PSI - Issue 66

Johannes L. Otto et al. / Procedia Structural Integrity 66 (2024) 256–264

261

6

Johannes L. Otto et al. / Structural Integrity Procedia 00 (2025) 000–000

shifted at all stress amplitudes to a lower number of cycles. In this case, the NaCl and the acids caused together with the plastic and elastic deformation a damage of the passive layer and local anodic polarizations occurred. However, the cracking of the brittle Cr-rich borides in the diffusion zone will have led to crevice corrosion in addition to the microstructural notch effect that was already present in the test series before. The formation of hydrogen and its embrittling influence due to acidification at the crack tip must also be considered, as described by Bland and Locke (2017), especially when it comes to strain-induced martensite formation of the metastable austenite. To understand whether the acids or the NaCl played a more important role, a one-molar NaCl solution was used in a further series of test. The NaCl content was higher by a factor of 35 than in the synthetic exhaust gas condensate and it did not contain any acids. It was found that there was no change in fatigue life compared to the synthetic exhaust gas condensate. Therefore, the small amounts of acids appear to have hardly any influence, while even a small amount of chloride ions appears to be sufficient to disrupt the passive layer and promote critical crevice corrosion. In the last series of experiments, the synthetic exhaust gas condensate was used again, but at 80°C and with a pressure of 0.5 bar (p e ). This is a more common industrial application environment, even if such high stress amplitudes rarely occur. The tests showed a further significant drop in fatigue life, which at 350 MPa was only about one-tenth of the number of cycles to failure compared to air and pure water environments. In this environment, corrosion products could be visually identified after the test. A significant difference compared to the previous tests was that only a few specimens failed in the brazing seam, with multiple cracks forming all over the specimen surface in the testing area, which indicates stress corrosion cracking, Fujii et al. (2024). All SN-curves are shown in Figure 5.

Fig. 5. SN-curves of the brazed specimens for fatigue and corrosion fatigue loading at different environments.

Using the three-electrode setup and a potentiostat, the electrochemical open circuit potential (E OCP ) was measured during the tests in synthetic exhaust gas condensate and in the one-molar NaCl solution for most of the tests. It provides information about the state of the passive layer and crack initiation, since it drops when non-passivated material is exposed to the surrounding medium and dissolved metal ions influence the conductivity of the electrolyte, Nazarov et al. (2019). By normalizing the fatigue life, it was possible to compare the curves for different stress amplitudes, which then have a very similar course, as can be seen in Figures 6a) and 6b). The potential in the NaCl solution is lower due to the better conductivity. At the beginning of the fatigue tests, the high deformation in the first cycles shows a clear drop in the potential due to the rupture of the passive layer. Subsequently, repassivation occurs, which causes the potential to stabilize at around 0 mV in the exhaust gas condensate and at around -100 mV in the NaCl solution. This repassivation appears to be faster in the NaCl solution. Scatter in the further course can be explained by the formation of intrusions/extrusions and repassivation processes. After 60% to 80% of the fatigue life, a decrease in the potential sets in for all specimens. This marks the formation of microcracks and an increase in the corrosion processes. After 90% of the fatigue life, there is a strong drop in potential, which can be explained by the formation of macro-cracks.