A D V A N C E D M A T E R I A L S & P R O C E S S E S | J U L Y / A U G U S T 2 0 2 1 2 2 and large consistently the leading edge of a raster melt pool. Though research into the fluid dynamics in point melting is scarce, parallels can be drawn from publications on laser spot welding[19], where Marangoni convection (i.e., thermocapillary forces) dominates melt pool thermodynamics. Thermal gradients within individual melt pools are therefore higher in point-melting strategies compared to linear raster with elongated melt pools. Higher thermal gradients allow thermocapillary forces to dominate the fluid flow through a greater relative volume of each melt pool and eliminate some of the gas porosity present in the melt pools by dragging them to the surface where they are eliminated. Melt pool fluid dynamics and the ability of gas bubbles to escape themelt pool are not influenced by location (in a plane) or by height, as observed in the L, D, and R samples, only by varying the scanning strategy. Thermal gradients within AM parts are known to vary along the height of a part, based on thermal conductivity through the build plate as opposed to the surrounding powder bed, pre-heat temperatures, time at temperature, and the number of layers above a selected region, among other factors[2,14]. These thermal gradient variations affect thermal (and thermo- mechanical) gyrations and corresponding phase transformations, which results in notable differences in the resulting microstructure[2,14]. The fact that no significant variation in spherical porosity was observed along the build height suggests that the best time for gas pores to escape is during initial melting, either while the electron beam is centered on the melt pool or immediately after, particularly for point-melting strategies. INDUSTRIAL RELEVANCE While some of the typically undesirable attributes of AM builds (e.g., residual stress and surface roughness) can be mitigated with post-processing or concurrent hybrid-processing steps (e.g., annealing, in-situ or ex- situ machining)[2], porosity, on the other hand, is difficult to eliminate with these methods[4,16]. Porosity is detrimental in all structural metals, as all pores act as stress concentrators and can lead to premature failure[1,2,6,20]. Investigations into optimizing processing parameters to reduce porosity in powder bed fusion (PDF) AM techniques are common[1,4,5], but tend to focus on power and scan speed and rarely venture into differing scan strategies. In AM builds, the size of the gas pores ranges considerably, from tens of microns (Fig. 3a) to less than 5 microns (Fig. 3b). The smaller pore sizes are more commonly observed in the point-melting scans. With low volume fractions of porosity, it can be difficult to determine accurately the amount and size of porosity. Knowledge of the size of gas pores is critical, as large pores (Fig. 3a) will lead to a larger local loss of tensile properties than smaller pores. As mentioned, post-processing steps, typically in the form of heat treatments, are usually not successful at removing all pores in as-built parts[2,4,16]; thus, methods of reducing and eliminating porosity during the build process itself is preferable and the most effective route to ensure the final mechanical performance of a part is as expected during the design process. This work provides analysis on strategies which mitigate porosity (i.e., point-melting strategies are better than raster scans at eliminating porosity and reducing the size of the pores present) as a result of the forces dominating different stages of the melt pool fluid dynamics. CONCLUSIONS Observations of the spherical porosity in EBM Ti-6Al-4V builds have shown a clear difference in the resulting fraction and the size of pores present between a linear raster-melting strategy and point-melting strategies (Dehoff and random). Whether as a result of retained gas porosity from the powder, or vaporized elements in the build, samples from all three melting strategies present gas porosity. Fewer spherical gas pores were observed in both of the point-melting EBM scan strategies as opposed to the prototypical, and widely used, linear raster melting scan strategy. Point-melting strategies also resulted in spherical gas pores with an average diameter of half the size of the average diameter of pores in the raster-melting strategy. Pores with a diameter greater than 25 µm were only observed in the L sample. Gas pores are neither preferentially formed nor retained at any particular height of the sample, leading to the Fig. 3 — Spherical pores observed in EBM samples obtained with different scanning strategies: (a) linear raster scan-L and (b) point-melting scan-R. (a) (b)
RkJQdWJsaXNoZXIy MTYyMzk3NQ==