Page - 48 - in 3D Printing of Metals

Metals 2016,6, 284 Table1.Comparisonbetweenthecalculatedtheoreticaldelaminationwearrate (Wd)andtheactual globalwearrate (WG(exp)). Samples Wd (×10−12m3/m) WG(exp) (×10−12m3/m) 0.5mm 198.3 1.1 1mm 195.6 1.2 5mm 200.9 1.3 10mm 195.3 1.3 20mm 194.4 1.3 As-cast 208.7 1.4 The tribo-layer is commonly formedonametal surfacewhenrubbedbyacounterfaceduring aslidingwear test [28]. Inorder to furtherexplore thewearcharacteristicsandmechanismsof the EBM-builtandas-castTi64, it isfirstnecessary todelineate the termsof thenon-oxide tribo-layerand tribo-oxide layer. In the caseofnon-oxide tribo-layer, itmerely refers to theoxides that couldnot be identifiedbyXRDandarenotcompletelyabsentofanoxideontheTi64samplesurface.Onthe otherhand, the tribo-oxide layerrefers to the formationofastrongandcompactoxide layerduring slidingprocess thatservesasaprotectionto thewornsurface [28–30]. Inherently, inanEBMprocess where thebuilt condition isunderhighvacuumenvironment, the thicknessof surfaceoxide layeronly ranges from5to7nm[31].Moreover,all thewear testswereconductedat roomtemperature.Asa result, the formedoxide layerwasweakandincompact, thusallowingtheeaseofdelaminationof the oxide layer. Inourexperiment,white fragmentedparticleswereobservedfromboththeSEMimages of the0.5mmsample inFigure11a,where thewhiteparticles indicate thebreakdownof theoxide layer intooxideparticles. Thisobservationcanbereinforced fromtheEDXspectra inFigure11c,d. Moreover,due to theharshnatureduringsliding, theweardebris formedwasblackandpowdery innature. SubsequentXRDanalysisof theweardebris suggests that theoxideparticlesweremainly amorphousinnature.Onthecontrary,EDXwasunabletodetect thepresenceofoxygenonthetransfer layer (seeFigure11c) thatwasremovedfromthebulksample lyingbeneath theoxidesurface. Makingreference toall theresultsobtained in theanalysis, awearmechanismwasproposedand exemplified inFigure12. Becauseof thereactivenatureof titanium,TiO2wassubsequently formedas anoxide layeronthesamplesurface,whichhasbeenprovenbytheXPSresults inFigure8.During drysliding, theTiO2 layerwhichservesasaprotectionlayerof thesubstratesurfacewasfirst removed, thusgeneratingflakesofoxidedebris.Consequently,under loadingandrepeatedrubbing,coupled withthebrittlenatureof theoxidedebris, someof thesedebriswerefurtherbrokenupintofragmented oxideparticles. Further slidingresults in thecompete removalof theoxide layer, therebyallowing the directmetal-metal contact between the asperities of the counter ball and the sample surface. Accordingly,bulkmetalmaterialwas removed,which formedthemetaldebris. As the removalof thebulkmetal fromthesubstratesurface iscontinuousandinstantaneous,oxidemaynotbeable to formreadilyon themetaldebris surface andas suchallowing thesemetaldebris layers to adhere (transfer layer) together.However,as theslidingprocess in theexperimentwasrelativelyslowtothe onesreported in the literature, thecentrifugal force inducedduringslidingwasnotable toremove theoxidedebris, particles, and transfer layer fromthe sample surface. Therefore, these remaining materialsallowfor the formationofacompact tribo-layer that subsequentlyprevents further removal of themetalmaterial fromthesubstratesurface. Inourpresentexperimentof lowspeedsliding(2cm/s), thenon-oxide tribo-layerwasformed asaresultofbothoxideparticlesandmetaldebris.Additionally,ourcurrentexperimentresultsalso reveal that theglobalwearrate is twoorders lowercomparedto that fromthetheoretical calculation. This couldbeowing to the fact that the theoretical equationmaynot take into account the actual physical conditionduring sliding. With reference to theproposedwearmechanism illustrated in Figure12,atavery lowspeedof2cm/s, itwasobservedthat theoxideparticlesandmetaldebris still remainonthewear tracksurface.Hence, this tribo-layerwasnotpushedawaythroughcentrifugation 48