Effective Substrate Thermal Conductivity Modeling Method Extracted from Detailed Pattern for Premium SOC Packages

As the semiconductor size becomes smaller, it is important to improve the temperature prediction accuracy. In particular, the accuracy of temperature prediction in complex heat spreading environment (e.g. PCB pattern) is closely related to the overall simulation accuracy. In this study, an orthogonal K
eff
prediction model is proposed that reflects the 3-dimensional thermal behavior of PCB/substrate using detail pattern simulation results. A large number of sample simulations (n = 2700) are performed using the detail simulation tool to formulate a K
eff
prediction model that cannot be identified from a single PCB/substrate analysis. According to the metal continuity of each sample, samples are classified into three types and K
eff
prediction model is proposed for each category. In order to verify the simulation accuracy of the newly proposed model, numerical verification including the existing analytical model is conducted, and the predictive capability of the model is confirmed. Further analysis is required to improve the accuracy of the model and it is expected to be used as the basis for K
eff
prediction model through machine learning.