A new capability to model both crack initiation and growth in eutectic Sn-Pb solder joints was developed and implemented into finite element analysis codes. Two significant developments were needed to create this new capability. First, an ability to accelerate the simulations such that the effects of hundreds or thousands of thermal cycles could be modeled in a reasonable amount of time was needed. This was accomplished by applying a user prescribed acceleration factor to specific terms in the solder model’s damage evolution equation; then, the damage generated by an acceleration factor of cycles could be captured by the numerical simulation of a single thermal cycle. Second, an ability to capture the geometric effects of crack initiation and growth was needed. This was accomplished by replacing material in finite elements that had met the cracking failure criterion with very flexible elastic material. This diffuse crack modeling approach with local finite elements is known to generate mesh-dependent solutions. However, mesh refinement studies revealed that for thermal mechanical fatigue simulations, the mesh dependency is small and has a small effect on predictions for cycles to generate an electrical open. The new crack modeling approach will be described. Finally, crack predictions are compared with experimental observations.

This content is only available via PDF.
You do not currently have access to this content.