicalcompositionofthealloy.Oneof theTAcups was emptywhenit wasfilledwhilethe other onecontained 0.35gof acommercial inoculant (grain size

0.2–0.5mm,Si=69.9,Al=0.93,Ca=1.38,Bi=0.49,RE=0.37and Febalance,wt.%),i.e.about0.10wt.%ofthesampleweight pouredinthecup.Timesatwhichsamplingwascarriedout were controlledsoastomonitortheevolution ofthe alloy duringholdingfor8hinthepress-pour.The19castingswere identifiedwithaletterfromAtoSandasubscript“no-inoc” and“inoc”fornotinoculatedandinoculatedalloys, respec-tively. During holding, the temperatureof theliquid metal increased so that the first temperature recorded with the thermalcup–denotedpeaktemperature,Tpeak-variedfrom

1283to1367◦C.Theprecisevaluesarereportedwith metallo-graphicresults.

The composition of the collected medals was analysed by combustion(LECOCS300) forcarbon and sulfurand by sparkspectrometry(SPECTROLAB)forallotherelements.The initialcomposition(wt.%)inthemainelementsislistedin

Table1,whichdoesnotincludethecontributionofthe inoc-ulant addition; the alloy contained also <0.01 Mo, <0.01V, <0.01 Al and <0.005 Co, and had about 0.05 Cr and 0.006 Sn. The whole set of compositions is listed in Appendix

A.

Duringholding,thecarbonandsiliconcontentsdecreased respectivelyto3.65and2.39wt.%,whiletheevolutionofMg, CeandLawasmoremarked.Fig.1presentstheevolutionof thesefiveelementsasnormalizedwiththevaluesinTable1. Thecontentinallotherelementswasunchangedduringthe holdingtime.ThecarbonequivalentCEofthemeltwas calcu-latedbyusingthetwofollowingexpressions,CE99asdescribed

inAppendixB[20]andCEASM[21]:

CE99= wC+ 0.28·wSi+ 0.007·wMn

+ 0.092·wCu+ 0.303·wP (1)

CEASM= wC+ 0.31·wSi− 0.028·wMn

+ 0.076·wCu+ 0.331·wP (1’)

Fig.1–Effectofholdingtimeonthecontentofthemeltin C,Si,Mg,CeandLarelativetothecompositionofthefirst sample,andofthecalculatedcarbonequivalentCE accordingtothetwoEqs(1)and(1’).

Table1–Chemicalcompositionofthefirstmedalsample(wt.%).

C Si Mn S Cu Mg Ti Ce La

3.75 2.45 0.64 <0.005 0.85 0.043 0.021 0.0130 0.0051

Fig.2–Typicalexampleofcoolingcurveanditstime derivative(non-inoculatedQalloy)withdefinitionofthe characteristictemperatures(openarrows),seetext.The solidarrowindicatesanabruptrecalescencewhichis referencedtolaterinthemaintext.

wherewiisthecontentinwt.%ofelement“i”.ThetwoCE

val-uesareplottedinFig.1whereitisseenthattheyrunparallelto eachotherwithCEASMshiftedupwardsbyanamountofabout

0.04wt.%C.Furthercomparisonofthesetwoexpressionsis providedinAppendixB.BothCE99andCEASMdecreased

dur-ingtheholding,thoughremainingabovetheeutecticvalue whichmeansthemeltremainedhypereutecticallalongthe experiment.

TheThermolan® softwarewasusedtorecordthecooling curveswhichwerethenredrawnandanalysedasillustrated withFig.2.Relevantcharacteristictemperatures werethen evaluated, namelythe maximumor peak temperaturejust afterpouring,Tpeak,theso-calledliquidustemperature,TLA,

the minimum eutectic temperature, TE,min, the maximum

eutectictemperatureduringtheeutecticplateau,TER,andthe

solidustemperature, Tsolidus. Thisevaluationwas made by

directreadingofthecoolingcurverecordforTpeak,TE,minand

TER.Thedifferencebetweentheselasttwotemperaturesisthe

recalescence:R=TER-TE,min.TheTLAtemperaturecorresponds

tothetemperatureatwhichthesolidificationisfirstsensedby thethermocouple.Forhyper-eutecticalloysasthose investi-gatedhere,itcorrespondstotheappearanceofausteniteafter primarydepositionofgraphite.Inpractice,TLAwasevaluated

asthetemperatureatwhichthederivativeofthecoolingcurve showsaslopechangeattheendofliquidcooling.Thelocation ofthischangeisindicatedwiththedashedlineinFig.2.Inthis case,whichisshown,theslopechangeispositive,butitcould aswellbenegativeincaseswherethecoolingrateiscloseto0 whensolidificationissensed.Accordingly,thedetermination ofTLAmaysometimesbeinaccurate.Finally,thesolidus

tem-peraturewasdeterminedascorrespondingtoaminimumin thesecondderivativeofthecoolingcurve.

TheTAcupsampleswerethenpreparedformetallographic inspections.Micrographsofthreedifferentfieldsweretakenat amagnificationof×100inthecenterpartofthesamples,close tothethermocouplejunction.Thegraphitedistributionand shapewerethenevaluatedwithanimageanalysissoftware. Thegraphiteparticlesweresortedaccordingtothestandard forspheroidalgraphite,classIIIforirregularprecipitatesand classVandVIforirregularandwell-shapedspheroids, respec-tively.Inthepresentstudy,classIIIparticlesstandforcompact graphite.Thecount(indexC)andarea(indexA)fractionsof eachclasswerethenevaluatedandnormalizedwiththetotal countandareaofgraphite,i.e.onehasfIIIC+fVC+fVIC=1and

fIII A+fV A+fVI A=1.Nodularitycouldbeexpressedasthesum

fV C+fVI C.Then,the structureofthe sampleswaschecked

byetchingthepolishedsurfaceswithNital5%tolookforthe presenceofeutecticcementite.Whencarbideswereobserved, theirareafraction(fcarbides)wasdeterminedontheetched

sur-faces. However,forgetting theproper contrast,theimages wereprocessedinsuchawaythatferriteinledeburitewas countedascementite.Dependingonthemicrostructureofthe whiteeutectic,thefractionofcarbidesreportedinthepresent workmaythusaswellrepresenttheamountofwhiteeutectic. Anattemptwascarriedout tocharacterizethesizeand number of theeutectic cells onmetallographic sectionsof the non-inoculated samplesafter etching. Incase of sam-pleswithoutcarbides,threemicrographsat50×magnification were usedanddelimitationofthecellswasbasedon iden-tification ofeutecticcells boundaries. When carbides were present, they delineated the contours of the eutectic cells whichwerethenidentifiedonthreemicrographsat25× mag-nification.ThemaximumdiameterDCellofalleutecticcells

andtheirsurfacecountNCellwerethendetermined.Because

thenumberofcellswasquitelowinnon-inoculatedsamples, theirsizecouldbebestrepresentedbytheaverageofthefive largestdiametervaluesfound,whichwasusedasDCellvalue.

3.

Results

Solidificationofallinoculatedalloysshowedasingle eutec-tic plateau with a maximum temperature, TER, which did

notchangemuchduringtheholdingtime.Thisisillustrated withtherecordsforAinocandSinocsamplesinFig.3-awhere

thestable,TEUT,andmetastable,TEW,eutectictemperatures

determinedforalloyAinochavealsobeendrawnasdashed

hor-izontallines.Thesereferencetemperaturesarerespectively equalto1165.4and1119.4◦Canddidnotchangemuchwith holdingtime, seeAppendixBfortheirevaluation. CurveA inFig.3-afurtherillustratesthatbulkeutecticsolidification proceededwithsomerecalescencebeforetheeutecticplateau wasreached.Whenholdingtimeincreased,therecordswere seen tofirstflattenandfinallypresentaminimumlocated inthemiddlepartoftheeutecticplateauasclearlyseenin

Fig.3–Inoculatedsamples:coolingcurvesoftrialsAandS(a),andmicrographofsamplesAinoc(b)andSinoc(c).Theopen

arrowshowstheminimumtemperatureintheplateauofcurveS(seetextfordetails).

Fig.3-aforalloySinoc(openarrow).Thiscouldevidencethat

thebulkeutecticreactionofinoculatedalloystookplacein twosuccessivesteps,seebelow.

Typicalmicrographsofthefirst(Ainoc)andlast(Sinoc)

sam-plesareshowninFig.3-bandc,respectively.Theyillustrate anevolutionfromafullyspheroidalgraphiteirontoanearly compactonewiththe holdingtimeduetothe decrease in the content ofnodularizing elements,see Fig. 1. Itshould benotedthatthesmallnodulesinsampleAinoc have

com-pletelydisappearedandhavebeenreplacedintheSinocsample

bycompactgraphiteparticles.Thissuggeststhatthe precip-itationofgraphiteininoculated alloysbeginswithprimary spheroidsthatgivethelargenodulesthatappearwith sim-ilarsize inallinoculated samples.Compact graphitecould thusappearlaterduringthe bulkeutecticreactionleading, eventually,tothetwostepplateauobservedforalloySinoc.

Theabovequalitativeanalysisissustainedbythe quan-titativeresultsillustratedinFig. 4wherethe changes with holdingtimeoftherelativeareafractionsfIIIA,fVAandfVIA

arereportedforthewholesetofinoculatedsamples.Itisseen thatfVA,whichcouldbeassociatedwiththelargegraphite

spheroids,isnearlyconstant.Thisthusconfirmsthatthese arethesmallnoduleswhichareprogressivelyreplacedwith compactgraphite.Inthisseriesofinoculatedalloys,thetotal graphite fraction, fgraphite, wasnearly constant at0.08–0.10

andnocementitewasobserved.Allmicrostructuredataare collectedinAppendixC.

Thecharacteristictemperaturesandrecalescenceforthe TArecordsofinoculatedalloyshavebeenplottedinFig.5as functionofholdingtime.ItisseenthatTLAandTE,minremained

significantlylowerthan TEUT inlinewiththedescriptionof

thermal analysisrecordsforhyper-eutectic alloys byHeine

[22].BothTLAandTE,minincreasedwithholdingtime,whichis

furtherdetailedinSection4.Forthepresentseriesofresults on inoculated alloys, it isalso noticed anoverall decrease of the recalescence amplitude. For all samples, Tsolidus is

aboveTEW inagreementwiththefactthatnocementitewas

observedforanyoftheinoculatedalloys.Owingtothe uncer-taintyrelatedtotheuseofthesecondderivativeforevaluating Tsolidus,theslightvariationofthistemperatureseeninFig.5

maynotbesignificant.TheresultsinFig.5willbediscussed furtherlater.

Fig.4–Changewithholdingtimeoftherelativefractionsof graphite,fIII-A,fV-AandfVI-A,forallinoculatedsamples. AfterRegordosaetal.[19].

Fig.5–Evolutionofthecharacteristictemperaturesandof recalescenceduringholdingforinoculatedalloys.

Fig.6-ashowstheTArecordsfromafewnon-inoculated alloys,namelysamplesAno-inoc,Cno-inocandSno-inoc.In

con-trastwiththecaseofinoculatedalloys,theserecordschange asafunctionofholdingtime.Thisisinlinewiththe signifi-cantmicrostructurechangesobservedonthemicrographsin

Fig.6-b,candd.Itshouldbenotedthatthemicrostructure oftheAno-inocsampleconsistsmainlyofspheroidal

precip-itatesandthat itsTA recordissimilar tothoseinFig. 3-a. Solidificationofthefollowingnon-inoculatedsamplesfrom BtoSoccurred intwostages, withafirstshort plateauat 1139−1149◦_{Candamainplateauatlowertemperature.This}

latter temperature was observed to decrease with holding time.AfterNitaletching,samplesfromCno-inoctoSno-inocdid

containanincreasingamountofcarbidesasillustratedwith

Fig.6-eandf.

Quantitative microstructuredataare presentedinFig.7. InFig.7-a,thesurfacefractionsofgraphiteandcarbidesare plottedasafunctionofholdingtimeandshowthegeneral trendthatisexpected,namelyadecreaseintheamountof graphiteascementiteisincreasinglypresent.InFig.7-bare seen theevolutionsofthe areafractionsofthethreetypes ofgraphite.Duringthefirst100min,theamountoftypesV andVIspheroidalgraphitedecreaseswhilethatofcompact graphitesignificantlyincreases.Atlargerholdingtimes,the amountsofthevarioustypesofgraphiteremainnearly con-stant.Accordingly,thechangesseenintheTAcurvesaredue tothe decreaseofthe graphiteamountand theassociated increaseofcementitedepictedinFig.7-a.

Fig. 8 shows the evolution with holding time of the characteristictemperaturesandofrecalescenceforthe non-inoculatedsamples.ItisseenthatTLAincreasessignificantly

asinthecaseofinoculatedsampleswhileTE,minfirstdecreases

rapidlyduringthefirst100minandthenmoreslowlywhenit takesvaluesbelowTEW.TsolidusisbelowTEWforallalloysand

doesnotchangemuchwithholdingtime.However,itisworth stressingthatnocementitewasobservedinalloysAno-inocand

Bno-inocthoughTsoliduswasbelowTEW.

Recalescence first increases as graphite gets more and more compactinstead ofspheroidal, and it shows a max-imum corresponding toalloy Dno-inoc. Such amaximum is

inagreementwithpreviousreports[15].Afterfurther hold-ing,recalescencedecreasescontinuouslytonearlyzerowhen thestructureismostlywhite.ForalloysCno-inocandDno-inoc,

TE,minwasaboveTEW,meaningcarbidesshouldhaveappeared

in thesesamplesatthe end ofsolidification. Inall follow-ing samples, TE,min was belowTEW so that cementite may

have appeared at the beginning of the second plateau or later towardstheendofsolidification. ForsamplesHno-inoc

toSno-inoc,thesecondplateauwasentirelylocatedbelowTEW

andasmallbutabruptrecalescencecouldoftenbeobserved suchasthatindicatedbythesolidarrowinFig.2.Thisthermal arrestcouldpossiblyberelatedtotheappearanceofledeburite asdiscussedelsewhere[23].

4.

Discussion

Due to the hyper-eutectic composition of the alloys, their solidificationbeganwithaprimaryprecipitationofgraphite whenthetemperaturedroppedbelowthegraphiteliquidus. However,thisprecipitationdoesnotleadtoamarkedthermal arrestonTArecordsbecausetheamountofprimarygraphite remainsquitelimitedevenforinoculatedalloys[24].Whenthe temperaturedecreases further,conditionsforaustenite for-mationarefinallyreachedandausteniteprecipitationbegins. Thisleadstoathermalarrestthathasbeen labelledTLA in

Fig.2.ThevaluesforTLA forinoculatedandnon-inoculated

alloysarecomparedinFig.9wherehavealsobeenplottedthe TE,minvaluesforinoculatedalloysandthecalculatedeutectic

temperature,TEUT.

ItisfirstnoticedinFig.9thatTEUT remainednearly

con-stantallalongtheexperiments,inagreementwiththefact thatthecontentinsiliconofthemeltdidnotchange

signif-Fig.6–Non-inoculatedsamples:coolingcurvesoftrialsA,CandS(a);micrographsofsamplesAno-inoc(b),Cno-inoc(c)and

Sno-inoc(d)beforeetching,andofsamplesCno-inoc(e)andSno-inoc(f)afterNitaletching.Thescalebaris200␮mlongforall

micrographs.

icantly,seeFig.1and thecalculationformulaeinappendix

B.ConsideringtheTLAvalues,itisthusseenthattheliquid

wasstronglyundercooledwithrespecttotheeutecticwhen austenitestarteddevelopingatthecenteroftheTAcups,with anundercoolingamountingto20−25◦_{Catshortholdingtimes}

andmorethan10◦Cattheendoftheexperiments.Two rea-sonsrelatedtothe presentexperimentsleadtoexpectthe observedincreaseinTLAwithholdingtime.Thesereasonsare,

first,thedecreaseofthealloycarboncontentand,second,the decreaseofthecoolingratebecauseoftheincreaseinpeak temperature.

Asa matteroffact,the 0.1wt.%decrease inthe carbon contentfromthebeginningtotheendofthetrialsaccounts forabouta10◦Cchangealongtheausteniteliquidus[22],i.e. itexplainsmostofthechangeinTLA recordedforthe

Fig.7–Evolutionwithholdingtimeoftheamountofgraphiteandcementite(a)andoftherelativefractionsf_IIIA,f_{V A}and

fVI A(b)inthenon-inoculatedsamples.AfterRegordosaetal.[19].

Fig.8–Evolutionwithholdingtimeofthecharacteristic temperaturesandofrecalescenceforthenon-inoculated samples.

could possiblybeaccountedforbythe decrease incooling rate,whichgivesmoretimeforprimarygraphitegrowth,thus movingthesolidificationpathclosertothegraphiteliquidus andhenceincreasingthetemperatureatwhichaustenitemay appear.

Forinoculatedalloys,theTLAvaluesaresignificantlyhigher

thanthosefornot-inoculatedalloysforshortholdingtimes.1

Thisdifferencehascertainlytodowiththefactthat inocu-lationincreasestheamountofgraphiteprecipitatedduring primarysolidification,thusovertakingpartofthecoolingrate

1 _{Inoculationleadstoan0.07wt.%increaseinsiliconandthus}

toa1.6◦Cdecreaseoftheausteniteliquidus,seeAppendixB, whichisnotconsideredinthepresentdiscussion.

Fig.9– EvolutionwithholdingtimeofTEUT(calculatedfrom thenon-inoculatedalloys’composition),ofT_E,minandT_LA

forinoculatedsamples,andofT_LAfornon-inoculated samples.

effect againbymovingthe solidificationpath closertothe graphiteliquidus.Atincreasingholdingtime,the solidifica-tion path ofinoculated alloys duringprimary precipitation ofgraphiteisthusmuchlesssensitivetochangeinthe car-boncontentandinthecoolingrateasseenwiththeslower increaseinTLAascomparedtonon-inoculatedalloys.

Finally,partofthedifferencebetweenTEUTandTLAmustbe

relatedtoausteniteundercoolingaspointedoutbyHeine[22]. Indeed,calculatingthe(metastable)austeniteliquidusas indi-catedinAppendixBforthecompositionofthetwoextreme alloysAandSgives1148.0and1159.0◦C,respectively.These calculatedtemperaturesareabout10◦Cabovethemeasured

TLA temperatures, thusdemonstratingaustenite

undercool-ing.

Oncetheextrapolationoftheausteniteliquidushasbeen reached,primaryprecipitatesofgraphiteandaustenitemay combinetogiverisetotheeutecticreaction.Duringthisstage, thecompositionoftheliquidisexpectedtocloselyfollowthe austeniteliquidus.Thismayleadtoprecipitationor dissolu-tionofso-calledoff-eutecticausteniteforanyformofgraphite, thoughthis hasbeen quantitatively demonstrated onlyfor spheroidalgraphiteiron[25,26].Thealloyfurtherundercools untilTE,min isreached, which relates to the set-up of bulk

eutecticsolidification. In the caseofinoculated alloys, the TE,minvaluesinFig.9followatrend,whichnearlyparallels

thatofTLA.Thiswasexpectedowingtothefactthatthe

inoc-ulationprocesswasthesameallalongtheexperiments.For non-inoculatedalloys,theevolutionistotallydifferentas dis-cussedbelow.

ThesampleAno-inoc showedthesame solidification

pro-cessastheinoculatedalloyscharacterizedbyasingleeutectic plateau.Thissuggeststhat,forthisshortholdingtime,there weresufficientexogenousparticlesremaininginthemeltto triggergraphite nucleation when this samplewas poured, evenwithouttheadditionofinoculant.Thenumberof exoge-nousparticlesthengraduallydecreasedduringmeltholding, resultinginacontinuousdecreaseinthenumberofprimary graphiteprecipitatesinthenon-inoculatedsamples,ascanbe noticedinthemicrographsinFig.6.Thisdecreaseisalso asso-ciatedtoachangefromspheroidaltocompactgraphiteandto adecreaseofTE,min,whichisrapidatfirstandthenslowsdown

when TE,min islower than TEW. Focusingon alloys Bno-inoc,

Cno-inocandDno-inocwhichallshowedaTE,minvalueaboveTEW,

itthusseemsthatearlydevelopmentofthecompactgraphite cellsishindereduntilsomehigh enoughundercooling has beenreachedatwhichtheirgrowthbecomessignificantand leadstomarkedrecalescence.

The above characteristic of the solidification of CGI as recorded by TA has been stressed since long [15]. It is here associated with a growth hindrance during the early stageofeutecticgrowth.Somesupportofthisanalysiswas previously gained by the observation of deep-etched non-inoculatedsamples[19].Ithasbeenseenthatsomeprimary graphitespheroidscouldevolveincompactgraphitecells dur-ing the first stageof eutecticsolidification, beforeTE,min is

reached.However,someotherspheroids remainedgrowing withoutdevelopingprotuberancesthatwouldleadtocompact graphite,and thus appearedunchanged untilbulkeutectic solidification.

Innon-inoculatedalloysthatshowedsignificantamounts ofcementite,it was foundthat the compactgraphitecells show a rounded or an elongated shape. It may easily be concludedthat cells appearinground got locked whenthe metastable eutectictook place, while elongated cells have grownforawhiletogetherwithledeburite.Thisinteraction betweengrowthofCGcellsandledeburitecouldbe quanti-tativelyillustratedbyrecordingthenumber andsizeofthe compactgraphitecells.Fig.10showsthesedataasfunction ofholdingtime.Averticalinterruptedlinehasbeenaddedto showwhentheeutecticplateaustartedtobeentirelybelow TEW.Itisseenthatthistransitioncorrespondstowhenthe

sizeofthecellsstartedtodecrease,thusconfirmingtheabove

Fig.10–Evolutionwithholdingtimeofthenumberand sizeofcompactgraphitecellsinnon-inoculatedalloys.

Fig.11–Correlationbetweentherelativeamountof compactgraphite,fIIIA,ininoculatedsampleswiththe valuesofT_E,minandRinthecorrespondingnon-inoculated sample.Thegreyedareacorrespondstoinoculatedalloys withcompactgraphite.

schematic.Itisworthnotingthatthesizeofthecellsreported herearequitelargerthanthosemeasuredbyPanetal.[24].

Finally, it is of interest to go back to the question of the capability of thermal analysis as a means to control meltpreparationbeforecasting.Duringholdingofthemelt, thequantityofnodularizingelementsdecreasesuntil inoc-ulated alloys solidify mainly as compact graphite while non-inoculatedalloyssolidifymainlyinthemetastable sys-tem. This suggested to plot the fraction fIIIA of compact

graphiteininoculatedsamplesasfunctionofTE,minandR

mea-suredforthenon-inoculatedones.ThisisdoneinFig.11where itisseenthatinoculatedalloysthatcanbeconsideredas com-pactgraphiteiron,i.e.havingfIIIAhigherthan0.60,refertothe

simultaneouslylowestvaluesofbothTE,minandRforthe

cor-respondingnon-inoculatedalloys.Thesevaluescorrespondto anearlyfullymetastablesolidification.FromFig.11,itmaybe inferredthatalowinoculation,muchlowerthanthatusedin

thepresentstudy,wouldgivethemostappropriateresultsfor characterizingmeltpreparationforCGIcasting.

5.

Conclusions

Aquantitativeanalysis ofboththe coolingcurves and the microstructurewascarriedoutonthermalanalysissamples castwithvaryingnodularizingcontentconductingtoachange fromfullyspheroidaltomainlycompactgraphite.The sam-pleswereeitheror notinoculatedleadingtoverydifferent coolingcurvesandmicrostructuresforeachmeltsampling.

Forinoculatedsamples,thechangesinthecoolingcurves weretooweakwhengraphiteevolvedfromspheroidalto com-pacttobeuseful forany microstructurepredictionormelt control.

Onthecontrary,non-inoculatedsamplesprovidedmuch moreinformationonbothmicrostructureandcoolingcurve characteristics.Forthesesamples,eutecticsolidificationtakes placewithasignificantundercooling,whichincreasesasthe amount of nodularizer decreases and as graphite changes fromspheroidaltocompact.Thisleadstoabulk solidifica-tion takingplacepartlyin themetastable systemwiththe formationofcarbides.

Theresults thus obtainedon non-inoculated alloys are in line with previous descriptions while providing some moreinsight through quantitative microstructureanalysis, e.g.growthhindranceofcompactgraphitecellsduringthefirst

stageofeutecticgrowthand competitionbetweencompact andwhiteeutecticcells.

Thepresentworkprovidesthewholesetofquantitative datanecessaryforcheckingtheappropriatenessofany mod-ellingapproachofthesolidificationofcompactgraphitecast iron: chemical analysis, cooling curves (in particular mini-mumeutectictemperatureandrecalescence)andquantitative microstructuredata(amountofphases,numberandsizeof compacteutecticcells).

Onapracticalpointofview,thepresentworksuggeststhat thermalanalysiscouldcertainlybeausefulmeansforcontrol ofmeltpreparationforCGIcastingbyaddingverylowlevel ofinoculantinthethermalcups.Experimentsinthislineare on-going.

Acknowledgments

TheauthorswouldliketothankL.A.HurtadoandBetsaide S.A.L.foundryforallthecollaboratingeffortsmadetoobtain thesamplesusedinthepresentwork.

Appendix

A.

SeeTableA1

TableA1–Compositionofthe19alloys,notincludingthecontributionofinoculation(wt.%).

Alloy C Si Mn P S Cr Mo Ni Cu Mg Ti Ce La A 3.75 2.45 0.64 0.022 <0.005 0.049 <0.010 0.027 0.85 0.043 0.021 0.0130 0.0051 B 3.76 2.42 0.63 0.023 <0.005 0.047 <0.010 0.030 0.85 0.040 0.021 0.0130 0.0044 C 3.75 2.45 0.63 0.024 <0.005 0.049 <0.010 0.028 0.85 0.038 0.021 0.0120 0.0037 D 3.74 2.43 0.64 0.022 <0.005 0.050 <0.010 0.028 0.85 0.034 0.021 0.0110 0.0032 E 3.72 2.42 0.63 0.023 <0.005 0.048 <0.010 0.030 0.84 0.035 0.021 0.0100 0.0027 F 3.71 2.45 0.63 0.025 <0.005 0.051 <0.010 0.030 0.84 0.031 0.021 0.0081 0.0021 G 3.72 2.42 0.63 0.022 <0.005 0.049 <0.010 0.030 0.84 0.028 0.021 0.0076 0.0019 H 3.71 2.41 0.64 0.022 <0.005 0.050 <0.010 0.028 0.84 0.021 0.020 0.0070 0.0018 I 3.72 2.44 0.63 0.024 <0.005 0.052 <0.010 0.028 0.84 0.019 0.021 0.0061 0.0015 J 3.69 2.43 0.64 0.023 <0.005 0.054 <0.010 0.028 0.83 0.019 0.021 0.0052 0.0013 K 3.70 2.43 0.62 0.025 <0.005 0.052 <0.010 0.031 0.83 0.018 0.021 0.0047 0.0012 L 3.69 2.43 0.64 0.026 <0.005 0.054 <0.010 0.032 0.83 0.018 0.022 0.0044 0.0011 M 3.67 2.42 0.63 0.021 <0.005 0.051 <0.010 0.026 0.83 0.016 0.020 0.0040 0.0011 N 3.69 2.45 0.63 0.022 <0.005 0.053 <0.010 0.028 0.83 0.015 0.021 0.0036 0.0010 O 3.66 2.43 0.63 0.025 <0.005 0.051 <0.010 0.028 0.83 0.015 0.022 0.0031 0.0009 P 3.67 2.45 0.62 0.025 <0.005 0.050 <0.010 0.028 0.82 0.013 0.022 0.0028 0.0008 Q 3.66 2.40 0.62 0.023 <0.005 0.049 <0.010 0.028 0.83 0.013 0.021 0.0023 0.0007 R 3.67 2.40 0.62 0.023 <0.005 0.049 <0.010 0.027 0.83 0.010 0.021 0.0018 0.0006 S 3.65 2.39 0.62 0.022 <0.005 0.050 <0.010 0.027 0.83 0.008 0.021 0.0014 <0.0005

TableB1–DatausedtocharacterizetheeffectofthirdelementsonthebinaryFe-Cstablesystem(accordingtoCastro etal.[20]).

ispecies Solidphasesinequilibriumwithliquid wC wi T(◦C) mi m g i

Cr Austenite,graphiteandcementite 4.2 4.30 1156 −2.71 13.14

Cu Austeniteandgraphite 4.0 3.7 1172 −4.08 40.62

Mn Austenite,graphiteandcementite 4.32 3.0 1139 −5.66 −2.40

Mo Austeniteandgraphite 5.0 12.6 1350 −10.3 −4.84

Ni Austeniteandgraphite 3.8 10.0 1128 −7.86 18.41

P Austenite,cementiteandFe3P 2.2 7.1 954 −57.8 89.6

Si Austeniteandgraphite 3.78 2.0 1162.5 −23.0 113.2

Appendix

B.

Inalimitedrangeofsiliconcontent,theausteniteandgraphite liquidussurfacescouldberepresentedbyhyperplanesinthe compositionspace.Accordingly,theausteniteliquidus tem-perature,T_L,andthegraphiteliquidustemperature,Tg_L,could beexpressedbylinearrelationsofalloycomposition: T_L=T₀+m_C·wC+





inwhichT₀andTg₀areconstants,m_i andmg_i areliquidusslopes relativetoelementiforausteniteandgraphite,respectively, andwiisthecontentinelementiofthealloy(wt.%).

UsingtheassessmentoftheFe-CsystembyGustafson[27], thestableeutecticisgivenbytheinvariantpoint(4.34wt.%C; 1154◦C).Combiningthisdatawiththeslopeoftheaustenite andgraphiteliquidusassessedbyHeine[28]leadstothe fol-lowingexpressionswherethetemperatureisgiveninCelsius:

T_L=1576.3−97.3·wC+





Toestimatethem_i values,pointswereselectedinthe rel-evantFe-C-iphasediagramsassessedbyRaynorand Rivlin

[29]orbyRaghavan[30],buttheFe-C-Sisystemforwhichthe pointwastakenfromapreviousassessmentofthissystem

[31].InTableB1areindicatedtheselectedpointsandthe cal-culatedvaluesoftheausteniteandgraphiteliquidusslopes. Theexpressionsderivedfromthisselectionareexpectedto bevalidforsiliconcontentsupto3wt.%andforanyother alloyingelementupto1wt.%.

The intersection ofthe two hyper-plans describing the austeniteandgraphiteliquiduscorrespondstotheeutectic trough.Thus,equatingEqs.B3andB4givestheeutecticcarbon content,weut C : weut_C =4.34−



Thecorrespondingeutectictemperature,TEUT,isobtained

byinsertingB5ineitherofEqs.B3orB4,e.g.B3:

TEUT=1154.02+







For thealloysunderinvestigationwhich containsilicon, copperandmanganeseasalloyingelements,onegets:

TEUT= 1154.02+4.246·wSi+4.86·wCu–5.00·wMn (B7)

For thealloysinvestigatedinthepresent study,onehas TEUT changing from 1165.1◦C to 1165.4◦C along the series

of samples. The calculated metastable eutectic tempera-ture isgivenas TW=1150–12.5·wSi [32], and thus increases

slightly from 1119.4 to1120.1◦C during the series of cast-ings.

Finally,itisworthnotingthatthecarbonequivalentCEof thecastironisobtainedfromEq.B5as:

CE=wC+



i

(mg_i−m_i)·wi

mg_C−m_C (B8)

UsingtheparametervalueslistedinTableB1,this expres-siongivesthecarbonequivalentnotedCE99inthemaintext,

Eq.(1).ThecorrespondingexpressionCEASMsuggestedinthe

ASMhandbook[21],Eq.(1’)inthemaintext,isdueto Neu-mann[33].NeumannestimatedthecoefficientsinCEASMusing

thefollowingtwo-stepprocedure:1.Recordingexperimental solubilityvaluesofcarboninFe-C-imeltsat1500◦C;2.Using thesevaluesasestimatesofthechangeincarboncontentof thebinaryFe-Ceutecticinducedbyalloyingwithelement“i”. Hence,theprocedurewefollowedisformerlythesamebut shouldgiveabetterestimateofthealloyingeffectoncastirons because:

1 C-iinteractionsarecertainlytemperaturedependent.The largedifferencebetweentheeutectictemperatureinFe-C-i systemsandthetemperatureof1500◦Cselectedby Neu-mannmayrelatetosignificantchangesinthequantitative effectofelement“i”oncarbonsolubilityintheliquid. 2 ThecarboncontentoftheFe-Ceutecticwassetat4.26wt.%

byNeumannwhileitisnowadmitteditis4.34wt.%.

ThedifferencebetweentheCE99 andCEASMvalues

illus-trated inFig.1isforalarge partduetothechange inthe Fe-Ceutecticcomposition.Further,Fig.2binNeumann’spaper

showsthattheeffectofcarbideformerelements,e.g.Crand Mn,oncarbonsolubilityisquitesmall.Hence,theaboveshift oftheassessedeutecticcompositioninthebinaryFe-Csystem maywellexplainthatthesecoefficientsaresmallandnegative inCEASMwhiletheyaresmallandpositiveinCE99.

Appendix

C.

Microstructuredata SeeTablesC1andC2

TableC1–Microstructuredataforinoculatedalloysandmaximumrecordedtemperature,Tpeak(◦C).

Sample fIIIC fVC fVIC fIIIA fVA fVIA fgraphite Tpeak

Ainoc 0.07 0.25 0.68 0.03 0.26 0.71 0.091 1282 Binoc 0.07 0.23 0.70 0.02 0.22 0.76 0.088 1283 Cinoc 0.03 0.20 0.77 0.02 0.18 0.80 0.084 1296 Dinoc 0.07 0.23 0.70 0.05 0.28 0.67 0.082 1294 Einoc 0.07 0.14 0.79 0.05 0.16 0.79 0.095 1311 Finoc 0.07 0.15 0.78 0.05 0.15 0.80 0.086 1302 Ginoc 0.06 0.17 0.77 0.04 0.20 0.76 0.086 1313 Hinoc 0.10 0.20 0.70 0.06 0.25 0.69 0.091 1247 Iinoc 0.15 0.27 0.58 0.15 0.36 0.49 0.084 1310 Jinoc 0.25 0.16 0.59 0.29 0.19 0.52 0.086 1307 Kinoc 0.33 0.13 0.54 0.41 0.16 0.43 0.084 1312 Linoc 0.41 0.16 0.43 0.53 0.16 0.31 0.082 1325 Minoc 0.35 0.19 0.46 0.42 0.17 0.41 0.084 1333 Ninoc 0.41 0.21 0.38 0.45 0.22 0.33 0.082 1344 Oinoc 0.47 0.21 0.32 0.57 0.19 0.24 0.093 1358 Pinoc 0.46 0.17 0.37 0.65 0.13 0.22 0.094 1346 Qinoc 0.54 0.18 0.28 0.63 0.17 0.20 0.088 1367 Rinoc 0.62 0.19 0.19 0.69 0.16 0.15 0.091 1357 Sinoc 0.63 0.18 0.19 0.69 0.15 0.16 0.087 1367

TableC2–Microstructuredatafornon-inoculatedalloysandmaximumrecordedtemperature,Tpeak(◦C).

Sample fIIIC fVC fVIC fIIIA fVA fVIA fcarbides fgraphite DCell(mm) NCell(mm−2) Tpeak

Ano-inoc 0.15 0.35 0.50 0.13 0.37 0.50 0.00 0.066 0.15 98.06 1283 Bno-inoc 0.42 0.36 0.22 0.47 0.37 0.16 0.00 0.069 0.21 50.16 1291 Cno-inoc 0.54 0.27 0.19 0.73 0.20 0.07 0.09 0.075 0.32 24.01 1290 Dno-inoc 0.65 0.28 0.08 0.77 0.21 0.02 0.16 0.057 0.34 16.46 1298 Eno-inoc 0.63 0.22 0.15 0.80 0.15 0.05 0.09 0.075 0.54 6.38 1294 Fno-inoc 0.64 0.23 0.13 0.80 0.17 0.03 0.09 0.071 0.68 4.97 1299 Gno-inoc 0.67 0.19 0.14 0.88 0.10 0.02 0.22 0.062 0.55 7.08 1308 Hno-inoc 0.40 0.20 0.40 0.77 0.14 0.09 0.30 0.046 0.68 3.80 1294 Ino-inoc 0.60 0.24 0.16 0.84 0.12 0.04 0.20 0.074 0.53 6.64 1303 Jno-inoc 0.58 0.21 0.21 0.84 0.10 0.06 0.35 0.055 0.55 4.58 1290 Kno-inoc 0.69 0.19 0.11 0.88 0.10 0.02 0.34 0.043 0.53 4.11 1300 Lno-inoc 0.62 0.18 0.20 0.86 0.10 0.04 0.19 0.078 0.47 5.03 1310 Mno-inoc 0.57 0.26 0.17 0.80 0.16 0.04 0.29 0.062 0.67 2.73 1321 Nno-inoc 0.64 0.20 0.17 0.81 0.15 0.04 0.18 0.077 0.45 3.36 1328 Ono-inoc 0.58 0.23 0.19 0.83 0.13 0.04 0.32 0.047 0.33 2.73 1337 Pno-inoc 0.60 0.22 0.18 0.86 0.11 0.03 0.25 0.050 0.49 4.11 1334 Qno-inoc 0.43 0.22 0.35 0.75 0.16 0.09 0.42 0.028 0.36 1.46 1348 Rno-inoc 0.47 0.22 0.30 0.87 0.09 0.04 0.38 0.027 0.28 1.22 1338 Sno-inoc 0.59 0.23 0.18 0.83 0.12 0.05 0.39 0.037 0.32 1.77 1346

r

e

f

e

r

e

n

c

e

s

[1]DawsonS,SchroederT.Practicalapplicationsforcompacted graphiteiron.AFSTrans2004;112,paper04-047.

[2]KönigM,WessénM.Influenceofalloyingelementson microstructureandmechanicalpropertiesofCGI.IntJCast MetRes2010;23:97–110.

[3]BazdarM,AbbasiHR,YaghtinAH,RassizadehghaniJ.Effect ofsulfurongraphiteaspectratioandtensilepropertiesin compactedgraphiteirons.JMaterProcTech2009;209:1701–5.

[4]CharoenvilaisiriS,StefanescuDM,RuxandaR,PiwonkaTS. Thinwallcompactedgraphiteironcastings.AFSTrans 2002;110,paper02-176.

[5]LekakhSN,QingJ,RichardsVL.Investigationofvastiron processingtoproducecontrolleddualgraphitestructurein casting.AFSTrans2012;120:297–306.

[6]GornyM,KawalecM.Effectsoftitaniumadditionon microstructureandmechanicalpropertiesofthin-walled compactedgraphiteironcastings.JMEPEG2013;22:1519–24.

[7]FiricanMC,RiposanI.Graphitephasecharacteristicsin compacted/vermiculargraphitecastironinoculatedinthe mould.AdvMaterRes2015;1128:72–9.

[8]LoizagaA,Larra ˜nagaP,AsenjoI,SertuchaJ,SuárezR.An experimentalmethodologybasedonsolidificationcurvesfor controllingthemanufactureofcompactedgraphiteirons. AFSTrans2012;120,paper12-006.

[9]GornyM.Castiron:compactedgraphite.In:Encyclopediaof iron,steel,andtheiralloys.TaylorandFrancis;2015.p. 718–34.

[10]MampaeyF,HabetsD,PlessersJ,SeutensF.Onlineoxygen activitymeasurementstodetermineoptimalgraphiteform duringcompactedgraphiteironproduction.IntJMet 2010;42:25–40.

[11]DawsonS.Controllingtheproductionofcompactedgraphite iron.ModernCasting1998;88:38–41.

[12]StefanescuDM,MartínezF,ChenIG.Solidificationbehaviour ofhypoeutecticandeutecticgraphitecastirons.Chilling tendencyandeutecticcells.AFSTrans1983;91:

205–16.

[13]JinhaiLiu,LitaoYi,GuoluLi,ChangqiLiu,YinguoLi,Zhaoyu Yang.Influenceoffadingoncharacteristicsofthermal analysiscurveofcompactedgraphiteiron.ChinaFoundry 2011;8:295–9.

[14]BackerudL,NilssonK,SteenM.Studyofnucleationand growthofgraphiteinmagnesium-treatedcastironbymeans ofthermalanalysis.In:ThemetallurgyofcastIron.Georgi Pub.Co.;1975.p.625–37.

[15]StefanescuDM,LoperCR,VoigtRC,ChenIG.Coolingcurve structureanalysisofcompacted/vermiculargraphitecast ironsproducedbydifferentmelttreatments.AFSTrans 1982;90:333–48.

[16]RodriguezS,CastroM,HerreraM,MendezJ,EscalanteJI, GonzalezC.Thermalanalysisofcompactedgraphitecast

ironscontainingvariouslevelsofMgandS.IntJCastMet Res1999;11:381–6.

[17]SunXJ,LiYX,ChenX.Identificationandevaluationof modificationlevelforcompactedgraphitecastiron.JMater ProcessTechnol2008;200:471–80.

[18]HernandoJC,DomeijB,GonzalezD,AmievaJM,DioszegiA. NewexperimentaltechniquefornodularityandMgfading controlincompactedgraphiteironproductiononlaboratory scale.MetallMaterTransA2017;48A:5432–41.

[19]RegordosaA,delaTorreU,LoizagaA,SertuchaJ,LacazeJ. Microstructurechangesduringsolidificationofcastirons– effectofchemicalcompositionandinoculationon

competitivespheroidalandcompactedgraphitegrowth.IntJ Met2020;14:681–8.

[20]CastroM,HerreraM,CisnerosMM,LesoultG,LacazeJ. Simulationofthermalanalysisappliedtothedescriptionof thesolidificationofhypereutecticSGcastirons.IntJCast MetRes1999;11:369–74.

[21]StefanescuDM,LacazeJ.Thermodynamicsprinciplesas appliedtocastiron.In:ASMhandbook,volume1A,Castiron scienceandtechnology.ASMInternational;2017.p.31–42.

[22]HeineRW.Austeniteliquidus,carbideeutecticand undercoolinginprocesscontrolofductilebaseiron.AFS Trans1995;103:199–206.

[23]LacazeJ,RegordosaA,SertuchaJ.Quantitativeanalysisofthe effectofinoculationandmagnesiumcontentoncompacted graphiteirons–amodellingapproach.Workinprogress. [24]PanEN,OgiK,LoperCR.Analysisofthesolidification

processofcompacted/vermiculargraphitecastiron.AFS Trans1982;90:509–27.

[25]LesoultG,CastroM,LacazeJ.Solidificationofspheroidal graphitecastiron.PartI:physicalmodelling.ActaMater 1998;46:983–95.

[26]LacazeJ,LesoultG,CastroM.Solidificationofspheroidal graphitecastiron.PartII:numericalsimulation.ActaMater 1998;46:997–1010.

[27]GustafsonPA.AthermodynamicevaluationoftheFe-C system.ScandJMetall1985;14:259–67.

[28]HeineRW.TheFe-C-Sisolidificationdiagramforcastirons. AFSTrans1986;94:391–402.

[29]RaynorGV,RivlinVG.Phaseequilibriainironternary systems.TheInstituteofMetals;1988.

[30]RaghavanV.Phasediagramsofternaryironalloys.The IndianInstituteofMetals;1987.

[31]LacazeJ,SundmanB.AnassessmentoftheFe-C-Sisystem. MetallTrans1991;22A:2211–23.

[32]LacazeJ.Solidificationofspheroidalgraphitecastirons.Part III:microsegregationrelatedeffects.ActaMater

1999;47:3779–92.

[33]NeumannF.Theinfluenceofadditionalelementsonthe physico-chemicalbehaviorofcarbonincarbonsaturated molteniron.In:TheAmericanSocietyofMetals,organizer. Proceedingsofrecentresearchoncastiron,Detroit,1964 June16–18.GordonandBreach;1968.p.659–705.