Thermal-Mechanical Optimization of 2.5D Flip-Chip Packages With Glass and Silicon Interposers (Univ. of Ottawa)

Researchers at the University of Ottawa have published a technical paper on the thermal-mechanical optimization of 2.5D flip-chip packages using silicon and glass interposers. The paper introduces a co-design framework that combines finite element analysis (FEA) with machine learning to address trade-offs in these packages. The study focuses on minimizing thermal resistance and plastic strain-energy density at various points in the package. By using machine learning models like Random Forest and XGBoost, the researchers were able to optimize the design for balanced thermo-mechanical performance. The paper highlights the importance of decoupling heat-path optimization from interconnect reliability in these advanced packages.