PSI - Issue 7

Jean-Yves Buffiere / Procedia Structural Integrity 7 (2017) 27 – 32 ffi ere / Structural Integrity Procedia 00 (2017) 000–000

29 3

Jean-Yves Bu

a)

b)

Fig. 1. Resolution issues for fatigue crack observation with tomography (a) 2D slice of a fatigue crack in a Ti alloy (voxel size = 0.7 µ m synchrotron tomography); the crack appears in dark, the colour indicates the direction of the normal to the crack surface as shown in the stereographic triangle (b) SEM image of a fatigue crack in a di ff erent specimen of the same alloy showing deflections that are too small to be imaged by tomography(Herbig (2011))

if the latter is of much smaller size (a factor ten in equivalent diameter). In the rare cases where internal initiation and propagation were observed, the morphology and paths of the cracks were typical of a growth in a vacuum-like environment; the growth rate of this internal crack was also much lower than that of a crack initiated from the surface and growing in the same sample. From our observations and some recent results obtained for a Ti alloy, however, it is not very clear if the environment where an internal crack grows corresponds to vacuum or not (see the discussion in Chapman et al. (2015) and Yoshinaka et al. (2016) for example). Nevertheless, the fact that internal cracks propagate at lower rates than surface ones help to understand why the presence of artificial internal defects has hardly any influence on the Wo¨hler curve of a cast Al alloy, in spite of the fact that crack initiation did occur on those defects (Serrano-Munoz et al. (2016)). One can rationalize the higher probability of surface crack initiation from a defect from the reduction of the local stress level when the defect moves from the surface towards the bulk of the specimen (Borbe´ly et al. (2002)) but the global level of plasticity induced within the sample might also be important: very low levels of plasticity promoting internal initiation (an argument often evoked to explain internal crack initiation in giga cycle fatigue). At the other end of this scale (large level of plastic strain) recent observations of crack initiation and propagation during low cycle fatigue of Al alloys at 250 ◦ C tend to confirm this point as, in that case, internal initiation was found to be as frequent as surface initiation (Dezecot et al. (2016, 2017)). In order to spare the time spent to initiate a crack during synchrotron experiments and also to localise crack initia tion within a restricted field of view (typically from 1 to 3 mm depending on the voxel size) artificial defects / notches can be used. They can be produced by focused ion beam machining, or, for larger notch sizes, by electro discharge machining or femtosecond laser (see for example Lachambre et al. (2015)). In metallic alloys, cracks initiated form such defects with a size of the order of a few grains have a relatively irregular crack front and the crack roughness is of the order of the grain size. This has been reported for cubic centered (Herbig et al. (2011); Herbig (2011)), hexagonal (King et al. (2010)) or cubic face centered structures (Proudhon et al. (2012)). Both characteristics are due to the crack interaction with the local microstructure. Di ff raction contrast tomography can be used to obtain non destructively the shape and orientation of the grains forming a polycrystalline aggregate. Combined with phase contrast tomography it allows to monitor the growth of a short fatigue crack within 4. Crack initiation and propagation from artificial defects