Application of AI-enabled Simulation in Power Package Development

The design of power packages with both high efficiency and high-power density performance while maintaining the highest possible reliability is a challenge. During development for power package, design of experiment (DOE) by simulation can effectively shorten the design cycle, reduce costs, and effectively optimize the packaging structure. However, the simulation analysis results are highly dependent on the individual researcher and are only part of overview of factor. Especially, as the package size continuously decrease for high-density designs, there are lots of challenge areas, like thin die pick-up process, material optimization to enhance process and reliability performance. Maching learning (ML) within the field of Artificial Intelligence (AI) simulation can be very effective tools to help address these challenges.In this paper, application of ML algorithms and enabled simulation for die pick-up process and reliability testing related to power device are discussed. For thin die pick-up process, Bayesian optimization was used to optimize the parameters to achieve the lowest die stress under a given condition. An automatic molding compound selection framework is proposed to generate the optimal material properties of epoxy molding compound (EMC) properties for solder joint reliability. Through these cases, it has been demonstrated that the AI-enabled simulation can be utilized to significantly improve parameter optimization and material selection for robust power device development.