PSI - Issue 38

Hugo Roirand et al. / Procedia Structural Integrity 38 (2022) 149–158 Author name / Structural Integrity Procedia 00 (2021) 000 – 000

154

6

3.2. Defect study LPBF process is known to cause many defects such as lack of fusion or gas pores in as-printed specimens (Andreau et al., 2019b; Garlea et al., 2019; Liverani et al., 2017; Ronneberg et al., 2020; Vayssette, 2019). In order to quantify the size and number of defects in the four specimens studied, X-Ray tomography analysis have been done on cylinders (10 mm height, 4 mm diameter) built with the same strategy and on the same plateau as for the fatigue specimens. Defect data have been analyzed using Dream3D and Paraview software in order to plot the 2D projections images in Figure 4. They correspond to the projection of all the defects found inside the volume in the plane perpendicular to the building direction (ie: screen plane). They are colored depending on their equivalent volumic diameter. The first obvious result is that SLM_SR sample has a very low defect population compared to other samples. Defect are less numerous, smaller and spherical. These defects are certainly gas pores (Gunenthiram, 2018). On contrary, SLM_Chess sample shows a high density of defects, with many large size defects. By looking more precisely, it can be seen that these defect are elongated, with their maximum dimension contained into the plane perpendicular to building direction. This indicates that these are lacks of fusion, which probably occur between the islands of the chessboard pattern (Ronneberg et al., 2020). The scan of SLM_Rot sample shows a more numerous population of gas pores than the sample without rotation. This could be explained by the interaction between the laser and the metallic vapors that can occurs when laser comes into the direction of argon flux (Gunenthiram, 2018). However, even if pores are more numerous, doing rotations between layers does not modify the size distribution. The Orlas specimen seems to exhibits a comparable number of defects than SLM_Rot sample. However, there are bigger and their tortuous shapes suggest that there are lacks of fusion.

(b)

(a)

(c)

(d)

Fig. 4. 2D projection of defects in the plane perpendicular to the building direction for (a) SLM_SR, (b) SLM_Chess, (c) SLM_Rot and (d) Orlas

The effect of defect ratio, the differences in type of defects (gas pores or lack of fusion), and the effect of defect size at comparable defect population will be studied in following parts. Furthermore, the competition between defect and microstructure for fatigue crack initiation will be particularly pointed out.