Development of a Jet Impingement Thermal Management System for a Semiconductor Device with the Implementation of Dielectric Nanofluids

Abstract

This study investigates the viability of a liquid jet-impingement (JI) thermal management system (TMS). The JI-TMS was used as a heat dissipation tactic to ensure a silicon semiconductor-based pulse-power devices temperature was maintained below 80 ◦Cduring operation. The JI-TMS’s design, which is comprised of 27 nozzles (1 mm nozzle diameter), is analyzed by computational fluid dynamics (CFD) modeling, with 100 W ofaverage power (i.e., heat generation rate). The heat transfer fluid (HTF) is pure silicone fluid with a viscosity of 20 centiStokes. The selected HTF has the dielectric constant which provides better electrical insulation than other common HTFs (water, water-glycol, etc.). Validation of the model was performed experimentally to ensure the accuracy of the CFD results. The results show that devices temperature is maintained well below the maximum desired operating temperature (57.5 ◦C when using 0.06 kg s), showing average deviation of 5.89% and 0.99% between numerical and experimental data for the 100 W and 200 W average power tests, respectively. However, the predicted temperatures by the CFD modeling are somewhat lower than the experimental values. To better understand the achieved error, heat capacity analysis of the HTF was investigated with a differential scanning calorimeter which shows an average deviation of 13.12% with the reported vendor value. To increase the performance of the system, 18 nanofluid samples were created using various nanoparticles and surfactants. The nanolfuids were created with the goal of improving the thermal conductivity of the HTF. The samples were tested using a DSC and the specific heat capacity of each sample was found. A 0.25 wt% mixture of silicon carbide with surfactants was found to have the most significant heat capacity reduction which correlated with the greatest thermal conductivity increase of all the sample. From the heat capacity measurements, the thermal conductivity was inferred. Using the new thermophysical property values, CFD models were run. The nanofluids showed a 12.6% 13.3% maximum PCSS temperature reduction for the 100 W and 200 W average power simulation, respectively. While the 200 W average power simulation still showed the PCSS exceeded the maximum threshold temperature of 80◦C, the significant improvement in temperature reduction of the PCSS shows that nanofluids should be considered for future TMS designs.