Thermal Conductivity Depth Profiling of Hardened Solids using Infrared Photothermal Radiometry Technique

Being very sensitive to heterogeneities, the thermal conductivity of materials is a quite appropriate thermophysical quantity to probe their microstructure. In this paper, we focus on hardened steel samples, in which a depth profile of varying microstructure is induced by a hardening treatment. An artificial neural network recognition approach is used to extract the thermal conductivity depth profile from infrared photothermal radiometry (PTR) signals obtained in a non-contact way on cylindrical steel rods. The quality of the reconstructed thermal conductivity depth profile is confirmed by a good correspondence between the signals calculated a posteriori from the profile and the experimental data that were fed to the neural network. The results confirm the expected anti-correlation between the thermal conductivity and the hardness, which was determined from a classical indentation procedure at different depths.