HAL Id: hal-00836837
https://hal.archives-ouvertes.fr/hal-00836837
Submitted on 21 Jun 2013
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.
Reinforced sol-gel thermal barrier coatings and their
cyclic oxidation life
Lisa Pin, Florence Ansart, Jean-Pierre Bonino, Yannick Le Maoult, Vanessa
Vidal, Philippe Lours
To cite this version:
Lisa Pin, Florence Ansart, Jean-Pierre Bonino, Yannick Le Maoult, Vanessa Vidal, et al.. Reinforced
sol-gel thermal barrier coatings and their cyclic oxidation life. Journal of the European Ceramic
Society, Elsevier, 2013, vol. 33, pp. 269-276. �10.1016/j.jeurceramsoc.2012.07.037�. �hal-00836837�
Any correspondence concerning this service should be sent to the repository administrator:
staff-oatao@inp-toulouse.fr
To link to this article:
DOI:10.1016/j.jeurceramsoc.2012.07.037
Official URL: http://dx.doi.org/10.1016/j.jeurceramsoc.2012.07.037
This is an author-deposited version published in:
http://oatao.univ-toulouse.fr/
Eprints ID: 8733
To cite this version:
Pin, Lisa and Ansart, Florence and Bonino, Jean-Pierre and Le Maoult, Yannick
and Vidal, Vanessa and Lours, Philippe
Reinforced sol–gel thermal barrier
coatings and their cyclic oxidation life. (2013) Journal of the European Ceramic
Society, vol. 33 (n° 2). pp. 269-276. ISSN 0955-2219
O
pen
A
rchive
T
oulouse
A
rchive
O
uverte (
OATAO
)
OATAO is an open access repository that collects the work of Toulouse researchers and
makes it freely available over the web where possible.
Reinforced
sol– gel
thermal
barrier
coatings
and
their
cyclic
oxidation
life
Lisa
Pin
a,b,∗,
Florence
Ansart
a,
Jean-Pierre
Bonino
a,
Yannick
Le
Maoult
b,
Vanessa
Vidal
b,
Philippe
Lours
baUniversityofToulouse,UPS- INP- CNRS,InstitutCarnotCIRIMAT,118RoutedeNarbonne,31062ToulouseCedex09,France bUniversityofToulouse,MinesAlbi–InstitutClémentAder,CampusJarlard,F- 81013AlbiCedex09,France
Abstract
Cyclicoxidationlifeenhancementofsol– gelthermalbarriercoatingsisobtainedviathereinforcementofthecontrolledmicro-cracknetwork thatformsduringtheinitialsinteringofthedeposit.Twodifferentsol– gelmethodsareusedtofillintheprocess-inducedcracks,namely dip-coatingandspray-coating.Fillingparameters,forinstancethenumberofpassesortheviscosityofthesolareadjusted,usingvarioustechniques suchasprofilometryandmicrostructuralanalysis,tooptimisecrackfilling.Cyclicoxidationtestsareimplementedatboth1100◦Cand1150◦C toinvestigatetheefficiencyofthevariousreinforcementproceduresdevelopedandaddresstheinfluenceofthespecificmicrostructureonthe oxidationbehaviour.
Keywords:Sol– gel;Oxidation;Reinforcement;Thermalbarriercoating
1. Introduction
Improving turbojet engine performances requires to con-tinuously increase the turbine inlet temperature. Nickel base superalloys,mainmaterialsusedfortheapplication,show supe-rior creep and fatigue strength. However, superalloy turbine bladescannotsustainthehightemperatureimposedtomodern engineswithoutthermalprotectionandinternalcooling. Ther-mal barrier coatings(TBCs) are deposited onhollow turbine bladesinordertoestablishathermalgradientpronetoprotect against the detrimentaleffects of long-term hightemperature exposure.ThickerTBCsarealsobeneficiallyusedtocoatand thermallyinsulatetheinternalpartsofthecombustionchambers. Currently,TBCs–consistedofyttria-stabilised-zirconia(YSZ) –aremanufacturedusingtheso-calledElectronBeam–Physical VapourDeposition(EB-PVD),typicallyforturbineblades,or PlasmaSpraying(PS),forcombustionchambers.Thethermal andmechanical propertiesof TBCs strongly depend ontheir
∗Correspondingauthorat:UniversityofToulouse,UPS-INP-CNRS,Institut
CarnotCIRIMAT,118RoutedeNarbonne,31062ToulouseCedex09,France. Tel.:+33561556285;fax:+33561556163.
E- mailaddress:pin@chimie.ups-tlse.fr(L.Pin).
microstructure that straightforwardlyresults fromthe deposi-tionprocess.EB-PVDcoatingsexhibitcolumnarmicrostructure wherethe presence of elongated grainsandgrainboundaries normaltothesubstratesurfaceallowsaconvenient accommoda-tionofin-servicelateralthermomechanicalstressesandrelated strains.However,thethermalconductivityofEB-PVDcoatings is not fullyoptimised as the presence of columnar porosities favoursextensiveheattransferbetweentheoutersurfaceofthe multi-materialsystemandthesuperalloysubstrate.Nevertheless someworks aimingatadjustingthe chemicalcomposition1– 4 andthemicrostructureofelongatedgrains1allowtoimprovethe insulationpropertiesofsuchEB-PVDcoatings.ForPScoating, the typicalmicrostructureishighlylamellar, resultinginboth alowerthermalconductivityastheporosityisdistributed per-pendiculartotheheatflux,andalowerabilitytoaccommodate lateral thermomechanical stresses.Variousadditionalwaysto synthesizeTBCshavebeenproposedintheliterature,basedon soft chemicalprocessing.5– 7In thispaper,anovelmethodfor depositingandreinforcingTBCs,isinvestigated.Derivedfrom the sol– gelroute, itincludeseitherdip-coating,for theinitial depositionandthepost-processingreinforcementoftheTBCs, orspray-coating,usedafterdip-coating,forthepost-processing reinforcementof theTBCsonly. Contrarytootherprocesses, sol– gelTBCs show a specific isotropic microstructure with
randomly distributedporosities. Propertiesandinduced char-acteristics are very different compared to those of EB-PVD and PScoatings. Thisshould result in an attractive compro-misebetweenthermalconductivityandmechanicalstrength.In apreviouspaper,8themechanismsresponsibleforthedamage of sol– gelcoatings subject tocyclic oxidation wasidentified anddiscussed as afunctionof the oxidationtemperature and the TBCthickness.Initially,the degradationof sol– gelTBCs proceedsthroughtheformationofaregularcracknetwork occur-ring,eitherduringthepost-depositionthoroughlycontrolledby thermaltreatmentorduringtheveryfirstcyclesofoxidation,in bothcasesastheresultof thethermally-activatedsinteringof thezirconiadeposit.Thisnetworkcanfurtherdevelopitself dur-ingthecumulativeoxidationcyclespromotingtheenlargement andthecoalescence of cracksthat finallyletsappear individ-ualTBCcells pronetospalloffundertheeffectofthecyclic thermomechanical stresses.In ordertoimprovethe durability of sol– gelcoatings upon oxidation, it is proposed tocontrol andstabilizetheprocess-inducedcracknetworkbyfillingthe crackswithfreshmaterialbroughteitherbysupplementary dip-coatingpasses9oradditionalspray-coatingpasses.Byproviding an appropriate adjustment and control of the process, it has been shownthat the crack fillingbydip-coating significantly improvesthecyclicoxidationbehaviouroftheTBCs9throughan overallreinforcementofthebarrierresultingfromanoptimised mechanicalpegging.
The present paper investigates (i) the feasibilityof filling cracks using spray-coating and the principal difference with crackfillingbydip-coating,(ii)theresultingcharacteristicsof themicrostructureandmacrostructuredeveloppedand(iii)the durabilityof spray-coatedreinforcedTBCsduringcyclic oxi-dation at1100◦C and1150◦C. The overall performances of
spray-filledanddip-filledreinforcedTBCsarecompared.
2. Experimental
2.1. Synthesisofsol–gelthermalbarrier
AM1 superalloys substratesare initiallycoated with NiP-tAlbondcoattoenhanceTBCadhesion.Thesubstrateisthen grit-blastedat2.5barsduring10swithcorundomparticles.To processTBCsbysol– gelroute,precursorYSZsolwitha com-positionof9.7mol%yttria,issynthesizedfromzirconium(IV) propoxide Zr(OPr)4 (Aldrich) and yttrium (III) nitrate
hexa-hydrate(AcrosOrganics).10 Acetylacetone(AcAc) isusedas complexing agent toreduce the hydrolysis rate of zirconium alkoxide,7 and the solvent is the 1-propanol. Volume ratios betweenAcAcandZr(OPr)4andbetweenH2OandZr(OPr)4are
respectively0.8and9.5.Zirconiumconcentrationiskept con-stantat0.5molL−1.SuchasolofYSZisusedtoproduceYSZ
powder,obtainedaftersupercriticdryingofthepropanolusedas solvent.11Followingthisstep,anaerogelisfirstobtained. Ulti-mately,thepowderisheattreatedat700◦C.YSZpowdersare
dispersedintoaslurrycomposedofthestartingsol(9.7mol% YO1.5) loadedwith40wt.%of powder directlycoming from
the supercriticdrying describedabove. Then, superalloy sub-stratesaredippedintotheslurryandwithdrawnatacontrolled
Fig.1.SEMcross-sectionmicrographofasol– gelthermalbarriercoating.
rate(250mm/min)toshape thecoatings.Aftereach dipping, thelayerisdryingat50◦Cduring5min.Theseoperationsare
repeateduntiltherequiredTBCthicknessisreached.Processed specimensaresinteredat1100◦Cduring2husingheatingand
coolingratesof100◦C/h,topromotetheformationofthe
ini-tialcracknetworkandsomewhatstabilizetheoverallstructure of the barrier by releasing residual sintering stresses. Subse-quently,specimenswithcontrolledcracksarereinforcedusing additionalfillingzirconiabrought upinsidecrackgroovesby eitherdiporspray-coating.
2.2. Fillingprocess
Thesuspensionsusedforsol– gelfillingarecomposedof a commercialYSZpowder(TOSOH8Y),in1-propanolsolvent. YSZcommercialpowdersaredispersedintoaslurrycomposed ofthestartingsol(9.7mol%YO1.5)loadedwith10or20wt.%of
thiswelldispersedandmonodispersedpowder(sizeof elemen-taryparticlesaround30nm).Twodifferentmethodsareusedto fillincracks:dip-coatingandspray-coating.Dip-coatingmethod consists insuccessively dippingand withdrawing atconstant speed the substrate out of aslurry. Spray-coatingconsists in sprayingtheslurryfromthetopsideofspecimens,thus allow-ingthecontrolofboththepressureandthedisplacementrate ofthespraynozzle.Betweeneachpass,thecoatingisdryingat 50◦Cduring5min.When,reinforcementsynthesizeprocessis
finished,finallythespecimensareheattreatedat1100◦Cduring
2h,usingheatingandcoolingratesof50◦C/h.Across-section
micrographof asol– geldeposit,filled withadditional spray-coatingispresentedinFig.1.Aspowdersusedrespectivelyfor the initialdip-coating andfor the spray filling havedifferent densities,thecontrastbetweentheYSZbroughtbysprayfilling andthelayer ofinitiallydip-coated YSZis clearlyobserved.
Table1givesthedetailsofthevariousspecimensinvestigated intermsoffillingprocessemployed(DC:dip-coatingandSC: spray-coating),numbersofpassesandweightpercentofpowder intheslurry.
In order to characterize surface topography of the as-processed as well as the filledsamples andto quantify both the depth and the width of cracks, confocal interferometry microscopy,withadynamicrangeof15nm– 350mmisused.In
Table1
Referenceandprocessparametersfortheinvestigatedspecimens. Sample reference Numberofdipor sprayat20wt.% Numberofdipor sprayat10wt.% DC1 4 0 DC2 6 0 SC1 4 0 SC2 2 4 SC3 4 2 DC3 4 2 DC4 2 4
addition,imageswithhighfielddepthofthespecimensurface areperformedusinganumericalmicroscope.
2.3. Cyclicoxidationtests
Cyclicoxidationtestsareperformedinaspecificdedicated automatedriginstrumentedwithCCDcameratomonitorina realtimebasistheevolutionofthespecimensurfaceduringthe air-forced5mincoolingfollowingthevariouscumulated5min heatingplus1-hholdingat1100◦Candat1150◦C.Individual
CCDimagesareextractedfromthevideorecordingtobetreated usingimageanalysissoftware.Fromthose“ in-situ” experimen-taldata,thesurfacefractionofcracksisderivedasafunction
of timefollowing each cycle toapproach the overall surface damagekinetics.
3. Resultsanddiscussion
3.1. Reinforcementofthermalbarrierssynthesizedby sol–gelroute
TheprincipleofTBCreinforcementisbasedonthedeliberate establishmentofacontrolledmicro-cracknetworkthroughout theentirecoatingdepositedbythesol– geltechnique.As indi-cated above, the formation of such cracks is inherent to the deposition process itself,but its geometricalcharacteristic is directlyrelativetotheparametersusedduringthesol– gel pro-cessing.Asamatteroffact,thecontrolofthemicro-crackdepth andspacedistributioncanbeachievedusingwell-defined pro-cessparameters,e.g.thenumberofdip-coatingpassesthatkeeps constant,theTBCthickness,andthetimeandtemperature of heattreatmentthatdeterminetheextentofsintering,namelyin termsof densificationandresidualstress.Oncethe crack net-workisregularlyformed,fillingofthecrackgroovesusingeither dip-coatingorspray-coatingcanbebeneficiallyusedtostiffen and strengthenthe deposit so as toimprove the resistanceto cyclicoxidation.
Fig.2ashowstheprofileoftheoutersurfaceofaTBCinitially depositedby4successivedip-coatingpassesandheattreatedfor
Fig.2.Microstructureandcross-sectionprofileofcracksof(a)theinitialTBC,(b)theTBCreinforcedusingdip-coatingprocess,and(c)theTBCreinforcedusing spray-coatingprocess.
Table2
Surfaceanddepthfillingratesforspecimenwithvariousprocessparameters. Fillingofsample 1D/D(%) 1S/S(%) DC1 14 35 SC1 44 45 SC2 44 53 DC4 12 37 DC3 7 47
2hat1100◦C.Theprofileischaracterisedbyaratherregular
distributionofcrackswithvariousdepthandwidth.Fig.2band cpresentsdetailsof anindividualcrackrespectivelyfilledby dip-coatingandspray-coating,wherethegrey-colouredzones correspond tothe filled parts wherematerial, brought up by eitherdiporspray-coating,penetratesthecrack,thensolidifies andsintersastheconsequenceofpost-processingheating.Both processesarequiteefficientforpluggingupcracksbut,asclearly showninFig.2,thequalityoffillingishighlyprocess depend-ent. Unidirectional dip-coating based on axial withdrawal of specimens, resultsin aheterogeneous,non-symmetric filling, whilespray-coating,depositedfromthetopandalonga direc-tiongloballynormaltothespecimen,resultsinahomogeneous andsymmetricfilling.
To quantifyprocessefficiency,the surface(1S/S) (respec-tivelydepth(1D/D))fillingratesaredeterminedbyestimating theratiobetweentheincrementincracksurfaceDS(respectively depth DD) after fillingand the initialsurface S (respectively depthD)beforefilling.
Itisinterestingtonotethatthesurfaceanddepthfillingrates are bothhigher for specimens treatedby spray-coating, indi-catingabetter fillingefficiency for thisprocess(Table 2).In addition,thedifferenceoffilling,beneficialtothespray-coating technique,ismuchmorepronouncedforthedepthfillingrate thanforthesurfacefillingrate.Thisunambiguouslyremark con-firmsthelowcapabilityofdip-coatingtosatisfactorilywetthe bottomofthecrackgrooves.
Asaconsequenceofthisvariationinthehomogeneityand inthesymmetryoftheYSZfillings,thedistributionand mag-nitudeofstressinthereinforceddepositarelikelytobeprocess dependentalso.Resultingfromthesedifferentinitialconditions, theevolutionofthermo-mechanicalstressesduetocumulated oxidation cyclesisexpectedtobedifferent,leading to differ-ent thermo-mechanicalbehaviour.The oxidationdurabilityof specimens,respectivelyreinforcedusingdip-coatingand spray-coatingisinvestigatedinthenextsection.
Asimplewaytoimprovesurfaceanddepthfillingratescould bedonebyincreasingtheamountofmaterialbroughtupby addi-tionaldeposition.So,toinvestigatethemicrostructuralpossible benefits,thenumberofpasseshasbeenincreasedbydip-coating technique, fromfour tosix. The 3D profilometryimageof a typical100mmthickTBC depositedby tensuccessive initial
dip-coatings andheat-treatedat1100◦Cis showninFig.3a.
Theresulting2Dprofilesand3DmapoftheTBCaftersix sup-plementarydip-coatingpassesusingaslurrywith20wt.%YSZ aregivenrespectivelyinFig.3bandc.Anadditionalcracks for-mation,revealedontheverytopsurfaceofTBCbythe3Dprofile
map(Fig.3c),clearlyindicatesthatsinteringhasoccurredduring thepost-filingheattreatment.Itisalsoevidencedthatnew mate-rialhasbeenbroughtupduringfilling,notonlyinsidethecracks butalsoontopoftheexternalsurfaceofthebarrier.Inaddition topenetratecrack grooves,furtherfillingbeyond fourpasses tendtofavourthedevelopmentofanover-layerwhosethickness canbeestimatedbymeasuringthedepthof sintering-induced cracks(Fig.3b).Typically,thethicknessoftheoverlayer result-ingfrom6passesisabout30mm(notethatitisonly2– 5mmfor
fourpasses).5Thepropensitytospallationofover-filledsol– gel TBCsduringcyclic oxidation isclearlyshownanddiscussed in the paper.5 In this case, following four filling passes, the cyclicoxidationbehavioursignificantlyimproves,while follow-ingeightfillingpasses,itisverysimilartothatofthenon-filled asprocessedTBC.Thissuggeststhatbelowathreshold num-berofpasses,typicallyfourwith20wt.%ofpowderinslurry, theprevalentresultremainstheoptimumfillingofcracksand couldleadtoabeneficialstrengtheningofthereinforcedTBC. Beyond thisthreshold, theoverspreading prevailsandalmost nofurtherfillingofcracksoccurs.Thisdrasticallyincreasesthe overallthicknessof the TBC atthe expenseof a satisfactory fillingof thecracks. Insteadof reducing their depth,existing cracksfurtherdeepen andadditionalnewcracks form,which somewhatresultsinareturnbacktotheinitialpoorlyadherent microstructure.
Anattractivewaytoimprovetheefficiencyofthecrack fill-ingprocessconsistsinadjustingtheviscosityoftheslurryby modifyingtheweightpercentofpowderdependingonthe con-sideredpass.Initial(respectivelyfinal)passescanbeperformed using high weight percent (20wt.%), high viscosity (respec-tivelylowweightpercent(10wt.%),lowviscosity)slurrytofill inthelargest(respectivelythesmallest)cracks.Insuchaway, theformationofthedetrimentalover-layercanbeavoidedwhile theoverallstiffnessandthemechanicalstrengthoftheTBCare enhanced.ExamplesofoptimisedcrackfillingaregiveninFig.4
fortwospecimens,filledinwithsixpasseswithtwodifferent viscositiesofslurry(SC2andDC5asreferencedinTable2).It isshownthatwhateverthefillingprocess,i.e.spray-coatingor dip-coating,allcracksareatleastpartiallyfilledasindicatedby thegrey-colouredareasseparatingtheprofilesbeforeandafter filling.Inaddition,notethat nolargeover-thickness,proneto degrademechanicalresistanceoftheTBC,isobserved.Thisis particulartruefortheTBCfilledusingspray-coating.
3.2. Cyclicoxidationofreinforcedthermalbarriercoatings
Toestimatetheefficiencyofsol– gelthermalbarriercoating reinforcedbyfillingofthemicro-cracksnetwork,oxidationtests havebeenperformedonthefullyautomatedcyclicoxidationrig described inSection 2.3.Coating damage resulting from the successivecoolingsequencesofthecumulatedoxidationcycles ismonitored insituto derivethe degradation kinetics of the TBC.
First,theoxidationbehaviour oftwo specimensreinforced respectivelyusingdip-coatingandspray-coatingiscompared. Thisallowstoaddressintermsof resistancetocyclic oxida-tiontheadvantagesanddrawbacksofthetwofillingtechniques,
Fig.3.MorphologyoftheTBCsurface(a)as-processed,beforefilling,(b)cracksprofileofthesurfacerepresentedbylineshowninc,(c)following6dip-coatings (DC3).
Fig.4.Crackprofilesofspecimensfilledusing(a)spray-coating(SC2)and(b)dip-coating(DC3).
eachleadingtoaspecificmicrostructure,eitherdirectionaland asymmetricfor dip-coatingandhomogeneousandsymmetric for spray-coating. Then, the specificdamage mechanisms of spray-coating reinforcementswill be discussed by investigat-ingtheimpactofdifferentmaterial-intrinsicparameters,i.e.the numberofpassesandtheweightpercentofpowderintheslurry, andmaterial-extrinsicparameters,i.e.thenumberofoxidation cyclesandtheoxidationtemperature.
3.3. Cyclicoxidationresistance:dip- coatingversus spray- coatingreinforcement
Fig.5acomparestheinitialsurfaceandsurfaceevolutionof spray-coated(leftcolumn)anddip-coated(rightcolumn)filled TBCs during cumulative one-h oxidation cycles at 1100◦C,
i.e. before oxidation (i),following 50 one-hoxidation cycles at 1100◦C (ii), andfollowing 100 one-h oxidation cycles at
1100◦C (iii). The aspect of the surface prior to oxidation is
quite different as cracks are clearly observed in the case of the dip-coated TBC while only ridges, delineating the initial sintering-induced cracks, are on the spray-coated TBC. This confirms the limited capabilityof dip-coating to achieve sat-isfactoryfillingofcrackgrooves,duetotheuseofdirectional fillingprocess.Thesurfacemorphologyofthespray-coatedTBC does notprogressfollowingeither50or100oxidation cycles as no evolution of the ridgesnetwork occurs confirming the remarkableefficiencyoftheTBCmechanicalstrengthening.
Forthedip-coatingreinforcedTBC,oxidationdamagetakes placethroughthepercolationoftheinitialcracknetworkcoupled toanextensivewideningofthecracks.
In bothcases, the surface fraction of cracks,estimated by imageanalysisofroomtemperaturemicrographsextractedfrom thevideotaperecordingfollowingsuccessivecompleted cool-ingsequences,isplottedasafunctionofthenumberofoxidation
Fig.5.SurfaceofTBCreinforcedwithspray-coatingSC3(leftcolumn)andwithdip-coatingDC4(rightcolumn)duringthecooling(5min)afterone-hoxidation cyclesat1100◦C.
Fig.6.Evolutionofthesurfacefractionofcrackversusthenumberofone-h oxidationcycles.
cyclesinFig.6(blackcirclesandsquares).Monitoringthe sur-face fraction of cracks, which is the difference between the fraction before oxidation (cycle0) andthe fraction after oxi-dation(cycle10– 100),allowstoaccuratelydeterminetheextent ofadditionalcrackingresultingfromeachindividualcycle.For the dip-coatedreinforced TBC,the surfacefraction of cracks rapidlyincreasesduringtheveryfirstcyclestoreach7.5%at70 cyclesandthenstabilisesfrom70to100cycles.Notethatfor thespray-coatedreinforcedTBC,thesurfacefractionofcrack remainslowerthan1%upto100cycles.Thehigherresistanceto crackcoalescenceandenlargementofthespray-coatreinforced TBCisessentiallyduetothehigheramountofmaterialthatfills thecrack networkandthebetterhomogeneousdistributionof materialswithin thecracks.Thisprocessspecificity undoubt-edlyminimizestheextentandhomogenizesthedistributionof residualstressresultingfromthepost-fillingheattreatment.
In the next section, spray-coated reinforced TBCs filled using variousparameters, namely the sequenceof fillingand the weight percentof YSZ inthe slurry for each passes, are thermally cycledat both1100◦C and1150◦Ctofinally
pro-poseanoptimisedprocessingrouteregardingcyclicoxidation performance.
3.4. Theinfluenceofprocessparametersandtestconditions onthecyclicoxidationresistanceofspray- coatTBCs
Fig.7detailsthemicrostructuralbehaviourofspray-coated reinforced TBCs subject to 350 oxidation cycles at 1100◦C
(Fig.7a),followedby22additionaloxidationcyclesat1150◦C
(Fig.7b).Onthelefthandsidecolumn(respectivelytheright handsidecolumn),isreportedtheevolutionofthesurface degra-dation foraTBCreinforcedbytwo successivefillingswitha 20wt.%YSZslurryplusfourfillingswitha10wt.%YSZslurry (respectivelyfoursuccessivefillingswitha20wt.%YSZslurry). NamelythesurfaceofthereinforcedTBCsbeforeoxidation(i), following150one-hoxidationcyclesat1100◦C(ii),following
350one-hoxidationcyclesat1100◦C(iii),following350one-h
oxidation cyclesat1100◦Cplus15one-hoxidation cyclesat
1150◦Candbefore(iv)andjustafter(v)the372ndone-h
oxi-dationcycles(350cyclesat1100◦Cplus22oxidationcyclesat
1150◦C)arepresented.
Forbothspray-coatfillingroutes,oxidationat1100◦Cdoes
not degrade the surface as the typical ridged microstructure presentnocrackingorspallationupto350cycles.Fig.6reports thesurfacefractionofcrackversusthenumberofcycles(white circles and triangles) that remains lower than 1% up to 350 cycles. In order to enhance oxidation damage, the oxidation temperatureisincreasedto1150◦Cfromthe360thcycle.Such
ahighoxidation temperatureacceleratesboththesintering of zirconiaandthegrowthofthealuminathermallygrownoxide (TGO). Thisresults in arapid andextended surface damage through crack coalescence andenlargement, giving rise to a suddenincreaseof thecrackssurfacepercentvalue inFig.6, typicalof breakaway oxidation. Forthe TBCs filledusing 6 spray-coatingswithtwodifferentslurryviscosities,thesurface fraction of cracks remains however low and no spallation is observed.Conversely,fortheTBCsreinforcedusingfour spray-coatingsandonlyoneslurryviscosity,degradationismuchmore pronouncedandendsup,rightafterthe372ndcycle,bya com-pletespallationofthewholebarrier,asindicatedbytheverylast whitecircleinthegraphinFig.6andtheimageinFig.7b(v). ThehigherstrengthoftheTBCsreinforcedbyslurrywith appro-priateviscosityisrelatedtotheoptimumfillingofcrackswith differentsizes,promotingabetterhomogeneityofthesurface subjectedtooxidationcycles.
ThecyclicoxidationbehaviourofreinforcedTBCsisquite different tothat of sol– gelas-processed TBCs, where crack-ing and subsequent spallation are progressive as cracks first coalescetofinallydelineateindividualYSZcellspronetolocal delamination.8,9Anattemptismadetomodelthemechanismsof YSZdetachmentfromthesubstratesurfaceduringtheoxidation cyclesanddifferentiatethebehaviourofnon-reinforced(Fig.8a) andreinforcedTBCs(Fig.8b).Foras-processed,non-reinforced TBCs,horizontalfailureinitiatesatthebottomofthe vertical sintering-inducedcracksastheresultofthethermo-mechanical stressesduetocumulativeoxidationcycles.Onceformed,such ahorizontalcrackpropagatesparalleltotheinterfacebetween thesubstrateandthebarrier.Asthecrackfullycrossesthewidth ofanelementaryzirconiacell,thebarrierloosesitsadherence tothebondcoatanddetachesintheformofanindividualspall (Fig.8a).Damagecanfurtherenhancethroughthespallationof supplementaryzirconiacellsastheoxidationcyclingprogresses andadditionalcracksdevelop.
Thereinforcementbyspray-coatfillingoftheinitialcracks allowstheconsolidationofTBCsbyprovidingstrongbonding betweenthe individual zirconia cells, whichdelays the initi-ation of horizontal failure.Because of the stiffbridging that occursastheresultoffilling,thehorizontalcracks,once initi-ated,propagateroughly attheinterfacebetweenthesubstrate andthebarrierwithoutanyindividualspalloff.Thecrackshave topropagatethroughouttheentireinterface beforeleading to thecompletespallationoftheTBC(Fig.8b).Thissuddenand totalspallationofsol– gelreinforcedTBCsisverysimilartothat observedtypicallyinEB-PVDTBCsasextensivelyreportedin theliterature.12– 14
Thefailureofsol– gelreinforcedTBCisascomplexasthe oneof EB-PVD barriersonPt-aluminised system, wherethe fracturemayeitherlocateattheinterfacebetweentheTGOand
Fig.7.SurfaceofTBCreinforcedwithoptimisedspray-coating(SC2)(leftcolumn)andnonoptimisedspray-coating(SC1)(rightcolumn)duringthecooling (5min)afterone-hoxidationcyclesat1100◦C(a)andone-hoxidationcyclesat1100◦Cplusone-hoxidationcyclesat1150◦C(b).
Fig.9.SEMmicrographofsol– gelTBCreinforcedwithspray-coating(SC 3)after372one-hoxidationcycles(350cyclesat1100◦Cplus22cyclesat
1150◦C).
thebondcoat15orbetweentheTGOandtheTBC.16Insol– gel TBC,thecrackpathresponsibleforthespallationisalternatively adhesive,resultinginadetachmentattheinterfacebetweenthe bond-coatandtheTGOorattheinterfacebetweentheTGOand theYSZ,orcohesive,leadingtothespallationinsidetheYSZ coating.ThisisillustratedinFig.9,wheretheSEMmicrograph oftheexternalsurfaceofacoatingreinforcedbytwosuccessive spray-coat fillings witha20wt.% YSZ slurry addedby four spray-coatfillingswitha10wt.%YSZslurryafter372oxidation cycles(350at1100◦Cand22at1150◦C)ispresented.Notethat
themainfailureislocatedwithintheYSZsol– gelcoatingover theTGO,whichsuggestsagoodadherenceofthecoatingtothe substrate.
4. Conclusion
Sol– gelthermal barrier coatings are intrinsically micro-crackedasthemanufacturingprocessincludesaheattreatment resulting inthesintering of theyttria-stabilizedzirconia. The occurrence of thiscrack network has the main advantage to release residualstressesandstabilizetheTBC.However,this specific microstructure alone shows poor resistance tocyclic oxidation and needsto bereinforced priorto long-termhigh temperatureexposure. Thefeasibilityofconsolidatingsol– gel TBCsbyfillingtheprocess-inducedcrackgroovesbyadditional YSZmaterialbroughtupeitherbydip-coatingorspray-coating, hasbeenaddressed.Theparametersusedtoproperlyconduct the fillingprocesshavebeenoptimised,regarding mainlythe cyclicoxidationresistanceofreinforcedTBC.Themainresults arelistedbelow.
• Thedip-coatingmethod,consistinginwithdrawingsubstrate
tobecoatedfromaslurry,isessentiallydirectionalandresults inanon-symmetric fillingofcrack,whilethefillingbythe spray-coatingmethodishomogeneous.
• The spray-coating technique, consisting in spraying
uni-formly YSZ, results in a more efficient and a more homogeneousfillingofcracks,
• Spray-coatedreinforcedTBCsshowlongercyclicoxidation
lifethandip-coatedreinforcedTBCsasaresultoftheirmore uniformmicrostructurewithlowerresidualstresses,
• Thefillingefficiencycanbeoptimisedbyaproperadjustment
oftheprocessparameters,namelytheviscosityoftheslurry thatmustbechangedasafunctionofthesizeofcrackstobe filledin,
• OptimisedreinforcedTBCsshowsahighresistanceto
oxi-dation cyclingas almostnocracks andnospallationoccur within350one-hcyclesat1100◦C,
• Increasing oxidation temperature from1100◦Cto1150◦C
dramaticallyenhancesthedegradationoftheTBCsthatmay failbytotalspallation(nonoptimisedreinforcement)orby detrimentalcrackenlargement(optimisedreinforcement),
• Failureoftheoptimisedspray-coatedreinforcedTBCduring
thermalcyclingmainlyoccurswithintheYSZ,suggestinga fairlygoodadhesionofthedepositontopofthebondcoat.
References
1.Nicholls JR, Lawson KJ, Johnstone A, Rickerby DS. Methods to reduce thethermalconductivity ofEB-PVD TBCs. SurfCoat Technol
2002;151–152:383– 91.
2.FenechJ,DalbinM,BarnabeA,BoninoJP,AnsartF.Sol– gelprocessingand characterizationof(RE-Y)-zirconiapowdersforthermalbarriercoatings.
PowderTechnol2011;208:480– 7.
3.ClarkeDR,PhillpotSR.Thermalbarriercoatingmaterials.MaterToday
2005;8:22– 9.
4.CaoXQ.Applicationofrareearthsinthermalbarriercoatingmaterials.J MaterSciTechnol2007;23:15– 35.
5.WangX,LanWH,XiaoP.Fabricationofyttriastabilizedzirconiacoatings byanovelslurrymethod.ThinSolidFilms2006;494:263– 7.
6.RenC,HeYD,WangDR.Cyclicoxidationbehaviorandthermalbarrier effectofYSZ– (Al2O3/YAG)double-layerTBCspreparedbythecomposite sol– gelmethod.SurfCoatTechnol2011;206:1461– 8.
7.ViazziC,BoninoJP,AnsartF.Synthesisbysol– gelrouteand characteri-zationofyttriastabilizedzirconiacoatingsforthermalbarrierapplications.
SurfCoatTechnol2006;201:3889– 93.
8. SniezewskiJ,LeMaoultY,LoursP,PinL,BekaleVM,MonceauD,etal. Sol– gelthermalbarriercoatings:optimizationofthemanufacturingroute anddurabilityundercyclicoxidation.SurfCoatTechnol2010;205:1256– 61. 9.PinL,AnsartF,BoninoJ-P,MaoultYL,VidalV,LoursP.Processing,
repairingandcyclicoxidationbehaviourofsol– gelthermalbarriercoatings.
SurfCoatTechnol2011;206:1609– 14.
10.ShaneM,MecartneyML.Sol– gelsynthesisofzirconiabarriercoating.J MaterSci1990;25:1537– 44.
11.FenechJ,ViazziC,BoninoJ-P,AnsartF,BarnabéA.Morphologyand structureofYSZpowders:comparisonbetweenxerogelandaerogel.Ceram Int2009;35:3427– 33.
12.TolpygoVK,etal.Morphologicalevolutionofthermalbarriercoatings inducedbycyclicoxidation.SurfCoatTechnol2003;163–164:81– 6. 13.MummDR,EvansAG,SpitsbergIT.Characterizationofacyclic
displace-mentinstabilityforathermallygrownoxideinathermalbarriersystem.
ActaMater2001;49:2329– 40.
14.TolpygoVK,ClarkeDR,MurphyKS.Oxidation-inducedfailureofEB-PVD thermalbarriercoatings.SurfCoatTechnol2001;146–147:124– 31. 15.Vande-PutA,OquabD,MonceauD.CharacterizationofTBCsystemswith
NiPtAlorNiCoCrAlYTabondcoatingsafterthermalcyclingat1100◦C:a
comparativestudyoffailuremechanisms.MaterSciForum2008:213– 21. 16. TolpygoVK,ClarkeDR.Surfacerumplingofa(Ni,Pt)Albondcoatinduced
