Hua Yuchao
Nantes Université, Laboratoire de thermique et énergie de Nantes, LTeN, F-44000 Nantes, France.
J Phys Condens Matter. 2025 Jul 21;37(29). doi: 10.1088/1361-648X/adeef2.
Embedded cooling has been regarded as an important innovation to address the thermal challenge of GaN power devices, which integrates fluidic cooling structures directly into substrate to reduce the hotspot-coolant distance. Nevertheless, the idea that improving thermal performance with decreasing hotspot-coolant distance is valid merely when serial heat conduction resistance dominates. In fact, the heat sources within GaN devices are highly-concentrated, leading to strong heat spreading resistance that increases with thickness decreasing. In this case, a fundamental question can emerge:? Regarding this issue, we conduct the optimization of hotspot-coolant distances for the GaN devices of a variety of configuration parameters by using a cost-effective approach to calculate the overall thermal resistance, with combining the analytical heat spreading model, size-dependent effective thermal conductivity model, and effective thermal boundary resistance based on diffuse mismatch model. Analysis of variance method is then used to specify the key parameters of significance that largely determine the optimal hotspot-coolant distance. A correlation is fitted in terms of those optimization results, which can server as a fast and convenient means to estimate the optimal distance between the hotspot and coolant for GaN devices based on the key parameters.