Conditions for retrieval of thermal conductivity profiles for surface hardened steel by radiometric measurements are discussed. Two different inversion procedures are applied to the same experimental data: a polygonal approximation and a neural network technique.

The experimentally verified connection between microhardness and local thermal conductivity enables the photothermal hardness depth profiling. This can be done by estimating the thermal conductivity depth profile from frequency dependent radiometric measurements. The accuracy of the retrieval depends strongly on signal-to-noise ratio. According to the underlying theoretical model the generation of plane thermal waves probing the material under test is required. A numerical inversion technique is presented to approximate the thermal conductivity profile by a set of linens slopes. Measurements are performed at case and laser hardened specimen to verify the feasibility of photothermal hardness depth profiling.

Surface problems in photothermal depth profiling of technical surfaces arise from the intrinsic optical absorptivity; therefore, it is not appropriate to use amplitude values for profile reconstruction. We report a reconstruction technique based only on phase data. The performance of this approach was tested successfully by measuring the thermal conductivity profiles of laser hardened steel samples.

Photothermal measurements at varied modulation frequencies allow for the estimation of depth profiles of thermal diffusivity or thermal conductivity in case of known thermal density. Knowledge of thermal conductivity k is important because of its correlation to relevant microstructural properties of the samples under test as for instance mechanical hardness of steel. We suggest an inversion algorithm performing a sequence of one-parameter fits in order to estimate k profiles. Surface problems in photothermal depth profiling and a approach to overcome them are discussed. Experimental results of thermal conductivity depth profiling of laser-hardened steel, based on this inversion algorithm are presented. The photothermally obtained k-profiles are interpreted and compared with results from conventional destructive measuring techniques.

An analytical solution for the photothermally measurable surface temperature of a sample with piecewise linearly inhomogeneous depth profile of the thermal conductivity k is presented. Based on this solution an inversion algorithm using a sequence of one-parameter fits is suggested in order to estimate the k profile. Numerical simulations demonstrate the performance of the approach and its insensitivity to random errors.

Experimental results of thermal conductivity depth profiling of case-hardened steel and mechanically loaded ceramics are presented, based on an inversion algorithm for a sample with piecewise linearly inhomogeneous depth profile of the thermal conductivity. The photothermally obtained profiles are interpreted or compared with results from conventional or destructive measuring techniques.