Page - 126 - in 3D Printing of Metals

Metals 2017,7, 64 energy is applied, resulting in spatter ejection [36]. In this study, spherical pores dominated the non-spherical ones, where rectangular-shaped poreswere visible only near the edges of the cut specimens. This indicates thatmostof theporositydefects in theSLMsampleswereduetogaspores duringthegasatomisationofthe316LSSpowders,similartotheworkcarriedoutbyTammas-Williams etal. [37]. Thegasporescouldbeproduceddueto thepresenceofmoistureorcontaminantsonthe surfaceof thepowderparticles [38]. TheseporescouldalsobeformedbythereactionsbetweenO2 andCwhicharepresent insmallamountsduringSLMprocessing,causingCOorCO2gasentrapment in theSLM-built parts [39]. Thenon-uniformporedistribution in thisprocess couldbe causedby oneof the followingfactors: (i) thevariationofsurfaceroughness [40]; and(ii) the layer-wisebuild manner of theAM[41]. Thevariation in surface roughness results in an inhomogeneouspowder distribution that leads to an inconsistentmelt flowandanunstablemoltenpool in the successive layers [42]. These, in turn, contribute to thediscontinuities in the scan track formation, andhence the irregularporedistributionobtained in this study.Nevertheless, thesedefects (poresandvoids) aredetrimental to thequality ofAM-fabricatedmetal parts, especially as they reducemechanical propertiessuchasyieldandtensile strength. 3.3.Microhardness Theresultsof theVickersmicrohardness (HV) testswereevaluatedby: (i) comparingHVvalues ofSLMsamples fordifferentcutplanes (x–y,x–zandy–z)asshowninFigure10;and(ii) comparing HVvaluesofSLMandWMsamples (Figure11). Figure10.Averagemicrohardness (HV)values forSLMspecimens in thex–y,x–zandy–zplanes. FromFigure10, it canbeobservedthat theaveragemicrohardnessvalues for theSLMspecimens inx–y,x–zandy–zplanesare262HV,237HV,and239HV, respectively. Themicrohardnessof the SLMsamplesat thex–yplane(scandirection)wasthehighestcomparedtotheother twoplaneswhich hadsimilarbutconsiderably lowerHVvalues. Thediscrepancyof themicrohardness ineachplane indicatesanisotropyinSLM,whichis typicalofAMprocesses formetalcomponents [28,43–45]. This is becauseofthelayer-wisebuildapproachwiththe“island”scanstrategyinAMprocesses,whichmeans localised melting of powder particles that often results in non-homogeneous morphologies and anisotropicgrainstructures [46,47].However, thesimilaraverageHVvalues in thex–zandy–zplanes indicatedamoreuniformmicrohardnessdistribution in thebuilddirection compared to the scan directionfor theSLM-processedsamples. Figure11showstheaveragemicrohardnessvaluesof the SLMsamples (228HV),whichwerehigher thanthoseof theWMsample (192HV).This is consistent withvarious literature, inwhichthemicrohardnessofAM316LSSparts is typicallyhigher thanthat ofconventionallymanufactured316LSSparts [28,48].Highermicrohardness inAM316LSSsamples 126