A.C.
Alves
a,*
,
F.
Wenger
b,
P.
Ponthiaux
b,
J.-P.
Celis
c,
A.M.
Pinto
a,d,
L.A.
Rocha
a,e,f,
J.C.S.
Fernandes
gaCMEMS-UMinho-CenterofMicroElectroMechanicalSystems-UniversidadedoMinho,Azurém,4800-058Guimarães,Portugal b
LGPM–LaboratoiredeGéniedesProcédésetMatériaux–ÉcoleCentraleParis,France c
MTM–MaterialsEngineering–KULeuven,Belgium d
Dep.MechanicalEngineering–UniversityofMinho,Portugal e
Dep.Physics,FaculdadedeCiênciasdeBauru,UNESP-UniversidadeEstadualPaulista,Brazil f
IBTN/Br–BrazilianBranchoftheInstituteofBiomaterials,TribocorrosionandNanomedicine,Bauru,Brazil gCQE/DEQ–InstitutoSuperiorTécnico,UniversidadedeLisboa,Lisboa,Portugal
A R T I C L E I N F O Articlehistory:
Received12October2016
Receivedinrevisedform2March2017 Accepted2March2017
Availableonline4March2017 Keywords: Titanium Biomaterial Anodictreatment Corrosion EIS A B S T R A C T
Thankstoitsexcellentcorrosionresistance,goodmechanicalpropertiesandbiocompatibility,titanium hasbeenwidelyusedasdentalimplantmaterial.Apassiveoxidefilmformedontitaniumsurfaceis responsibleforitshighcorrosionresistance.Thisstudyhasevaluatedthesurfacecharacteristicsofoxide layersformedoncommerciallypuretitaniumsamplesbyanodictreatmentandtheeffectofanodic treatmentontheircorrosionbehaviour.FEG-SEMandXRDwereusedtoevaluatethemicromorphology and crystalline structure of these oxide films. Their corrosion resistance was evaluated using electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization curves. EIS was performedfordifferenttimesofimmersionandanewequivalentelectricalcircuit(EEC)isproposedtofit theexperimentaldataoftheanodicoxidefilms.Itwasconcludedthatthemorphology,composition,and structureoftheouterporouslayeroftheanodiclayerdeterminethecorrosionprotectionofthematerial. ©2017ElsevierLtd.Allrightsreserved.
1.Introduction
In comparisonwithothermetallic biomaterials(Co-Cralloys and stainless steels), titanium (Ti) possess the most adequate balance of characteristics for several biomedical applications, namelydentalimplants.Highbiocompatibility,specificstrength, and corrosion resistance may be pointed out as the most interestingfeaturesforthoseapplications[1–4].However,being abio-inertmetal,titaniumdoesnothavetheabilitytochemically bondwiththenaturalbone[5–8].
Although itwas shown thatmetallic dentalimplantscoated withhydroxyapatite (HA)display enhanced biological response compared to uncoated implants [9], lack of adhesion of HA coatingstothemetallicsubstrateand/orfractureordelamination of the coating are liable to occur [10]. Thus, several surface modification techniques, as plasma-spraying [9,11,12], ion
implantation [13–15], sol-gel deposition [16,17] and anodic treatment[10,18–24]havebeeninvestigatedinordertoproduce surfacelayerswithimprovedbioactivity.Duetotheirsimplicity, anodictreatmentsappearasattractivetechniquesfortailoringTi surfacesintermsoftopography,porosity,andcomposition[19,21– 28].
Inthe90's,IshizawaandOgino[10,19–21,29,30]wereamong thefirstauthorstoperformsurfacemodificationofTiwiththe incorporation of bio-elements suchas Caand P.An electrolyte consistingofcalciumacetateand
b
-glycerophosphatedisodium saltpentahydratewas usedtoformanodictitaniumoxidefilms containing Ca and P in different amounts depending on the processingconditions.TheauthorsobservedprecipitationofHA crystalsafterahydrothermaltreatment[10,19,20,29,30].Several authors have characterized different porous anodic oxidefilmsformedbymicro-arcoxidation(MAO)onTioritsalloys, interms ofcrystallinestructure, topography,porosity, composi-tion, corrosion behaviour and biological response [7– 10,19,21,22,24,26–48]. Electrochemical impedance spectroscopy (EIS)wasalsousedbyseveralauthorstocharacterizethosefilms
[24,31,33,34,49,50].Nevertheless,discussiononthemostsuitable
*Correspondingauthorat:UniversityofMinho-DepartmentofMechanical Engineering-CampusdeAzurém,4800-058Guimarães,Portugal.Tel.:+351253510 220;fax:+351253516007.
E-mailaddress:alexandra@dem.uminho.pt(A.C. Alves).
http://dx.doi.org/10.1016/j.electacta.2017.03.011
equivalentelectricalcircuits(EEC)for interpretingEIS resultsis stillgoingon.Mostoftheauthorsassumethattheanodicoxide
filmiscomposedoftwo-layers,aninnerbarrierlayerfollowedbya porousouterlayer,thusadoptingtheequivalentcircuitproposed byPanet.al[51]fortwo-layersoxidefilmonanodisedTi.
An acceptable equivalent circuit should meet two main conditions:firstofallitshouldprovideagoodfittingtotheEIS experimentaldataandsecondlyitshouldhaveaclearandvalid physicalmeaning.Inparticular,acleardescriptionofthedifferent paths for the flow of current should be done, allowing to distinguish between electronic and ionic charge carriers and between the faradaic and non-faradaic paths, as proposed by Grahame[52].Asthenumberoftimeconstantspresentinthedata increases,thenumberof degenerateEEC'sthat fitthedataalso increases,butmostofthemhavenophysicalmeaning.
Someofthepresentauthorshadpreviouslyreported[45]that increased concentration of calcium acetate in the mixture of
b
-glycerophosphate andcalciumacetateanodictreatment elec-trolyteledtobettertribocorrosionbehaviour,duetothealteration ofthecrystallographicstructureoftheresultinganodiclayer.The presentworkaimstostudy,forthefirsttime,theelectrochemical propertiesofTianodicoxidelayersduringlong-timeimmersionin aphysiologicalsolutionusingEIS.Inparticular,anewEECtofitthe experimentaldataisproposed,basedonadetailed characteriza-tionoftheanodiclayers.2.Experimental
2.1.Materialsandsolutions
Commercialpuretitanium (CPTi grade2, Goodfellow Cam-bridgeLimited,England)sampleswerecutfromthesameoriginal plate in square forms of 10x10x1mm. Immediately before the anodictreatmentallthesampleswerecleaned inanultrasonic bathwithacetonefor3min,etchedinKroll'sreagent(1mlHFand 5mlHNO3in44mlH2O)during10min,andcleanedagaininan
ultrasonicbathfor10mininpropanolfollowedby5minrinsingin distilledwater,andthendriedwithwarmair.
Theelectrolyteusedfor theanodic treatmentconsistedin a solution of
b
-glycerophosphate disodium salt pentahydrate (b
-GP)atdifferentconcentrationsofcalciumacetatemonohydrate (CA) as depicted in Table 1. After the anodic treatment, electrochemicaltestswereperformed inan 8g/l NaClsolution. The reagents used in the preparation of these solutions were supplied by Fluka-BioChemika, Sigma-Aldrich and Panreac, respectively.2.2.Anodictreatment
Theanodictreatmentwasperformedunderconstantvoltage mode. A DC power supply (GPR-30H10D) was used and the treatmentwascarriedoutatroomtemperaturefor1minat300V, undera sparkingregime.Aplatinumplatewasusedascathode (2cm2). The distance between the cathode and the anode (Ti
plates)waskeptconstant(8cm)foreachtreatment.Thesurface areaofthetitaniumsamplesexposedtotheelectrolytesolution was 0.358cm2. All the anodic treatments were done under
agitationinaturbulentregimebyusingamagneticstirrerrotating at200rpm.
2.3.Electrochemicaltests
Electrochemicaltestsconsistedofopencircuitpotential(Eocp)
measurements,potentiodynamicpolarizationcurves,and electro-chemical impedance spectroscopy(EIS).Alltheelectrochemical testswereperformedusinga3-electrodearrangementwherethe sampleswereusedasworkingelectrode,withanexposedareaof 0.358cm2, a Ptelectrode was used as counter electrode and a
saturatedcalomelelectrode(SCE)wasusedasreferenceelectrode. AllthepotentialsaregivenwithrespecttoSCE.
AllelectrochemicaltestswereperformedusingaReference600 potentiostat/galvanostat fromGamryInstruments.Thepotential scanrateinthepotentiodynamictestswas1mV/s,startingat0.5V belowEocpandmovingintotheanodicdirectionupto2V.
TheEIStestsweredonefordifferenttimesofimmersioninthe testsolution(0h,1h,1day,2days,4days,8days,16days,and20 days)atEocp.BeforeeachEIStesttheEocpwasmeasuredduring
30min.Theimpedancedataacquisitionwasdonebyscanninga rangeoffrequenciesfrom63kHztill10mHz,with10pointsper frequencydecade,andtheamplitudeofthesinusoidalsignalwas 10mV,inordertoguaranteelinearityoftheelectroderesponse.
2.4.CharacterizationMethods
Thetopography,microstructureandchemical compositionof theanodicoxidelayersformedonthesurfacewereanalysedby scanningelectronmicroscope,FEINova200FieldEmissionGun ScanningElectronMicroscope(FEG-SEM)equippedwithEnergy Dispersive X-RaySpectroscopy (EDS).The cross sectionsof the anodicoxide filmswerepreparedbynano-machiningusingthe focused ion beam of a SEM-FIB (Nova 600 NanoLab) and characterizedbyFEG-SEM/EDS.
Thesurfaceroughnesswasmeasuredonfiverandomlyselected samples(usingfivedifferentareasoneachsample),usingahigh resolution optical sensor from STIL (model CHR150-N), a 3D motionmeasuringstation,acontrollingsystemfromSTILandan opticsensor(F3)fromZEISS.
The structure of the anodic layers was analysed by X-ray diffraction with Bragg-Brentano, Göbel mirror mode, CuK
a
radiation(Bruker D8Discover).A scanningrange(2u
)of10 to100 was used, and the step size was set to 0.02. The peaks identification was done using Diffrac Plus Evaluation (Package Release2006)software.Thephasepercentage(distribution)was calculatedfollowingEq.(1).
%phasea¼ P Iapeaks P Iallpeaks ð 1Þ
3.ResultsandDiscussion
3.1.Characterizationoftheanodicfilms 3.1.1.Surfacemorphology
Fig. 1 presents the current density evolution during the formationoftheanodicoxidelayersat300Vinthetwodifferent electrolytes.Afteraninitialcurrentpeak,adecreaseisobservedin theearlystagesoftheanodictreatment,beingattributedtothe formationofacompactandthinoxidelayer[37–39,42,53].When thethicknessoftheoxidefilmsreachacriticalvalue,theapplied voltage promotes the dielectric breakdown of these films and micro-arcdischargestarttooccuratthesurfaceofthematerial, leading to localised melting and formation of micro-pores
Table1
Samplesunderstudyandelectrolytecomposition.
Group Electrolyte
015CA 0.02mol/lb-GP+0.15mol/lCA
035CA 0.02mol/lb-GP+0.35mol/lCA
[28,42,49].Micro-arcdischargesareidentifiedinthecurvesbythe current oscillations during the treatment. These discharges acceleratethegrowthofthelayer,whichthenleadstoadecrease ofthecurrentdensity.
A representative SEM micrograph of the untreated Ti after etching(beforetheanodictreatment)isshowninFig.2(a),where thetypicalgranularstructureofTicanbeobserved.Fig.2(b)and(c) showthesurfacemorphologyafteranodictreatmentin0.15Mand 0.35MCA,respectively.Ascanbeobserved,amultiscaleporous structurewasobtainedwithsomecracksbeingobservedonthe surfaceofbothgroupsofsamples.ThearrowsshowninFig.2(c) identifysomeofthecrackspresentinthe035CAgroup.Thedensity ofcracksanddistributionof poressizewerecalculated usinga point-countingmethodadaptedfromASTME562.Agrid consist-ingofa15linesparalleltoeachotherwithaconstantspacingof 0.05mm was superimposed on the micrograph. All the pores intersecting the lines were counted and their diameters were measured.FiveSEMimagesweretakenforeachsampleindifferent zonesofthesurfaceandfivesampleswereusedpercondition.It couldbeobservedthatfor015CAgroupthecrackdensitywasof (1.10.4)x103cracks/mm2whilefor035CAitishigher,namely
(7.11.6)x103cracks/mm2.Anincreaseofthenumberofcracks
withtheincreaseofthecalciumacetateconcentrationwasalso observedbyS-DWuetal.[53].
Also, an increase in the amount of calcium acetate in the electrolyticsolutionresultedinhigherporediameters,asshownin
Fig.3.Finally,theanodictreatmentresultedinporeswithawide rangeofsizes[28].Regardingthe015CAgroup,mostofthepores haveadiameterbetween0.5and1
m
m,withamaximumsizeof 3.5m
m.Inthe035CAgroup,thesizeofporesincreased,andtheir diameter is between 1.5 and 2m
m, for most pores, with a maximumsizeof5m
m.Comparing the pores sizedispersion of both groups, it was observedthatthe035CAgrouppresentsalargerdispersionthan 015CA.Also,thenumberofporespersurfaceareawasdifferentin bothgroups,namely(343)x103pores/mm2forthe015CAand
(243)x103pores/mm2forthe035CA.Theincreaseonthepores
size with increasing electrolyte concentration has also been reportedbyIshizawa etal. [21]for anodictitaniumoxide films containing Ca and P, processed by similar anodic treatment conditions. The microstructural differences between the two anodizedsurfacesare attributedtothedielectric breakdownof theoxideslayers[32].Theincreaseofcalciumacetate concentra-tionintheelectrolytemayimprovetheelectricalconductivityof
theelectrolyte.S. Abbasi et al.[37] stated that thevariationin electricalresistanceoftheelectrolyteisthemainreasonforthe poresizeenlargementathigherconcentrationsofcalciumacetate. In terms of roughness, the acid etchedsurface (just before anodic treatment) presented a mean roughness (Ra) of
0.420.02
m
m. After the anodic treatment the roughnessFig.1.Currentdensitiesevolutionduringtheanodictreatmenttimeat300Vfor twodifferentelectrolyteconcentrations.
Fig.2. RepresentativeSESEMimagesofthetitaniumsamplesfordifferentsurface conditions:a)etched,andanodictreatedgroupsb)015CAandc)035CA.
increasedaround4times,upto1.640.36
m
mand1.660.22m
m forgroup015CAand035CA,respectively.3.1.2.Anodicfilmscompositionandstructure
TheCa/PweightratiooftheanodicfilmswasestimatedbyEDS analysis, being strongly influenced by the calcium acetate concentration.The Ca/P ratio for the015CA group (1.610.09) waslowerthanfor035CA(3.620.11).
Under sparking conditions, local heating occurs and high temperatures canbe reached, sowater vapour and oxygenare released[18].Duringthisprocess,specificionsasCaandPcanbe retainedintheporousoxidelayer[19,20,22].Thisisoneadvantage ofthisanodictreatment,asthepresenceoftheseionsmayimprove osteointegration[43,46,47].
ThepresenceofHAisknownasbeneficialtotheimplants,asit promotesadirectbondingtotheboneandenhancementofnew bone formation around it. Many studies have proved that HA coated implants show superior histological results than the uncoatedones[30].TheCa/PratioonHAis1.67. Inthisway,it may be expected that by changing the concentration of the electrolyteoftheanodictreatment,aCa/Pratiosimilartotheone inHAmaybereached.Ishizawaetal.[19,20,30]haveshownthat severalHAcrystalsprecipitatedintheanodictitaniumoxidelayer containingCaandPafterahydrothermaltreatment,whentheCa/P ratiowassimilartoHA.
Anyway,evenwithoutHAprecipitation,samplesinvestigated inthisworkshowedenhancedhumanosteoblastadhesionwhen testedinvitro[43,46,47].
ThecrosssectionoftheanodicfilmsisshowninFig.4a-d.Both anodicfilmspresentedthreedifferentlayers.Adjacenttothebulk material (Ti) a compact film was formed (A) followed by two porousouterlayers.Theinnerporouslayer,markedinFig.4asB, presentedsmallpores,whereastheouterporouslayer(C)formed neartheoutersurfacepresentedbiggerpores.
The thickness of the anodic films was measured by image analysis. The 035CA group has a slightly increased thickness (4.381.21
m
m)comparedtothe015CAgroup(3.391.04m
m).TheX-raypatternsobtainedfortheanodictreatedsamples,as wellasforuntreatedTi,areshowninFig.5(onlythe2
u
valuesfrom 20to60areshown,asthisistherangewherethecharacteristicspeaksfor TiandTioxidesarefound). Tidiffractionpeakswere observedonallgroups:015CApresentedanatase,and035CAboth anataseandrutile.
Kimetal.[25]havefoundthatanodictreatmentoftitaniumina calciumglycerophosphate /calciumacetateelectrolyteproduced layerswithtwocrystallinephases,asrutileandanatase.Onthe otherhand,Suletal.[54] haveshownthat thestructure ofthe anodiclayerischangedfromamorphoustocrystalline,meaning that formation of the anatase, rutile or brokite occurs abovea critical oxide thickness. The electrochemical parameters, as electrolyte concentration or the current density, influence the dielectricbreakdownpotential,whichisrelatedtothe crystallo-graphic transformation. It has been reported that after the dielectricbreakdown,anataseisformedatlowerformingvoltages, whilethecombinationofanataseandrutilephasesisformedat higherformingvoltages[36,55,56].However,theformingvoltages of both phases may decrease when the conductivity of the electrolyteincreases.Inthiscase,byincreasingtheconcentration ofcalciumacetatethedielectricbreakdownchangedand,probably due to this fact, the anodic oxide layer was formed as a combination of anataseand rutile phases onthe 035CAgroup. Since norutile was detected in thecase of 015CA group,it is assumedthatthecrystallinephaseoftheanodicoxidelayerwas 100%anatase.Thephasepercentages(phasedistribution) regard-ing035CAgroupwerecalculatedbyEq.(1),andtheresultsshowed thattheanodicoxidelayerwasacombinationof70%anataseand 30%rutile.
3.2.Electrochemicalbehaviour
3.2.1.PotentiodynamicPolarizationcurves
Fig.6showsthepolarizationcurvesoftheuntreatedtitanium samples and anodic treated groups of samples. The average corrosionpotential,E(i=0),andtheaveragepassivecurrentdensity,
ipass,obtainedfrom several experiments(uptoseven) in each
group,arepresentedinTable2.
Thecorrosionpotentialshowedaclearincreasefromuntreated Titothe015CAgroup,followedbyaslightdecreasefrom015CAto 035CA. A passivation plateau was observed for all types of specimens,starting at 0.250V for Ti,whereas for015CA it was registered from 0.191V to 0.652V, being followed by a slight increaseofthecurrentthatcorrespondstoadegradationofthe protectivefilm, possibly due to thedissolution of calcium and phosphorous.Forthe035CAgroup,thepassivedomainisobserved from0.064to0.580V.The ipassvaluesshoweda two ordersof
magnitude drop fromTi to 015Ca, with a slight increase from 015CAto035CA.Althoughthepassiverangeseemstobereduced fromtheuntreatedTitobothtreatedmaterials,evenabovethis primary passive range and up to 2V (the maximum applied potential)thevaluesof current densityarealwaysmuch lower thanthoseobservedforTi.Thus,boththeEi=0andtheipassvalues
indicated an evident tendency for the improvement of the corrosionresistanceontheMAO-processedmaterials,the015CA groupshowingthebestproperties.
Il.S.Parketal.[28]havestudiedthecorrosionresistanceofCPTi (grade2)anodizedinsimilarconditions.Theyhavealsoreported higher corrosion potential value and lower current density as comparedwithuntreatedTi.Theauthorsattributedtheseresults totheincreasedresistanceoftheoxidelayertocorrosionthrough theanodictreatment.Previousstudies[8,28]alsodefendedthat, beyond theincrease in roughness,which is desirable for many biomedical applications, the increase in thickness of theoxide layerobtainedthroughMAOalsoresultsinanimprovementofthe corrosionresistance,comparedtountreatedTi.
3.2.2.OpenCircuitPotentialevolutionversusimmersiontime
TheevolutionofOCPwithimmersiontimeforuntreatedand anodictreatedTisamplesin8g/lofNaClispresentedinFig.7.After exposingallsamplestothesolution,theOCPfortheuntreatedTi increasedwhileitdecreasedfortheanodictreatedsamples.Forthe untreatedTisamples,theincreaseonOCPvalueswithimmersion timecorrespondedtothedecreaseoftheanodicreactionkinetics duetotheformationofthenaturaloxidefilmoftitaniumincontact withtheelectrolyte[41,57].Regardingtheanodictreatedsamples, thedecreaseonOCPwithimmersiontimemayberelatedtopartial
dissolution of theanodic film. Banakh et al.[35] evaluatedthe responseofanodicoxidelayerscontainingCaandPbyimmersion inasimulatedbodyfluidandreportedthatCaandPcontentinthe anodiclayersdecreasedafter14daysofimmersion.Moreover,Ca dissolution occurs at higher rates than P. However, during the immersion period, the anodic treated samples presented more positivevaluesthanuntreatedTiindicatingalowertendencyto corrosion.
3.2.3.ElectrochemicalImpedanceSpectroscopy
Fig.8showstheelectrochemicalimpedancespectraforTiinthe form of Nyquist and Bode diagrams for different times of immersion. At high frequencies (102 Hz to 105 Hz), the Bode
diagramshowsconstantvaluesof|Z|wherethephaseanglewas near 0 which corresponds to the response of the electrolyte resistance. At low and middle frequencies, the phase angle presented values approaching 90 which is a typical capacitive behaviourofacompactoxidefilm.Theprotectivecharacterofthis
filmseemedtobeenhancedwithincreasedimmersion time,as alreadyconcludedfromtheOCPevolution.
Fig.9presentstheequivalentelectricalcircuitfor thenative oxidefilmformedonthesurfaceoftheCPTisamplesusedinthe
fitting of the experimental data, containing: Rs - electrolyte
resistance,Rbf(N)-nativebarrieroxidefilmresistanceandQbf(N)
-constant phase element (CPE), accounting for the non-ideal capacitanceofthenativebarrieroxidefilm.
Qbf(N)valueswereconvertedtoCbf(N)(capacitanceofthenative
oxidefilm)byusingEq.(2),derivedfromBrug'sequation[58].
CbfðNÞ¼ QbfðNÞRsð1nÞ
h i1
n
ð2Þ
TheequivalentcircuitparametersobtainedfromEISdataforthe untreatedTiatdifferentimmersiontimesareshowninTable3.As canbeseenfromthefittingresults,thevaluesofRbf(N)increaseand
those of Cbf(N) decrease with the increased immersion time,
indicatingahigherprotectionofthepassiveoxidefilmformedon thesurfaceatlongerimmersiontime.
Fig.10presentstheelectrochemicalimpedancespectrainthe form of Nyquist and Bode diagrams, obtained at different immersiontimesfortheanodictreatedsamples.The015CAgroup presented two time constants: one corresponding to the high frequencies being shiftedto higher frequencies values as time elapses.The035CApresentedasimilarbehaviourtothe015CAat thebeginningofimmersionbuttheresistancecorrespondingto thehighfrequenciestimeconstantwasconsiderablylowerthan thatofthe015CA.Bothgroupspresentedaphaseanglenear80, forthelowestfrequenciesrange(102–101Hz)duringallthe
immersionperiod.
Different circuits for fitting EIS spectra on MAO treated Ti surfaces were found in the literature. Among other authors, Shokouhfaretal.[49]andVenkateswarluetal.[33,50]usedEEC consistingoftwotimeconstants,oneathighfrequency(CPEowith
aresistanceRoinparallel),characteristicoftheouterporouslayer,
and the second corresponding to low frequencies and being characteristicoftheinnerbarrierlayerofthefilm,consistingina CPEbwitharesistanceRbinparallel.BothauthorspresentRoasthe
“outer porous layer resistance”, although obtaining quite low valuesforthisparameter,ascommentedbybothauthors.Infact,Ro
shouldnotbeassignedtotheresistanceoftheouteroxides,butto theadditionalsolutionresistanceintheirpores.Boththelayoutof theEEC,whereRoisinserieswiththecontributionoftheinner
barrierlayer,andthelowRovaluespointouttothisinterpretation.
Additionally,Panetal.[51]haveintroducedthisinterpretationin theirpublications,mentioningthatRo(referredasRpintheirwork)
could be either the outer layer resistance or the electrolyte resistanceinsidepores.
Ontheotherhand,Fazeletal.[42]usedasimilarcircuitforMAO treatedTisurfacesbuttheauthorsaddedaconstantphaseelement (Qd) in series withRb which was assigned todiffusion of ions
throughtheoxidelayer.Thisinterpretationisquestionable,asthe nvalueobtainedbytheseauthorsforthatconstantphaseelement is 0.845, far from the typical values expected for diffusion-controlledprocesses(ca.0.5)andclosertothoseofacapacitanceof a rough oxide. Thus, even if it fits the experimental data, the proposedEEClacksphysicalmeaning,whichisamajor require-mentforacceptanceofanEEC.
Severalstudies[56,59–61]usedEECcontainingRs(electrolyte's
solution),anouterporousoxidelayer(CPEpandRp),aninnerdense
barrieroxidelayer(CPEbandRb)andaWarburgdiffusionelement
Fig.5. X-raydiffractionpatternsobtainedforgroups015CA,035CA,togetherwith untreatedTi.
Fig.6.Potentiodynamicpolarizationcurvesofthedifferentgroupsofsamplesin NaCl(8g/l).
Table2
Averagevaluesofcorrosionpotential,E(i=0),andpassivationcurrentdensity,ipass, forallgroups.
Group E(i=0)(Vvs.SCE) ipass(Acm2)
CPTi 0.4750.017 (10.11.1)x106
015CA 0.0910.022 (2.81.0)x108
to account for the mass-transfer process inside the irregular shapedporesintheoxide.However,diffusioncontrolisnotvery much expected in these materials, as the rate of any faradaic reaction(especiallymetaloxidation)isexpectedtobeverylow.On theotherhand,Quinteroetal.[40]proposedadifferentequivalent circuitfor similaranodic treatedsurfaces, ina 3 time constant ladderarrangement,consistingofthecontributionsoftheouter layer,innerlayerandmetal/electrolyteinterfaceatthebottomof nanometricporesofa notcompletelyhomogeneousinnerlayer that would allow contact between the Ti substrate and the electrolyte.However,thefittingspresentedbytheseauthorsare verypoor,with
x
2 valuescloseto10-2,whichmayquestionthevalidityoftheircircuit.
In summary,allthecircuitspresentedintheliterature, both relativeto standard anodisingor MAO, assumed a two-layered oxidefilm,withaninnerbarrierlayerandanouterporouslayer, although some of them may include other elements, such as WarburgorCPE's,normallyassignedtodiffusioncontrol.However, bychoosingfromthesecircuitstheonesthathaveaclearphysical meaning,noneofthemwasabletoprovideacceptablefittingtothe impedancespectraofthepresentanodictreatedsamples.Infact, basedonthestructureoftheanodicfilm,asrevealedfromthecross
sections of the anodic oxide films prepared by SEM-FIB and characterizedbyFEG-SEM/EDS(Fig.3),andontheEISdiagrams (Fig.10),atriplexstructureisconsidered,withathinandcompact oxide film formed at the metal/oxide interface (barrier film) followed by two porous layers, namely an inner porous layer presentingsmallporesandanouterporouslayer,formedatthe surface,thatconsistedinlargerporesbutnotreachingthebarrier
film. Therefore, a new equivalent circuit is proposed (Fig. 11), where the barrier film is associated to a resistor Rbf(AT) and a
constantphaseelementQbf(AT),whereasthethickeranodiclayers
correspondingtotheintactporous wall andtotheporouswall under the outer pores are represented by Qwall and Q1/2wall
respectively.Intheselastcases,noresistorisconsideredinparallel withtheCPE'sas,duetothethicknessofthefilms,therespective valuewouldbeextremelyhigh.Threeresistorswerealsoadded, correspondingtotheoverallelectrolyteresistance,Rsandtothe
additionalresistancesofthesolutioninside theinnerporesand outerpores(Rs-ip,Rs-op).
Theimpedancespectrafortheanodisedsampleswerefittedto thisequivalentcircuitusingZViewsoftware(version2.9)andthe qualityofthefittingwasevaluatedthroughtheirchi-square(
x
2)values.
TheimpedanceofCPEisdefinedasZCPE¼ Y0ðjwÞn
1
,whereY0
is the CPE admittance,
v
is the angular frequency, n theFig.7.Opencircuitpotentialevolutionwithimmersiontimeofthedifferentgroup ofsamplesinNaCl(8g/l).
Fig.8.NyquistandBodediagramsrecordedontheuntreatedCPTi(Grade2)immersedinNaCl(8g/l)for0h,8daysand20daysofimmersion.
Fig.9. EquivalentcircuitproposedforthefittingofEISspectrafortheuntreatedCP Ti(grade2)samples.
exponential factor (1n1) and j¼pffiffiffiffiffiffiffi1 is the imaginary number.Whenn=1,n=0andn=–1,theCPEresponsescorrespond tothoseofacapacitor,aresistororaninductor,respectively.When
n1,anon-idealcapacitormaybedescribedbythiselement,then
valuebeinginfluenced by theroughness ofthe surface andits heterogeneity.AscanbeobservedinFig.2,allgroupsofsamples presentedroughsurfaces,resultinginarangeofnvaluesbetween 0.90and0.93.
Fig.12showstheBodediagramsofexperimentaldataandthe respectivefittedcurves,comparingtheproposedmodel(Fig.11b) and oneof themodelsthat hasbeenusedin theliterature for titaniumanodicfilmsproducedbyMAO[33,49,50](Fig. 11a).These diagramsshowdifferencesinthequalityofthesimulationbyusing the two different models. The proposed model describes
adequatelythebehaviouroftheanodicfilmincontactwithNaCl, withchi-squarevaluesoftenbelow1x103,whereastheduplex
modeldoesnotsatisfactorilyfitthebehaviouroftheanodicoxide
films,withchi-squarevaluesaround1x102.
The conversion of Q values into C (capacitance) is very important when experimental capacitance data are used to estimate the parameters as the thickness of anodized layers
[62].Thus,thecapacitancevalueswerecalculatedinthecaseof Cwall(3)andC1/2wall(4)usingEqs.(3)and(4)respectivelythatwere
extrapolated fromBrug'sequation [58].On theotherhand,the Eq. (5) was used tocalculate Cbf(AT), when a resistor exists in
parallelwiththeCPE[58].
Cwall¼ QwallRsð1nÞ
h i1
n
ð3Þ
Table3
EquivalentcircuitparametersobtainedfromEISdatafortheuntreatedTiatdifferentimmersiontimes.
0h 1h 24h(1day) 48h(2days) 96h(4days) 192h(8days) 384h(16days) 480h(20days)
Rs(Vcm2) 47.95.1 47.65.1 47.74.2 48.55.1 45.83.8 46.05.8 46.05.8 44.35.8 Rbf(N)(MVcm2 ) 0.370.06 0.530.02 2.541.39 5.330.01 6.511.01 6.600.89 8.151.30 11.303.46 Cbf(N)(mFcm2 ) 48.64.1 47.73.7 44.12.0 39.00.9 35.63.5 35.01.6 33.60.4 28.25.7 n 0.920.02 0.930.02 0.930.01 0.930.02 0.930.02 0.940.02 0.940.01 0.950.02 x2 < 1.35103 <1.15103 <1.58103 <1.87103 <1.53103 <1.16103 <0.87103 <1.13103
Fig.10.EISspectraintheform:a)andc)Nyquistdiagramsandb)andd)Bodediagramsrecordedonthe015CAand035CAsamplesimmersedinNaCl(8g/l)fordifferent immersiontimes.
C1=2wall¼ Q1=2wallRsþRsopð1nÞ h i1 n ð4Þ CbfðATÞ¼ QbfðATÞ R 1 sþRsipþ 1 RbfðATÞ !ðn1Þ 2 4 3 5 1 n ð5Þ
Fig.13showsthevariationofC1/2wallwithimmersiontimefor
theanodicallytreatedgroups.The035CAgrouppresentedhigher valuesofcapacitancethan015CAforalltimesofimmersion.Itis knownthatcapacitance is proportionaltothe area.The 035CA
group presented higher pore size and higher crack density ((7.11.6)x103cracks/mm2)than015CA((1.10.4)x103cracks/
mm2), which can explain this behaviour. Besides, both groups
presentedanincreaseonC1/2wallvaluesthroughouttheimmersion
time.
The variation of Cwall with immersion time is presented in
Fig.14.The 035CA grouppresented alwayshigher Cwall values,
increasingupto24hofimmersion and then remainingalmost constanttilltwodaysofimmersion,followedbyanincreasetillthe endofimmersion,whilethevaluesofCwallfor015CAarealmost
constantduringtheimmersionperiod.Infact,duringimmersion, thesolutioncanpenetrateorattacktheoxidefilmwhichmaylead toporesblockingthatmayexplainthedecreaseofCwallvalues.The
Fig.11.Equivalentcircuits:a)traditionallyusedforthefittingofEISresultsforporousmaterialsb)newconceptforthefittingofEISspectrafortheanodicallytreatedgroups.
electrolytepenetrationthroughtheporesonthe035CAgroupis facilitated,asthisgrouppresentedbiggerporesandhighernumber ofcracks.
ThebarrierfilmbehaviourisshowninFig. 15.Thebarrierfilmof the anodic treated samples(Cbf(AT)) was compared with theTi
nativeoxidefilm(Cbf(N))formedonitssurface.Takinginaccount
thestandarddeviationfortheanodictreatedgroup,itwaspossible toobservethattheCbfvaluesdidnotchangesignificantlyduring
theimmersiontime.Howeveradecreasewithtimewasobserved for the untreated Tithat may bedue tothe contact withH2O
leadingtothehydrationoftheoxidefilm.Ifthegrowthrateofthe oxidefilmislargeenough,itcanremaininhydratedform[63].
The values of Cwall are considered representative of the
capacitanceoftheintactporouswall,andmaybeaffectedboth byvariationoftheexposedareaofthislayerandbyvariationofits thickness.ItwasconcludedfromFig. 13thattheporesizeincreases withtheimmersion time,thusaslightdecreaseintheexposed areaofintactwallisexpected.However,takingintoaccountthat this area is quite large when compared to the pore area, its variationshouldnotbeveryimportant.
On theother hand,accordingtotheEq. (6) thecapacitance valuesdependonthethickness:
C¼
ee
d0A ð6Þd¼
ee
C0A ð7Þwhere,
e
is thedielectric constant of the oxide film,e
0 is thevacuum permittivity, A is the area and d is the thickness. Considering
e
0=8.8541014Fcm1ande
=100(typicaldielectricconstant for TiO2 [64,65]), it maypossible toestimatethefilm
(wall)thicknessatthebeginningoftheimmersionandatthelast immersiontime.Thus,adecreaseinthethicknessoftheoverall
film,resultingfromitsdissolution,wouldleadtoanincreaseofthe capacitance,asitoccursfor035CAgroup.Inthecaseofthe015CA group,theeffectofdecreasingareamayaffectthisbehaviour,asit isobservedfromFigs.13and14thattheenlargementofthepore sizewasmuchbiggerinthistypeofspecimens.
Togetabetterunderstandingoftheelectrochemicalbehaviour of the anodic oxide films under long-time immersion in physiological solution,the characterization of the anodic films cross-sectionshouldbedone,aswellpotentiodynamictestsafter differentimmersiontimesshouldbeperformed.
4.Conclusions
The effect of the calcium acetate concentration in the electrolyteusedtoanodizeCP Tionthecomposition,structure, andcorrosionbehaviouroftheanodiclayerswasinvestigated.The outcomesofthisworkarethefollowing:
–anewequivalentelectrochemicalcircuitisproposedtofitthe EIS spectraof anodicfilmsproduced byMAO. Theequivalent circuit was established in accordance with the complex structureofthefilmsandovercomesthelimitationsofalready existing circuits proposed for this kind of materials. The proposed circuit has a clear physical meaning, allows for a goodfittingoftheexperimentaldataandexplainsthechangesin theEISspectrarelatingthemwiththedifferentcharacteristicsof theanodicoxides.
–thecorrosionresistancewasimprovedbythesurfacetreatment, withbothconcentrationsof calciumacetate,when compared withuntreatedtitanium.Nevertheless,higherconcentrationsof calciumacetateresultsinadetrimentaleffectonthecorrosion resistanceofthematerial,essentiallyduetothecontributionof thecrackspresentintheoutermostporouslayeroftheanodic
film.Theobtainedknowledgeisusefulforthedevelopmentof
Fig.13.VariationofC1/2wallasfunctionofimmersiontime.
Fig.14.VariationofCwallasfunctionofimmersiontime.
References
[1]M.Niinomi,RecentMetallicMaterialsforBiomedicalApplications,Metall. Mater.Trans.A.33A(2002)477–486.
[2]M.Niinomi,Mechanicalbiocompatibilitiesoftitaniumalloysforbiomedical applications,J.Mech.Behav.Biomed.Mater.I(2008)30–42.
[3]H.J.Rack,J.I.Qazi,Titaniumalloysforbiomedicalapplications,Mater.Sci.Eng. C.26(2006)1269–1277.
[4]S.S.Dheda,Y.Kyung,C.Melnyk,W.Liu,F.A.Mohamed,Corrosionandinvitro biocompatibilitypropertiesofcryomilled-sparkplasmasintered
commerciallypuretitanium,J.Mater.ScineceMater.Med.24(2013)1239–
1249.
[5]Y.Ni,Z.Liu,W.Gao,S.Qu,J.Weng,B.Feng,Characterizationofself-assembled decylbisphosphonate–CollagenlayersontitaniumbyQCM-Dand osteoblast-compatibility,Appl.Surf.Sci.257(2011)9287–9292.
[6]Y.J.Chen,B.Feng,Y.P.Zhu,J.Weng,J.X.Wang,X.Lu,Fabricationofporous titaniumimplantswithbiomechanicalcompatibility,Mater.Lett.63(2009) 2659–2661.
[7]G.He,L.Xie,G.-F.Yin,X.-M.Liao,Y.-W.Zou,Z.-B.Huang,etal.,Synthesisand mechanismof(101)-preferredorientationrutiletitaniaviaanodicspark oxidation,Surf.Coat.Technol.228(2013)201–208.
[8]H.Song,M.Kim,G.Jung,M.Vang,Y.Park,Theeffectsofsparkanodizing treatmentofpuretitaniummetalsandtitaniumalloysoncorrosion characteristics,Surf.Coat.Technol.201(2007)8738–8745.
[9]L.LeGuéhennec,A.Soueidan,P.Layrolle,Y.Amouriq,Surfacetreatmentsof titaniumdentalimplantsforrapidosseointegration,Dent.Mater.3(2007) 844–854.
[10]H.Ishizawa,M.Fujino,M.Ogino,Histomorphometricevaluationofthethin hydroxyapatitelayerformedthroughanodizationfollowedbyhydrothermal treatment,J.Biomed.Mater.Res.35(1997)199–206.
[11]Y.Yang,J.L.Ong,J.Tian,Invivoevaluationofmodifiedtitaniumimplant surfacesproducedusingahybridplasmasprayingprocessing,Mater.Sci.Eng. C.20(2002)117–124.
[12]S.Vercaigne,J.G.Wolke,I.Naert,J.A.Jansen,Theeffectoftitanium plasma-sprayedimplantsontrabecularbonehealinginthegoat,Biomaterials.19 (1998)1093–1099.
[13]T.Hanawa,Invivometallicbiomaterialsandsurfacemodification,Mater.Sci. Eng.A.267(1999)260–266.
[14]T.Hanawa,H.Ukai,K.Murakami,K.Asaoka,StructureofSurface-Modified LayersofCalcium-Ion-ImplantedTi-6Al-4VandTi-56Ni,Mater.Trans.36 (1995)438–444.
[15]M.Pham,M.Maitz,W.Matz,H.Reuther,E.Richter,G.Steiner,Promoted hydroxyapatitenucleationontitaniumion-implantedwithsodium,ThinSolid Films.379(2000)50–56.
[16]A.Montenero,G.Gnappi,F.Ferrari,M.Cesari,E.Salvioli,L.Mattogno,etal., Sol-gelderivedporoushydroxyapatitecoatings,J.Mater.Sci.35(2000)2791–2797. [17]H.-W.Kim,Y.-H.Koh,L.-H.Li,S.Lee,H.-E.Kim,Hydroxyapatitecoatingon
titaniumsubstratewithtitaniabufferlayerprocessedbysol–gelmethod, Biomaterials.25(2004)2533–2538.
[18]Y.Wang,J.Wang,J.Zhang,Z.Zhang,Effectsofsparkdischargeontheanodic coatingsonmagnesiumalloy,Mater.Lett.60(2006)474–478.
[19]H.Ishizawa,M.Ogino,Characterizationofthinhydroxyapatitelayersformed onanodictitaniumoxidefilmscontainingCaandPbyhydrothermal treatment,J.Biomed.Mater.Res.29(1995)1071–1079.
[20]H.Ishizawa,M.Fujino,M.Ogino,Mechanicalandhistologicalinvestigationof hydrothermallytreatedanduntreatedanodictitaniumoxidefilmscontaining CaandP,J.Biomed.Mater.Res.29(1995)1459–1468.
[21]H.Ishizawa,M.Ogino,Formationandcharacterizationofanodictitanium oxidefilmscontainingCaandP,J.Biomed.Mater.Res.29(1995)65–72. [22]I.S. Park, T.S. Bae, K.W. Seol, Surface Characteristics of Anodized and
HydrothermallyTreatedTitaniumwithanIncreasingConcentrationof CalciumIon,Met.Mater.Int12(2006)399–406.
[23]E.M.Szesz,B.L.Pereira,N.K.Kuromoto,C.E.B.Marino,G.B.deSouza,P.Soares, Electrochemicalandmorphologicalanalysesonthetitaniumsurfacemodified
CaandP,J.Biomed.Mater.Res.29(1995)1459–1468.
[30]H.Ishizawa, M. Ogino, Hydrothermal precipitationof hydroxyapatite on anodictitaniumoxidefilmscontainingCaandP,J.Mater.Sci.34(1999)5893–
5898.
[31]S.A.Fadl-allah,Q.Mohsen,Characterizationofnativeandanodicoxidefilms formedoncommercialpuretitaniumusingelectrochemicalpropertiesand morphologytechniques,Appl.Surf.Sci.256(2010)5849–5855.
[32]L.-H.Li,Y.-M.Kong,H.-W.Kim,Y.-W.Kim,Improvedbiologicalperformanceof Tiimplantsduetosurfacemodificationbymicro-arcoxidation,Biomaterials. 25(2004)2867–2875.
[33]K.Venkateswarlu,N.Rameshbabu,D.Sreekanth,A.C.Bose,V.Muthupandi,S. Subramanian,Fabricationandcharacterizationofmicro-arcoxidizedfluoride containingtitaniafilmsonCpTi,Ceram.Int.39(2013)801–8012. [34]W.F.Cui,L.Jin,L.Zhou,Surfacecharacteristicsandelectrochemicalcorrosion
behaviorofapre-anodizedmicroarcoxidationcoatingontitaniumalloy, Mater.Sci.Eng.C.33(2013)3775–3779.
[35]O.Banakh,T.Journot,P.-A.Gay,J.Matthey,C.Csefalvay,O.Kalinichenko,etal., Synthesisbyanodic-sparkdepositionofCa-andP-containingfilmsonpure titaniumandtheirbiologicalresponse,Appl.Surf.Sci.378(2016)207–215. [36]P.-C.Chang,B.-Y.Liu,C.-M.Liu,H.-H.Chou,M.-H.Ho,H.-C.Liu,etal.,Purified
titaniumoxidewithnovelmorphologiesuponsparkanodizationofTialloysin mixedH2S04/H3PO4electrolytes,J.Biomed.Mater.Res.A.80(2007)283–296. [37]S.Abbasi,F.Golestani-Fard,S.M.M.Mirhosseini,A.Ziaee,M.Mehrjoo,Effectof
electrolyteconcentrationonmicrostructureandpropertiesofmicroarc oxidizedhydroxyapatite/titaniananostructuredcomposite,Mater.Sci.Eng.C. 33(2013)2555–2561.
[38]I.Han,J.H.Choi,B.H.Zhao,H.K.Baik,I.-S.Lee,Changesinanodizedtitanium surfacemorphologybyvirtueofdifferentunipolarDCpulsewaveform,Surf. CoatingsTechnol.201(2007)5533–5536.
[39]N.K.Kuromoto,R.A.Simão,G.A.Soares,Titaniumoxidefilmsproducedon commerciallypuretitaniumbyanodicoxidationwithdifferentvoltages, Mater.Charact.58(2007)114–121.
[40]D.Quintero, O.Galvis, J.A.Calderón, J.G.Castaño,F.Echeverría,Effectof electrochemicalparametersontheformationofanodicfilmsoncommercially puretitaniumbyplasmaelectrolyticoxidation,Surf.CoatingsTechnol.258 (2014)1223–1231.
[41]D.Veys-Renaux,Z.A.ElHaj,E.Rocca,Corrosionresistanceinartificialsalivaof titaniumanodizedbyplasmaelectrolyticoxidationinNa3PO4,Surf.Coatings Technol.285(2016)214–219.
[42]M.Fazel,H.R.Salimijazi,M.A.Golozar,M.R.Garsivazjazi,Acomparisonof corrosion,tribocorrosionandelectrochemicalimpedancepropertiesofpureTi andTi6Al4Valloytreatedbymicro-arcoxidationprocess,Appl.Surf.Sci.324 (2015)751–756.
[43]S.A.Alves,R.Bayón,V.S.deViteri,M.P.Garcia,A.Igartua,M.H.Fernandes,etal., TribocorrosionBehaviorofCalcium-andPhosphorous-EnrichedTitanium OxideFilmsandStudyofOsteoblastInteractionsforDentalImplants,J. Bio-Tribo-Corrosion1(2015)23.
[44]S.A.Alves,R.Bayón,A.Igartua,V.SaénzdeViteri,L.A.Rocha,Tribocorrosion behaviourofanodictitaniumoxidefilmsproducedbyplasmaelectrolytic oxidationfordentalimplants,Lubr.Sci.26(2014)500–513.
[45]A.C. Alves, F. Oliveira, F. Wenger, P. Ponthiaux, J.-P. Celis, L.A. Rocha, Tribocorrosionbehaviourofanodictreatedtitaniumsurfacesintendedfor dentalimplants,J.Phys.DAppl.Phys.46(2013)404001–404009. [46]H.P. Felgueiras, L. Castanheira, S. Changotade, F. Poirier, S. Oughlis, M.
Henriques,etal.,Biotribocorrosion(tribo-electrochemical)characterizationof anodizedtitaniumbiomaterialcontainingcalciumandphosphorusbeforeand afterosteoblasticcellculture,J.Biomed.Mater.Res.PartBAppl.Biomater.103 (2015)661–669.
[47]A.R. Ribeiro, F. Oliveira,L.C. Boldrini, P.E.Leite, P. Falagan-Lotsch,A.B.R. Linhares,etal.,Micro-arcoxidationasatooltodevelopmultifunctional calcium-richsurfacesfordentalimplantapplications,Mater.Sci.Eng.C.54 (2015)196–206.
[48]F.G.Oliveira,A.R.Ribeiro,G.Perez,B.S.Archanjo,C.P.Gouvea,J.R.Araújo,etal., UnderstandingGrowthMechanismsandTribocorrosionBehaviourofPorous TiO2AnodicFilmsContainingCalcium,PhosphorousandMagnesium,Appl. Surf.Sci.341(2015)1–12.
[49]M.Shokouhfar,C.Dehghanian,A.Montenero,A.Baradaran,Preparationof ceramiccoatingonTisubstratebyplasmaelectrolyticoxidationindifferent electrolytesandevaluationofitscorrosionresistance:PartII,Appl.Surf.Sci. 258(2012)2416–2423.
[50]K.Venkateswarlu,N.Rameshbabu,D.Sreekanth,A.C.Bose,V.Muthupandi, Roleofelectrolyteadditivesonin-vitroelectrochemicalbehaviorofmicroarc oxidizedtitaniafilmsonCpTi,Appl.Surf.Sci.258(2012)6853–6863. [51]J.Pan,D.Thierry,C.Leygraf,Electrochemicalimpedancespectroscopystudyof
thepassiveoxidefilmontitaniumforimplantapplication,Electrochim.Acta. (1996)1143–1153.
[52]D.C. Grahame, Mathematical Theory of the Faradaic Admittance (PseudocapacityandPolarizationResistance),J.Electrochem.Soc99(1952) 370–385.
[53]S.-D.Wu,H.Zhang,X.-D.Dong,C.-Y.Ning,A.S.L.Fok,Y.Wang,Physicochemical propertiesandinvitrocytocompatibilityofmodifiedtitaniumsurfaces preparedviamicro-arcoxidationwithdifferentcalciumconcentrations,Appl. Surf.Sci.329(2015)347–355.
[54]Y.-T.Sul,C.B.Johansson,Y.Jeong,T.Albrektsson,Theelectrochemicaloxide growthbehaviourontitaniuminacidandalkalineelectrolytes,Med.Eng.Phys. 23(2001)329–346.
[55]R.S.Williamson,J.Disegi,J.A.Griggs,M.D.Roach,Nanoporeformationonthe surfaceoxideofcommerciallypuretitaniumgrade4usingapulsed anodizationmethodinsulfuricacid,J.Mater.Sci.Mater.Med.24(2013)2327–
2335.
[56]M.D.Roach,R.S.Williamson,I.P.Blakely,L.M.Didier,Tuninganataseandrutile phaseratiosandnanoscalesurfacefeaturesbyanodizationprocessingonto titaniumsubstratesurfaces,Mater.Sci.Eng.C.58(2016)213–223.
[57]K.Elagli,M.Traisnel,H.F.Hildebrand,Electrochemicalbehaviouroftitanium anddentalalloysinartificialsaliva,Electrochim.Acta.38(1993)1769–1774. [58]M.E.Orazem, B. Tribollet,Electrochemical ImpedanceSpectroscopy, Jonh
Wiley&Sons,Inc.Publications,NewJersey,2008.
[59]E.Matykina,R. Arrabal, M.Mohedano,A. Pardo,M.C.Merino,E.Rivero, Stabilityofplasmaelectrolyticoxidationcoatingontitaniuminartificialsaliva, J.Mater.Sci.Med.(2013)37–51.
[60]M.Mohedano,E.Matykina,R.Arrabal,A.Pardo,M.C.Merino,Metalrelease fromceramiccoatingsfordentalimplants,Dent.Mater.30(2013)e28–e40. [61]E.Matykina,R.Arrabal,B.Mingo,M.Mohedano,A.Pardo,M.C.Merino,Invitro
corrosionperformanceofPEOcoatedTiandTi6Al4Vusedfordentaland orthopaedicimplants,Surf.Coat.Technol.307(2016)1255–1264. [62]C.H.Hsu, F.Mansfeld,TechnicalNote:Concerning theConversion ofthe
ConstantPhaseElementParameterYointoaCapacitance,Corrosion57(2001) 747–748.
[63]T.Ohtsuka,N.Nomura,ThedependenceoftheopticalpropertyofTianodic oxidefilmonitsgrowthratebyellipsometry,Corros.Sci.39(1997)1253–1263. [64]J. Lundstrom, L. Rinehart, R. Pate, T. Smith, M. Krogh, W. Huebner, MeasurementofthedielectricstrengthoftitaniumdioxideceramicsPulsed Power,Conf.Dig.TecnhicalPap.12thIEEEInt(1999)10–12.
[65]A.Wypych,I.Bobowska,M.Tracz,A.Opasinska,S.Kadlubowski,A. Krzywania-Kaliszewska,etal.,Dielectricpropertiesandcharacterisationoftitanium dioxideobtainedbydifferentchemistrymethods,J.Nanomater.2014(2014) 1–9.
