The effects of varying wall thickness on the surface roughness of Ti-6Al-4V by electron beam powder bed fusion

While the electron beam powder bed fusion (EB-PBF) process is recognized for generating parts with high surface roughness, the impact of wall thickness remains relatively understudied and shows inconsistent literature results. This study addresses this gap by performing a full factorial Design of Experiments (DoE) to examine Ti-6Al-4V specimen surfaces built in three distinct orientations (vertical, upskin, and downskin) and six wall thickness levels (0.5, 1.0, 2.0, 2.5, 3.0, and 4.0 mm). Findings reveal orientation-dependent effects of wall thickness on surface roughness, particularly for thicknesses below 1.0 mm. Vertical surfaces show increased roughness with decreasing wall thickness, in which 0.5 mm-thick specimens exhibited the highest Ra roughness values. In contrast, no particular trend was observed for upskin surfaces, with 0.5 mm specimens displaying roughness values comparable to the 2.0 to 4.0 mm range of specimens. Yet, downskin surfaces of 0.5 mm specimens showed significantly lower Ra roughness compared to their thicker counterparts, a trend opposite of that of vertical surfaces. These behaviors may relate to scanning strategies and electron beam parameters, particularly when building smaller cross-sectional areas.

Automated summary

This study investigates the impact of wall thickness on the surface roughness of Ti-6Al-4V specimens built using the EB-PBF process. The results show that wall thickness has an orientation-dependent effect on surface roughness, particularly for thicknesses below 1.0 mm. Vertical surfaces exhibit increased roughness with decreasing wall thickness, while upskin surfaces show no distinct trend. Interestingly, downskin surfaces of 0.5 mm specimens have lower roughness compared to thicker specimens, indicating an opposite trend to vertical surfaces. These findings can be attributed to scanning strategies and electron beam parameters, especially when building smaller cross-sectional areas.