PSI - Issue 68

R.J. Mostert et al. / Procedia Structural Integrity 68 (2025) 351–357 R.J. Mostert et al. / Structural Integrity Procedia 00 (2025) 000–000

356

6

For the AA 2198-T8 alloy, surface cracking after EXCO exposure was quantified for all levels of pre-stretching. A travelling microscope was used, and crack widths were determined. The crack width density, determined over the first 12 mm from the fracture plane and over the full gauge length width (assessment area = 144 mm 2 ), is expressed as a function of the percentage of pre-strain in Fig. 9. Accordingly, pre-stretching by 1.5 % resulted in approximately 64 % increase of the crack width density. At higher levels of pre-stretching, the crack width density dropped off again, so that pre-stretching by 7.5 % resulted in a similar crack width density compared to the as-received sample.

Fig. 9. (a) Crack width densities (Σ of crack widths divided by assessment area) for the four levels of pre-stretching. (b) Correspondence of the trends of tensile ductility degradation and crack width density as a function of pre-stretching level. The photographic insert demonstrates the measurement of an individual crack width. 4. Discussion The fractography demonstrated that the EXCO exposure resulted in the attack of the sub-surface grain boundaries, up to a depth of ~ 100 micron. Following exposure, the grain boundaries were rich in oxygen, indicating preferential attack. Due to the IGC, a small applied strain (2%), was sufficient to pull the grain boundaries apart, resulting in the observed shallow IG crack formation. These observations, together with the fact that the degassing heat treatment did not prevent the formation of these cracks, lead to the conclusion that the shallow surface cracking observed was due to IGC. The network of sub-surface grain boundary attack is believed to accelerate hydrogen diffusion into the matrix, due to the presence of a pathway facilitating the transport of H atoms into the matrix. The bulk H-embrittlement observed (Fig. 2 and Fig. 3) is therefore believed to be partly facilitated by the presence of the surface layer of degraded grain boundaries. The influence of the pre-stretching percentages on the AA2198-T8 surface embrittlement and hence, crack width densities, is shown in Fig. 9 (a), with a peak surface embrittlement after 1.5 % pre-stretching. This trend closely follows the trends of total elongation decrease and surface crack width density versus pre-stretching levels, as shown in Fig. 9 (b). The peak elongation degradation at 1.5 % pre-stretching level, corresponds to the peak in secondary crack-width density at the same pre-stretching level. This correspondence translates to a peak in the number of degraded grain boundaries present immediately after EXCO exposure which can be expected to peak at 1.5 % pre stretching. With an increased number of degraded grain boundaries, it is probable that the hydrogen diffusivity will also be at a peak at 1.5 % pre-stretching, explaining the peak in ductility degradation at this level of pre-stretching. 5. Conclusions • The shallow surface cracking observed upon straining EXCO-exposed AA 2024-T3 and AA 2198-T8 tensile and C(T) samples, is due to the intergranular attack of a shallow layer of the alloy during the two hours of EXCO exposure. Application of strain at levels of 2% and higher, results in localized fracture along the damaged grain boundaries, forming the shallow IG cracks observed.