Bayesian Optimization for Tailoring Filler Formulations in Polymer Composite for Efficient Heat Dissipation
발표자
()
초록
내용
Thermal interface materials (TIMs) with high thermal conductivity (TC) are crucial for efficiently dissipating heat from electronics to maintain optimal operating temperatures. Achieving a polymer composite-based TIM with high TC requires establishing a continuous thermal conduction pathway through filler packing within the TIM. Therefore, using multi-sized fillers with high TC is a common strategy to enhance the composite’s TC. However, determining the optimal formulation (i.e., size and ratio) of fillers to optimize the composite TC is challenging due to the numerous sizes and wide size distributions of commercially available fillers. In this study, using a spherical alumina model filler and a PDMS model matrix, machine learning-based Bayesian optimization was employed to find the optimal filler formulation. This approach is expected to enhance heat dissipation and improve the mechanical properties of TIM.