Effect of input powder attributes on optimized processing and as-built tensile properties in laser powder bed fusion of AlSi10Mg alloy

Abstract

A systematic material characterization approach is presented to demonstrate the critical impact of input powder attributes on the optimized parameters and the as-built properties in the Laser Powder Bed Fusion (LPBF) of AlSi10Mg alloy. Two batches of powders (conventional and specialty) with widely different attributes, including particles’ morphology, size distribution, internal oxidation, surface chemistry and roughness, were separately investigated for the optimized LPBF parameters. It is showed that the specialty powder with a spherical morphology and a larger mean particle size exhibits a higher flowability, as well as a higher apparent density. The latter, together with a smoother surface topography in the specialty powder particles (thus a lower laser absorptivity), necessitates a higher volumetric energy density (VED) requirement for a complete fusion during the LPBF. The as-built microstructures which resulted from both powders show a similar evolution of total porosity content, melt-pool boundaries and solidification structure, both qualitatively and quantitatively, which is consistent with their comparable YS values. The specialty powder, however, yields a higher UTS and %El in the as-built state, which is believed to be due to a scarce presence of O-rich particles (which otherwise preside primarily within the intercellular regions, as is the case for the conventional powder). The emergence of such O-rich particles in the as-built microstructure could be traced back to the powder source (i.e., in the form of internal O-rich phases and a thick oxide surface layer in the conventional powder). In other words, an O-lean powder translates into an O-lean printed part. One-time recycling of the specialty powder is shown to have a negligible effect on its morphological characteristics, and thus on the optimized set of parameters, consistently yielding as-built YS values similar to those of the original powder. The as-built UTS and %El values resulting from the recycled powder, however, are notably lower than those from the original powder, which are suggested to be due to the presence of contaminants and/or oxide inclusions resulting from the LPBF processing itself and/or during the recycling procedure (e.g., sieving). We believe the correlations established here between the input powder attributes and the LPBF performance of AlSi10Mg alloy can be applicable to many metallic systems.