Evaluation of image reconstruction algorithms for non-destructive characterization of thermal interfaces

Abstract

Thermal interfaces are encountered in many thermal management applications and interface materials are used to minimize thermal contact resistance resulting from solid–solid contact. For opto-electronic devices the quality of the thermal interface is critical for removing the generated heat for proper thermal management. Defects in the thermal interface introduce additional thermal resistance in the thermal path, and must be prevented. Detection of defects in the thermal interfaces becomes critical during the assembly process development. Imaging techniques such as X-ray computerized tomography, or scanning acoustic microscopy that require expensive equipment and significant processing time is necessary. Thermal tomography in conjunction to IR thermometry can be used as a lower cost alternative to these techniques. The feasibility of thermal tomography for non-destructive characterization of thermal interfaces is presented by considering different image reconstruction algorithms. The algorithms considered are the iterative perturbation algorithm, Levenberg–Marquardt algorithm and the regularized Newton–Gauss algorithm, and they were found to be capable of characterizing the thermal interface layer.