Isothermal cure study of epoxy mold compounds
Abstract
Capacity improvement in implementing shorter cure time during mold process has driven mold engineering to conduct optimization studies on manufacturability and reliability of a 40-second cure time to replace what used to be a 50-second cure time held isothermally at 180°C on different mold compounds. From capacity studies, the 10-second process improvement is translated to a capacity improvement of 15,520 units per shift or 46,560 units per day or an overall improvement of 325,920 units per week. In this paper, an isothermal cure study based on an autocatalytic model for epoxy resins was applied on various resin formulations: A – fast cure ortho-cresol novolac (OCN); B – slow cure OCN; C – dicyclopentadienyl (DCPD); and D – biphenyl (BP) using differential scanning calorimetry. Isothermal conversion curves were plotted against time for various mold temperatures to predict degree of cure of mold compound at specified isothermal temperature and time. Isothermal kinetics on slow cure OCN-based compound B indicated low conversion of 59.08% and 78.30% at 40s and 50s cure necessitating post-mold cure process or PMC. Samples A, C, D yielded higher degree of cure (85.30%, 83.97%, 85.13% and 94.48%, 86.92% and 96.55%) at 40s and 50s respectively making them good candidates for non-PMC system. Cost consideration eliminated biphenyl compound (sample D) despite high conversion percent. Confidence lots for samples A and C were ran and checked for quality attributes (sticking, pitted surface, foreign matter, blisters and bubbles, resin bleed, chips, wiresweep, mold flash, incomplete encapsulation and unmolded units). No defective unit was noted. Reliability testing for moisture sensitivity level 1 (MSL-1) was performed on SOIC 8L devices (Device# 34072, LM317, and TL431) at reflow temperatures of 220°C and 260°C with oil immersion for autoclave and temperature cycling tests respectively. Package cracking, delamination and electrical performance were confirmed. No failure was observed.
