Identification of key influencing factors and failure prediction for corrosion in sensitive interconnection structures of electronic devices

This is the authors' abstract. We don't add key points for this paper.

To address the critical issue of sudden short circuits caused by electrochemical migration in electronic products operating under complex environmental conditions, this study focuses on SAC305 solder joints. By integrating failure physics and machine learning methodologies, the stage-dominant mechanisms underlying their failure are elucidated. The research reveals that water vapor partial pressure governs the condensation and incubation stage, while the electric field strength dominates the subsequent ion migration and dendritic growth processes. Building on these insights, a physics-informed prediction model based on a stage-partitioned modeling and reintegration approach is developed. This model not only enhances prediction accuracy but also ensures strong physical interpretability. The study identifies environmental humidity and wiring spacing as the key controlling factors, thereby providing a quantitative theoretical foundation and a practical predictive tool for the design of highly reliable electronic packaging.