Characterization and modeling of solubility and diffusivity after JEDEC prescribed baked conditioning vs. vacuum desiccation for mold compounds

JEDEC identifies in J-STD-020C, that all devices should be baked for 24 hours minimum at 125°C +5°C/ -0°C to remove moisture from the package so that it will be "dry". This bake is required prior to other reliability stresses , again to "dry" the plastic encapsulated devices. When this preparation technique was first established, most of the mold compounds used were comprised of OCN / ECN novolac resins with glass transition temperatures (Tg) significantly greater than this 125°C bake. The mold compounds that are used today have a variety of chemistries and often have Tg regions near or below the 125°C bake temperature. This change in mold compound chemistry has resulted in two unpredicted responses to the 125°C bake. Some mold compounds will undergo additional cross linking during this conditioning (drying) bake and will result in a more fully cured matrix. Additional mold compounds that have their Tg regions at or near this 125°C bake temperature will undergo a change in their inherent equilibrium state, known as physical aging. This will result in a polymer network with a reduced free volume space between polymer chains that would ultimately result in performance behavioral changes when the material is exposed to the moisture soak conditioning, and reflow preconditioning required prior to reliability testing. This paper will describe the analytical methods for characterizing this change in equilibrium state, and explore the effect of this change in free volume space on the behavior of three mold compound systems. An alternative drying procedure utilizing vacuum desiccation will be used for comparative purposes during subsequent characterization experiments. ANSYS modeling will be used to demonstrate the moisture diffusion rate through the mold compound of the package.