At first, the popcorn cracking phenomena in plastic IC packages during reflow soldering are investigated by considering the heat transfer and moisture diffusion through the epoxy molding compound (EMC) along with the mechanics of interface delamination by assuming an inherent edge crack at the die pad/EMC interface and subsequent interface delamination under thermal and vapor pressure loadings. Heat transfer and moisture diffusion through EMC under die pad are analyzed by finite difference method during the preconditioning and subsequent reflow soldering process, and the amounts of moisture mass and vapor pressure at delaminated die pad/EMC interface are calculated as a function of the reflow soldering time. Using the finite element method and the methods of interface fracture mechanics, path independent energy release rate, stress intensity factor and phase angle are calculated and compared to the interface toughness which is assumed to be a function of the phase angle. Also, the effects of (i) thermal loading, (ii) crack face vapor pressure loading and (iii) mixed loading of both are estimated separately. The edge crack propagates toward the center leading to the delamination of the entire die pad/EMC interface most notably for the vapor pressure loading when thermal and vapor pressure loadings are applied simultaneously. The energy release rate increases parabolically with the crack length but proportionally with the vapor pressure while the interface toughness decreases with the crack length. Stress states near the crack tip were closer to mode II for thermal loading but closer to mode I for vapor pressure loading, and changed from mode II to mode I with the crack length for the mixed loading. Finally it was shown that thermal loading was the main driving force for crack propagation for small crack lengths, but vapor pressure loading played more significant role as the crack extended.
Secondly, the interfacial fracture energies of the Cr/polyimide interface were deduced for a Cu/Cr/polyimide/Cr/alumina structure under varying Cu film thickness and polyimide pretreatment conditions based on two methods: X-ray measurement and theoretical methods. The two methods showed reasonable agreement for most cases, rendering validity to both approaches. Estimated interfacial fracture energies were quite independent of the metal film thickness and increased with the rf plasma power density of polyimide pretreatment as expected. Estimated values of interfacial fracture energies were 46.8±17.8, 170.3±42.9 and 253.9±44.4 J/㎡ for the rf plasma power density of 0.03, 0.036 and 0.05 W/㎠, respectively. Then, the fracture mechanics solutions of the phase angle for the film/substrate structure were extended to the Cu/Cr/polyimide/Cr/alumina structure by assuming the path independence of J-integral, which turned out to be quite true with or without interlayers. Calculated phase angles were virtually independent of the metal film thickness, and the presence of the polyimide layer increased the mode mixity by shifting the phase angle 11˚~12˚, which indicates that ignoring the polyimide interlayer may result in substantial underestimation of the real phase angle of multilayer films under the peel test.