PSI - Issue 7

Theo Persenot et al. / Procedia Structural Integrity 7 (2017) 158–165 Persenot et al. / Structural Integrity Procedia 00 (2017) 000–000

161

4

Fig. 3. a) Radial tomographic slice of an as-built sample. The red arrow indicates an unmelted powder particle stuck to the surface. Blue arrows point out at the “plate-pile” stacking defects. #1 represents a microshrinkage porosity and #2 a pore due to entrapped gas from the powder atomization. b) Example of a profile extracted from the radial slice shown in a) and used to measure the roughness.

to Tammas-Williams et al. (2015), those small pores are entrapped gas coming from the powder atomization. Larger pores correspond to microshrinkage (both types can be visualized on Figure 3). As-built EBM samples exhibit a poor surface quality (see Figure 3) which is due to the relatively coarse powder used and to the specificity of the EBM process. The samples’ roughness was measured on tomographic images fol lowing the method used by Suard et al. (2015): a series of profiles of the surface was extracted from a radial slice every 10 ◦ around the sample circumference (Figure 3). The 36 profiles obtained for each sample were then used to measure the average values of the arithmetic roughness (Ra) as well as the maximum roughness (Rt). The values obtained for the 36 profiles are shown in Figure 4. The scans resolution of tomographic images enables us to identify two di ff erents types of surface defects which are responsible for the high roughness values measured. Because of the presence of thermal gradients, unmelted powder particles are stuck to the specimen surface (red arrow on Figure 3). Also, because of the beam versatility, some variations in the regularity of the layers stacking can occur and generate “plate-pile” stacking defects (blue arrows on Figure 3) (Suard et al. (2015), Lhuissier et al. (2016)). After chemical etching, the samples are once again characterised by tomography so that the evolution of the rough ness and of the surface defects can be studied. Figure 4 shows the evolution of the specimen volume and roughness with chemical etching duration. After 45 minutes, it can be seen that the surface roughness is still decreasing. Figure 5 shows an optical micrograph of the as-built microstructure which consists in alpha phase lamellae and beta phase at the grain boundaries (respectively in white and black in figure 5) with a characteristic size ( ∼ 2 µ m thickness) smaller than that obtained by casting (Kobryn and Semiatin (2003)) and coarser than the one generated by laser processing (Facchini et al. (2010)).

4. Results and discussion

4.1. As-built results

For as-built samples manufactured with the fresh powder, the fatigue strength at 10 7 cycles is of the order of 140MPa. With the recycled powder, the fatigue limit is reduced by 20%, clearly illustrating the detrimental e ff ect of recycling powder on the cyclic properties of the material. This can be linked to the increase of oxygen content in the recycled powder (see Table 1) as suggested by Nandwana et al. (2016). The fatigue limit of as-built samples is lower than that obtained with fatigue tests performed on machined cast TA6V with the same type of microstructure as that shown on figure 5 (Oh et al. (2004)) and also lower than the value obtained for EBM specimens also machined from samples built in the same condition as our samples (370MPa in both cases). The typical aspect of the fracture surface of an as-built sample submitted to low stress levels ( σ max < 160 MPa) is shown in Figure 6 and can be divided into two distinct zones. The first one corresponds to the stable crack propagation