Simulation driven physics-of-failure analysis for System-in-Package development

System-in-Package (SiP) which combines different chips and technologies into a single package is a viable solution to meet the rigorous requirements for today's mixed signal system integration. As the level of integration increases, challenges related to product manufacturability and reliability also increases. As a result, design for reliability using CAE (Computer-Aided-Engineering) or FEM (Finite Element Method) simulation is becoming an effective tool in recent years to reduce the development time, cost and manpower. A case study with simulation driven failure analysis during a SiP module development for RF transceiver is demonstrated. Due to structure complexity (27 components) and lack of detailed information from suppliers, some components are simplified in the global package simulation model. Based on simplified global model, design guidelines are provided for package geometry and material selection, which are validated during package qualification for package warpage control and overall reliability, except failure related SAW filters after TC (thermal cycling). With details shared from supplier, SAW filter is actually a small cavity package. First simulation trial focused on solder bump fatigue life estimation showed weak correlation with experimental data, which imply this is not a normal fatigue related issue (although failure seen after TC), and efforts for failure analysis (FA) should also be put on other factors affecting pre-mature failure, e.g. assembly process parameters. Then from substrate mapping, it is identified that failure units mostly located at mold vent side, not the gate side. Further process DOE trials highlighted that lower transfer pressure leads to higher failure rate. With a close collaboration among project leader, process engineers and suppliers, and a creative data digging on experimental results, final relevant model has been built up and inherent mechanism proposed, which linked all the experimental data observed. Based on the clear physics-of-failure analysis, solution and future development guideline provided, and the related product qualified.