Multiscale characterization of pitting corrosion and application to an aluminum alloy

This paper documents a novel method for characterizing pitting corrosion damage in structural materials such as Al 2024-T3. Specimens of such alloys are corroded in a controlled environment and the pits' geometry is captured digitally using white light interference microscopy. The digital data are then processed with wavelet-based analysis, thus making possible a multi-resolution description of the geometrical features. The analysis reveals several interesting features of the pits that are similar for all the experimental data analyzed herein, and independent of the process followed for creating them (time the material is exposed to corroding environment, concentration of the corroding agent, surface area exposed to the agent, etc.). The first property identified as common to all pits is their geometrical scaling with a (Hurst) exponent of 0.63 +/- 0.12. Furthermore, the ratio omega of the surface area of the pit as represented at coarse scales through the wavelet representation, over the area of its intersection with the plane at zero depth is found to be 1.17 +/- 0.07 consistently. The ratio of the total surface area over its intersection is found to be 1.6 +/- 0.2. Either one of these ratios together with the Hurst exponent provide sufficient information for obtaining a pit's geometry from images capturing its two-dimensional shape only, a capability important for efficient characterization. Additionally, such a characterization is paramount for rigorously addressing fatigue crack initiation and propagation. (C) 2000 Editions scientifiques et medicales Elsevier SAS.