Surface integrity after pickling and anodization of Ti–6Al–4V titanium alloy

HAL Id: hal-01150411

https://hal.archives-ouvertes.fr/hal-01150411

Submitted on 11 May 2015

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.

Surface integrity after pickling and anodization of

Ti–6Al–4V titanium alloy

Eric Vermesse, Catherine Mabru, Laurent Arurault

To cite this version:

Eric Vermesse, Catherine Mabru, Laurent Arurault. Surface integrity after pickling and anodization

of Ti–6Al–4V titanium alloy. Applied Surface Science, Elsevier, 2013, vol. 285 (Part B), pp. 629-637.

�10.1016/j.apsusc.2013.08.103�. �hal-01150411�

Any correspondence concerning this service should be sent to the repository administrator:

staff-oatao@inp-toulouse.fr

Identification number: DOI: 10.1016/j.apsusc.2013.08.103

Official URL:

http://dx.doi.org/10.1016/j.apsusc.2013.08.103

This is an author-deposited version published in:

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

Eprints ID: 13885

To cite this version:

Vermesse, Eric and Mabru, Catherine and Arurault, Laurent

Surface integrity

after pickling and anodization of Ti–6Al–4V titanium alloy

. (2013) Applied

Surface Science, vol. 285 (Part B). pp. 629-637. ISSN 0169-4332

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.

Surface

integrity

after

pickling

and

anodization

of

Ti–6Al–4V

titanium

alloy

Eric

Vermesse

_,

_Catherine

_Mabru

_,

_Laurent

_Arurault

b_{UniversitédeToulouse,InstitutCarnotCIRIMAT,UMRCNRS-UPS-INP5085,UniversitéPaulSabatier,118,routedeNarbonne,}

31062Toulousecedex9,France

Keywords: Titaniumalloy Pickling Anodization Roughness Internalstresses

a

b

s

t

r

a

c

t

ThesurfaceintegrityofTi–6Al–4Vtitaniumalloywasstudiedatdifferentstagesofsurfacetreatments, especiallypickling andcompact anodization,through surfacecharacteristicspotentially worsening fatigueresistance.

Nosignificantchangesoftheequiaxemicrostructureweredetectedbetweensamplecoreandsurface, orafterthepicklingandanodizationsteps.Surfacehydrogenandoxygensuperficialcontentswerefound toremainunchanged.Roughnesscharacteristics(i.e.Ra,RzbutalsolocalKtfactor)similarlyshowedonly

slightmodifications,althoughSPMandSEMrevealedcertainrandomlocalsurfacedefaults,i.e.pitsabout 400nmindepth.Finallyinternalstresses,evaluatedusingX-raydiffraction,highlightedasignificant decreaseofthecompressiveinternalstresses,potentiallydetrimentalforfatigueresistance.

1. Introduction

Titaniumalloysareincreasinglyusedintheaeronautical indus-tryduetotheirgoodmechanicalpropertiesandtheirlowdensity. The most commonly used is Ti–6Al–4V because it is a good compromisebetweentitanium’sproperties.However,additional surfacetreatmentisusuallyrequired[1,2]toincrease,forinstance, superficialmechanicalproperties(i.e.tribologicalproperties,wear resistance and superficial hardness), as well as to improve its behaviourwithrespecttocorrosionbyfluoridedacidicsolutions. Surfacetreatmentsontitaniumalloysusuallyinvolvethreemain steps.Thefirstisdegreasingstepwhichremovestheoiland impu-ritiesonthesurface,leftbypreviousmachiningstages.Thesecond ispicklingcommonlyperformedbychemicallycontrolled corro-siontypicallyinmixedhydrofluoric-nitricacidbath.Theaimofthis pre-treatmentistoremovethenaturalpassivatinglayer.Finally, thethirdstageisthemaintreatment,i.e.nowadays thermodiffu-siontreatments(nitriding,carburizing)orshotpeeningtreatments

[3–5]. Anothersimple and cost-effective treatment is anodiza-tion,creatingeitheracompacttopfilmoraporousone,including numerousmesopores[6,7].Porousanodicfilmsarecurrentlybeing widelystudiedtoprepareinnovatingphotovoltaiccells [8,9]or toenhanceosteointegrationforbioapplications[10].Incontrast,

∗ Correspondingauthor.Tel.:+33561556148;fax:+33561556163.

E-mailaddress:arurault@chimie.ups-tlse.fr(L.Arurault).

compactanodicfilmsareusedbothtocolourthetitaniumsurface

[11–13]andtoimprovepaintadhesionforaeronauticparts[2,14]. However,previousindustrialandacademicstudieshave unfor-tunately shown, especially for aluminium alloys (AA), that the anodizationprocesscausessignificantmodificationsoftheAA sur-faceintegrityandasubsequentdecreaseofthefatigueresistance

[15–18].To ourknowledge,noresearch workshave previously studiedandexplainedthepossibleinfluenceofsurfacetreatments, especiallypicklingortheanodization,onfatigueresistancein tita-niumalloys.InthecaseofAA,decreasedfatigueresistanceafter anodizationisunderinvestigationandhasbeenmainlyattributed to influent surface parameters such as: microstructure [19,20], uptakeofembrittlingchemicalspecies[20–22],roughness[23,24]

andinternalstresses[25,26].

Microstructurehasanimportanteffectonthefatigueresistance oftitaniumalloys[19,20].Themain microstructuralparameters areagrainsizeandthepercentageandmorphologyofbphase. For example,alamellarmicrostructuredoesnothavethesame mechanicalpropertiesasequiaxemicrostructure[20].Some sur-facetreatments,suchasnitridingorcarburizingtreatment,canalso affectthemicrostructureandconsequentlythefatigueresistance.

Hydrogen,carbon,nitrogenandoxygenarethefourmain poten-tialuptakingandembrittlingspecies.Theirrespectiveeffectsare different but hydrogen embrittlement is well-known to affect mechanicalpropertiesofmetalparts.Thispointisparticularly crit-icalasOgawaetal.[27]clearlyshowedusinghydrogenthermal desorptionthattheamountofabsorbedhydrogeninabetatitanium alloyincreasedwithimmersiontimeinfluoridedacidsolutions. http://dx.doi.org/10.1016/j.apsusc.2013.08.103

Theroleofroughnessonfatigueresistanceofmetalpartsiswell knownand hasbeenstudiedfora longtime.Roughnesscanbe significantlymodifiedbysurfacetreatments,suchasforinstance thepicklingstep. Thus, picklingof AAsubstrate oftenmodifies theroughness throughdissolution of microprecipitates initially includedinthemultiphaseAAmatrix[18].Oncommerciallypure titanium,picklingperformedinaconcentrated(48%)sulphuricacid bathissimilarlyusedtoincreaseroughnessinordertopromote osteo-integration[28].

Finally,internalstressescontributetotheaveragestressapplied tothematerialandconsequentlyitsfatigueresistance. Internal stressesmaybegeneratedbymechanicaltreatment(shotpeening, ballburnishing),thermaltreatment(quenching,annealing)and/or (electro)chemicaltreatment(nitriding,anodization).Forinstance, anodizationinducescompressiveortensileinternalstresses,whose intensitiesdependontheoperationalparametersofthe electro-chemicaltreatment[29–32].

Inthepresentwork,theaimwastostudytheinfluenceof sur-facetreatment(picklingandanodization)onsurfaceintegrityof Ti–6Al–4Vtitaniumalloy,widelyusedforaircraftpartssuchasfor exampleenginepylons.

Surfaceintegrity wasstudiedthrough surfacecharacteristics potentiallyinfluencingfatigueresistance.Firstlythe microstruc-tureofTi–6Al–4Valloywascarefullystudiedthroughtwotypes ofsubstrates,i.e.rolledsheetandforgedbar.Secondly,thiswork onlyfocussedonhydrogenandoxygenuptakesbecausecarbon andnitrogenarenotinvolved,eitherinpicklingtreatmentorin anodizationtreatment.ThenroughnesswasstudiedthroughRaand R_zvaluesbutalsoanapproachbasedonthelocalstress concentra-tionfactor(localKt).Erraticandpunctualdefaultswereobserved

andcharacterisedtocompletethemorphologicalanalysis.Finally, internalstressesweremeasuredusingXRDmethods.These dif-ferentcharacteristicsweredeeplystudiedatbothstepsofsurface treatment(pickling,anodization)whileendurancelimitwasfinally evaluatedinordertobothclarifythesurfacechangesandpredict theirpotentialimpactonfatigueresistance.

2. Experimental

2.1. Surfacepreparation

ThesubstratematerialwasTi–6Al–4Vtitaniumalloy.Its chemi-calcompositioninweightpercentwas:5.5<Al<6.5%,3.5<V<4.5%, C≤0.08%, O≤0.20%, N≤0.05%, Fe≤0.30%, H≤0.0125% with Ti accountingfortheremainder.Twodifferentsubstrateswereused inthisstudytofinallyobtainthreedifferentsurfacestates.

The first substrate was Ti–6Al–4V rolled sheet (45mm×60mm×1mm). From this substrate, two surfaces wereprepared:rawrolledsheetandpolishedsheet.Thepolished samples were obtained using 800, 1200, 2400 grade polishing paperdiscsand6,3then1mmdiamondpastepolishingpads.

The second substrate was Ti–6Al–4V forged bar (length: 105mm;diameter:16mm)turned(feedrate:0.1mm/revolution; cuttingspeed:25m/min)toobtainafatiguespecimen(working length:20mm;workingdiameter:8mm),thatmadeupa third surfacestate.

Allsamplesweredegreasedwithethanolandthenacetone.They werethenpickledin aqueous20w% HNO3 and2w% HFmixed

solutionat20◦_{Cfor200swithoutstirring.}

Finally,theanodizationwasperformedinanelectrochemical cell,wherethetitaniumsubstratewasusedasanodeandalead plateascounter-electrode.Theanodizationwasrunfor2minin thedirectvoltagemode(5–80V)usingasulfuricacidsolution(1M) thermallyregulatedat20◦_{C. Sampleswererinsedwithdistilled}

wateraftereachstep.

2.2. Characterizations

Thethicknessoftheanodicfilmwasmeasuredusinga Ben-thamPVE300reflectometer,withaTMc300monochromatorinthe 300–900nmwavelengthrange.Fortheanodicfilm,therefractive indexprovidedbyDiamantietal.[11]wasused.

Secondary Ion Mass Spectrometry(SIMS) surface and cross-sectionalanalysiswereperformedwithaCamecaIMS4F6device. Theareaanalysedwas150mm×150mmusinga10keVCs+

pri-marybeamwith10–50nAcurrentrange.Itisimportanttonote thatthisareaofanalysiswaslargerthantheaandbgrainsizes (maximum40mm2).

Priortomicrostructuralobservations,samplesweredippedfor 10sinamixedacidsolution(4%HF;3%HNO3)atambient

tem-perature.Samplemicrostructureswereobservedwithascanning electronmicroscope(XL30ESEM),whileaandbgrainsizeswere estimatedbyanalyzingSEMviewswiththefreesoftwareImageJ

[33].Additionalobservationsofthesamplemicrostructurewere carriedoutusinganopticalmicroscope(OM)(OlympusGX71).

Elementary chemical analyses were performed by energy-dispersiveX-rayspectroscopy(Rondec-EDX)coupledwithaSEM device,inordertoobtainasemi-quantitativeanalysisofV,Ti,Al contentsinbothtypesofgrains.

AMahrperthometerPGK120(contactmodeinambient atmo-sphere,diamondtipwith2mmradius)wasusedtoaccesstothe amplitudeparametersof theroughness,i.e.heretheroughness average(Ra)androughnessheight(Rz),definedby:

Ra=1 l

Z

0

|z(x)|dx (1)

withlthelengthofprofileandz(x)theprofileheightdistribution withrespecttomeanline

Rz=1 5





X

X





where(zi)maxand(zj)minarerespectivelythefivehigherlocal

max-imaandlowerlocalminimaoftheprofileheightdistribution(z). TheRaandRzvaluesshowninthispaperweretheaverage

val-ues,resultingfrom4to40measurements.Correspondingaccuracy waslow(about0.01mm)whilethestandarddeviationwasabout 0.05mmforRaand0.5mmforRz.OnlyRaandRzvalueswereused

inthispaperbutthefinalconclusionwouldremainthesamewith theotherroughnessindexes.

Unfortunately,roughnessindexesdonotalwaysprovide suf-ficient information to know the real effect of morphology on fatigueresistance[34,35].Consequently,Suraratchaïetal.[34]and Shahzadetal.[16]proposedtousethelocalKttodeterminethe

morphologyimpacton thefatigueresistance. Thespecificity of theK_tparameteristhatittakesdepthandsharpnessofthelocal defaultsintoaccount.Itscalculationisbasedonroughnessprofile analysisandfiniteelementssimulation[34].Inthismethod, rough-nessprofilewasfilteredtoextractits“useful”part,thecutlength, assimilatedtoa₀[23],beingherecloseto10mm.Standard devia-tionforKtisestimatedtobeequalto0.03formachinedsamples

and0.5forlaminatesamples.

Ascanningprobemicroscope(SPM-Bruker)incontactmodein ambientatmospherewitha10nmradiusanda10mmlong can-tilevertipwasusedtocharacteriseerraticlocalsurfacedefaultsat themicro-andnano-levels(50mm×50mm).

TheinternalstressesweremeasuredbyX-raydiffraction (X-pertPhilipsdevice)usingthewell-known“sin2 _{method”firstly}

introducedbyMacherauch[36,37].Measurementswereonly car-riedoutonunpolishedrolledsamplesbecausetheXRDtechnique isunsuitableforcylindricalsamples(likemachinedforgedbarused

forfatiguespecimen)andpolishingcaninduceadditional uncon-trolledstresses.Allmeasurementswereperformedinthelaminate directionusinga213diffractionraywithscanningangle(2)from 136◦_to146◦_{.Themeasurementofinternalstresseswasperformed}

in a material depth depending on the XRD analysis operating parameters(angle,X-raytube),butdepthisusuallyconsideredas aboutthefirst14mm[38].

Preliminaryfatiguetestswereperformedonaservohydraulic machineMTS-810at10Hzusingaloadratioof0.1.Specimen geom-etry,alreadyusedinpreviousworks,wasdescribedforinstancein Ref.[16]:gaugelengthwas20mmanddiameterwas8mm.The specimensweremachinedfromauniquebar.Threegroupsofseven specimensweretested:thefirstonewithoutsurfacetreatment, anotheroneonlypickledandthelastonepickledandthenanodized at80V.

3. Resultsanddiscussion

Asmentionedintheintroduction,surfaceintegritywasstudied throughsurfacecharacteristicspotentiallyinfluencingthefatigue resistance,i.e.microstructure,speciesuptake,roughnessand inter-nalstresses.

3.1. Microstructure

Ti–6Al–4Vrolledsheet(Firstsubstrate)showedafineisotropic equiaxemicrostructure(Fig.1),homogeneousthroughthewhole thickness.Thesecond substrate(Ti–6Al–4V forgedbar)showed similar fineequiaxe-type microstructure in theradial direction (Fig.2a)butalsospecificallylongfinebgrainsorientedalongthe longitudinaldirection(Fig.2b).Theaveragediameterofbgrains wasabout1mmwhilethediameterofagrainswasestimatedto beabout5mm.

Image analysis showed that the surface areaof b grains is 10±2%forbothrolledsamplesandradialobservationoftheforged sample,and7±2%forthelongitudinaldirectionoftheforged sam-ple.Theseresultsagreewiththefactthat(a+b)titaniumalloys showbphaseinthe5–20%range,especiallycloseto12%forthe Ti–6Al–4Valloy[39].

Therewerenosignificantdifferencesofmicrostructurebetween samplecores andtheirinterfaces,i.e.inthelastsuperficialone thousand microns. Moreover, no changes were experimentally observedafterthepicklingandanodizationsteps.These results werefirstlyexplainedbythesimilarchemicalcontentandchemical distributioninbothtypesofsubstrates.Secondly,bothprocesses are indeed known to usually induce no changes of superficial microstructurebecauseoftheambientormoderate operational temperatures used,contrarytothermodiffusion treatments(i.e. nitriding [5]) or high energy treatments (i.e. micro-arc oxida-tion,lasertreatment[40,41]).So,usingpicklingandanodization,

microstructurecannotbeconsideredasakeyparameterfor poten-tialfatigueresistancemodificationofTi–6Al–4Vtitaniumalloy.

3.2. Hydrogenandoxygenuptake

Thisworkfocussedontheuptakeofhydrogenandoxygenboth knowntoaffectmechanicalproperties[42,43]andbothinvolved inthepicklingandtheanodizationprocesses,unlikecarbonand nitrogen.TostudythehydrogenstoragecapacityoftheTi–6Al–4V, temperatureprogrammeddesorption(TPD)andavolumetric sorp-tiontechniquehavebeenperformed[44].Thediffusionkinetics ofoxygenintoTi–6Al–4Vsubstratewasmeasuredbyweightgain measurements,oxygendiffusionzone(ODZ)depthandhardness measurementsonthecross-sectionsofoxidisedalloy[45].Using SIMS,Thairetal.[46]obtainedtheelemental(Ti,Al,V,N,O)depth profilesofpassivelayersof(un)implantedspecimensofTi–6Al–4V, whileLamolleetal.[47]studiedtheimpactofthepickling dura-tionontheSIMSdepthprofilesofoxide,fluorideandhydrideinthe surfaceofcommerciallypure(cp)titanium.

Inthepresentstudy,SIMSwasselectedasitisverysensitiveand providestheelementalin-depthprofilesofthepassivelayerandthe beginningofthemetalmatrix.However,quantitativevalueswere notobtainedbutthemethodwasefficienttocomparesignificant variationsoftheelementspresent.

Fig.3showstheoxygenprofilesaftereachtreatment,i.e. pol-ishing, pickling (20w% HNO3–2w% HF solution for 200s) and

anodization(at80V).Forpolishedandpickledsamples,theinitial peaksattheextremesurface(sputerringtimelowerthan200s) wereattributedtopassivelayers.Sittigetal.[48]highlightedthat thenaturalpassivelayer(beforepickling)onTi–6Al–4Valloyis usuallymadeupofTiO2,butalsoincludesAl2O3andV2O5.

Lam-olleetal.[47]pointedoutthatprolongingthepicklingtime(upto 150s)in0.2%HFcausedahigheroxidecontentatthecptitanium passivationlayer.ForTi–6Al–4Valloy,ourresultsdidnotshowsuch anincrease,maybeduetothedifferenttitaniumsubstrateandthe lackofoxidisingcompound(i.e.HNO3)inthepicklingsolutionused

byLamolleetal.[47].

Foranodizedsamples,theprofilefirstpresentsanincreaseand thenreachesaplateauduetotheformationoftheanodicfilm,about 200nmthick,accordingto:

Ti+2H₂O→ TiO₂+4H++4e (3) Thenthereisa drasticdecreaseoftheSIMSprofileatabout 1500sofsputtering,clearlyindicatingtheinterfacebetweenthe bottomoftheanodicfilmandthetopofthemetalalloy.

Aftereach treatment(polishing,pickling,anodizing), oxygen intensityinbulkmaterialwasverylow, indicatingthatno oxy-genuptakeoccurredinthebulkalloy.So,apotentialdecreaseof fatigueresistanceduetooxygenuptake[20],usuallyattributedto titaniumhardening[49],isnotexpectedintheseconditions.

Fig.2.SEMviews(a)inradialdirectionand(b)inlongitudinaldirection(arrow)offorgedbarmicrostructure,previouslypickledinamixed4w%HF–3w%HNO3solutionfor 10satambienttemperature.

Fig.3alsoshowsthehydrogenSIMSprofileaftereachtreatment, i.e.polishing,picklingandanodization.Afterpolishing,theSIMS profileshowedahighinitialpeakattributedtothehydrationofthe naturalpassivelayer,mainlymadeupofTiO2.

Afterpickling(20w% HNO₃–2w%HFsolution for200s),the interfacialpeakwasclearlylowerthanthepreviousone(after pol-ishing).Thisresultcouldbeexplainedbyconsideringthepossible picklingreactionsoccurringin20w%HNO3and2w%HFaqueous

mixedsolution:

6HF+TiO2→ 2H++2H2O+TiF62− (4)

3HF+Ti→ 3/2H₂+3F−+Ti3+ (5)

Fig.3.(a)Hydrogenand(b)oxygenSIMSprofilesobtainedonrolledsampleafter threedifferentsteps:polishingthenpickling(HNO3–HF(20–2w%)solution,20◦C, 200s)andfinallyanodization(80V,1MH2SO4,20◦C,2min).

Eq.(5)clearlyshowsthatH2 gasevolutioncanoccurbutalso

thattitaniumoxidereactswithfluorideanions.Thus,thedecrease oftheinterfacialpeakafterpicklingcouldperhapsbeexplained bythemodificationof theinitialhydratedpassivelayerbythe fluorideions.Furthermore,Bijlmer[50]demonstratedthat hydro-genuptakedecreaseswithincreasingnitricacidconcentration.He showedinparticularthat,whenthenitricacidcontentishigher than20%inthepicklingbath,hydrogenuptakebecamecloseto zero,due toapredominantre-passivationphenomenon.So,our results,obtainedusing20w%HNO3–2w%HFaqueousmixed

pick-lingsolution,areingoodagreementwithBijlmer’swork.

For an anodized sample (80V, i.e. about 200nm thick), the hydrogenprofile increased from750sreaching a maximum at about1500sofsputtering.Thismaximumcorresponded,as pre-viouslydemonstratedfortheoxygendepthprofile,totheinterface betweenthebottomoftheanodicfilmandthetopofthebulkmetal alloy.Moreover,theareaintegrationunderthehydrogenprofiles for pickledand anodized sampleswere thesame, proving that anodizationdidnotinduceadditionalhydrogenuptake. Finally, hydrogenintensityinthebulkofthetitaniumalloyappearedlow andidenticalaftereachtreatment,confirmingthatnoadditional hydrogeninclusion wasinduced.In contrast,Lamolleet al.[47]

showedasignificantpenetrationdepthofhydride (deeperthan onemicron)butused0.2%HFalone(withoutnitricacid)aqueous solutionaspicklingbath,i.e.withoutre-passivationphenomenon. AdditionalSIMSsurfacemapping(Fig.4)wasthenperformedto clarifythehydrogenstate,whichisakeyfactorinhydrogen embrit-tlement[51]Diffusivehydrogenindeedhasadifferenteffectthan

Fig.4.SIMSsurfacemappingofhydrogenonapolishedandthenpickledsample (HNO3–HF(20–2w%)solution,20◦C,200s).Inthemiddleofthesampleappearsa sputteredcraterafewmicronsindepth.

Table1

RoughnessandlocalKtvaluesforrolledandpolished,rolledandmachinedsamplesbeforeandafterpickling(HNO3–HF(20–2w%)solution,20◦C,200s).

Polished Rolled Machined

Initial(mm) Etched(mm) Variation(%) Initial(mm) Etched(mm) Variation(%) Initial(mm) Etched(mm) Variation(%)

Ra 0.06 0.08 33 0.42 0.42 0 1.01 0.96 −5.3

Rz 0.28 0.58 107 2.72 2.70 −1.8 4.91 4.62 −6.4

Kt 1 1 0 2.31 2.16 −6.5 1.15 1.16 0.9

hydrideprecipitates[52,53],thatdecreasetheductilityofthe mate-rial[54].Thesamplewasinfactdividedintotwozonesinorderto comparethetopsurfaceandthe“bulk”material(Fig.4):inthe middlewaslocatedasputteredcraterafewmicronsindepthand aroundthetopsurfaceoftheanodicfilm.Thecolourhomogeneity attestedthathydrogenwasuniformlydistributedthroughthefilm, i.e.bothonthesurfaceandinthebulkoftheanodicfilm. Consid-eringthatnochemicalheterogeneitywasdetected,itisassumed thatinthepresentcasetherewerenohydrideprecipitatesbutonly diffusivehydrogen.NakasaandSatoh[21]claimedthatdiffusive hydrogenisusuallyincorporatedintotheb-phasebutisless detri-mentalthanthehydridefromthea-phase,thatcanbecomeacrack nucleationandpropagationpathunderloading.

Thus,theoxygenandhydrogenuptakeswereverylimitedin Ti–6Al–4Valloyusingsuchsurfacetreatments(polishing,pickling, anodization)inouroperatingconditions.So,theeffectof varia-tionsinthehydrogenandoxygencontentswillbeconsideredas negligiblefromthefatigueresistancepointofview.

3.3. Averageroughnessvaluesandlocaldefaultmeasurements

The morphological modifications of the surface which can impactthefatigueresistancewerethenanalysed.Average rough-ness values (Ra and Rz) were firstlyanalysed to have a global

descriptionofthesurfacemodifications.Consideringthatthereal impactofsurfacemorphologyonfatigueresistanceisincompletely describedthroughroughnessindexes[34,35],localKtvalueswere

secondly obtained. However, these values do not directly take localandrandomdefaultsintoaccount.ThatiswhySEMandSPM analysiswereperformed(onpolishedsurfaces)todeterminethe occurrenceofanypossiblelocaliseddefaults.

3.3.1. Onpickledsamples

Roughnessvalues(RaandRz)weremeasuredbeforeandafter

picklingonthree typesofsamples(rolled,rolledthenpolished, machined)withdifferentinitialroughness(Table1).

For rolled then polished samples, Ra and Rz changed from

0.06±0.01mmand0.28±0.01mmrespectivelybeforepicklingto 0.08±0.01mmand0.58±0.02mmafterpickling.So,Raincreased

about33%and Rzabout107% (Table1).Thesearehighrelative

increasesbutthefinalroughnessvaluesstayedverylow.Whenthe initialroughnesswashigher,i.e.forrolledormachinedsamples, roughnessindexes(measuredperpendicularlytotherolldirection andinthelongitudinaldirectionrespectively)didnotincreaseor evenshowedaslightdecrease(Table1).Forexample,changeswere closeto0%forrolledsamplewithinitialRaof0.42±0.07mm,while

Ra and Rz decreased, about−5.3%and −6.4% respectively,after

picklingthemachinedsample.However,theseroughnessindex changeswerelowerthanstandarddeviation,i.e.SD>15%.

Roughnesschanges(increaseordecrease)canresultfrom differ-entphenomena:preferentialmorphologicaldissolutionofspecific zones (tip/hollow), preferential chemical dissolution of specific phasesofthealloymatrix,formationoflocalcorrosionpits,etc.

Firstly,preferentialmorphologicaldissolutionofthepeaksthat weremoreexposedtothepicklingbathcanoccur.Indeed,the dis-solutionratewouldbeprobablyhigheronpeakscomparedwith

Fig.5.SEMsurfaceviewofrolledsample,polishedandthenpickled(HNO3–HF (20–2w%)solution,20◦_{C,200s)showingpitsandb-phase(whiteparts).} therateobservedinhollowssincemasstransportandbathrenewal wouldbemoredifficultinhollows.

Secondly,thereispreferentialchemicaldissolutionofaphase, asshownbySEM(Fig.5)wherebgrainsarehigherthantheaphase. SPM analysis(Fig.6) provided anapproximate height ofabout 800nmforthebphase,i.e.lowerthantheaveragediameterofthe b-grain(≈1mm).Thispreferentialchemicaldissolutioncouldbe explainedbyconsideringtherespectivechemicalcompositionsof bothaandbphases.Indeedtheb-phasehasahigherconcentration invanadiumandthea-phaseahigheraluminiumcontent(Table2). Knowingthatthestandardpotentialofaluminiumandvanadium arerespectively−1.66V/SHEand−1.13V/SHE,theb-phasecanbe consideredasmorenoble(containingmorevanadium)thanthe a-phasewithlessvanadiumandmorealuminium.

Thirdly,anotherwaytoincreaseroughnessispitting.However,

Fig.6showsthatpitswerelocalandlessthan1mmindiameter, andthepitdensitywasevaluatedataboutonepitfor200mm2. Pitscanbefoundatthegrainboundaries,i.e.theinterfacebetween aandbphases,oralsointhemiddleofagrains.SPManalysis revealedpitdepthofabout400nm(Fig.6).So,preferentialchemical dissolutionandpittingphenomenaresultingfromthepicklingstep inducedfaults(respectivelyabout800and400nm)lowerthanthe rollroughness(Rz=2.70mminTable1).Boththesetypesofdefaults

couldexplaintheroughnesschanges(Rz=0.58mminTable1)in

polishedsubstrate.

Surprisingly, local Kt remained unchanged in polished and

machinedsamples, while rolledsamplesshoweda Kt decrease

(−6.5%). Eventhough itsvariation seemsto besignificant, it is clearlylowerthanstandarddeviationonlocalK_t analysis,which reached30%fortherolledsample.Thelowinfluenceoffaultson Table2

EDXevaluationofaluminiumandvanadiumcontentsintheaandbphasesinrolled Ti–6Al–4Vtitaniumalloy.

%Al %V

aPhase 5.6 2.9

Fig.6.SPMsurfacemappingandprofileofpolishedthenpickledsample(HNO3–HF(20–2w%)solution,20◦C,200s)showingapit. localKtvaluecanbeexplainedbytooshallowsharpnessanddepth

offaults.Forinstance,inrolledsample,defaultshadalowerdensity and/oralowerdepththantheinitialguidemarksfromtherolling process(Table1).

3.3.2. Onanodizedsamples

Ra, Rz, Kt changes after anodization of rolled samples,

per-formedatdifferentconstantvoltages(5–80V)arereportedinFig.7, i.e. as function of the thickness of the compact anodic film in the16–195nmrange (Table3).Ra values changedsignificantly,

maximalvariationsbeingabout7%betweenthethinnestandthe thickestfilms.Althoughsignificant,thevariationsstayedverylow. InthecaseofRzvalues,thechangeswerelarger(upto8%)butthey

dependedmoreonlocalfaults.LocalKtchangeswerehigherthan Rz,butthecorrespondingstandarddeviationwaslarge(30%)due

toirregularroughnessprofiles.

So,noeffectoftheanodizationvoltage(i.e.thecompactfilm thicknessinthe16–195nmrange)wasnotedontheroughnessor localKt.

ComplementarySEMsurfaceviewsshowedthatthe morphol-ogyoftheanodizedsurfacewasunchangedwhatevertheanodic

filmthickness.Forinstance,SEMobservations(Fig.8)ofthetop sur-faceofthethickestanodicoxidefilmclearlyshowedpitssimilarto thepreviousonesresultingfrompicklingstep(Fig.5);besidesthe pitdensityremainedthesame.However,additionalconcretions appeared(Fig.8)onthesurfaceofthisanodicfilmbutonlywhen preparedatthehighestvoltage(80V).Theconcretionsoccurringat thehighvoltagerequiredtoobtainthicklayers,probablyresulted fromtheoccurrenceofmicro-sparksduringtheanodization.

Table3

Thicknessvariationasafunctionofdirectanodizationvoltage(1MH2SO4,20◦C, 2min.)onrolledthenpickledsamples(HNO3–HF(20–2w%)solution,20◦C,200s). Thicknesseswereobtainedusingreflectometryanalysis.

U(V) e(nm) 5 16 20 60 35 94 50 130 65 164 80 195

Fig.7.Changesof(a)Ra(b)Rzaverageroughnessvaluesand(c)localKtvaluesasafunctionoftheanodizationvoltage(5–80V),i.e.thethicknessofthefilm(16–195nm) preparedonrolledsamplepickled(HNO3–HF(20–2w%)solution,20◦C,200s)thenanodized(1MH2SO4,20◦C,2min).Thedottedlinescorrespondtostandarddeviation.

Tosumup,thegrowthoftheanodicfilmonTi–6Al–4Valloy inducednoorslightmodificationsoftheroughness(Ra,Rz)and

localdefault(Kt)parameters.Theyappearedalmostunchanged,

compared with those obtained before anodization, i.e. after

Fig.8. SEMsurfaceviewofrolledsample,polishedandpickled(HNO3–HF(20–2w%) solution,20◦_{C,200s)thenanodized(80V,1MH}

2SO4,20◦C,2min).

pickling.So,thegrowthofthecompactfilmcouldbeconsidered assimplyarisingfromthepreviousroughnessandlocaldefault characteristics,andcouldhavenoadditionalinfluenceonfatigue resistancefromthispointofview.

3.4. Internalstresses

Therearemanytechniquestomeasureinternalstress.Among them,themostwidespreadareX-raydiffraction(XRD)analysis

[55,56],incrementalholedrillinganalysis[57]orbeambending analysis[58,59].TheXRDanalysiswasselectedinourstudyfor threemainreasons:itistheonlynon-destructivetechnique allow-ing analysisof the internal stress and then subsequentfatigue testingonthesamesample;thesecondreasonisthat,unlikebeam bendinganalysis,themeasurementcanbecarriedoutonthick sam-plesandthirdlythedataforstresscalculationareeasilyavailable.

Rolled sheets showed initial compressive internal stresses (−198±50MPa)inthelaminatedirection(Fig.9)in agreement withAbdelkhaleketal.[60].Thenthepicklingstephalvedthe inter-nalstresses(−100±50MPa).Knowingthatinthiscasethepickling removedabout3mm thicknessfromtheoutersurfacesuggests thattheinternalstresses weremainlylocatedattheouter sur-faceandthatastressgradientcouldbelocatedinthefirstmicrons

Fig.9.Evolutionofinternalstressesobtainedonfourrolledsamplesafterdifferent surfacetreatments:rolled,rolledthenpickled(HNO3–HF(20–2w%)solution,20◦C, 200s),rolledthenpickledandfinallyanodizedat5V(1MH2SO4,20◦C,2min)and arolledthenpickledandfinallyanodizedat80V(1MH2SO4,20◦C,2min). ofdepth.Moreover,stressisseentoresultfromanodization,for thethinnest(16nmat5V)andthickest(195nmat80V)films,i.e. −50±40MPaand−39±35MParespectivelyFig.9;thevaluescan beconsideredassimilar,andwereclearlylowerthantheprevious onesobtainedafterpickling(−100±50MPa).Moreover,theanodic filmwasmainlyamorphousandconsequentlydidnotcontributeto theXRDanalysis.So,themeasuredremainingcompressivestresses wereprobablylocatedatthemetal/oxideinterface,inagreement withthemechanismproposedbyNelsonand Oriani [31].They explainthattheanodizationinducesadditionaltensilestress(up to50MPa)duetovacanciesgeneratedbytheformationandthe migrationoftitaniumionsatthemetal/oxideinterface.Such com-pressivestresses,detectedateachstepofthesurfacetreatments (especiallypicklingandanodization)couldclearlyhavea detrimen-taleffectonthefatigueresistance[25,61,62].

3.5. Endurancelimit

Fatiguelifeis dividedin threeparts:crack nucleation, crack propagationand ductile failure.Crack nucleationoccurs princi-pallyatthesurface.Thus,itishighlysensitivetosurfaceintegrity.

Fig.10.Endurancelimitofthreesamples:turned,turnedthenpickled(HNO3–HF (20–2w%)solution,20◦_{C,200s),turned,pickledandfinallyanodizedat80V(1M} H2SO4,20◦C,2min).

Furthermore,itrepresentsthemainpartoffatiguelifewhenthe numberofcyclesishigh.Thatiswhy,endurancelimitswhich cor-respondingtofatiguefailureat107_{cyclesarecloselylinkedwith}

surfaceintegrity.Theexperimentalvaluesofendurancelimitasa functionofsurfacetreatmentsarereportedonFig.10.Adecrease ofendurancelimitwasobservedafterpicklingandanodizing.The decreasewasabout20MPaafterpicklingand10MPaafter anodiz-ing.Thesevariationsweremeaningfulbutappearedquitelowin comparisonwithinternalstressesmodificationsmeasuredinrolled specimens. It must benoted that theseresults arepreliminary resultsandthat furtherworkis neededtopreciselyinvestigate fatiguebehaviourandfatiguemodelsincludingtheeffectof resid-ualstresses.

Nevertheless,thispreliminaryfatiguestudyisingood agree-mentwiththepreviousresultsonsurfaceintegrityandconfirms thatinthiscasechangesofinternalstressesmodifytheendurance limit.Theseresultsshouldbetakenintoaccounttoimprovethe surfacetreatmentsfromafatiguepointofview.

4. Conclusion

ThesurfaceintegrityofTi–6Al–4Valloywasstudiedthrough surface characteristics (microstructure, hydrogen and oxygen uptakes,roughnessand localsurface defaults,internalstresses) potentially influencing fatigueresistance. These surface charac-teristics were studied in detail after both surface treatments (pickling, anodization). The main conclusions were the follow-ing:

BothrolledandforgedTi–6Al–4Vsubstratesshowedequiaxe microstructures,includingabout10±2%bgrains.Nosignificant microstructuralchangesweredetectedbetweensamplecoreand surface,orafterthepicklingandanodizationsteps.

SIMS analysis revealed then that no oxygen and hydrogen uptakeoccurredinthebulkalloy,irrespectiveofthesurface treat-ment(pickling,anodizing).Itwasalsoassumedthatnohydride precipitateswereformedbutonlydiffusivehydrogenwaspresent inourexperimentalconditions.

Themorphologicalmodificationsofthesurfacewerethen ana-lysedthroughaverageroughness(Ra,Rz)orlocalKt parameters,

aswellasusingSEMandSPMtechniquestodetectsomepossible localisedandrandomfaults.Afterpickling,localK_twasnotaffected whiletherewereslightroughnesschanges,attributedto differ-entphenomena:preferentialmorphologicaldissolutionofpeaks, preferential chemicaldissolution of thelessnoble a-phase and formationoflocalpits.Noeffectoftheanodizationcompactfilm thicknessinthe16–195nmrangewasfoundoneitherthe rough-nessparametersorlocalKt.

In contrast, XRD analysis clearly showed that there were compressiveinternalstressesand thattheirvalues significantly decreased(from−198±50MPato−39±30MPa)aftereachstep ofthesurfacetreatment.Theinternalstressesweremainlylocated atthemetal/oxideinterface.

Moreover, preliminary fatigue tests revealed a significant decreaseofendurancelimitaftereachsurfacetreatment.Finally, thissurfaceintegrityanalysisshowedthatthemodificationofthe initialcompressive internal stresses could be a key-parameter, potentiallyexplaining thedecreaseoftheendurance limitafter eachsurfacetreatment(picklingandanodization)ontheTi–6Al–4V titaniumalloy.

Acknowledgments

TheauthorsthankClaudeArmandforhishelpinperformingthe SIMSanalysisandPascalLameslefortheinternalstress measure-ments,aswellasPeterWintertonforhishelpfulcomments.

References

[1]A.Zhecheva,W.Sha,S.Malinov,A.Long,Enhancingthemicrostructureand propertiesoftitaniumalloysthroughnitridingandothersurfaceengineering methods,Surf.Coat.Technol.200(2005)2192–2207.

[2]G.W.Critchlow,D.M.Brewis,Reviewofsurfacepretreatmentsfortitanium alloys,Int.J.Adhes.Adhes.15(1995)161–172.

[3]D.G.Bansal,O.L.Eryilmaz,P.J.Blau,Surfaceengineeringtoimprovethe dura-bilityandlubricityofTi–6Al–4Valloy,Wear271(2011)2006–2015.

[4]O.I.Yaskiv,I.M.Pohrelyuk,V.M.Fedirko,D.BokLee,O.V.Tkachuk,Formation ofoxynitridesontitaniumalloysbygasdiffusiontreatment,ThinSolidFilms 519(2011)6508–6514.

[5]I.M.Pohrelyuk,V.M.Fedirko,O.V.Tkachuk,R.V.Proskurnyak,Corrosion resis-tanceofTi–6Al–4ValloywithnitridecoatingsinRinger’ssolution,Corros.Sci. 66(2013)392–398.

[6]J.M.Macak,H.Tsuchiya,L.Taveira,A.Ghicov,P.Schmuki,Self-organized nan-otubularoxidelayeronTi–6Al–7NbandTi–6Al–4Vformedbyanodizationin NH4Fsolution,J.Biomed.Mat.Res.75(2005)928–933.

[7]V.Zwilling,M.Aucouturier,E.Darque-Ceretti,Anodicoxidationoftitaniumand TA6Valloyinchromicmedia:anelectrochemicalapproach,Electrochem.Acta 45(1999)921–929.

[8]K.Y.Lee,J.Y.Kim,H.Y.Kim,Y.J.Lee,Y.S.Tak,D.Kim,P.Schmuki,Effectof elec-trolyteconductivityontheformationofnanotubularTiO2photoanodefora dye-sensitizedsolarcell,J.KoreanPhys.Soc.54(2009)1027–1031.

[9]L.Cao,C.Wu,Q.Hu,T.Jin,B.Chi,J.Pu,L.Jian,Double-layerstructurephotoanode withTiO2nanotubesandnanoparticlesfordye-sensitizedsolarcells,J.Am. Ceram.Soc.96(2013)549–554.

[10]R.Chiesa,E.Sandrini,M.Santin,G.Rondelli,A.Cigada,Osteointegrationof titaniumanditsalloysbyanodicsparkdepositionandotherelectrochemical techniques:areview,J.Appl.Biomater.Biomech.1(2003)91–107.

[11]M.V.Diamanti,B.DelCurto,M.P.Pedeferri,Interferencecolorsofthinoxide layersontitanium,ColorRes.Appl.33(3)(2008)221–228.

[12]S.VanGils,P.Mast,E.Stijns,H.Terryn,Colourpropertiesofbarrieranodic oxidefilmsonaluminiumandtitaniumstudiedwithtotalreflectanceand spectroscopicellipsometry,Surf.Coat.Technol.185(2004)303–310.

[13]Y.T.Sul,C.B.Johansson,Y.Jeong,T.Albrektsson,Theelectrochemicaloxide growthbehaviourontitaniuminacidandalkalineelectrolytes,Med.Eng.Phys. 23(2001)329–346.

[14]F.Dalard,C.Montella,J.Gandon,Adherenceofpaintsontitanium:studyand characterizationofoxidefilms,Surf.Technol.8(1979)203–224.

[15]B.Lonyuk,I.Apachitei,J.Duszczyk,Theeffectofoxidecoatingsonfatigue prop-ertiesof7475-T6aluminumalloy,Surf.Coat.Technol.201(2007)8688–8694.

[16]M.Shahzad,M.Chaussumier,R.Chieragatti,C.Mabru,F.Rezai-Aria,Influence ofsurfacetreatmentsonfatiguelifeofAl7010alloy,J.Mater.Process.Technol. 210(2010)1821–1826.

[17]M.Shahzad,M.Chaussumier,R.Chieragatti,C.Mabru,F.Rezai-Aria,Surface characterizationandinfluenceofanodizingprocessonfatiguelifeofAl7050 alloy,Mater.Des.32(2011)3328–3335.

[18]M.Shahzad,M.Chaussumier,R.Chieragatti,C.Mabru,F.Rezai-Aria,Effectof sealedanodicfilmonfatigueperformanceof2214-T6aluminiumalloy,Surf. Coat.Technol.206(2012)2733–2739.

[19]R.Boyer,G.Welsch,E.W.Collins(Eds.),MaterialsPropertiesHandbook: Tita-niumAlloys,fourthedition,ASMInternational,MaterialsPark,OH,2007.

[20]G.Lütjering,J.C.Williams,Titanium,Springer,Berlin,2003.

[21]K.Nakasa,H.Satoh,Theeffectofhydrogen-chargingonthefatiguecrack prop-agationbehaviorofb-titaniumalloys,Corros.Sci.38(1996)457–468.

[22]W.J.Evans,M.R.Bathe,Hydrogenandfatiguebehaviourinanearalphatitanium alloy,Scr.Metall.Mater.32(1995)1019–1024.

[23]M.H.ElHaddad,T.H.Topper,K.N.Smith,Predictionofnonpropagatingcracks, Eng.Frac.Mech.11(1979)573–584.

[24]M.Chaussumier,C.Mabru,M.Shahzad,R.Chieragatti,F.Rezai-Aria,Apredictive fatiguelifemodelforanodized7050aluminiumalloy,Int.J.Fatigue48(2013) 205–213.

[25]L.Wagner,G.Luetjering,Influenceofshotpeeningonthefatiguebehaviorof titaniumalloys,in:1stInternationalConferenceonShotPeening,Paris,1981, pp.453–460.

[26]M.N.James,D.J.Hughes,Z.Chen,H.Lombard,D.G.Hattingh,D.Asquith,J.R. Yates,P.J.Webster,Residualstressesandfatigueperformance,Eng.Fail.Anal. 14(2007)384–395.

[27]T.Ogawa,K.Yokoyama,K.Asaoka,J. Sakai, Hydrogenabsorption behav-iorofbetatitaniumalloyinacidfluoridesolutions,Biomaterials25(2004) 2419–2425.

[28]S.Ban,Y.Iwaya,H.Kono,H.Sato,Surfacemodificationoftitaniumbypickling inconcentratedsulfuricacid,Dent.Mater.22(2006)1115–1120.

[29]N.B.Pilling,R.E.Bedworth,Theoxidationofmetalsathightemperatures,J.Inst. Metals29(1923)529.

[30]L.C.Archibald,Internalstressesformedduringtheanodicoxidationoftitanium, Electrochim.Acta22(1977)657.

[31]J.C.Nelson,R.A.Oriani,Stressgenerationduringanodicoxidationoftitanium andaluminum,Corros.Sci.34(1993)307–326.

[32]J.-D.Kim,S.-I.Puyn,M.Seo,Effectofhydrogenonstressesinanodicoxidefilm ontitanium,Electrochim.Acta48(2003)1123–1130.

[33]C.A.Schneider,W.S.Rasband,K.W.Eliceiri,NIHImagetoImageJ:25yearsof imageanalysis,Nat.Methods9(2012)671–675.

[34]M.Suraratchaï,J.Limido,C.Mabru,R.Chieragatti,Modellingtheinfluenceof machinedsurfaceroughnessonthefatiguelifeofaluminiumalloy,Int.J.Fatigue 30(2008)2119–2126.

[35]S.K.Ås,B. Skallerud,B.W.Tveiten, Surfaceroughness characterizationfor fatiguelifepredictionsusingfiniteelementanalysis,Int.J.Fatigue30(2008) 2200–2209.

[36]H.K.Tonshoff,B.Karpuschewski,A.Mohlfeld,H.Seegers,Influenceofstress distributiononadhesionstrengthofsputteredhardcoatings,ThinSolidFilms 332(1998)146–150.

[37]I.S.Jawahir,E.Brinksmeier,R.M’Saoubi,D.K.Aspinwall,J.C.Outeiro,D.Meyer, D.Umbrello,A.D.Jayal,Surfaceintegrityinmaterialremovalprocesses:recent advances,CIRPAnn.–Manuf.Technol.60(2011)603–626.

[38]D.PuertaVelasquez,A.Tidu,B.Bolle,P.Chevrier,J.J.Fundenberger,Sub-surface andsurfaceanalysisofhighspeedmachinedTi–6Al–4Valloy,Mater.Sci.Eng. A527(2010)2572–2578.

[39]J.W.Elmer,T.A.Palmer,S.S.Babu,E.D.Specht,Insituobservationsoflattice expansionandtransformationratesofaandbphasesinTi–6Al–4V,Mater. Sci.Eng.A391(2005)104–113.

[40]A.L.Yerokhin,X.Nie,A.Leyland,A.Matthews,Characterisationofoxidefilms producedbyplasmaelectrolyticoxidationofaTi–6Al–4Valloy,Surf.Coat. Technol.130(2000)195–206.

[41]L.E.Murr,S.A.Quinones,S.M.Gaytan,M.I.Lopez,A.Rodela,E.Y.Martinez,D.H. Hernandez,E.Martinez,F.Medina,R.B.Wicker,Microstructureandmechanical behaviorofTi–6Al–4Vproducedbyrapid-layermanufacturing,forbiomedical applications,J.Mechan.Behav.Biomed.Mater.2(2009)20–32.

[42]R.W.Judy,I.L.Caplan,F.D.Bogar,Effectsofoxygenandirononthe environ-mentalandmechanicalpropertiesofunalloyedtitanium,in:Proceedingof7th WorldTitaniumConference,SanDiego,1992,pp.2073–2081.

[43]M.R.Bache,W.J.Evans,M.McElhone,Theeffectsofenvironmentandinternal oxygenonfatiguecrackpropagationinTi–6Al–4V,Mater.Sci.Eng.A234–236 (1997)918–922.

[44]J.L.Blackburn,P.A.Parilla,T.Gennett,K.E.Hurst,A.C.Dillon,M.J.Heben, Mea-surementofthereversiblehydrogenstoragecapacityofmilligramTi–6Al–4V alloysampleswithtemperatureprogrammeddesorptionandvolumetric tech-niques,J.AlloysCompd.454(2008)483–490.

[45]H.Guleryuz,H.Cimenoglu,OxidationofTi–6Al–4Valloy,J.AlloysCompd.472 (2009)241–246.

[46]L.Thair,U.K.Mudali,S.Rajagopalan,R.Asokamanic,B.Raj,Surface charac-terizationofpassivefilmformedonnitrogenionimplantedTi–6Al–4Vand Ti–6Al–7NballoysusingSIMS,Corros.Sci.45(2003)1951–1967.

[47]S.F.Lamolle,M.Marta,M.Rubert,H.J.Haugen,S.P.Lyngstadaas,J.E.Ellingsen, Theeffectofhydrofluoricacidtreatmentoftitaniumsurfaceon nanostruc-turalandchemicalchangesandthegrowthofMC3T3-E1cells,Biomaterials30 (2009)736–742.

[48]C.Sittig,M.Textor,D.Spencer,M.Wieland,P.H.Valloton,Surface characteri-zationofimplantmaterialsc.p.Ti,Ti–6Al–7NbandTi–6Al–4Vwithdifferent pretreatment,J.Mater.Scie.Mater.Med.10(1999)35–46.

[49]W.Liu,G.Welsch,Effectsofoxygenandheattreatmentonthe mechani-calpropertiesofalphaandbetatitaniumalloys,Metall.Trans.A19A(1988) 1121–1125.

[50]P.F.A.Bijlmer,Picklingtitaniuminhydrofluoric-nitricacid,Met.Finishing68 (1970)64–72.

[51]K.Yokoyama,T.Ogama,K.Asaoka,J.Sakai,Hydrogenabsorptionoftitaniumand nickel-titaniumalloysduringlong-termimmersioninneutralfluoridesolution, J.Biomed.Mater.Res.78B(2006)204–210.

[52]A.Nagaoka,K.Yokoyama,J.Sakai,Evaluationofhydrogenabsorptionbehaviour duringacidpicklingforsurfacemodificationofcommercialpureTi,Ti–6Al–4V andNi–Tisuperelasticalloys,Corros.Sci.52(2010)1130–1138.

[53]C.L.Briant,Z.F.Wang,N.Chollocoop,Hydrogenembrittlementofcommercial puritytitanium,Corros.Sci.44(2002)1875–1888.

[54]D.Hardie,S.Ouyang,Effectofhydrogenandstrainrateupontheductilityof mill-annealedTi-6Al-4V,Corros.Sci.41(1999)155–177.

[55]B.R.Sridhar,D.Mehta,K.A.Padmanabhan,Effectofrecrystallisationonthe residualstresspatternandfatiguebehaviouroftitaniumalloyIMI-685,J.Mater. Process.Technol.138(2003)284–290.

[56]H.Lee,S.Mall,S.Sathish,M.P.Blodgett,Evolutionofresidualstressesina stress-freetitaniumalloysubjectedtofrettingfatigue,Mater.Lett.60(2006) 2222–2226.

[57]C.Cellard,D.Retrainta,M.Franc¸ois,E.Rouhaud,D.LeSaunier,Lasershock peeningofTi-17titaniumalloy:influenceofprocessparameters,Mater.Sci. Eng.A532(2012)362–372.

[58]Y.Goueffon,C.Mabru,M.Labarrère,L.Arurault,C.Tonon,P.Guigue,Mechanical behaviorofblackanodicfilmson7175aluminiumalloyforspaceapplications, Surf.Coat.Technol.204(2009)1013–1017.

[59]X.L.Peng,Y.C.Tsui,T.W.Clyne,Stiffness,residualstressesandinterfacial frac-tureenergyofdiamondfilmsontitanium,DiamondRelat.Mater.6(1997) 1612–1621.

[60]S.Abdelkhalek,P.Montmitonnet,N.Legrand,P.Buessler,Coupledapproachfor flatnesspredictionincoldrollingofthinstrip,Int.J.Mechan.Sci.53(2011) 661–675.

[61]D.W.Schwach,Y.B.Guo,Afundamentalstudyontheimpactofsurfaceintegrity byhardturningonrollingcontactfatigue,Int.J.Fatigue28(2006)1838–1844.

[62]G.A.Webster,A.N.Ezeilo,Residualstressdistributionsandtheirinfluenceon fatiguelifetimes,Int.J.Fatigue23(2001)S375–S383.