High Heat Fluxes Analysis Using Microfluidic Cooling

Abstract

The exponential development in computational systems in the last decades along with new emerging 2.5 and 3D integrated circuits (ICs) geometries allow to overcome some limitations from conventional computers. Due to the small dimensions and the number of micro-circuits contained in new micro-chip geometries, overheating is a significant issue that needs to be solved. Conventional cooling methods are not enough to keep acceptable temperatures on these 3D IC systems. A potential solution is to develop alternatives cooling methods like micro embedded pin fins with pin-fin density variation for hotspot zones. However, experiments to analyze this potential solution have high costs due to manufacturing; hence, numeric simulation is a convenient tool to analyze and to develop different microfluidic cooling models for 3D IC. The present work shows a numerical methodology for microfluidic cooling, using micro embedded pin-fins on a thin silicon plate. A parametric analysis was performed, using the theory of design of experiments (DoE). The proposed factorial design considers 64 numerical microfluidic cooling models. The pressure drop and the average temperatures were obtained for each model to determine the importance of input parameters utilizing a statistical approach. Through the suggestions made, useful for future experimental and numerical works, some optimized geometries are proposed using statistic tools. Promissory microfluidic cooling systems and conclusions are reported in this thesis work.