Optical macrograph of the cylindrical porous NiTi alloy sample.thecorrosionresistanceofporousNiTialloysdecreasedwhenthe porosityofthealloyincreasedfrom50.2%to60.4%.. A schematic diagram
Trang 1Electrochimica Acta 56 (2011) 6389– 6396
j o ur na l h o me p ag e :w w w e l s e v i e r c o m / l o c a t e / e l e c t a c t a
A comparative study on the corrosion behavior of porous and dense NiTi shape memory alloys in NaCl solution
a School of Materials Science and Engineering, South China University of Technology, Guangzhou 510640, China
b School of Mechanical and Automotive Engineering, South China University of Technology, Guangzhou 510640, China
c MEMS Center, Harbin Institute of Technology, Harbin 150001, China
a r t i c l e i n f o
Article history:
Received 12 October 2010
Received in revised form 4 May 2011
Accepted 6 May 2011
Available online 13 May 2011
Keywords:
Porous NiTi alloy
Porous electrode
Corrosion behavior
Potential distribution
Interconnectivity
a b s t r a c t
© 2011 Elsevier Ltd All rights reserved
1 Introduction
Inrecentdecades,porousNiTishapememoryalloyshavedrawn
agreatdealofattentionasoneofthepromisingbiomaterialsfor
orthopedicimplantsandhard-tissuereplacementsbecauseofthe
combinedvirtueoftheshapememoryeffect,superelasticityand
adjustablemechanicalproperties,inparticularthetailoredpore
structureofpromotingtissuein-growth[1,2]
However,thehighnickelcontentoftheNiTialloysmightresult
inpotentiallynegativeeffectsonthesurroundingtissueby
induc-ing allergic responses [3,4] It has been reported that Ni ions
releasedduetothecorrosionprocesscanmaintainhighlevelsfor
upto8weeksorevenforseveralmonths[5,6].Inparticular,
com-paredwithconventionaldenseNiTialloys,thecomplexinterface
structureandlargerexposedsurfaceareasofporousNiTialloys
posea moreseriousissuewithregardtoleachingof Ni[7–10],
whichindicatestheneedforabetterunderstandingofthecorrosion
behaviorofporousNiTialloys
Thecorrosionofmetalimplantsinthehumanbodyisprimarily
drivenbyelectrochemicalreactions.Consequently,
electrochem-icaltestshavebeenusedtoevaluatethecorrosionresistanceof
NiTialloys[7–13].Incontrastwiththedenseelectrode,theporous
electrodegenerallysuffersfromnon-uniformpotential
distribu-∗ Corresponding author Tel.: +86 20 22236396; fax: +86 20 22236393.
E-mail address: mexzhang@scut.edu.cn (X.P Zhang).
tionbecauseoftheohmicpotentialdropandtheconcentration decayoftheelectroactivespecies[14–20].Itisknownthattherole
ofpotentialdistributioninsidecavitiesorrecessesinametal sur-faceisquiteimportanttoitscrevicecorrosionandpittingcorrosion behavior,sincethepotentialdistributionhasastrongeffectonthe distributionofcathodicandanodicreactionsatvariousdistances intothecavityorrecess[21,22].Thus,theelectrodepotential dis-tributionshouldbeanimportantconsiderationinunderstanding theelectrochemicalcorrosionbehaviorofporousNiTialloy Furthermore,theelectrochemicalreactionisessentiallya het-erogeneouselectrontransferreactionoccurringata solid-liquid interface.Thethree-dimensionalnatureoftheporousNiTialloys significantlyincreasestheinterfacialstructurecomplexity.A previ-ousstudy[23]hasshownthatthesinteredporoustitaniumwitha higherporosityratioundergoesmorecorrosionthandoesasample withalowporositybecauseofthelargerrealsurfacearea.However, theunsinteredsample(i.e.,thegreensampleofcoldcompacted titaniumpowder)withahigherporosityratio(producedunderlow compactionpressure)experiencedlesscorrosionthantheonewith
alow porosity(producedunderhighcompactionpressure).The discrepancywasrelatedtotheporecharacteristicsoftheporous titanium
Currently,therearemanypublishedresultsonthecorrosion behaviorofdenseNiTialloys[7–11],whileonlyverylimitedstudies havebeenpublishedonthecorrosioncharacteristicsofporousNiTi alloys.Inapreviousstudy,porousNiTialloysweredeterminedtobe lesscorrosionresistantthanthedenseNiTialloys[24].Additionally, 0013-4686/$ – see front matter © 2011 Elsevier Ltd All rights reserved.
Trang 2Fig 1. Optical macrograph of the cylindrical porous NiTi alloy sample.
thecorrosionresistanceofporousNiTialloysdecreasedwhenthe
porosityofthealloyincreasedfrom50.2%to60.4%
Thepresentstudyaimedtoclarifytheinfluencesoftheelectrode
potentialdistributionwithintheporeand thepore
characteris-ticsonthecorrosion behavior of theporousNiTialloys during
electrochemicaltests; thiswasaccomplishedthrougha
system-aticcomparisonofthecorrosioncharacteristicsoftheporousand
denseNiTialloys.BoththeporousNiTialloysampleswith
poros-ityrangingfrom35.5%to63.8%andthedenseNiTialloysamples
havingthesamenominalatomiccompositionastheporous
sam-pleswereprepared,andtheircorrosioncharacteristicsina0.9%
aqueousNaClsolutionat37◦Cwereinvestigated
2 Experimental
2.1 PreparationofporousanddenseNiTialloysamples
TheporousNiTialloysampleswerefabricatedbyapore-forming
technique and powder metallurgy method using a high-purity
ammoniumbicarbonate(NH4HCO3)powderanda blendof
ele-mentaltitaniumandnickelpowderswithanominalatomicratio
of50.8at.%Nito49.2at.%Ti[25–27].Theporecharacteristicsand
porosityratiosoftheNiTialloysamplesweretailoredbyadjusting
theamountofNH4HCO3powderaddedtothesamples.Fig.1shows
anopticalmacrographof theporousNiTialloysample
Dimen-sionsof thecylindricalsamplesare 14–15mmindiameter and
10–20mminheight.Therewerefourtypesofsampleswith
poros-ityratiosof35.5%,44.9%,55.8%and63.8%,respectively,andhad
averageporesizesofapproximately100–200m
ThegeneralporosityoftheporousNiTialloysamples,P,canbe
calculatedbythefollowingequation:
P(%)=
0V
wheremandVarethemassandvolumeoftheporoussamples,
respectively,and0 isthetheoreticaldensityof NiTialloy(i.e.,
6.45g/cm3forthebulknear-equiatomicNiTialloy)
Foracomparativestudy,thedenseNiTialloywaspreparedin
anon-consumablevacuumarc-meltingfurnace.Inthepreparation
process,theelementaltitaniumandnickelpowders,withthesame
nominalcompositionastheporoussample(i.e.,Ni50.8at.%–Ti
49.2at.%),wereblendedandcoldpressedintogreencompactswith
ageometryof15× 10(diameter×length,mm).Thegreensample
wasfirstsmeltedandthenremeltedfivetimesinanon-consumable
Fig 2. A schematic diagram of porous and dense NiTi alloy samples for electro-chemical tests.
vacuumarc-meltingfurnacewiththeprotectionofargonunder normalpressure.Finally,themoltensamplewasfurnace-cooled andadenseNiTialloyingot(nearlyahemi-sphere,withabase diameterof18mmandaheightof6mm)wasobtainedfor
follow-upstudies
2.2 Samplepreparationforcorrosiontesting Thecorrosionmayhaveoccurredthroughoutthethicknessof theporousalloy.Toaccurately characterizethecorrosion, both sidesofthesamplewereexaminedasthetestsurface Adense NiTialloyrod(2.0mm diameter)wasusedastheconnect elec-trode,whichwasdirectlyscrewedintothethreadedblindhole
onthecylindricalsideofthetestsample,asshowninFig.2.The severewearanddeformationproducedapore-freetappedhole Therefore,theelectrolytewashardlyincontactwiththescrewed rod.Theedgesoftheworkingelectrodeswerenotcoveredbythe non-conductingresinbecausetheresincouldpenetratetheporous sampleswithdifferentporosities
Thedisc-shapedporousNiTialloysampleswithadiameterof
15mmandathicknessof3mmwerecutfromtheporousNiTialloy rods(as-fabricated)andthedenseNiTialloyingotsbyelectrical dischargemachining
TheexposedsurfacesofboththeporousanddenseNiTialloy sampleswerewellpolished.Thesampleswerethendegreasedwith acetoneinanultrasoniccleanerfor20min,followedbyrinsingin distilledwater.Finally,thedenseNiTisamplesweredriedinair
atroomtemperature;theporoussampleswereplacedinadrying ovenataconstanttemperatureof80◦Cfor4h
2.3 Electrochemicaltestprocedure ElectrochemicalmeasurementswereperformedusingaZahner (modelIM6ex)potentiostat,accordingtotheASTMG5[28];the standardthree-electrodesystemwasadopted.Theworking elec-trodewasaporousor denseNiTialloysample withbothsides exposedtothesolution.Thereferenceelectrodewasasaturated calomel electrode (SCE), which was connected to the working electrodeviaaLuggincapillary,anddoublesymmetricalgraphite electrodeswereusedascounterelectrodes.Thetest electrolyte was0.9wt.%NaCl(analyticalreagent)inde-ionizedwater.Prior
Trang 3X.T Sun et al / Electrochimica Acta56 (2011) 6389– 6396 6391
Fig 3.Evolution of open circuit potential over time for dense and porous NiTi alloys
in a 0.9% NaCl solution.
toimmersionoftheelectrodes,theelectrolytecellwasheatedtoa
constanttemperatureof37◦Cusingawaterbath
The potentiodynamic polarization tests and electrochemical
impedancespectroscopy(EIS)wereperformed24hafter
immer-sionatopencircuit Potentiodynamiccurves weremeasuredby
scanningthepotentialfrom−0.4Vbelowtheopencircuitpotential
to+1.0Vatascanrateof0.02mV/s.Eachelectrochemical
exper-imentwasrepeatedthreetimeswithafreshspecimenforeach
test.Thecorrosionpotentialandthecorrosioncurrentdensitywere
obtainedthroughTafelapproximation.TheEISmeasurementswere
obtainedusingapolarizationof±10mVinthefrequencyrangeof
100kHzto1mHzand6pointsmeasuringperdecade.EISspectra
wereinterpretedbythesoftwareZSimpWin3.10
2.4 MorphologyofporousNiTialloysamples
TheporefeaturesofthefabricatedNiTialloysampleswere
char-acterizedbyanopticalmicroscope(Leica,DM2500P)andviadigital
imageanalysis.ThesurfacemorphologiesoftheporousNiTi
sam-pleswereexaminedbeforeandaftertheelectrochemicaltestsusing
ascanningelectronmicroscopy(SEMQuanta200,FEI)
3 Results
3.1 Opencircuitpotentialmeasurement
Theopencircuitpotentials(OCP)ofdenseandporousNiTialloys
in0.9%NaClsolutionweremeasuredoveraperiodof24h.Fig.3
presentstheevolutionoftheOCPasafunctionoftime,whereitis
obvioustheOCPofporousNiTialloysexperiencedasharperrateof
changecomparedtothedenseNiTialloys.Inallcases,theporous
NiTialloysexhibitedamorepositiveOCPthandidthedenseNiTi
alloys
3.2 Potentiodynamicpolarizationmeasurements
Thecorrespondingchangesinthebehaviorofthe
potentiody-namicpolarizationoftheNiTialloyswiththestructurevariation
fromdensetoporousareshowninFig.4a,andpresentationofthe
extractedelectrochemicalparametersareshowninFig.4b.Both
thedenseandporousNiTialloysexhibitedatypicalpassiveregion
andwerepittingattackedatthechemicalbreakdownpotential,
wherethecurrentdensitysharplyincreased.Thecorrosioncurrent
densitiesofthedenseandporousNiTialloysseeninFig.4 are
cal-Fig 4. (a) Polarization curves of porous and dense NiTi alloy samples after immers-ing in a 0.9% NaCl solution for 24 h and (b) evolutions of extracted electrochemical parameters from polarization curves with porosity ratio.
culatedfromthegeometricalareaforcomparison(i.e.theapparent area,notnecessarilytherealarea).Thevaluessuggestedthatthe currentdensities(icorr)ofporousNiTialloysweremarkedlyhigher thantheicorrvaluesofthedenseones,andthebreakdown poten-tials(Eb)ofporousNiTialloyswereclearlylowerthantheEbvalues
ofthedensesamples.However,itwasworthnotingthatthe cor-rosionpotentials(Ecorr)ofporousNiTialloysexhibitedapositive shiftbymorethan200mVfromtheEcorrvalueofthedenseNiTi alloy
ThetypicalsurfacemorphologiesoftheporousNiTialloy sam-plespolarizedtoapotentialof1.0Vfrom−1.0Vwereexamined
bySEM,asshowninFig.5.Itwasshownthattheobservedpitting primarilyexistedontheedgeoftheporousNiTialloysamples,as indicatedbythewhitearrowsinFigs.5(b),(d)and(f).Theedge attackmayhavebeenrelatedtoacombinationofgeometryand metallurgicalconditions(e.g.,inclusionstringers[26,27]),which cancomplicate thetest results.However, itwaslikely that the uncertaintiesinthesamplegeometryandmetallurgicalconditions wereconsistentacrossthesamplesand,hence,conclusionsbased
onthecomparisonsbetweenthedamageattheedgesshouldstill
bevalid.Intherepeatedtests,theobserveddamageoftheedgewas reproducible
Inparticular,thecharacteristicsofthesurfacemorphologiesdid notshowanobvioustendencyofdeteriorationofcorrosion resis-tancewithanincreaseintheporosityratio.AsshowninFigs.5(g) and(h),comparedwiththesampleswiththeporosityof35.5%and
Trang 4Fig 5.Typical SEM surface morphologies of the edge region of disc-shaped porous NiTi alloy samples with different porosity ratios before (a, c, e and g) and after (b, d, f and h) potentiodynamic polarization measurement in a 0.9% NaCl solution, terminated at 0.8 V: (a and b) 35.5%; (c and d) 43.9%; (e and f) 55.8%; and (g and h) 63.8%.
Trang 5X.T Sun et al / Electrochimica Acta56 (2011) 6389– 6396 6393
Fig 6. Bode spectra for porous and dense NiTi alloys: (a) impedance modulus plot;
and (b) phase angle plot.
43.9%,itwasdifficulttodistinguishthemorphologiesofthesample
withtheporosityof63.8%beforeandafterthepolarizationtest
3.3 Electrochemicalimpedancespectroscopy
ImpedancespectraforporousanddenseNiTialloysamplesafter
immersionina0.9%NaClsolutionat37◦Cfor24harepresentedas
Bodeplots(Fig.6 wherebothporousanddenseNiTialloysamples
exhibitedsimilarspectralfeatures.Acapacitivebehavior,whichis
representedbythephaseangleapproaching−90◦andtypicalofin
passivematerials[7,29],appearedinamediumtolowfrequency
range,Fig.6(b).Thismeantthatapassivefilmhadformedonall
samplesintheelectrolyte,whichwasconsistentwiththepassive
regiondeterminedinthepolarizationtests(Fig.4a)
Thelargephaseanglepeakcouldbeindicativeoftheinteraction
ofatleasttwotimeconstants.Therefore,anequivalentcircuitwas
proposedtomodeltheEISdataobtainedfromboththeporousand
denseNiTialloys,asshownFig.7a.Thismodelwaswidelyaccepted
forTiandTi-richalloysonwhichapassivefilmwithadouble-layer
structurewasformed[7,29,30].Inthemodel,Rsistheresistance
ofthesolution,Rpistheadditionalresistanceofthesolutioninside
thepores,R isthechargetransferresistanceofthebarrierlayer,
Fig 7. (a) Equivalent circuits used for fitting the experimental data and (b and c) experimental results and simulated data for the porous NiTi alloy with a poros-ity of 63.8% and dense NiTi alloy after immersion in a 0.9% NaCl solution for 24 h, respectively.
QpisthecapacitanceoftheporewallandQbisthecapacitanceof thebarrierlayer.Here,Qisthemagnitudeoftheconstant-phase element(CPE),representingadeviationfromtheidealcapacitor, theimpedanceofwhichisdefinedasZCPE=
Y0(jw)n−1
,where
−1≤n≤1.Thevalueofnisassociatedwiththenon-uniform distri-butionofcurrentasaresultofroughnessandsurfacedefects.This wasthecasefortheresultspresentedinthisstudy,wherethen valuesoftheporousNiTialloysamplesrangedfrom0.81to0.87, whilethedenseNiTialloysampleswasapproximately0.94 Theresistance,capacitanceandnvaluesfrommeasurements upontheNiTialloys,extractedusingequivalentcircuitillustrated
inFig.7(a),aretabulatedinTable1.Theimpedanceresultswere
Trang 64 Discussion
4.1 Effectofporousstructureonthecorrosionresistanceof
porousNiTialloy
Thecorrosion current density ofthe porousNiTialloy
sam-pleswereabouttwoordersofmagnitudehighercomparedtothe
densesamples.Thecorrespondinglysmallerpassivewindowfor
theporousNiTialloyisalsoclearlyseeninthepolarizationcurves
AssuchtheporousNiTialloywasqualitativelymoresusceptible
topittingcorrosionthanthedenseNiTialloy.Thisistoacertain
extentexpectedfromthelarger‘real’surfacearea.Moreover,itis
knownthatcorrosionresistanceoftheNiTialloyisassociatedwith
theformationoftheresistanttitaniumoxidefilmsonthealloy’s
surface[7,31]
Thesurface-finishquality,theamountoftheresiduesleftonthe
surfaceandthemicrostructuresinhomogeneityoftheporousNiTi
alloywerethecriticalfactorsthataffectedtheformationand
qual-ityoftheoxidefilms.Thesefactorswerestronglyinfluencedbythe
preparationprocessofthealloy.However,thepreparationofthe
porousNiTialloysgenerallyconsistedofcomplex
thermomechan-icalprocesses.Fig.8showstheSEMphotographsofporestructure
forporousNiTialloywiththeporosityratioof55.8%.Thepores
Fig 8.SEM image of pore morphology for porous NiTi alloy formed by the
pore-inanypassivefilms,whichisconsistentwiththeEISresults(The resistancevalueRbofporousNiTialloyisthreeordersofmagnitude lowerthanthatofdenseNiTialloy.)
4.2 Effectofporosityratioonthecorrosionbehaviorofporous NiTialloys
Thechangesinelectrochemicalparametersobtainedfromthe polarizationcurvesandimpedanceresultsoftheporousNiTialloys with increasing porosity ratio are summarized in Table 1 and abridgedintoFig.4b.Itwasdifficulttoidentifyatrendinthese parameterswithvaryingporosityratio.Infact,therealsurfacearea
oftheporousNiTialloysgenerallyincreaseswithincreasing poros-ityratiowithincertainlimits;however, thesurface qualityand structureuniformityoftheporousNiTialloysdoesnotnecessarily getworsewiththeporosityincreasing
Fig.9showstheopticalmicrographsforthepolishedporousNiTi alloysampleswithdifferentporosityratios.Clearly,thereweretwo typesofpores.Thelargeporeswithsizesof100–200mare pri-marilyformedbythedecompositionofthespace-holder,NH4HCO3 particles[26,27].Thesmallpores(alsocalledsecondarypores)with sizesbelow10mcouldbeattributedtothetrappedresidualsof
NH4HCO3decompositionandthenon-metallicimpuritiesexisting
intherawpowders,aswellasthevolumeshrinkageandthe Kirk-endalldiffusionbetweenNiandTiatoms.Thesmallporesshoweda bettersize-uniformityandamoreuniformdistributionthandidthe largeones.Whentheporosityratioincreased,theinterconnectivity
oflargerporeswasclearlyincreased
Forthesampleswithalowerporosity,themajorityofthepores weretheisolatedsecondarypores(seeFig.9a),whichisunlikely
to have trapped appreciable volumes of solution This caused thesamplesrelativeresistanttoinducedpittingcorrosion.When theporosity ratio increased, theinterconnectivity of the larger poreshadbeenmarkedlyimproved.Theinterconnectedchannels allowedthefreeflowoftheliquidandfewersiteswereavailable
toinducepittingcorrosion.Whentheporosityratioreached63.8%, theporemorphologywasdominatedbywiderandmorehighly interconnectedporestructures(Fig.9d).AsshowninFig.5(g)and (h),therewasnovisibledamageontheedgeofthesample Furthermore,withthechangeofporosityratio,poresize,pore sizedistributionandporeshapeoftheporousNiTialloyschange correspondingly.Itwasdifficulttoprovideacomplete characteri-zationoftheporesbysolelymeasuringthesampleporosity.This alsoaccountedfornoobvioustrendinthecorrosionbehaviorwith thesinglefactorchange
Trang 7X.T Sun et al / Electrochimica Acta56 (2011) 6389– 6396 6395
Fig 9.Optical micrographs of porous NiTi alloy samples with different porosity ratios: (a) 35.5%; (b) 43.9%; (c) 55.8%; and (d) 63.8%.
4.3 Effectofpotentialdistributiononthecorrosionbehaviorof
porousNiTialloys
FromtheEISresults,theRpvalueofporousNiTialloyswasfour
ordersofmagnitudehighercomparingtothedenseNiTialloys
Thiscanprovethatthetypicalpotentialdistributionexistsonthe
internalporesurfaceasaresultofcurrentflowwithinthepore
electrolyte[22,32].Thetypicalpotentialdistributionfollowsthat
E(x)=E(0)−
x
0
whereE(x)istheelectrodepotentialatadistancexintothepore,E(0)
isthepotentialattheporeopening,andI(x)andR(x)arethe
magni-tudesrespectivelyofthecurrentflowingatadistancexthroughthe
poreelectrolyteandoftheresistanceoftheelectrolyticpathwithin
thepore[32,33].Sincetheelectrodepotentialvarieswithdistance
xintothepore,therateofmetaldissolutionalsovariesoverthe
internalporesurfaceinaccordwiththepolarizationbehaviorof
themetalunderthelocalelectrochemicalconditions
Intheactiveregionofanodicpolarization,E(x)becamelessnoble
withincreasingdistanceintotheinternalpore,andthustherateof
theNiTialloydissolutionreactiondecreasedwithincreasing
dis-tancex.ThisexplainsthediscrepancythattheporousNiTialloys
withthepoorprotectiveoxidefilmsexhibitedmuchnobler
corro-sionpotentialthanthedenseNiTialloys.Inthepassiveregionof
theanodicpolarization,withtheexternalsurfaceinsidethepore
polarizedintothepassiveregion,E(x)decreasedwithincreasing
dis-tancexandmaybestillintheactiveregion,whichdependsonthe
dimensionsoftheporeandtheelectrochemicalconditions
prevail-ingatthewholeporousstructureofporousNiTialloyelectrode[33]
Thus,thepotentialdistributionisplayingarolethatstrengthensthe
differencesamongthedifferentpartsoftheporousmetalsurface
Thisistheessentialelectrochemicalconditionthatresultsinthe
localizedmetalcorrosion.Therefore,theroleofpotential
distribu-tioninsidetheporeandporousstructureintheelectrochemical
corrosionbehavioroftheporousNiTialloysisanimportantfactor
5 Conclusions
FromthepotentiodynamicpolarizationandEISresults,itwas concludedthattheporousNiTialloywasmoresusceptibleto local-izedcorrosionthanwasthedenseNiTialloy.However,theporous NiTialloysamplewithahigherporositydidnotsufferobviously morecorrosionthantheonewithalowerporosity
Thepotentialdistributionshouldexist ontheporewallasa resultofcurrentflowwithintheporeelectrolyteduringthe elec-trochemicaltests.Therefore,theroleofpotentialdistributioninthe electrochemicalcorrosionbehavioroftheporousNiTialloysisan importantfactor,sincethenon-uniformpotentialdistributioncan strengthenthedifferencesamongthedifferentpartsoftheporous metalsurface
Acknowledgements
ThisresearchwassupportedbytheNationalNaturalScience FoundationofChina(Nos.50871039and51001050),China Post-doctoralScienceFoundation(20100470917)andtheFundamental ResearchFundsfortheCentralUniversitiesallocatedinSouthChina UniversityofTechnology(No.2009ZM0160)
References
[1] A Bansiddhi, T.D Sargeant, S.I Stupp, D.C Dunand, Acta Biomaterialia 4 (2008) 773.
[2] A.C Jones, C.H Arns, D.W Hutmacher, B.K Milthorpe, A.P Sheppard, M.A Knackstedt, Biomaterials 30 (2009) 1440.
[3] D.J Wever, A.G Veldhuizen, M.M Sanders, J.M Schakenraad, J.R Horn, Bioma-terials 18 (1997) 1115.
[4] M Es-Souni, M Es-Souni, H Fischer-Brandies, Analytical and Bioanalytical Chemistry 381 (2005) 557.
[5] H.H Huang, Y.H Chiu, T.H Lee, S.C Wu, H.W Yang, K.H Su, C.C Hsu, Biomate-rials 24 (2003) 3585.
[6] M Cioffi, D Gilliland, G Ceccone, R Chiesa, A Cigada, Acta Biomaterialia 1 (2005) 717.
[7] N Figueira, T.M Silva, M.J Carmezim, J.C.S Fernandes, Electrochimica Acta 54 (2009) 921.