The particle size and shape distributions of metal powders used in additive manufacturing powder bed fusion processes are of technological importance for the final built product. Current three-dimensional (3D) measurements of these distributions assume a spherical shape, while techniques that measure both size and shape are always two-dimensional (2D) measurements of particle projections. This paper describes a set of techniques using X-ray computed tomography, combined with various mathematical algorithms, to measure the 3D size, shape, and internal porosity of individual particles. Calibrated by a limited amount of visual examination of 3D images of individual particles, these techniques can classify powder particles as single near-spherical (SnS) particles, and non-spherical (NS) particles, which consist of either single highly non-spherical particles or multi-particles, where two or more smaller particles have been joined together. From this 3D data, other algorithms can generate 2D particle size and shape information to compare with the results of 2D measurement techniques. These techniques are applied to two metal powders composed of a specific alloy of titanium with aluminum and vanadium, denoted as Ti64, which is in common use as a powder for selective laser or electron beam melting powder bed additive manufacturing. One powder was made with a gas-atomization process, the other with a plasma-atomization process, both have been recycled, and both pass the specifications for additive manufacturing use. The powders differ in the fraction of NS particles and porous particles, in their size and shape distributions, and in average shape and size statistics. The SnS/NS classification enables one to show how these classes contribute to the overall particle size distributions, even for a single powder type, and is useful for comparing different sources of powder as well as studying how the size/shape distributions of a powder might change over multiple recycling events.
Read full abstract