PSI - Issue 68

A. Barabi et al. / Procedia Structural Integrity 68 (2025) 285–291 A. Barabi / Structural Integrity Procedia 00 (2025) 000–000

288

4

Upon completion of the tests, the CT specimens were fractured to observe the crack surfaces. The surface area covered by corrosion products was examined with an optical microscope (Keyence VHX-7000). Image analysis was used to quantify the area fraction of corrosion products. Pixels with an RGB (red, green and blue) value exceeding 110 in the blue channel were identified as rust. X-ray photoelectron spectroscopy (XPS) was performed on the fracture surfaces to further analyze the oxide layer. An area 100 µm diameter was examined with a ULVAC-PHI Versa Probe III, using a monochromatic aluminum source with radiation at 1486.6 eV and a power of 25 W. The carbon peak at 284.8 eV was used for peak calibration. Peak identification and deconvolution were performed with CasaXPS software, employing a Shirley background. A Gaussian-Lorentzian line shape was applied to model chromium ( Cr ) and iron ( Fe ) peaks, while oxygen ( O ) peaks were fitted with a Lorentzian asymmetric line shape. Thermodynamic calculations were conducted using FactSage software to plot a Pourbaix diagram ( E vs pH ) for thermodynamic prediction of corrosion products formed on the fractured surface. 3. Results and discussion Figure 1a shows FCGR and OCP measured in both air and water (CFCGR) as a function of f . CFCGR increases significantly with a decrease in f . No such frequency dependency of the FCGR is observed for tests conducted in air. The red data points indicate the OCP measured near the crack tip. A higher OCP is observed at lower f , with a decreasing trend as f increases. As for the Figure 1b, sections of the CT sample subjected to different frequencies, delineated by black dashed lines, are highlighted with their corresponding rust fractions. Unlike the CFCGR and OCP, the rust fraction increases with rising f , showing a direct correlation. Higher frequencies result in greater rust formation. Based on the OCP measurements and rust fraction, it can be inferred that the crack tip is more susceptible to anodic dissolution at higher frequencies, as a lower OCP correlates with increased electrochemical activity.

Figure 1 (a) FCGR measured in the air and water at different f plotted as the main axis and on the secondary axis the OCP measured during the CFCG is plotted, and (b) the optical observation of the fractured surface of the sample tested in the water along with the measured oxide surface area fraction. Interestingly, the crack propagates at a faster rate in water compared with the results in air at low f . The combination of these results indicates that an additional mechanism contributes to the acceleration of the crack growth kinetic. The additional mechanism apparently does not involve the formation of rust. (Lou et al., 2017) and (Seifert et al., 2015) also observed that less corrosion product was formed on the fracture surface of an austenitic stainless steel when f was reduced. Their tests were nevertheless performed in boiling water and no explanation were supporting these observations. For further explanation of the results described above, pH at the crack tip is estimated by combining OCP measurements and the chemistry of the corrosion product. The results of XPS analysis performed on segments of the fracture surface tested at frequencies of 10 Hz and 0.1 Hz are shown in Figure 2. At both frequencies, they are mainly composed of Fe 2 O 3 and FeCr 2 O 4 as corrosion product. However, at 0.1 Hz their fraction is less than 10 Hz and there is a higher fraction of Cr₂O₃ (passive film) at 0.1 Hz. This finding supports the lower rust fraction covered by oxide at 0.1 Hz. The pourbaix diagram thermodynamically predict the stable corrosion products at each pH and E for a specific alloy in a solution. Therefore, with pourbaix diagram along with the measured E , the solution pH can be estimated.