Thermally enhanced PBGA: Effect of heat spreader design on plasma cleaning efficiency and wire sweep

In a thermally enhanced PBGA (TePBGA), heat spreader is commonly used to fulfill the heat dissipation requirement of the package. The presence of heat spreader reduces plasma cleaning efficiency and increases wire sweep due to narrower gap for plasma gas penetration and mold flow. The impact of insufficient plasma cleaning was found to be greater with certain die surface passivation which could even lead to die top delamination after reliability stressing. As for wire sweep, the issue could lead to high yield loss due to wire shorting especially for ultra fine pitch devices with wire diameter below 25um and multi-tier looping. These challenges can be overcome if the heat spreader design is carefully done and well characterized. A study was conducted on TePBGA with 4 new heat spreader designs to compare with a conventional design as control. In this study, the main response was plasma cleaning efficiency which was measured through contact angle. The 2 nd key response was wire sweep which was measured using x-ray machine. The 3 rd response was mold flow balance which was studied through molding short shots. The benchmark or target responses were those of a PBGA without heat spreader. Since the key requirement of a heat spreader is the area on top of the die surface which was needed to provide the intended heat dissipation capability during field application, hence it was not advisable to make any design modification on the heat spreader top surface. Therefore in all the 4 new heat spreader designs, the strategy was to reduce heat spreader peripheral metal volume to allow bigger gap for plasma gas penetration and a more balance mold flow. Each design varies in terms of lid foot design, with the purpose to prevent lid foot delamination, to allow sufficient support to the heat spreader to withstand the compression force during molding, and to provide the least restriction to mold flow. The study started with New Design A with 4×8mm lid foot which showed 0.43% reduction in wire sweep and 18 degree lower in contact angle which were great improvements. However, the design resulted in lid foot delamination due to big lid foot area that caused air trapped under the lid foot. This had lead to New Design B with 4×2mm lid foot which showed greater improvement, with 1.3% reduction in wire sweep and 26 degree lower in contact angle. However, the design suffered mold compound bleeding on top of heat spreader surface due to insufficient lid foot support to withstand the co