PSI - Issue 51

Nils Wegner et al. / Procedia Structural Integrity 51 (2023) 122–128 N. Wegner et al. / Structural Integrity Procedia 00 (2022) 000–000

126

5

Exemplary µCT images of the macroscopic corrosion morphologies are shown in Fig. 3. Images a)–c) represent corrosion at OCP, whereas images d)–f) show conditions after GSP. The majority of the WE43 specimens without polarization exhibit homogeneous corrosive attack (a), with deep, branched pitting occurring in individual cases (c). On the other hand, PEO specimens do not exhibit any localized corrosion; only a slight decrease in layer thickness and minor damage within the layer are evident (b). Under polarization, all specimens, regardless of the surface condition, show increased local corrosion. In longitudinal section (f), pitting is arranged parallel. In longitudinal and cross sections, it can be seen that the pitting propagates along the accumulation of the alloying elements (lighter areas). As an example, the results of a MAT on a PEO modified specimen in the initial condition and after GSP are shown in Fig. 4 and Fig. 5 a). Plotted are the stress amplitude σ a as controlled variable and both hysteresis characteristics (ε a,p , w L ) as well as the temperature change ∆T as measured variables over the number of cycles N. For the initial condition, the course of the three measured variables is similar: Initially, they take a constant or slightly decreasing course before changing to a linear and then to an exponential increase (until failure). The specimen fails after N f = 17.4∙10 4 at a stress amplitude of σ a,f = 224 MPa. In contrast, the specimen pre-corroded by polarization fails at σ a,f = 51 MPa. After polarization, the measured variables assume a flatter course with lower values. Before the specimen fails, a slight exponential increase in ε a,p and w L can be seen. Using this increase, the residual fatigue strengths σ a,fs,e are estimated and plotted in Fig. 5 b) against the immersion time t in weeks for OCP and in days for GSP. In the initial condition, the estimated fatigue strength of the unmodified specimens is slightly higher than the PEO modified specimens (152.8 to 142.5 MPa). With increasing immersion time at OCP, the residual fatigue strength of both materials decreases, with lower values estimated for WE43. For GSP, a drastic decrease in estimated residual fatigue strength occurs after one day of immersion, followed by a moderate decrease after day 2 and 3. Differences between both variants are barely noticeable. The estimated fatigue strengths of the test strategy with GSP are significantly lower than those of the tests at OCP. 4. Discussion The immersion tests with GSP show that the hydrogen volume assumes a linear course for a constant current density. The progressive course at the beginning of each day is explained by the formation of hydrogen bubbles on the surface, which only detach from the surface and can be detected once they reach a critical size (Yang et al., 2016). According to previous investigations (Wegner and Walther, 2021) and the literature (Shi et al., 2014), the corrosion rate is proportional to the current density, resulting in a change in HER with changing current density. The lower hydrogen volumes for PEO modified specimens correspond to the lower calculated current densities (Table 1). On this basis, the tests show that a qualitative control of HER is possible by applying polarization. The increased corrosion rates compared to the calculation function (Fig. 2 a) can be justified by the respective characteristics of the microstructure. By using embedded samples (2D) to generate the calculation function, the corrosion behavior of the cross section was investigated. As a result of using cylindrical specimens (3D) for the present tests, the longitudinal section of the microstructure dominates the material behavior. In this case, elongated precipitates with increased Volta potentials (Coy et al., 2010) are present due to the extrusion process, favoring pitting corrosion and leading to an increased corrosion rate. For immersion tests at OCP, the occurrence of pitting corrosion is subjected to a statistical probability (Ascencio et al., 2014) and can only be observed for individual specimens and, in particular, for longer immersion time. As a result of GSP, this effect is enhanced (Thomas et al., 2015), and pronounced pitting corrosion occurs along the precipitates, which can be observed for all specimens. According to the changed corrosion mechanisms, differences in fatigue behavior result. Due to the intensive pre-corrosion under polarization, there is significant embrittlement of the material, reflected in lower values of hysteresis characteristics (ε a,p , w L in Fig. 5 b). According to Klein et al. (2017), the first exponential increase estimates the (residual) fatigue strength. Compared to the substrate material, the PEO modification leads to slightly lower fatigue strengths, resulting from the porous and brittle nature of the oxide ceramic and the associated premature crack initiation (Klein et al., 2017). Due to the higher corrosion rates at OCP for WE43 (Table 1), the residual fatigue strengths decrease significantly compared to the PEO modification. Furthermore, the coating leads to a more uniform corrosion morphology (Fig. 3 b), favoring increased fatigue strength. As a result of GSP and the associated increased pitting tendency, there is a drastic decrease in estimated residual fatigue strength after one day of immersion. Due to the moderate decrease after day 2 and 3, respectively, it is assumed that the number and size of pitting have reached a critical value after one day and hardly