Micro-galvanic corrosion of Ti-6Al-4V in situ alloyed with molybdenum via selective ... · 2018....

Vol.4 No.1 2018 3|Page

Micro-galvanic corrosion of Ti-6Al-4V in situ alloyed with molybdenum via selective laser melting

C.MadikizelaCouncilforScientificandIndustrialResearch,NationalLaserCentre,Pretoria;andUniversityoftheWitwatersrand,andDST‐NRFCentreofExcellenceinStrongMaterials,Johannesburg,SouthAfricaL.H.ChownUniversityoftheWitwatersrand,andDST‐NRFCentreofExcellenceinStrongMaterials1,JanSmutsAvenueJohannesburg,SouthAfrica

L.A.CornishUniversityoftheWitwatersrand,andDST‐NRFCentreofExcellenceinStrongMaterials1,JanSmutsAvenueJohannesburg,SouthAfricaH.MollerCouncilforScientificandIndustrialResearch,NationalLaserCentre,Pretoria;Nowat:DepartmentofMaterialsScienceandMetallurgicalEngineering,UniversityofPretoria,Pretoria,0001,SouthAfrica

ARTICLEINFO

Keywords:AdditivemanufacturingSelectivelasermeltingTi‐6Al‐4VMolybdenum

ABSTRACT

Powder‐bedlasermetaladditivemanufacturingviainsitualloyingiscurrentlyofinterestforalloydevelopment.Molybdenumisabeta‐stabilizerwhichisaddedtoTi‐6Al‐4Vtoproduceamoreductilebetaalloy.Previousstudieshaveshownthatinsitu alloying of Ti‐6Al‐4V withMo produced a composite consisting of a beta‐stabilizedmatrixwithsomeun‐meltedMoparticles.AlthoughthismicrostructuredidgivebettermechanicalpropertiesthanwithoutMo,therewasaconcernthatthecompositecouldexperiencepossiblemicro‐galvaniccorrosioneffects,sinceTi‐6Al‐4VismorenobleonthegalvanicseriesthanMo.Immersiontestsweredoneona composite sample placed in a 3.5% NaCl solution, and a scanning electronmicroscopewasusedtoexaminethecorrosioneffectsweekly.CorrosionattackoftheanodicMoparticleswas found,whereas therewas littleeffecton thematrix.The resulting composite microstructure might therefore be beneficial forimproving mechanical properties, but its corrosion resistance might beproblematicincertainapplications.

1. IntroductionAdditive manufacturing has rapidly developedsince the invention of lasers in the 1960s. Thistechnologyhasbenefitedfromextensiveresearch,especially for applications in the manufacturingindustry where the benefits of lower materialwastage, higher production efficiencies and

infinite possible geometries can maximize profit[1].For thisreason,selective lasermelting(SLM)processes have gained preference overconventional methods (e.g. casting) in themanufacturing industry. These processes involvethe use of a CADmodel to design and “slice” thepart into layers. A laser is then used tomelt the

Micro‐galvaniccorrosionofTi‐6Al‐4Vinsitualloyedwithmolybdenumviaselectivelasermelting


powder deposited, and raked according to thethicknessofeachlayer,accordingtothedesign[1,2].In recent studies [3, 4, 5], selective lasermeltinghas been used to investigate its capabilities indirectly developing new complex materials toimprovemechanicalproperties.Factorswhichareessential in the improvement of mechanicalproperties include strengthening throughreinforcement, incitement of nucleation,refinement of the microstructure to producemanufactured parts with no porosity or defects[6]. Vrancken et al. [4] added 10wt%Mo to Ti‐6Al‐4V to stabilize the beta phase and thereforeincrease the ductility and fracture toughness ofthealloy.Processingat higher speedsandpowercompared toVranckenetal. [4],Madikizelaet al.[7] manufactured the same alloy. Themicrostructure of alloys produced in situ via theSLM process often had some inhomogeneities,especially un‐melted Mo particles in the as‐builtstate[3,4,5,7].The resultant (composite‐like) as‐builtmicrostructuregave rise toquestionsaround thecorrosion resistance of such a microstructure.This topic has not been reported by otherresearchersinthefield[4,5].Galvanic corrosion results from the contactbetween two different conducting materials in acorrosiveenvironment[8].Thistypeofcorrosionis commonly found in ships where the partsimmersed in the sea water are made of variousmetalalloysormetalmatrixcompositematerials,where the reinforcing material is spreadthroughout the metal, e.g. graphite [8]. Galvaniccorrosion may result in the accelerateddeterioration of one metal or the corrosionprotection of a connected metal (anodicprotection) [8]. Galvanic corrosion is not onlyaffected by the potential difference between thetwo metals in contact, and other metallurgicalfactors such as surface conditions, electrolyteproperties, reactions, environmental effects andgeometricfactorsalsoplayapart[8].Considering Ti‐6Al‐4V and molybdenum, agalvanicseriespreparedbytheU.S.ArmyMissileCommand [9] revealed a significant difference in

position between Ti‐6Al‐4V and Mo, which is acause for concern. On this specific series [9], Ti‐6Al‐4Vwasnumber85whileMowasnumber58,therefore the contact between these twodissimilarmetalswouldbeconduciveforgalvaniccorrosion in the right environment, thuspotentiallyresultinginthecorrosionoftheanodicMo.Thiswork investigated themicro‐galvanic effectsbetweenTi‐6Al‐4VandMoina3.5%NaClsolutionfora sample insitualloyedby theselective lasermeltingprocess.2. ExperimentalProcedure2.1. Powder mixing and SLM sample

manufacturingA 200gpowdermixturewaspreparedbymixing190g of Ti‐6Al‐4V powder (20‐60µm) and 10gmolybdenum powder (20‐45µm) in a tubularmixer for 30 minutes [7]. An SLM machinecontaininga5kWNdYagfiberlaserwitha250µmspotsizewasusedtoproduce10x10mmsamplesunder a protective argon atmosphere [7]. Thesamples were manufactured at a laser power of1.3kW,atalaserscanspeedof4m/s,withalayerthicknessof50µmandahatchoverlapof50%[7].2.2. Micro‐galvanictests108.8g of sodium chloride was dissolved in3000ml of distilled water to make a 3.5% NaClsolution in a 5 litre plastic bucket. The area ofstudy was identified by drawing a circle on thesample using the diamond indenter of a Vicker’shardnessmachinepriortoimmersion.Thismeantthat the same area could be studied after eachimmersion. The sample was tied with non‐conductivethreadandtheotherendofthethreadwastiedtoarodwhichwasplacedacrossthetopof the bucket. The polished sample was thensubmerged in the prepared 3.5% NaCl solution.The experiment was evaluated weekly for 20weeks,accompaniedbystirringthesolutiontwicea day in order tomake sure therewas sufficientoxygen for the reaction, and the solution wasreplenished weekly to compensate forevaporation. After removing the sample, an SEMwas used to examine the weekly changes, andthen the sample was replaced in the bucket.Initially, the sample was weighed, but this wasdiscontinued, because no difference was found,



since the proportion of theMo particleswas toosmallfortheeffecttobeclearlyobserved.2.3. MicrostructuralcharacterizationTobetter reveal thegrain structureofun‐meltedMoparticle,thesamplewasetchedwithasolutionof 10ml glycerol, 10mlHNO3 and 10mlHF for 5minutes. The sample was investigated using anOlympus BX51M optical microscope (OM) and aJEOL JSM‐6010PLUS/LA Scanning electronmicroscope(SEM).

3.0 ResultsandDiscussions3.1 PowdermixingMixingofTi‐6Al‐4Vandmolybdenumpowdersgaveanear‐homogenous mixture as shown in Figure 1. Themorphology of both Ti‐6Al‐4V and Mo powders wasspherical. This is an idealmorphology, since particleswith a spherical morphology are associated withforminggoodbondingbetweenindividuallasertracks,which leads todenserpartsand increasedcomponentquality [6]. Segregation was found at a smaller scale,e.g. 200x magnification which is expected as 100%homogeneitycannotbeobtainedbymixing.

Figure 1: SEM‐BEC images showing particle shape andmorphology of Ti‐6Al‐4V+10M: (a) at lower magnification,and(b)athighermagnification,wheretheMoparticleshavelightercontrast.3.2 MicrostructuralanalysisAlthough adding 10 wt% Mo was supposed tofullystabilizethebetaphase[4],Madikizelaetal.[4] found from XRD that alpha peaks were stillpresent for the microstructure (Figure 2). Thiswas attributed to the poor melting of Mo at themanufacturingparameters,which gaveMopeaksin the XRD pattern, aswell as some alpha peaks(Figure2).Thisindicatedthathigherlaserpowerand especially the higher scanning speeds thanVranckenetal.[4],werenotsufficienttoproperlymelt the Mo powder. The microstructure alsoshowed pronounced melt pool boundariesshowing the separate tracks as the sample wasbuilt, and darker contrast areas in backscatteredelectronmodewith lowerMo contents, and lightcontrastareaswhereMowasdissolved(Figure2)

[7]. Also, although the starting Mo powderparticleswerespherical initially (Figure1), somehad circular cross‐sections in Figure 2(b),whereas some did not, indicating that somereaction had occurred, but not homogeneouslythroughoutthesample.Xuetal.[10]discussedtheXRD pattern of as‐built Ti‐6Al‐4V produced viaSLM.Theάpeakswerefoundat36°,38°,40°,53°,57°and64°2θangles[10].Thepeaksinthisworkwerefoundat35°,37°,38°,40°,49°,63°,73°and74°2θangles.Althoughsomeάpeakscorrespondto the Ti‐6Al‐4V peaks found by Xu et al. [10],somehaveshiftedandάwasfoundatnewpeaksfor Ti‐6Al‐4V +10wt%Mo. The shift in peaks isattributedtothepartialdissolutionofMointothematrix.

Figure 2: X‐ray diffraction pattern and SEM‐BEC imageshowingthemicrostructureofthesample[7].Figure3(a)showsasectionthroughanentireMopowder particle in the matrix of the composite,andthecircularcross‐sectionwasagainsimilartothe shape of the starting powder in Figure 1.Therewasalsoalargeholeinthecentre(Figure3(a),andeven in theun‐etchedcondition,someofthegrainboundariescouldberesolved.Notallofthe surface was smooth as in Figure 1 (albeitFigure 1 was at a much lower magnification),whichsuggestedthatpossiblysomemeltingoftheMoparticlehadoccurredafterall,butnotenoughtodissolvetheentireparticleinthematrix.Figure3(b) isahighermagnification imageof thesameMo particle in the alloy, and reveals grains ofdifferent sizes and some porosity within the un‐melted Mo‐particle. There were also visible sub‐grainsinsidethelargergrains.TheporosityintheMo particle mainly occurred at the grainboundaries,andwhereitoccurredwithinagrain,itwasassociatedwiththesub‐grainboundaries.Asimilar grain structure was found forcommerciallypureMo[11].



Figure3:SEM‐SE imagesshowinganun‐meltedMoparticle:(a)wholeun‐etchedMoparticle, and (b)etchedMoparticlewithgrainsandsub‐grainswithintheparticle.3.3 Micro‐galvaniccorrosionFigure4showstheprogressionofcorrosionfromweek 1 to week 20. Comparing Figure 3(a) andFigure 4 (a), significant changeswere evident asthe particle had visible grain boundaries after aweek,duetothecorrosionattack.Therewasalsoan area of darker appearance in Figure 4 (a)where the corrosion was more severe than inother areas. After 20 weeks (Figure 4(b)), theparticlewasbeginning todetach fromthematrixand the prior grain boundaries were morepronouncedanddeeper.Therewasalsoevidenceofsub‐grainboundaryattack(Figure4(b)).Thegrainsizeoftheetchedparticle(Figure3(b))and the corroded particle Figure 4(a) wasmeasured using the line intercept method. Theaverage grain size for the etched sample was2.9µm;thiswaslowerthantheaveragegrainsizefound for the corroded particle from week 1(3.6µm) (Figure 4(a)). This could be because theetchantworksfasterthantheNaClsolution,hencemore grainswere revealed in the etched sample,or that the etchant just revealed the realmicrostructurebetter.Thedifferenceinareasoftheanodicandcathodicmetalsmust also be considered. In this case, thetotal surfaceareaof the corrodingMometalwassmallcompared to theareaof theelectropositiveTi‐6Al‐4V matrix. The corrosion reactiontherefore occurred rapidly because of theincreasedoxygenreduction,whichledtoagreatergalvanic current [12]. The smaller particles (1, 2and3)formFigure4(candd)hadcorrodedtotheextent that they had fully detached from thematrix.Thedegradationof the largerparticles (4and 5 in Figure 4(c) and (d)) occurred due todetachment of grains/sub‐grains within the Moparticles. This was followed by corrosion attackon the newly‐exposed grain boundaries. This

process repeated until the whole particle wasdetached from the matrix. Although the Moparticlescorroded,thedifferenceinthemasslossofthesamplewastoosmalltodiscern.Daietal.[13]studiedthecorrosionbehaviourofaselective laser melted Ti‐6Al‐4V alloy in NaClsolution.Intheirwork,theSLM‐producedTi‐6Al‐4Valloyshowedinferiorcorrosionresistancethantraditionally manufactured commercial Grade 5Ti‐6Al‐4V [13]. This was attributed to thepresenceof the large fractionof acicularαphasecompared to the β‐Ti phase [13]. In the currentstudy, the presence of Mo and its ability tostabilizetheβphasewassupposedtoimprovethecorrosion properties of Ti‐6Al‐4V, but with theinhomogeneity in the microstructure, itintroduced new problems. Therefore, furthercorrosion work needs to be done on the Ti‐6Al‐4V+10 wt% Mo alloy e.g. electrochemicalmeasurementsandpolarizationtests.

Figure 4: SEM‐BEC and SEM‐SE images showing theprogression of corrosion: (a) and (c) high and lowmagnificationimagesafterimmersionfor1week,and(b)and(d)highandlowmagnificationimagesafterimmersionfor20weeks.4. ConclusionsAlthoughbeneficial for the improvementofsomemechanicalpropertiesintheas‐builtstate,Ti‐6Al‐4V+10Mo in situ alloyedby SLMwas susceptibleto micro‐galvanic corrosion in a 3.5% NaClsolution. Thismight limit the applications of thismaterial in certain corrosive environments, andshould be taken into consideration duringmaterialselectionforspecificapplications.



AcknowledgementsThe authorswould like to thank theDepartmentof Science and Technology, and the NationalResearchFoundation,SouthAfrica.References1. J. Kruth, M. Leu, T. Nakagawa, “Progress inAdditive Manufacturing and Rapid Prototyping”,CIRP Annals‐Manufacturing Technology, 47(2)(1998):525‐540.2. S. Kumar, S. Pityana, “Laser‐Based AdditiveManufacturing of Metals”,http://researchspace.csir.co.za/dspace/bitstream/10204/4758/1/kumar6_2010.pdf (Sep. 14,2016)3. E. Chlebus, B. Kuznicka, R. Dziedzic, T.Kurzynowski, “Titaniumalloyedwithrheniumbyselective laser melting”, Materials Science &EngineeringA,620,(2015):155‐163.4. B. Vrancken, L. Thijs, J. Kruth, J. VanHumbeeck, “Microstructure and MechanicalProperties of a Novel β Titanium MetallicComposite by Selective Laser Melting”, ActaMaterialia,68(2014):150‐158.5. M. Fischer, D. Joguet, G. Robin, L. Peltier, P.Laheurte,“InsituElaborationofaBinaryTi–26NbAlloy by Selective Laser Melting of ElementalTitaniumandNiobiumMixedPowders”,MaterialsScienceandEngineering:C,62(2016):852‐859.6. H, Attar, M. Bӧnisch, M. Calin, L. Zhang, S.Scudino, J. Eckert, “Selective laser melting of insitu titanium–titanium boride composites:Processing, microstructure and mechanicalproperties”,ActaMaterialia,76(2014):13‐22.7. C. Madikizela, L.A. Cornish, L.H. Chown, H.Moller,D.Louw,“MicrostructureofinsituAlloyedTi‐6Al‐4V and 10Mo as a Function of ProcessParameters”, Frontiers in Optics / Laser Science,RochesterNY,2016.8. X.G.Zhang,GalvanicCorrosion, in:R.W.Revie(Ed.) Uhlig's Corrosion Handbook, 3rd ed. (JohnWiley&Sons,Inc.:Canada,2011).

9. M.C. Farman and E.A. Verchot, “Practicalgalvanic series,” U.S. Army missile command,Alabama,1967.10.W.Xu,M.Brandt,S.Sun,J.Elambasseril,Q.Liu,K. Latham, K. Xia, M. Qian, “AdditivemanufacturingofstrongandductileTi–6Al–4Vbyselective laser melting via in situ martensitedecomposition”,ActaMaterialia,85(2015):74‐84.11.Y. Hiraoka, T. Hoshika, “Parameterrepresentinglow‐temperaturefracturestrengthinmolybdenumhavinganelongatedandlargegrainstructure”, (1999),http://www.sciencedirect.com/science/article/pii/S0263436899000189(Sep.14,2016).12.P.H. Hack, Galvanic Corrosion, illustratededition,(ASTMInternational:OH,USA,1988).13.N. Dai, L. Zhang, J. Zhang, Q. Chen, M. Wu,“CorrosionbehaviourofselectivelasermeltedTi‐6Al‐4Valloy inNaCl solution”,CorrosionScience,102(2016):484‐489.


Temporary coating for in process or storage applications

MarkusBieberCortecCorporation4119WhiteBearParkwaySt.Paul,MN55110USA

DarioDell’OrtoCortecCorporation4119WhiteBearParkwaySt.Paul,MN55110USA

ARTICLEINFO

Keywords:CorrosionProtectionInhibitorTemporaryCoating

ABSTRACT

There is a tremendous need for the protection of assets during processing,shipmentandstorage.Criticalsparesmustbereadilyavailableandeasytoplaceinto service with minimal impact to the operation of equipment. The proventechnology of VCIs (Vapor Corrosion Inhibitors) has long been used to preservemany of these assets using traditional mechanisms such as papers and films.Neweradvanceshaveallowed theuseof thisVCI technology tobe incorporatedinto temporary coatingswhich are designed to provide corrosion protection inextreme environments, yet still be easy to remove compared to the oldertraditionalwax type coatingsorpermanent coatings. By incorporating theVCItechnology into these temporary coatings, it allows the use of thinner filmthicknessesand lessrelianceona thickbarrier tokeepcontaminantsaway fromthesurface1,2.

1. IntroductionFourLearningObjectivesWhatisaVCI?VCI (Vapor Corrosion Inhibitor) is a corrosioninhibiting technology which consists of variousorganic and inorganic compoundswhichhave anattraction to a metal surface. When properlyformulatedandcompounded, this technologycanbeusedinamultitudeofcarriersystems,suchascoatings, which bring the inhibitors close to themetalsubstrate.HowVCIsworkinacoatingVCIs are formulated into a coating by a complexdevelopmentprocesswhichinvolvesdeterminingchemical compatibilityof theVCIswith theothercomponents of the coating such as the resin,solvents,pigmentsandotheradditivesused foravarietyof reasons. VCIsworkbyadsorbingonto

the metal surface in a non‐reactive attractivecapacity, i.e., they are attracted to the metalthroughtheparticlecharge3.HowVCIscomparetotraditionalinhibitorsVCIs compare with traditional inhibitor systemsby using smaller particles, aswell as relying notonly on contact inhibition but also vapor phaseinhibition, thus providing more completecoverage and protection of the surface. This isillustratedinFigure1.The larger platelets are representative oftraditional inhibitorswhich are unable to fill themicro‐crevices, leavinggapswhere corrosioncanstartand/orgrow4.

Temporarycoatingforinprocessorstorageapplications


Figure1:SchematicDiagramofaVCICoatingWhattypeofcoatingsystemscanuseVCIs?VCIs can be used with most coating systems.There are many variations of VCIs and it isnecessary to choose the correct VCI for thecorresponding coating system by checkingcompatibility,effectivenessandprocessability.The environmental advantages of VCIs overtraditionalinhibitorsTraditionalinhibitorscontainingheavymetalsarebecoming increasingly more regulated and oftenare no longer allowed to be used due to thenegative impact they have on the environmentand carcinogenic effects on workers exposed tothem5. The environmental advantages of usingVCIs are that they are non‐toxic, do not containheavymetals, and have no adverse effect due totheir low usage concentrations. VCIs have longbeen used in other products such as PE films,foams, powders and liquids to provide a vaporphase of corrosion protection without impactingtheenvironment6,7,8,9.2. ExperimentalProceduresThese studies examined the effectiveness ofvarious types of corrosion inhibitors in solventandwaterborneremovablecoatings,basedonsaltfog results, (ASTM B117) and humidity results(ASTMD1748).Each coating was applied on cold rolled steel(CRS) panels (SAE 1010), or on actual partsprovided by customers. Dry film thicknesses(DFTs) were according to manufacturers’recommendations.

3 ResultsSalt spray and humidity chamber tests areconsidered industry standards for corrosiontesting. As shown by Figures 2‐6 below, the

differenceincorrosionprotectionwithVCIaddedis quite noticeable. The differences are notedbelow.

SampleA350HoursWithoutVCI

SampleB50Hours

WithoutVCI

SampleA400HoursWithVCI

SampleB>400HoursWithVCI

Figure 2: Salt Spray Testing for competitive solvent‐bornewaxtypecoatingwithandwithoutVCIinhibitors

SampleA3170HoursWithoutVCI

SampleB1100HoursWithoutVCI

SampleA>3170HoursWithVCI

SampleB>3170HoursWithVCI

Figure3:HumidityTestingforcompetitivesolvent‐bornewaxtypecoatingwithandwithoutVCIinhibitors

SampleA

Solvent‐borneCoatingwithoutVCI

SampleBSolvent‐borneCoatingwithVCI

SampleCWaterborneCoating

withVCIFigure 4: Salt Spray (ASTM B117) Testing for VariousSystems(600hours)


Vol.4 No.1 2018 10|Page

VCICoating Non‐VCICoating

Figure5:Self‐HealingEffectof theVCIon theScribe inSaltSprayTesting

Figure 6: Humidity Testing for solvent‐borne wax typecoatingandwaterbornewaxtypecoatingwithVCIinhibitors(768hours)This testing shows that waterborne systems cancompetewithsolvent‐bornesystemsbytheuseofVCI inhibitors. However, there are somedistinctadvantages of using a waterborne system whichinclude:• MoreEnvironmentallyfriendly• LowerVOCs• Easiercleanup4 ExamplesofOtherTechnologiesUsingVCI

Inhibitors

FilmUsingInferiorInhibitors FilmUsingVCIInhibitorsFigure7:EngineComponentsPackaginginVCIfilm

PreservativewithoutInhibitorsPreservativewithVCIInhibitorsFigure8:TransmissionComponentsinPreservativeOils5 CaseHistory15.1ProblemA manufacturer of large construction gradersneeded an effective alternative to preventcorrosion on their products. There were severaldisadvantagestotheheavywaxtypeproducttheywere using. First, it did not always work if theequipment was stored outdoors for extendedperiodsoftimebeforeshipping.Secondly,itleftagreasy and slippery film on the graders, whichmadeitdifficulttoclimbintothemformovingandshipping.Finally,theproductwashardtoremoveandhadtobedisposedofashazardouswaste.

5.2ApplicationThe manufacturer sprayed the VCI containingcoating and solvent in a 3:1 ratio on theequipment,whichresultedindrycoatthicknessesbetween0.8and1mil(20‐25microns).Thenthemachinesweretransportedbyrailtotheseaports.A few of the graders were placed in containers,butmostwereleftuncovered.

The VCI containing coating at 1mil (25microns)outperformedtheheavywaxtypecoatingat3mil(75 microns). The manufacturer also found theVCI containing coating easier to spray, free ofoffensiveodorsandmucheasiertoremove.Aftertwo years of export shipments, they hadexperienced no corrosion problems when usingthethinnerfilmVCIcontainingcoating.

Solvent‐borne Waterborn

StandardOil StandardOilwithVCIInhibitor


Vol.4 No.1 2018 11|Page

6 CaseHistory26.1ProblemA manufacturer of industrial equipment wasstruggling with protecting critically machinedsurfaces from corrosion and damage duringtransport. Due to the nature of the equipment,theywerenot able todealwith spray applying acoating,norwashingitdownforremoval.

5.2ApplicationThemanufacturer brush applied the waterborneVCIcontainingcoatingontotheequipment,whichresultedindrycoatthicknessesbetween2‐10mil(50‐250microns).Theequipmentwas transportto the final destination where the coating wasremovedbysimplypeelingitoff.

4. ConclusionsWithcorrosioncostsinthebillionsofdollarsandmakingupasmuchas3%ofUSA’sGDP10,thereisa need in the market place, for environmentallyfriendly,lowVOC,removablecoatingsthatcanbeapplied at a thin film thickness (1.0mils) whichprovideexcellentcorrosionprotectionandyetcanstill be easily removed. Compared to permanentcoatings where removal requires blasting or theuseofheavydutysolvents,orthickheavybarriertype wax coatings which are difficult to removeanddisposeof,manyremovablethinfilmcoatingswithVCI technologycanbeeasily removedusinganalkalinesolutionandhighpressurewater.

By research, this paper shows that systemsenhancedwithVCIinhibitorscangreatlyimprovethe corrosion resistance of both solvent andwaterborne coatings. In addition, waterbornecoatingswith VCI inhibitors can compete from aperformance aspect with solvent‐borne systems,while at the same time being moreenvironmentally friendly, easier to cleanup andlowerinVOCs.References1. C.W, Lea, FUCHS Lubricants, 2003 NationalCorrosionService.2. S.Prabhu,“TemporaryCorrosionProtectionduringStorage,TransportationandHandling”(2016).3. B.A., Miksic, “Use of Vapor Phase Inhibitors forCorrosion Protection of Metal Products,”CORROSION/83 paper no.308 (Houston, TX: NACE,1983).4. Y.I., Kuznetsov,, “InhibitingAction andAbsorptionofBeta‐AminoketonesonMetals”,ZasshchitaMetallov,32,5(1996)pp.528‐533.5. Technical Rules for Hazardous Substances, “TRGS615”,(2007).6. 48‐hourStaticAcuteD.pulexWaterAccomodatingFraction (WAF) Toxicity Testing, EnvironmentalEnterprisesUSA,Inc.,March1,20177. 96‐hour Static Acute P. Promelas WaterAccomodating Fraction (WAF) Toxicity Testing,EnvironmentalEnterprisesUSA,Inc.,March1,20178. 48‐HourStatic Acute Seawater, A‐bahia WaterAccomodating Fraction (WAF) Toxicity Testing,EnvironmentalEnterprisesUSA,Inc.,March1,20179. 96‐hour Static Acute Seawater, M.beryllinaWaterAccomodating Fraction (WAF) Toxicity Testing,EnvironmentalEnterprisesUSA,Inc.,March1,201710. NACE,“CostofCorrosionStudy,”(2002).

Vol.4 No.1 2018 12|Page

Corrosion characterization of ruthenium implanted AISI 304L in proton exchange membrane fuel cell environments

FortunateMoyoUniversityoftheWitwatersrand1JamSmutsJohannesburg2050SouthAfricaDanielWamwangiUniversityoftheWitwatersrand1JanSmutsJohannesburg2050SouthAfrica

JosiasvanderMerweUniversityoftheWitwatersrand1JanSmutsJohannesburg2050SouthAfrica

ARTICLEINFO

Keywords:CorrosionRutheniumProtonexchangemembranefuelcellsIonimplantation

ABSTRACT

Rutheniumwasimplantedinto304Lstainlesssteeltoincreasecorrosionresistanceinasyntheticprotonexchangemembranefuelcells(PEMFCs)simulationsolutionconsistingof10‐4MH2SO4and0.5M sodium sulfate (Na2SO4)at60°C.Corrosioncharacterization was done using potentiodynamic and potentiostaticmeasurement techniques, and metallic contaminants in the corrosion mediumweremeasuredusingatomicabsorptionspectroscopy.PotentiostaticpolarizationwasdoneatpotentialstypicaloftheanodeandcathodeenvironmentsofPEMFCs,with immersion times of up to 10 days. The corrosion rate of the rutheniumimplanted stainless steelwas

CorrosioncharacterizationofrutheniumimplantedAISI304Linprotonexchangemembranefuelcellenvironments

Vol.4 No.1 2018 13|Page

60 and 80 °C, and (5) distributing fuel gas andair.6‐8

Figure1:SchematicdiagramofaPEMFC.6The multifunctional characteristic of bipolarplates requires amaterialwith an assortment ofproperties, including good thermal and electricalconductivity, high strength, good machinabilityand most importantly, zero permeability to thereactant gases used in PEMFCs. Stainless steelssuch as AISI 304L easily satisfy theserequirements. Furthermore, the wide availabilityofstainlesssteelspresentsanopportunityforcosteffectivemassproductionofPEMFCs.However,stainlesssteelsarepronetocorrosioninthe reducing acidic conditions generated inPEMFCs.7,8 Reactions at the anode (Figure 1)involvetheoxidationofhydrogen,Equation1,andtheionsthusproducedarereducedatthecathodeaccording to Equation 2.6,9 Corrosion of stainlesssteel bipolar plates deteriorates the performanceof PEMFCs because corrosion products increaseinterfacial contact resistance between the platesand reduce power output. In addition, dissolvedmetallic ionspoison thecatalyst in thepolymericmembrane,10,11 further reducing cell efficiencybylowering ionicconductivityof themembraneandincreasingchargetransferresistance.

(Equation1) (Equation2)

Surface modification is undoubtedly a costeffectivemeansofimprovingcorrosionresistanceof stainless steels in PEMFCs environments.

Several researchers have focused on coatingstainless steelswithprotective layersof titaniumnitride,12‐14 chromium‐carbon15,16 and carbon17produced mostly by physical vapor deposition.Although some of the coatings studied exhibitedmetal‐like electrical conductivity and/oradequately inhibitedmetal dissolution, there arestill concerns around intrinsic defects such aspinholes and poor adhesion. These defectsmakelong‐termapplicabilityof thecoatings inPEMFCsdoubtful, and necessitate consideration ofalternativesurfacemodificationtechniques.Inthisstudy,rutheniumwasimplantedinto304Lstainless steel and the effect on corrosionresistanceinPEMFCsimulatedenvironmentswasinvestigated. Ion implantation is largely used fordopingsemiconductors forelectronicapplication,buthasattractedsubstantialinterestinthefieldofcorrosion in recent years,18‐23 and has beenconsidered for application in PEMFCs.24,25Implanted Ru was demonstrated by Tjong andChu22 to improve corrosion resistance of ferriticstainless steel in0.5M sulfuric acid (H2SO4), andcouldpotentiallybebeneficialinPEMFCs.2. ExperimentalProcedureSamples of AISI 304L stainless steel (18.32 Cr,8.20Ni(wt%))measuring20x20mm2werecutfroma6mmthickplateandmechanicallygroundwith a series of silicon carbide papers ofdecreasing grit size to a finish of 120 grit (~125μm). The samples were then cleanedultrasonically:firstwithdetergentwaterandthenwith ethanol for 20 minutes to remove anyembedded silicon carbide particles and othercontaminantsduetosamplehandling.Thereafter, the304Lstainlesssteelsampleswereimplanted at room temperature with 50 keV Ruions inaVarian350Dion implanterusingadoseof 1x1016 ions/cm2 at a zero degree angle.Simulation of the implantation process usingStopping Range of Ions in Matter (SRIM) and90x103 ionsrevealedthat thepeakconcentrationofRushouldlie13nmbeneaththestainlesssteelsurface.To analyze and confirm Ru distribution in thestainless steel, Rutherford BackscatteringSpectroscopy(RBS)wascarriedoutusing3.6MeVheliumionswithaspotsizeof500x500μm2.Theback scattered particles were detected with asolid‐statedetectorlocatedat150°totheprimary


Vol.4 No.1 2018 14|Page

ion beam, and the recorded spectra weresimulatedusingtheRUMPcode.Anelectrolyte consistingof10‐4MH2SO4and0.5Msodiumsulfate(Na2SO4)26wasusedtosimulatethe PEMFC medium. Corrosion characterizationwas done at different potentials for periodsrangingfrom2to240h:atthecorrosionpotential(Ecorr),at‐195mVAg/AgCl(‐240mVSCE)theoperationpotentialinsimulatedanodeenvironments,andat645mVAg/AgCl (600mVSCE) theoperationpotentialin simulated cathodic environments.12Potentiodynamicpolarizationwasdoneat a scanrate of 0.5 mV/s from cathodic to anodicpotentials, and the recorded data were used tocalculate current densities (icorr) by the Stern‐Geary method as described in ASTM G102‐89(2010).27Allelectrochemicaltestsweredoneat60±1°Cina thermally controlled 3‐electrodeelectrochemicalcellcomprisingthetestsampleasthe working electrode, a Ag/AgCl referenceelectrode(~199mVSHE),andalargeareagraphiteelectrode. The reference electrode had anapplication range of 0‐80 °C. A PGSTAT302Autolab potentiostat interfaced with Nova 1.70data acquisition software was used. Theelectrolyte used to simulate the PEMFC mediumwas highly conductive; hence, no potential drop(IR) compensation was done. Fresh solutionswere used for every test, and pristine (non‐implanted) 304L stainless steel was used as acontrol.ThechemicalcompositionofthePEMFCsyntheticsolution was analyzed after corrosion exposureusing an atomic absorption spectroscopy (AAS).Analysiswas limited to iron and chromium ions,whichcanpotentiallycontaminatethemembraneand catalyst in PEMFCs.10 In all instances, freshsolutions and deionized water were used ascontrols.Toguardagainstsystematicerrors,all testswererandomized. The precision of the potentiostatused in this study was determined as describedelsewhere.28,29 Ecorr values measured by thepotentiostatwereaccuratetowithin±7.4mV,andicorr values to within ± 7.1 µA/cm2. Since the

number of replicates per test condition waslimited by financial constraints, it was assumedthat the Ecorr and icorr measurements on thesampleshadsimilarprecisionandrepeatabilityasthepotentiostat.3. ResultsFigure 2 shows the RBS spectra obtained on thepristineandRuimplanted304Lstainlesssteel.InFigure2(b), there is adistinctpeakat~3.1MeV,correspondingtotheimplantedRu.

Figure 2: RBS spectra of (a) pristine and (b) Ru implanted304L stainless steel after ion implantation at 50 keV using1x106ions/cm2.Potentiodynamic polarization curves for pristineand Ru implanted 304L stainless steel in thesimulatedfuelcellsolutionareshowninFigure3.Onthewhole,bothspecimenshadcorrosionratesthatwere<16µA/cm², the target corrosion ratefor bipolar plates used in PEMFCs.7 In fact, the

(a)

(b)


Vol.4 No.1 2018 15|Page

corrosion rates of the stainless steel before andafter Ru implantation were similar: ~0.04 and0.035µA/cm²respectively.Figure 3 shows a higher value of Ecorr for the Ruimplanted stainless steel. It was ~ ‐60 mVcomparedto~‐160mVforpristine304Lstainlesssteel.ThisriseinEcorrwaslikelyaconsequenceofcathodic modification,30,31 where Ru lowered thehydrogen overvoltage of 304L stainless steel,increased theefficiencyof the cathodic corrosionreactionandshiftedEcorrtomoreanodicvalues.Forbothtestsamples,theEcorrwasgreaterthan‐195 mV, the operation potential of the anodeenvironmentofPEMFCs.Thismeansthatboththepristine and Ru implanted 304L stainless steelwould be corrosion resistant in the anodeenvironmentof the fuel cell.Atpotentials typicalof the cathode environment of PEMFCs (~645mVAg/AgCl), pristine 304L stainless steel could besusceptible to significant dissolution. It is clearfromFigure3thatthesepotentialscoincidedwiththeonsetoftranspassivecorrosionofthestainlesssteel, where trivalent chromium responsible forpassivity isoxidizedtonon‐protectivehexavalentCr.The curve corresponding to Ru implantedstainless steel clearly showed the absence of theanodic ‘nose’ typicalofactive‐passivealloys.Thisimplies that the implanted stainless steel waspronetospontaneouspassivationinthesimulatedsolution. The anodic ‘nose’ on the curvecorrespondingtopristine304Lstainlesssteelwasvery small but significantly wide, suggestingdifficultytopassivateinthetestsolution.

Figure 3: Polarization curves of pristine and Ru implanted304L stainless steels after 24 h immersion in 10‐4MH2SO4

with 0.5 M Na2SO4 at 60°C, at a scan rate of 0.5 mV/s vsAg/AgCl.Figure4showsthatresultsobtainedwhenthetestsamples were exposed to the simulated fuel cellsolutionfor10days.Opencircuitpotential(OCP)forbothpristineandRuimplanted304Lstainlesssteel shifted towards positive potentials (Figure4(a)).Thisisconsistentwithmovingtopotentialswhere Cr (III) species are predominant andthermodynamically stable,32 and points toincreasedabilitytoformstableprotectiveCr(III)filmswithincreaseinexposuretime.Noticeably,OCPvaluesonRu implantedstainlesssteel were nearly steady, varying from ~100 to140 mV in the last 8 days of exposure. For thesame period, the OCP values recorded on thepristine304Lstainlesssteelrangedfrom≈100to230mV.ThiswiderangeinOCPcouldhavebeencaused by a thickening passive layer, which islikely to increase interfacial contact resistance(ICR)ofthefuelcellduringoperationandresultinreducedcellefficiency.The amount of iron dissolved during the 10 dayexposure ispresented inFigure4(b). IntroducingRu via ion implantation clearly reducedsusceptibilityforthestainlesssteeltodissolutionby > 50% in some instances. Nonetheless, theamount of dissolved iron in the corrosionsolutionsincreasedwithincreasedexposuretime,suggesting that with time, cell efficiency couldpotentially decrease due to poisoning of thecatalystinthepolymericmembrane.Current‐timecurvesforpristineandRuimplanted304Lstainlesssteelat645and‐195mVAg/AgClarepresented in Figure 5. For clarity, data > |1µA/cm²| were omitted from the analyses in thefigure.Current density values recorded on the pristine304Lstainlesssteelat645mVwerequiteerraticas can be seen in Figure 5(a). This is consistentwith the expectation that 645 mV, the typicalpotential of PEMFC cathode environment,coincidedwiththeonsetoftranspassivecorrosionfor the pristine stainless steel (Figure 3).Transpassive corrosion is typically characterized


Vol.4 No.1 2018 16|Page

by localized dissolution of Cr (III) films thatpresumablycoverthestainlesssteelsurfacewhenitispassive,andthisdissolutionresultsincurrenttransients,suchas thoseobserved inFigure5(a).On the other hand, current densities on the Ruimplanted 304L stainless steel were very stableover the sameexposureperiod, andweremostly<0.1 µA/cm². It is likely therefore that Ruimplanted 304L stainless steel would be muchmore stable in the cathode regionof the fuel cellthanpristinestainlesssteel.

Figure4: Variationof(a)OCPvsAg/AgCl,and(b)dissolvedironwith time forpristineandRu implanted304Lstainlesssteelduringexposure to10‐4MH2SO4with0.5MNa2SO4at60°Cfor10days.At potentials typical of the PEFMC anodeenvironment (~ ‐195 mVAg/AgCl), the currentdensities recorded on the two samples werenegative (Figure 5(b)). Consistent with Figure 3,nometaldissolutionwouldbeexpectedfromboththe pristine and Ru implanted stainless steel atthesepotentials.ThecurrentdensitiesrecordedinFigure5(b)arepossiblyassociatedwithhydrogenevolution by Equation 3, which is a typicalsupporting cathodic reaction in corrosionoccurring in acidicmedia. Currentdensity valueson the Ru implanted samples were numericallylarger, suggesting higher reaction rates on the

alloysurface.Thisisconsistentwiththeprincipleofcathodicmodificationmentionedearlier.

(Equation3)

Figure5:CurrentvstimeplotsforpristineandRuimplanted304 stainless steel potentiostatically polarized in 10‐4 MH2SO4with0.5MNa2SO4at60°Cfor24hat:(a)645mVand(b)‐195mVvsAg/AgCl.After potentiostatic exposure at 645 mV, thechemical composition of the solution wasanalyzed via AAS, and the results obtained aregiven in Figure 6. The high iron dissolutionassociated with the pristine 304L stainless steelconfirmedthatmetaldissolutionwouldbehigheras 645 mV coincided with the onset oftranspassive corrosion. Interestingly, the amountofCrdissolvedduringexposurewasthesameforboth pristine and Ru implanted 304L stainlesssteels.ItshouldbenotedthattheAASusedinthisanalysishadnocapacitytodistinguishCr(III)andCr(VI)ions.4. DiscussionThe corrosion resistance of 304L stainless steelreliesontheformationofaself‐healingtenaciousCr (III) oxide on its surface. Although generallyprotective,Cr2O3filmsarepronetodissolutionin

(a)

(b)


Vol.4 No.1 2018 17|Page

acidicmediaandatpotentials>800mVSHE(~600mVAg/AgCl).32Atthispotentialrange,alsoknownasthe transpassive region, Cr (III) is oxidized tosolubleCr(VI)speciessuchasCrO42‐andCr2O72‐,whicharenotprotective.33

Figure 6: Concentration of iron and chromium dissolvedfrompristineandRuimplanted304Lafter24hexposureto10‐4 M H2SO4 with 0.5 M Na2SO4 at 60°C and 645 mV vsAg/AgCl.As shown in Figure 3, the onset of transpassivecorrosionforpristine304Lstainlesssteelexposedto 10‐4 M H2SO4 with 0.5 M Na2SO4 occurred at~650 mV, which coincided with the typicalpotential for the PEMFC cathode environment.Inevitably,304Lstainlesssteelwouldbepronetosignificant dissolution in this environment asdemonstrated in Figure 6, which would in turnpoisonthecatalystinthepolymericmembrane.Adding Ru to the surface of 304L stainless steelvia ion implantation seemed to increase thetolerance of the stainless steel in the fuel cellsimulation solution by shifting the onset oftranspassive potentials to higher values. Thelower iron dissolution and steady currentvariations in Figures 6 and 5(a) confirmed thissupposition. Ru had the effect of increasing therate of hydrogen evolution by Equation 3, andshifting Ecorr to more anodic potentials asdemonstrated in Figures 3 and 5(b) and byprevious researchers in acidic media.22,34,35However, there is no documentation to supportthe shift in transpassive potentials observed inFigure 3, implying that an additionalmechanismofcorrosionprotectionoccurred.Ion implantation is a process known to createnearly amorphic surface layers.18,36 Transpassivecorrosion is typically intergranular, consistent

with the fact that grain boundaries are highenergy regions, and are easily prone to anodicdissolution. Amorphisation therefore reducestendency to undergo intergranular attack, andcould potentially increase resistance of 304Lstainlesssteeltodissolutionat~645mVAg/AgCl.TheOCPofpristine304Lstainlesssteel in10‐4MH2SO4 with 0.5 M Na2SO4 markedly increasedduring exposure (Figure 4(a)). Decreasing OCPvalues are typically associatedwith dissolving orthinning passive films,37,38 and similarly,increasingOCPvaluesindicatethickeningpassivefilms. Considering that passive films areessentially products of corrosion, thick passivefilms could indicate large amounts of corrosionproducts on the stainless steel surface, whichcouldsignificantlyincreaseICRofthePEMFCsandreducecellefficiency.In the presence of Ru, OCP variations on 304Lstainless were relatively steadier throughout theanalysis period (Figure 4(a)) indicating passivefilms of nearly constant thickness. In addition,measured OCP values were mainly lower thanthose recorded on pristine 304L stainless steel.This suggests the formation of thinner passivefilms and accumulation of less corrosionproducton themetalsurface,whichcouldresult in lowerICR and better cell efficiency. Reduced metaldissolution reported in Figures 4(b) and 6demonstratedtheprotectivenessofthesefilms.5. ConclusionsThe corrosion behavior of Ru implanted 304Lstainless steelwas characterized in10‐4MH2SO4with 0.5 M Na2SO4 at 60°C with the view ofpotential application in PEMFCs. Although thepresence of implanted Ru did not necessarilyresult in lower corrosion rates, it certainlyincreased tolerance of 304L stainless steel in thecathodeenvironmentofPEMFCs(~645mVAg/AgCl)by shifting transpassive potentials to highervalues. Long term, Ru implanted stainless steelexhibited stable performance, which is essentialforlongtermstabilityofPEMFCs.While the results of this study show that Ruimplanted stainless steel holds promise for


Vol.4 No.1 2018 18|Page

potentialas‐castapplicationinPEMFCsasbipolarplates,itisnecessaryto:1. optimize Ru implantation dose to maximize

performance2. investigate ICR of the Ru implanted stainless

steel.

AcknowledgementsThesupportofDST‐NRFCentreofExcellenceinStrongMaterials (CoE‐SM) towards this research is herebyacknowledged. The Department of Science andTechnology, andNational Research Foundation, SouthAfrica are thanked for financial support. The authorsare grateful to Anglo Platinum for providing thepowderused toprepare the ruthenium targetused inion implantation. Many thanks to iThemba LABS(North) for generously allowing the authors access totheionimplanter.References1. A. Sakhare, A. Davari, A. Feliachi, “Fuzzy logic

control of fuel cell for stand‐alone and gridconnection,” Journal of Power Sources 13 (2004):pp.165‐176.

2. W. Rothengatter, “Climate change and thecontributionof transport:Basic facts and the roleof aviation,” Transportation Research Part D 15(2010):pp.5‐13.

3. R.Sims,R.Schaeffer,F.Creutzig,X.Cruz‐Núñez,M.D’Agosto,D.Dimitriu,M.J.FigueroaMeza,L.Fulton,S.Kobayashi,O.Lah,A.McKinnon,P.Newman,M.Ouyang, J.J. Schauer, D. Sperling, and G. Tiwari,“Transport,”IPCCClimateChange2014:MitigationofClimateChange.ContributionofworkinggroupIII to the Fifth Assessment Report of theIntergovernmental Panel on Climate Change[Edenhofer, O., R. Pichs‐Madruga, Y. Sokona, E.Farahani,S.Kadner,K.Seyboth,A.Adler,I.Baum,S.Brunner,P.Eickemeier,B.Kriemann,J.Savolainen,S. Schlömer, C. von Stechow, T. Zwickel and J.C.Minx (eds.)] (Cambridge, UK, New York, USA:CambridgeUniversityPress,2014),p.603.

4. C. Wang, S. Wang, L. Peng, J. Zhang, Z. Shao, J.Huang,C.Sun,M.Ouyan,X.He,“Recentprogressonthe key materials and components for protonexchange membrane fuel cells in vehicleapplications,”Energies9,603(2016):pp.1‐39.

5. W.R.W.Daud,R.E.Rosli,E.H.Majlan,S.A.A.Hamid,R. Mohamed, T. Husaini, “PEM fuel cell systemcontrol:A review,”RenewableEnergy 113 (2017):pp.620‐638.

6. A. Hermann, T. Chaudhuri, P. Spagnol, “Bipolarplates for PEM fuel cells: A review,” InternationalJournal ofHydrogen Energy 30 (2005): pp. 1297‐1302.

7. D.J.L.Brett,N.P.Brandon,“Reviewofmaterialsandcharacterization methods for polymer electrolyte


Vol.4 No.1 2018 19|Page

fuel cell flow‐field plates,” Journal of Fuel CellScienceandTechnology4(2007):pp.29‐44.

8. R.A. Antunes, M.C.L. Oliveria, G. Ett, V. Ett,“Corrosion of metal bipolar plates for PEM fuelcells:A review,” International JournalofHydrogenEnergy35(2010):pp.3632‐3647.

9. S.‐J. Lee, C.‐H. Haung, J.‐J. Lai, Y.‐P. Chen,“Corrosion‐resistant component for PEM fuelcells,” Journal of Power Sources 131 (2004): pp.162‐168.

10. A. Pozio, R.F. Silva, M. De Francesco, L. Giorgi,“Nafiondegradation inPEFCs fromendplate ironcontamination,”ElectrochimicaActa48(2003):pp.1543‐1549.

11. P. Ju, Y. Zuo, Y. Tang, X. Zhao, “The enhancedpassivation of 316L stainless steel in a simulatedfuel cell environment by surface plating withpalladium,”Corrosion Science 66 (2013): pp. 330‐336.

12. M. Li, S. Luo, C. Zeng, J. Shen, H. Liu, C. Cao,“Corrosion behavior of TiN coated type 316stainlesssteelinsimulatedPEMFCenvironments,”CorrosionScience46(2004):pp.1369‐1380.

13. E.A.Cho,U.‐S. Jeon,S.‐A.Hong,I.‐H.Oh,S.‐G.Kang,“Performance of a 1 kW‐class PEMFC stack usingTiN‐coated 316 stainless steel bipolar plates,”JournalofPowerSources142(2005):pp.177‐183.

14. Y. Wang, D.O. Northwood, “An investigation intoTiN‐coated 316L stainless steel as a bipolar platematerial for PEM fuel cell,” Journal of PowerSources165(2007):pp.293‐298.

15. H.‐C. Wang, H.‐H. Shen, C.‐E. Lu, K.‐H. Hou, M.‐D.Ger, “Preparation of corrosion‐resistant andconductivetrivalentCr‐Ccoatingson304stainlesssteel for use as bipolar plates in proton exchangemembranefuelcellsbyelectrodeposition,” JournalofPowerSources293(2015):pp.475‐483.

16. W. Mingee, L. Congda, H. Tao, C. Guohai, W.Doughui,Z.Haifeng,Z.Doug,W.Aiying,“Chromiuminterlayeramorphouscarbonfilmfor304stainlesssteelbipolarplateforprotonexchangemembranefuel cell,” Surface and Coatings Technology 307(2016):pp.374‐381.

17. T.Fukutsuka,T.Yamaguchi,S.‐I.Miyano,Y.Matsuo,Y.Sugie,Z.Ogumi,“Carbon‐coatedstainlesssteelasPEFC bipolar plate material,” Journal of PowerSources174(2007):pp.199‐205.

18. S.C. Tjong, M.C. Wong, S.P. Wong, “Corrosionbehaviorofboron‐andphosphorous‐implantedFe‐40Cr alloy in reducing acid solutions,” AppliedSurfaceScience64(1993):pp.127‐132.

19. M. Wang, Q.Y. Zhnag, W.D. Shi, T.C. Ma, S.Z. Hou,“Effects of Mo and Mo + N implantation oncorrosion resistance of austenitic stainless steel,”Surface and Coatings Technology 65 (1994): pp.171‐174.

20. C.M.Abreu,M.J.Cristóbal,X.R.Nóvoa,G.Pena,M.C.Pérez, R.J. Rodríguez, “Modifications of thestainless steel passive film induced by ceriumimplantation,” Surface and Coatings Technology158‐159(2002):pp.582‐587.

21. C.M.Abreu,M.J.Cristóbal,P.Merino,X.R.Nóvoa,G.Pena, M.C. Pérez, “Electrochemical behavior of anAISI304Lstainlesssteelimplantedwithnitrogen,”ElectrochimicaActa53(2008):pp.6000‐6007.

22. S.C. Tjong, P.K. Chu, “Corrosion properties of Fe‐24Cr stainless steel alloy modified by plasmaimmersion ion implantation in0.5M sulfuric acidsolution,” Surface and Coatings Technology 201(2007):pp.6781‐6784.

23. N.Padhy, S.Ningshen,B.K.Panigrahi,U.K.Mudali,“CorrosionbehaviorofnitrogenionimplantedAISItype 304L stainless steel in nitric acid medium,”CorrosionScience52(2010):pp.104‐112.

24. K. Feng, Y. Sheu, J. Mai, D. Liu, X. Cai, “ Aninvestigation into nickel implanted 316L stainlesssteelasabipolarplateforPEMfuelcell,”JournalofPowerSources182(2008):pp.145‐152.

25. K.Feng,Y.Sheu,D.Liu,P.K.Chu,X.Cai,“Ni‐CrCo‐implanted 316L stainless steel as bipolar plate inpolymer electrolyte membrane fuel cells,”International Journal of Hydrogen Energy 35(2010):pp.690‐700.

26. J.S. Kim, W.H.A. Peelen, K. Hemmes, R.C. Makkus,“Effects of alloying elements on the contactresistanceandthepassivationbehaviorofstainlesssteels,”CorrosionScience44(2002):pp.635‐655.

27. ASTM G102‐89 (2010), “Standard practice forcalculating of corrosion rates and relatedinformation from electrochemical measurements”(WestConshohocken,PA:ASTM)

28. F. Moyo, J.W. van der Merwe, D. Wamwangi,“Corrosionperformanceofpulseplatedruthenium:Dependence on pulse‐off time,” Surface andCoatingsTechnology307(2016):pp.971‐977.


Vol.4 No.1 2018 20|Page

29. ASTMG16‐95(2010),“Standardguideforapplyingstatistics to analysis of corrosion data” (WestConshohocken,PA:ASTM)

30. N.D.Tomashov,“Passivityandcorrosionresistanceofmetal systems,”CorrosionScience 4 (1964):pp.315‐334.

31. J.H. Potgieter, “Alloys cathodically modified withnoble metals,” Journal of Applied Electrochemistry21(1991):pp.471‐482.

32. S. Haupt, H.‐H. Strehblow, “A combined surfaceanalytical and electrochemical study of theformationofpassivelayersonFe/Cralloysin0.5MH2SO4,”CorrosionScience37(1)(1995):pp.43‐54.

33. P. Schmuki, “FormBacon to barriers: a reviewonthepassivityofmetalsandalloys,” JournalofSolidStateElectrochemistry6(2002):pp.145‐164.

34. E.M.Sherif,J.H.Potgieter,L.Cornish,P.A.Olubambi,C.N. Machio, “Effects of minor additions ofruthenium on the passivation of duplex stainlesssteel corrosion in concentrated hydrochloric acid

solution,” Journal of Applied Electrochemistry 39(2009):pp.1385‐1392.

35. J.H.Potgieter,W.O.Barnard,G.Myburg,K.Varga,P.Baradlai, L. Tomcsanyi, “Corrosion behavior ofduplexstainlesssteelscontainingminorrutheniumadditions in reducing acid media,” Journal ofApplied Electrochemistry 26 (1996): pp. 1103‐1110.

36. T.R. Rautray, R. Narayanau, K.‐H. Kim, “Ionimplantation of titanium based biomaterials,”Progress inMaterialsScience56(2011):pp.1137‐1177.

37. E.M. Sherif, J.H. Potgieter, J.D. Comins, L. Cornish,P.A.Olubambi,C.N.Machio,“Thebeneficialeffectofruthenium additions on the passivation of duplexstainless steel corrosion in sodium chloridesolution,” Corrosion Science 51 (2009): pp. 1364‐1371.

38. H.Luo,H. Su,B. Li,G. Ying, “Electrochemical andpassive behavior of tin alloyed ferritic stainlesssteel in concrete environment,” Applied SurfaceScience439(2018)pp.232‐239.

Vol.4 No.1 2018 21|Page

The effect of SiC content on the corrosion and tribocorrosion performance of SPS produced Al-SiC nanocomposites

P.V.CobbinahVaalUniversityofTechnologyDepartmentofMetallurgicalEngineeringAndriesPotgieterBlvdVanderbijlpark,Gauteng,1911SouthAfrica

W.R.MatizamhukaVaalUniversityofTechnologyDepartmentofMetallurgicalEngineeringAndriesPotgieterBlvdVanderbijlpark,Gauteng,1911SouthAfrica

ARTICLEINFO

Keywords:Sparkplasmasintering(SPS)AnodicandcathodicpolarisationCurrentdensityLinearsweepvoltammetry(LSV)Tribocorrosion

ABSTRACT

Using the spark plasma sintering (SPS) technique, five Al‐SiC composites wereproducedwith differingproportions of SiC ranging from 15wt% to 40wt%.TherelativedensitiesoftheSPScompositeswereabove97%,includingfortheAlalloy.IncreasingtheSiCcontentincreasedthehardnessofthecomposites.Thecorrosionbehaviour of the SPS Al‐SiC composites immersed in 3.5wt% NaCl at roomtemperaturewerestudiedelectrochemicallybypotentiodynamicpolarisation.TheTafel polarisation plots showed that the corrosion resistance of compositesreinforced with 15 and 20wt% SiC improved compared to the Al alloy. In allsamples, pits were also observed on both the Al alloy and composites afterpolarisation.Thesynergisticeffectsofwearandcorrosiononthecompositeswerealsoassessed.

1. IntroductionGood corrosion resistance, superior wearresistance, lowercoefficientofthermalexpansion(CTE), improvements in both mechanicalpropertiesandstructuralefficiencyarepropertiesascribed to metal matrix composites (MMCs).1Researchers have studied the mechanical andphysical properties of aluminium MMCs 2–5 andalsotheircorrosionbehaviour.6–8Although Al alloys are known to form a passivefilm9, interactions between the reinforcing phasesuchasSiCandthemonolithicAlalloy inevitablymake the composite prone to corrosion.Metallurgicalfactorsinthematrix,suchasdefectsand composition inhomogeneities, contribute tocorrosionofthecomposite,andthesecanserveassusceptiblesitesfortheinitiationofcorrosion.At

theirinterfaces,localisedformsofcorrosion,suchaspreferentialorcreviceattackandmicrogalvanicset‐ups promote the degradation of thecomposites.10,11CorrosioncharacteristicsofAl‐SiCcompositesareinfluenced by the volume fractions of thereinforcing phase.12–14 Conversely, findings ofother investigators7,8,15 suggest the corrosionbehaviourof thecomposite is independentof theamount of SiC making up the composite. Thedifference in findings can be as a result of thedifferentprocessingroutes16,which influence themicrostructures, and Al alloys17 employed asmatrix.However,despitethedisagreement,alltheinvestigators observed the formation of pits inboththecompositesandunreinforcedmatrix.

TheeffectofSiCcontentonthecorrosionandtribocorrosionperformanceofSPSproducedAl‐SiCnanocomposites

Vol.4 No.1 2018 22|Page

The phenomenon of tribocorrosion is thepermanent deterioration of a material, resultingfrom simultaneous mechanical and chemicalsurfaceinteractionsexistingintribologicalcontactin a corrosive environment. In a tribologicalsystem, the nature of wear, friction andlubrication influence tribocorrosion. Thus, thetotalmateriallossincludesdegradationfromwearonly, corrosion only and synergistic effects.18 Ingeneral,asdesired fromcomposites,MMCsboastof a goodbalanceofproperties in relation to theindividual inherent properties of the materialsthat make up the composite. In this respect, theaddition of SiC to the Al alloy is to improve itswear properties with the aim of not losing itsinherentcorrosionresistance.The studywas on the corrosion behaviour of Al‐SiC composites produced using spark plasmasintering (SPS). Not only is this method offabrication energy efficient, it also achievesconsolidationandhomogenisationinashorttimethrough volumetric joule heating of the samplesaided by pressure leading to no grain growth.19Potentiodynamic polarisation was used tomonitortheresponseofthecompositestomarineenvironments, and their corrosion resistancewere subsequently assessed. The tribocorrosionperformanceofthecompositeswerealsostudied.2. ExperimentalProcedure2.1. MaterialsandSamplePreparationAs‐received powders include aluminium (99.5%purity, 7‐15µm) and silicon carbide (SiC) (99%purity, 44µm). Composites reinforced with SiCrangingfrom15to40wt%wereprepared.Usingaluminaballs,ball‐to‐powderratio(BPR)of5:1,aReeves attritor mill and a rotating speed of 400rpm, SiC was first wetmilled in hexane for 16.5hoursandtheAlalloyfor8hours.Therespectivecompositeswerealsowetmilledfor8hourseachand 2 wt% of polyethylene glycol (PEG) wasadded as a process control agent (PCA). Thecomposites were then dried using a rotaryevaporator. The PCA was burnt off in a tubefurnace, under an argon atmosphere, at 400°Cusingaheatingrateof5°C/minandanisothermalholdtimeof30mins.Sparkplasmasintering(SPS)was carried out on each composite using a FCTHP‐D5sinteringfurnaceatapressureof50MPa,a

heating and cooling rate of 200oC/min and100oC/minrespectively.Samplesweresinteredat600oC and held for 10mins in a vacuum. Agraphitedieofinnerdiameter20mmandgraphitepunches were used. The die was lined withgraphite foil along the cylindrical surface and atthe flat ends, and hexagonal boron nitride (hBN)coatedgraphitefoilswereused.2.2. PowderandCompositeCharacterisationUsingaMalversizer, theparticle sizedistributionoftheas‐receivedpowders(AlalloyandSiC)andmilled powders were analysed before sintering.For metallography, sintered composites wereground with SiC paper through to 1200 grit,polished with diamond paste suspensions up to1µm and etched with Keller’s reagent. Phaseidentification and morphological studies ofcomposites were carried out using a ShimadzuXRD‐7000diffractometerandanOlympusBX51Moptical microscope. A Carl Zeiss Sigma fieldemission SEM was used to determine thecomposition of the various microstructuralfeatures.2.3. DensityandHardnessDeterminationSinteredsamplesweresandblastedtoremovetheattached coating and graphite foil. UsingArchimedes’ principle to calculate the actualdensityinEquation1,sinteredsampleswerefirstweighedinair(Wα)andafterweighedagainwhilesuspendedindistilledwater(Ww).

(1)wherea=actualdensity,Wα=massofthesamplein air, Ww = mass in distilled water and w =densityofdistilledwater.Hardness testing was carried out using anInnovast Falcon 500 by subjecting the sinteredcomposites to a load of 5N for 10 seconds forsevendifferentindentations.2.4. ElectrochemicalCorrosionTestingSintered samples for corrosion testing were cutinto dimensions of 10mm X 10mm. With an


Vol.4 No.1 2018 23|Page

aluminiumtape,anelectricalwirewasattachedtothe surface of each cut sample which was thenmounted. The 1cm2 exposed surface were thenground to 1200 grit, rinsedwith ethanol, hot airdried and kept in a desiccator. The mountedsamples with the wire protrusion served as theworkingelectrode,anAg/AgClelectrodewasusedas the reference electrode and platinum wire asthe counter electrode. Using the Digi‐Ivypotentiostat and DY2300 series software, theexposedsurfaceswereimmersedin3.5wt%NaClsolutionandpolarised.The corrosion behaviour of the compositeswereassessed from the Tafel polarization plotsobtainedbylinearsweepvoltammetry(LSV)test,atascanrateof0.1mV/sandthenanodicallyandcathodicallypolarisedfrom+200mVto–1000mV. 3 ResultsandDiscussions3.1 PowderandCompositeCharacterisation3.1.1 ParticleSizeDistributionFrom the particle size distribution curve, theparticlesizesoftheas‐receivedpowdersatd(50)for the Al alloy and SiC were found to be11.993µm and 48.012µm respectively. Aftermilling, the particle sizes then decreased to4.439µmforAlalloyand0.12µmforSiC.Figure1shows that the particle size distribution at d(50)for the compositewith30wt%SiC reinforcementwas0.489µmbefore sintering.TheSiCmilled for16 hours and Al‐30wt% SiC composite exhibitedbimodaldistributions.

Figure1:Particlesizedistributioncurvesofas‐receivedandmilledpowdersatd(50).3.1.2 XRDAnalysis

XRD identified no new phases, especiallydetrimental carbides such as Al4C3 and Al4SiC4.This can be attributed to solid‐state sintering ofthe composites. Figure 2 shows the peaksintensities of SiC increased with increased SiCproportions in the composite, whereas the Almatrixpeakintensitiesdecreasedless.

Figure1:XRDpatternsofmilledAl‐SiCcomposites3.1.3 Microstructural Characterisation of Al‐SiCCompositesFigure 3 shows optical micrographs of the AlcompositesreinforcedwithSiCof15wt%,20wt%,25wt%and40wt%.SiCwerefairlydistributedinthematrix.InareaswhereSiCwasless,poolsofAlalloy were noticed. This is probably due to thewide range of proportions of Al alloy in thesamples.

Figure3:Opticalmicrographsofcompositesaftersinteringat600oC: (a)Al‐15wt%SiC, (b)Al‐20wt%SiC, (c)Al‐25wt%SiC,and(d)Al‐40wt%SiC,whereSiChasthedarkercontrast.


Vol.4 No.1 2018 24|Page

Figure4(a)ofthemilledAl‐20wt%SiCcompositeshowsthatSiCparticleswereoftenstuckontheAlalloypowders.SPS is a pressure‐aided solid‐state sinteringtechnique, andFigure 4(b) shows the sinteringprocessduringconsolidationofthemilledpowdercomposites. An instantaneous decrease intemperature was observed when pressure wasapplied during the process, and a constantpressurewasmaintainedduringthe10minsdwelltimeat600C.3.1.4 DensityandHardnessMeasurementsThe actual densities of the composites increasedwith increased SiC content (Figure 5(a)). Therelative densities were all above 97%, henceexhibiting the effectiveness of the SPS method.Furthermore, the increased proportions of SiCadded to the ductile Al matrix, improved thehardnessof thecompositeasexpected,asshowninFigure5(b).

Figure4:(a)SEMsecondaryelectronimageofAl‐20wt%SiCpowdermilledfor8h,(b)sinteringanddisplacementcurvesofanAl‐SiCcomposite.

Figure 5: Properties of different composition SPS Al‐SiCcomposites:(a)densities,and(b)averageVicker’shardness3.2 ElectrochemicalMeasurements3.2.1 CorrosionPotential‐TimeTrendsComposites were immersed for 3200s at a scanrate of 0.05V/s, and their open circuit potentials(OCP) are presented in Table 1. The composites

reinforcedwith15and40wt%SiCexhibitedshiftsofpotential inthemorenegativedirectionbeforestabilising at potentials of about ‐800mV and ‐700mV respectively. For 30wt% SiC containingcomposite, the potential approached the positivepotential region.The trendsobserved for20 and25wt% SiC reinforced composites and thereferencemonolithicAlalloywerefluctuationsofthe applied potentials with time. This shows atendency of localised corrosion, like pitting, totakeplace.20Table1Opencircuitpotential(OCP),corrosionpotential(vsAg/AgCl) and corrosion current density of Al‐SiC SPScompositesimmersedin3.5wt%NaCl.Composite

Opencircuitpotential,OCP(mV)

Corrosionpotential,Ecorr(mVAg/AgCl)

Corrosioncurrent,Icorr(A/cm2)

Al‐0%SiC ‐739 ‐897 2.379x10‐5Al‐15%SiC ‐621 ‐645 9.447x10‐7Al‐20%SiC ‐668 ‐738 9.225x10‐5Al‐25%SiC ‐687 ‐701 1.487x10‐3Al‐30%SiC ‐717 ‐726 7.304x10‐4Al‐40%SiC ‐627 ‐729 2.061x10‐33.2.2 TafelPolarisationPlotsThecorrosionpotentialsobtainedfromthelinearsweep voltammetry (LSV) test listed in Table 1show the reference unreinforced Al alloy had anEcorrmorenegativethanthecomposites.However,the corrosion resistance of the compositesreinforcedwith15and20wt%SiCimprovedwithrespect to its corrosion current density (Icorr)compared to the unreinforced Al alloy. Thecorrosion resistance of the other compositesslightly decreased. Al‐15wt% SiC exhibited thehighestpolarizationresistance.A grey layer formed on all the composites afterexposure,withsometinywhitespots,likelytobepits. With the Al alloy matrix being reactive, itformed a protective oxide layer, and it isspeculated that the Cl‐ ions in the 3.5wt% NaClwere adsorbed and occupied flaws in theprotectivefilm.Theseanionsthenformedsolublecompoundswhichpenetrated the film, leading tothe pits. It is likely that sequentially, passivecompounds Al2O3 and Al(OH)3 were formed oflower energy and reacted with the Cl‐ ions,forming aluminium complex ions and transitorycomplexes such as AlCl++, AlCl4‐, Al(OH)Cl2 andAl(OH)2Cl,culminatinginthethinningoftheoxidefilm.9Thefundamentalstepsaresummarizedas9:


Vol.4 No.1 2018 25|Page

Figure6:TafelpolarisationplotsofSPSAl/SiCcompositesimmersedin3.5wt%NaClsolution.4. ConclusionUsing the results obtained, the followingconclusionsweredrawn: Al‐SiC composites produced using SPS

exhibited increased density with increasedproportionsofSiC.

The hardness of the Al‐SiC compositesincreasedcomparedtotheAlalloy,duetotheSiC ceramic reinforcement to the ductile Almatrix.

AcknowledgementsSpecial thanks go to Dr Patrick Rokebrand of theUniversity of the Witwatersrand (Wits) for all hisassistanceduringexperimentation.References1. D.B.Miracle,"MetalMatrixComposite–FromSciencetoTechnologicalSignificance",CompositeScienceandTechnology65(2005):pp.2526–2540.2. C.Sun,M.Song,Z.Wang,Y.He,"EffectofParticleSizeontheMicrostructuresandMechanicalPropertiesofSiC‐ReinforcedPureAluminumComposites",JournalofMaterialsEngineeringandPerformance20(2011):pp.1606–1612.3. S.Min,"EffectsofVolumeFractionofSiCParticlesonMechanicalPropertiesofSiC/AlComposites",TransactionsofNonferrousMetalsSocietyofChina19(2009):pp.1400–1404.

4. S.Bathula,R.C.Anandani,A.Dhar,A.K.Srivastava,"MicrostructuralFeaturesandMechanicalPropertiesofAl5083/SiCpMetalMatrixNanocompositesProducedbyHighEnergyBallMillingandSparkPlasmaSintering",MaterialsScienceandEngineeringA545(2012):pp.97–102.5. Z.Zhang,F.Wang,J.Luo,S.Lee,L.Wang,"MicrostructuresandMechanicalPropertiesofSparkPlasmaSinteredAl–SiCCompositesContainingHighVolumeFractionofSiC",MaterialsScienceandEngineeringA527(2010):pp.7235–7240.6. T.S.Mahmoud,E.Y.El‐kady,A.Al‐Shihri,"CorrosionBehaviourofAl/SiCandAl/Al2O3Nanocomposites",CorrosionEngineering,ScienceandTechnology47(2012):pp.45–54.7. P.C.R.Nunes,L.VRamanathan,"CorrosionBehaviorofAlumina‐AluminumandSiliconCarbide‐AluminumMetal‐MatrixComposites",Corrosion51(1995):pp.610–617.8. Y.Shimizu,T.Nishimura,I.Matsushima,"CorrosionResistanceofAl‐BasedMetalMatrixComposites",MaterialsScienceandEngineeringA198(1995):pp.113–118.9. R.T.Foley,"LocalizedCorrosionofAluminumAlloys—AReview",Corrosion42(1986):pp.277–288.10.A.J.Trowsdale,B.Noble,S.J.Harris,I.S.R.Gibbins,G.E.Thompson,G.C.Woods,"TheInfluenceofSiliconCarbideReinforcementonthePittingBehaviourofAluminium",CorrosionScience38(1996):pp.177–191.11.M.Metzger,S.G.Fishman,"CorrosionofAluminum‐MatrixComposites.Statusreport",IndustrialandEngineeringChemistryProdustResearchandDevelopment22(1983):pp.296–302.12.M.M.Buarzaiga,S.J.Thorpe,"CorrosionBehaviorofAs‐cast,SiliconCarbideParticulate‐AluminumAlloyMetal‐MatrixComposites",Corrosion50(1994):pp.176–185.13. E.Maahn,S.Roepstorff,“CorrosionResistanceofAluminium‐SiliconCarbideCompositeMaterials”,in12thRisoSymposiumofMetalMatrixComposites,eds.N.Hansen,D.J.Jensen,T.Leffers,H.Lilholt,T.Lorentzen,A.S.Pedersen,O.B.Pedersen,andB.Ralph(Roskilde,Denmark:RisoNationalLaboratory,1990),p.497.14.I.BSingh,D.P.Mandal,M.Singh,S.Das,"InfluenceofSiCParticlesAdditionontheCorrosionBehaviorof2014Al‐CuAlloyin3.5%NaClsolution",CorrosionScience51(2009):pp.234–241.


Vol.4 No.1 2018 26|Page

15.H.Sun,E.Y.Koo,H.G.Wheat,"CorrosionBehaviorofSiCp/6061AlMetalMatrixComposites",Corrosion47(1991):pp.741–753.16.M.S.N.Bhat,M.Surappa,H..SudhakerNayak,"CorrosionBehaviourofSiliconCarbideParticleReinforced6061/AIalloycomposites",JournalofMaterialsScience26(1991):pp.4991–4996.17.S.Coleman,V.D.Scott,B.McEnaney,"CorrosionBehaviourofAluminium‐BasedMetalMatrixComposites",JournalofMaterialsScience29(1994):pp.2826–2834.18.D.Landolt,S.Mischler,M.Stemp,"ElectrochemicalMethodsinTribocorrosion:ACriticalAppraisal",ElectrochimicaActa46(2001):pp.3913–3929.

19.W.R.Matizamhuka,"SparkPlasmaSintering(SPS)–AnAdvancedSinteringTechniqueforStructuralNanocompositeMaterials",TheJournaloftheSouthernAfricanInstituteofMiningandMetallurgy116(2016):pp.1171–1180.20.N.Diomidis,P.Celis,J,P.Ponthiaux,F.Wenger,"AMethodologyfortheAssessmentoftheTribocorrosionofPassivatingMetallicMaterials",LubricationScience21(2009):pp.53–67.

Vol.4 No.1 2018 27|Page

The electrochemistry and the prediction of stress corrosion cracking

JosiasvanderMerweUniversityoftheWitwatersrandandDST‐NRFCentreofExcellenceinStrongMaterials1JanSmutsAveJohannesburg,Gauteng,2001SouthAfrica

ARTICLEINFO

Keywords:Stress‐corrosioncrackingPassivityStressCorrosion

ABSTRACT

Stress‐corrosion cracking is very dependent upon the susceptible alloy andenvironmental interactions to either sustain cracking, ormitigate or inhibit it.Therefore, to understand this interaction, it is important to characterise theelectrochemistry, and in particular the passivation characteristics of the alloy.Crack propagation occurs when passivation is disrupted by the strain of theunderlyingmetalbecauseoftheapplicationofanexternalforceorresidualstress.Whenthepassivefilmonthemetalisnotverystable,thisdisruptionismorelikelyandcrackpropagationbecomesverydependentuponpassivationcharacteristics,such as the passivation rate and the extent of passivation. Therefore, byunderstanding the passivation characteristics, it is possible to determine themechanismof crackpropagationand identify theparameters that influence thepropagationrate. Inaddition, itbecomespossibletopredictthesusceptibilitytocracking,aswellaspossiblymonitortheelectrochemicalnoise inordertodetectcracking.Thiswasinvestigatedinvariousstress‐corrosionsensitivesystemswherethe passivation characteristics were compared to stress‐corrosion susceptibilitywhile parameters, such as temperature, were altered to influence crackpropagationrates.Thecorrelationbetweenpassivityandcracksensitivitycanbeusedtopredicttheextentofcrackingoncearelationshiphasbeenestablishedforaparticular system. However, it remains difficult to make an exact predictionbecauseofvariousparametersthatinfluencecracking,especiallyduringthecrackinitiation stage. Although detecting and monitoring crack activity in certainsystemswaspossible,itshouldbeevaluatedfurther.

1. IntroductionStress corrosionhas beendifficult topredict andin many instances the onset and propagation ofcracks are not easily predictable1. This isdependentuponanumberof factorsthathavetobedescribed,butcrackingoccursgenerallyintwodistinct stages, crack initiation and propagation2.Theinitiationstagecouldoccuroveranextendedperiod of time and might require very specificconditions to enable the crack to form. Whenthese conditions are not met, cracking will notoccur,andmanytimesunderplantconditionsthis

is exactly the reason why the system might notlead to cracking3. This of course could lead tofindingwaystomitigatethecrackingprocess,butfirst of all it is important to understand theparameters that would enhance the crackinitiation process. Stress corrosion occurswhenthereexistsafinebalancebetweenthetransitionfrompassivationtolocalisedcorrosion4.Withtheapplicationofstress,thepassivefilmisdisruptedwhenaslipstepemergesonthesurfacewherethepassive film has formed. This disruption can beenhanced by the slow passivation of the freshly

Theelectrochemistryandthepredictionofstresscorrosioncracking

Vol.4 No.1 2018 28|Page

exposed metal, or can be reformed to ensurecontinuous passivation. Therefore, when thebalance between passivation and local corrosionis at a certain level, the formation of a crackbecomes possible. If corrosion is enhanced, thiscould lead to pitting corrosion, or in the mostsevere case, uniform corrosion. However, whenpassivation isveryeffectivecompletepassivationwilloccur nocrackingwill result. Generally, thepassive‐active transition zones have beenidentified as electrochemical potential regionswhere cracking will occur with greater ease. Itwas the aim of this study to investigate thispassivationcharacterthroughtheelectrochemicalnature, and then to compare this with the cracksensitivity as determined with slow strain‐ratetest results. Therefore, would it be possible topredict the extent of cracking from theelectrochemical characteristics of the crackingprocess? Two systems were investigated todetermine various sensitivities and the extent ofcracking of very different material‐environmentcombinations,andprotectivemechanisms.Oneofthese systems was investigated previously todetermine the extent of passivation required toinhibitcrackingcompletely5.Thisishelpful,sinceat thepointofsufficientpassivation, thecrackingmightevenbeenhanced. Theembrittlementofahigh strength steel exposed to low pH aqueousconditionswasanalysedoverawidetemperature

range. Theseresultswerecompared toasystemfoundinlowcarbonpressurevesselsteelexposedto a carbon monoxide‐carbon dioxideenvironment, where the effect of corrosion wasalteredbytheadditionofaninhibitor5.2. ExperimentalProcedureA high strength steel was evaluated and thechemicalanalysesofthesteelusedwereobtainedby optical emission spectroscopy and the resultsareshowninTable1.3.0 ResultsandDiscussionsThe slow strain‐rate test results of the samplesexposedtothelowpHenvironmentareshowninTable2.Thenatureofthecracksfoundwereintergranularin places, as shown in the tempered martensiticmicrostructureinFigure1.ThenatureofthecracksonthesurfaceofthesteelsampleseenwithastereomicroscopeareshowninFigure2.The polarisation characteristics of the steel at45Cwas investigated,andtheresultsareshowninFigure3. Heretheabsenceofapassiveregionwasevident.

Table1Bulkcompositionofthehighstrengthsteel

C Mn S P Si Cr Mo Ni Cu Al V Nb Co FeWeight% 0.14 1.25 ≤0.005 0.01 0.23 0.41 0.05 0.08 0.02 0.074 0.015 0.006 0.019 Balance

Table2Slowstrain‐ratetestresultsatvarioustemperatures

Temperature d0 l0 df A0 Af% Red.inArea

35°C 5.0 25.2 4.0 19.8 12.6 36.745°C 5.1 24.7 4.5 20.4 15.8 22.555°C 5.0 24.0 4.6 19.6 17.0 13.765°C 5.0 24.0 4.6 19.6 16.8 14.6

Figure1:Temperedmartensiticstructureofthesteelwithatypicalcrackrunningthroughthecentre.

InvestigatingwaterpipelinesusingIR‐freepotentialmeasurementsunderstraycurrentinterferenceconditions

Vol.4 No.1 2018 29|Page

Figure2: Secondary surface crackson the sample testedat65°C.

Figure 3: Polarisation characteristics of the steel in theenvironmentat45°C.3.1 Passivationandsensitivitytocrackingof

steelinCO‐CO2‐H2OIn the carbon monoxide‐carbon dioxide‐watersystem, the effect of passivationwasmuchmoresignificantasshowninFigure4.Here,apressurevesselA516grade70steelwasusedforthetestsas described previously6. The nature ofpassivation was significantly affected bytemperature and was especially pronounced atthehighcarbonmonoxideconcentration. Carbonmonoxide falls in the category of an adsorptioninhibitor (Heusler7) and therefore inhibits thecorrosion process of the steel by strongchemisorbtion84.Thiswasconfirmedforiron,onwhich it reacts slowly to formcarbonyl,which isnot stable in acid solutions7, making carbonmonoxide not a very effective inhibitor, but onethatcouldbedisruptedbyslipsteps.

Cotton and Wilkinson9 indicated that carbonmonoxide possessed an electronic structuresimilar to cyanide, in that it is a reproducingcomplexing ligand which bonds with the metalatom via the carbon atom. Here, the weakelectrondonorpropertiesofthemoleculeandthevacantorbitalsofthemetalareinvolvedwiththecorrosionreaction,assuggestedbyTrabanelli10.The carbon monoxide molecule has a pair ofelectronsstronglydirectedawayfromthecarbon‐oxygen bond7, which facilitate adsorption on ametal with unfilled electron levels. Carbonmonoxide is not adsorbed on zinc, and does notinhibit zinc dissolution as found by King andRau12.Theyfoundthatcarbonmonoxidewasnotcompletely effective as an inhibitor, since duringweight loss experiments, therewas someweightloss, even after short exposure times, which isusually not true for pure adsorption inhibitors.The reason for this ineffectiveness might be theslow adsorption of carbon monoxide to protectcertain regions, as found by Heusler7. Carbonmonoxideadsorbsreversiblyonironinsulphuricacid. The adsorption caused a slight shift inpotential, but the overall effect was seen on theanodic polarisation curve, with a considerablediminution of the corrosion rate. This slowadsorption of the carbon monoxide stronglyinfluenced the stress‐corrosion sensitivity of thesteel.

Figure4:PolarisationcharacteristicsofA516pressurevesselsteelexposedto50%CO–50%CO2at800kPa,atascanrateof 0.1mV/s after being held at open circuit potential (OCP)


Vol.4 No.1 2018 30|Page

forvarioustimesfrom1minuteto5hours.Bluecurveswereperformedat25°Candtheredcurvesat45°C13.From slow strain‐rate results, it was shown thatthesteelexposed to50%CO‐50%CO2 (800kPa)at 45°C was considerably more susceptible tocrackingthanforthesamemixtureat25°C.Fromthe curves shown in Figure 4, it is clear that theextentofpassivation increasedby several ordersof magnitude with the increase in temperature.Notonlywasthepassiveregionmorepronouncedat45°C,butalsothepassivationwasinfluencedbythetimeexposedtotheenvironment.Thecurvesshown in Figure 4were determined in the sameenvironment, although not only at twotemperatures,butalsodifferentexposuretimesof10and15minutesaswellas45minutes. Whiletheexposure timewasnotexactly thesame timefortheshorterexposure,theextentandinfluenceon inhibition was clear. At the short exposuretime and 25°C the passive region was notsignificantlydifferentfromwherethecurvewouldhavebeen in the active state. However, after 45minutes, thepassiveregionhadextendedslightlyandevenintheactiveregion,beyond‐390mV,thecurve showed an overall drop in current densityvalues. This time dependent passivation of thesteel surface was different at the highertemperature where the stress corrosionsusceptibilitywasalsomoresignificant.For45°C,thecurrentdensitydifferencewasmoreapparentbetween the active region, beyond ‐430mV, andthe passive region, which was where the stresscorrosion susceptibility also increased. Inaddition,after45minutesthedifferencebetweenthepassive regionand theactiveprojectedcurve(notshown)wasalsomoreafter15minutes.Itisimportant to note that the severity of theenvironment‐metal interaction was increasedwith a faster passivation rate. Therefore, a finebalance exists between the passivation rate andthecrackpropagationrate.3.2 Inhibited steel in CO‐CO2‐H2O

environmentAsshownbyVanderMerweandDuToit6forsteelin the carbon monoxide‐carbon dioxide system,the passivation required for complete inhibitionwith no crack initiation or propagation was

significant. When passivation was improved bythe addition of a corrosion inhibitor, thesusceptibilitytocrackingevenincreasedwhentheinhibitor concentration was too low. Thepolarization characteristics of the steel showedthatwhen the passive current density decreasedfrom 100 µA/cm2 to approximately 10µA/cm2,theinhibitionwassufficienttoremovesensitivityto cracking. However, the extent of the passiveregionintheinhibitedstatewasoveralongrangeand not the short passive potential range ofaround 80mV found on the 25% CO‐75% CO2systemat 45°C. The shape of the passive regionactuallychangedcompletely,andtheactiveregionwas removed for the potential region beingevaluated.4. ConclusionFrom these three systemsdescribed, itwas clearthat the dependence of cracking on thepassivation of the steel varied widely. The highstrength steel did not show any signs ofpassivation, but its sensitivity to crackingincreased with temperature. The crackingmechanism in this case related to hydrogenembrittlement, considering the high strength ofthe steel and the presence of the hydrogenreactionat lowpHlevels. Thecarbonmonoxide‐carbon dioxide system showed a definiterelationship between crack propagation andpassivation.Whenpassivationwassignificantandoccurringatahighenoughrate,crackpropagationwas sustained. However, to completely inhibitcracking, it was necessary to add an inhibitor,which extended the passive range over a muchlarger potential region. Therefore, to predictcracking from the passivation would be difficultand this should be considered only byunderstanding the passivation rate and loadingrates, and especially the mechanism causingembrittlement. However,oncecrackinghasbeenestablished,themeasureofpassivationmightgiveaqualitativeindicationofcrackingsusceptibility.


Vol.4 No.1 2018 31|Page

References1. M. Kowaka and S. Nagata, “Transgranular stress‐corrosioncrackingofmildsteelsandlow‐alloysteelsinH2‐CO‐CO2systems,”Corrosion24,12(1968):p.427–429.2. R.N. Parkins, “Significance of pits, crevices, andcracks in environment‐sensitive crack growth,”MaterialScienceTechnology1,(1985):p.480.3. R.H. Jones and R.E. Ricker, “Mechanism of stress‐corrosioncracking,”1sted. (MaterialsPark,Ohio:ASMInternational,1992):p.8.4. A. Brown, J.T. Harrison and R. Wilkins, CorrosionScience, 10 (1970), p. 547.5. J. W. van der Merwe,"EnvironmentalandMaterial InfluencesontheStress‐Corrosion Cracking of Steel inH2O‐CO‐CO2 Solutions,"InternationalJournalofCorrosion,vol.2012,ArticleID414156,(2012):13pages.6. J.W. van der Merwe and M. du Toit, “Mitigation ofstress corrosion cracking of carbon steel exposed toCO‐CO2‐H2Oenvironmentsthroughinhibitoraddition,”Anti‐CorrosionMethodsandMaterials60,6(2013):p.295–300.7.K.E.Heusler,G.H.Cartledge,“Theinfluenceofiodideionsandcarbonmonoxideontheanodicdissolutionofactiveiron,”JournaloftheElectrochemicalSociety108,8(1961):p.732‐740.8. C.DeWaard,D.E.Milliams,Corrosion31, 5 (1975):p.177.9. F.A. Cotton, G. Wilikinson, Advanced inorganicchemistry,611,Interscience,(1968).10.G.Trabanelli,F.Zucchi,G.I.Zucchini, “Influencedumono‐oxyde de carbone sur les processus dedissolution des mètaux,” Corrosion ‐ Traitementprotectionfinition16,7(1968):p.335‐345.11.H.B.Urbach,L.G.Adams,R.E. Smith, Journalof theElectrochemicalSociety55,(1974):p391.12. C.V. King, E. Rau, Journal of the ElectrochemicalSociety,103,(1956):p.331.13. J.W.vanderMerwe,Thestress‐corrosioncrackingofcarbonsteelinCO‐CO2‐H2O,PhDthesis,UniversityofPretoria(Pretoria,2013),p.118.

Vol.4 No.1 2018 32|Page

Benchmarking international cathodic protection and alternating current mitigation standards to determine suitable protection criteria

CraigBothaReignite(Pty)LtdDurban,SouthAfrica

ARTICLEINFO

Keywords:BenchmarkingCathodicProtectionAlternating;CurrentSP0169EN50162EN13509ISO15589‐1TM0497RP0104SP0177EN15280SP21424CurrentDensityCorrosion

ABSTRACT

From time to time international standards are withdrawn, amended,updatedornewstandardsareintroducedthatchangethecurrentthinkingonexistingsubjectmatter.Thispaperdemonstratesthecomplexprocessofreviewing a number of prominent Cathodic Protection and AlternatingCurrentMitigation standardsand thenattempting tobenchmark suitableprotection criteria. Today common Right ofWay issues exist due to theconvergenceofpipelines,highvoltagepowerlines,telecommunicationsandotherservicesthatproduceanenvironmentconducivetorapidcorrosiononpreviouslysecurepipelineassets.In this paper we review aspects of ISO15589‐1 Part 1, NACE SP0169,EN50162,EN13509,NACETM0497,NACERP0104and theworkofNACEInternationalTaskGroups210and211.InadditiontotheseclassicalworksrelatingtoCathodicProtection,thispapergivesattentiontoaspectsofthecontemporary Alternating Current standards and reports namely NACESP0177, BS EN 15280, NACE SP21424‐2018 and NACE InternationalTechnicalReport35110.AsthesedocumentsareunpackedandassesseditsoonbecomesevidentthattheexclusiveuseofCathodicProtectioninahighlycongestedRightofWaydoesnotguarantee thatcorrosionwillbecurtailedonwellcoated,buriedsteelpipelines.Some commentsand recommendationsaremade toguide the thinkingofengineers,pipelineownersandoperatorstowardsapplyingsuitabledesignsandmore importantlymonitoring themostcriticalvariables inpredictingthe probability of corrosion in complex Alternating Current and DirectCurrentenvironments.

1. IntroductionThe past 20 years in the field of CathodicProtection (CP)haveushered in thedistinct shift

from the assessment of conventional DirectCurrent (DC) corrosionmechanisms to themorecomplexworld of electrochemistry influenced byAlternatingCurrentcorrosionmechanisms.Infacttraditional assessment of CP protection levels

Benchmarkinginternationalcathodicprotectionandalternatingcurrentmitigationstandardstodeterminesuitableprotectioncriteria

Vol.4 No.1 2018 33|Page

without taking cognizance of theAC interferenceon a structure can have catastrophicconsequences for pipeline operators. The mostcommon misnomer over the past 20 years hasbeentheideathatapipelineundersuitablelevelsofCPcannotcorrodeinaHighVoltagePowerlineRightofWay.Extensive amounts of research and in‐fieldinvestigations have revealed thatAC corrosion isnow a significant element of pipeline integrityassessment and the emphasis over two decadeshasbeentoupdatestandards,theintroductionofACmitigationsystemsandthereallifeassessmentoftheprobabilityorlikelihoodofACcorrosioninareaswhere CP criteria appear to be acceptable.This paper will firstly assess recent thinking onthe measurement of DC structure‐to‐electrolytepotentials, secondly provide some insight intodevelopments around AC structure‐to‐electrolytepotential measurements and thirdly provideguidelines being used to assess the likelihood ofcorrosionbasedonratioofACcurrentdensity toDC current density at simulated coating defectsitesonatypicalburiedpipeline.2. Direct Current Structure to Electrolyte

PotentialMeasurementsA number of standards, reports and documentsare considered here and relevant portions arehighlighted from each to develop a picture ofcurrentthinking.2.1. ISO15589‐1 Petroleum and Natural Gas

Industries‐CathodicProtectionOfPipelineTransportation Systems. Part 1 on‐landpipelines1

Specific reference is made to paragraph 5.3Criteria for CP and the sub‐clauses containedtherein as we have extracted some key pointsfromthisstandard: The acceptable metal‐to‐electrolyte potential

at which the corrosion rate is less than0.01mm per year is the protection potential,Ep. This considered acceptable and withinlimitsforthedesignlife.ThecriterionforCPistherefore E ≤ Ep, where E is the polarizedstructurepotential.

Some metals can be subject to hydrogenembrittlementatverynegativepotentials,andcoating damage can also increase at verynegative potentials. For such metals, thepotential shall not be more negative than alimitingcriticalpotentialEl.Insuchcases,thecriterionforCPisEl≤E≤Ep.

The CP system shall be capable of polarizingall parts of the buried pipeline to potentialsmore negative than −850 mV referred to asaturated Copper Sulphate ReferenceElectrode (CSE), and to maintain suchpotentials throughout the design life of thepipeline. These potentials are those whichexist at the metal‐to‐environment interface,i.e.thepolarizedpotentials.

Topreventdamagetothecoating,thelimitingcriticalpotentialshouldnotbemorenegativethan −1200mV referred to CSE, to avoid thedetrimental effects of hydrogen productionand/orahighpHatthemetalsurface.Itmustbe noted that this criterion is unlikely to beachieved in SouthAfrica due to the influenceofStrayCurrentemanatingfromtheextensiveDC Traction Railway network. For highstrength steels (specified minimum yieldstrength greater than 550 MPa), the limitingcritical potential shall be determined withrespect to the detrimental effects in thematerial due to hydrogen formation at themetalsurface.

Forpipelinesoperatinginanaerobicsoilsandwhere there are known, or suspected,significant quantities of sulfate‐reducingbacteria (SRB) and/or other bacteria havingdetrimental effects on pipeline steels,potentials more negative than −950 mVreferred to CSE should be used to controlexternalcorrosion.

Forpipelinesoperatinginsoilswithveryhighresistivity, a protection potential morepositivethan−850mVreferredtoCSEmaybeconsidered,e.g.asfollows:−750mVfor100<ρ<1000;−650mVforρ>1000whereρisthesoilresistivity,expressedinohmmetres.

As an alternative to the protection potentialsgivenabove,aminimumof100mVofcathodicpolarizationbetweenthepipelinesurfaceanda reference electrode contacting theelectrolytemaybeused.

Benchmarkinginternationalcathodicprotectio

Micro-galvanic corrosion of Ti-6Al-4V in situ alloyed with molybdenum via selective ... · 2018....

Documents

Transcript of Micro-galvanic corrosion of Ti-6Al-4V in situ alloyed with molybdenum via selective ... · 2018....