Home

Home

Multi-Objective Optimization Algorithm for TSV-Embedded Microfluidic Cooling

2026.09.22 13:52 Editor:刘昊旻

High heat flux and local thermal stress have long been critical factors affecting the performance of three-dimensional integrated circuits (3D ICs). In this study, a fluid–solid coupled model incorporating embedded through-silicon via (TSV) structures is developed to design an enhanced microfluidic cooling system. Using computational fluid dynamics (CFD) simulations, the thermal performance of various TSV array configurations, microchannel cross-sections, flow channel topologies, and inlet designs is systematically analyzed. The results show that a variable-density microchannel layout employing wavy and cavity structures achieves a temperature reduction of approximately 15 K while maintaining a pressure drop below 20,000 Pa. The cavity-type microchannels are further optimized using response surface methodology (RSM) combined with a multi-objective genetic algorithm, yielding a performance enhancement coefficient of 1.12. In addition, the thermal stress distribution within the embedded TSV microfluidic cooling structure is examined. Heat source power and TSV current density are identified as the primary factors inducing stress concentration. The study further demonstrates that increasing the TSV array spacing and placing TSVs away from heat-generating chips effectively mitigate this issue. The developed numerical model provides a quantitative benchmark for optimizing microfluidic cooling in 3D ICs and lays a foundation for alleviating thermal challenges and improving overall system reliability. The results were published in Applied Thermal Engineering under the title “Study on microfluidic heat dissipation enhancement and thermal stress analysis in three-dimensional integrated circuit with through silicon via structures”.

School Email: weidianzi@nwpu.edu.cn

School Tel: 029-88430774

Address: Northwestern Polytechnical University, No. 127 Youyi West Road, Xi'an


集成电路学院(微电子学院)版权所有 2021-2022