Superior Drop Test Performance of BGA Assembly Using SAC105Ti Solder Sphere

Abstract

Board-level drop test performance was evaluated and compared for the following four different solder combinations in BGA/CSP assembly: 1) SnPb paste with SnPb balls, 2) SnPb paste with SAC105Ti balls, 3) SAC305 paste with SAC105Ti balls, and 4) SAC305 paste with SAC105 balls. Presence of Ti improved the drop test performance significantly, despite the voiding side effect caused by its oxidation tendency. It is anticipated that the voiding can be prevented with the development of a more oxidation resistant flux. The consistently poor drop test performance of 105Ti/SnPb is caused by the wide pasty range resulted from mixing SAC105 with Sn63 solder paste. The effect of Ti in this system is overshadowed by the high voiding outcome due to this wide pasty range material. In view of this, use of SAC105 BGA with SnPb solder paste is not recommended, with or without Ti addition. High reflow temperature drove fracture shift to interface at package side, presumably through building up IMC thickness beyond the threshold value. A lower reflow temperature is recommended. Electrical response is consistent with complete fracture data. But, complete fracture trend is inconsistent with that of partial fracture trend, and neither data can provide a full understanding about the failure mode. By integrating complete fracture and partial fracture into “Virtual Fracture”, the failure mechanism becomes obvious and data sets become consistent with each other.