Quantitative non-destructive characterization of defects in electronic packages using fuzzy inference based thermal tomography

Abstract

Thermal challenges have been a roadblock for electronic packaging with the in creasing number of transistors and smaller package sizes. Thermal interface material (TIM) layers play a key role in heat dissipation at all levels within an electronic pack age. The function of TIM is to minimize the thermal contact resistance by filling the microscale gaps between the die and the integrated heat spreader (IHS). For this reason, there have been intensive efforts to achieve a high-quality TIM in electronic packaging. Defects in TIM layers must be identified during the assembly process development to obtain dependable thermal management. Non-destructive characterization techniques such as scanning acoustic microscopy or X-ray tomography have been used to identify such defects and help to advance manufacturing procedures. Thermal tomography is proposed as a low-cost alternative to these qualitative imaging techniques, all of which require high-cost devices and a long processing time. The location and size of defects are identified by evaluating the measured thermal response of a heated electronic pack age. Fuzzy inference method (FIM) is used as an image reconstruction algorithm to solve the resulting ill-posed inverse problem. The feasibility of thermal tomography is studied numerically; therefore, simulated measurements are used in this study rather than experimental data. The results indicate that the fuzzy inference based thermal tomography can be a powerful tool for quantitative non-destructive characterization of defects in the electronic packages, with less cost and shorter processing time than other established methods.