The Challenge of Lower Temperature Soldering for Large Ball-Grid Array (BGA) Board-Level Assembly – Phase 1. Processing

Abstract

Low temperature soldering (reflow peak temperature at 200^oC and below) has been considered as one of the effective ways to reduce the processing defects caused by warpage of large BGA assembly. Most of the large BGAs are still surfaced with SnAgCu (SAC) and/or modified SAC balls, whose melting temperature ranges from 210^oC to 220^oC. Although BGA warpage can be reduced under low reflow temperature, the limited solder paste volume, because SnAgCu ball will not melt but dissolve during reflow, may still be insufficient to compensate for the mitigated warpage, possibly resulting in forming the defective joints, including non-contact, head-in-pillow and non-wet-open etc. Adding more paste would reduce the possibility of forming afore-mentioned defective joints but bring in the risk of joint bridging. Increasing the reflow peak temperature may enlarge the component warpage but allow more molten solder volume during soldering, which may compensate for the warpage displacement and help the joint formation as well. Hot-tearing from using low temperature solder is another concern related to the overlap of component warpage inversion temperature and solder solidification temperature. Thus, optimization of both paste volume and reflow temperature, as well as selection of the appropriate solder alloy are critical to succeed.

The current article has systematically, in the first time, investigated the joint formation through the selection of (1) solder alloys of different melting temperature (138^oC for BiSnAg eutectic, 205^oC for DFLT-1  and 196^oC for DFLT-2), (2) paste to ball volume ratio (PBVR of 0.13, 0.25 and 0.5),  and (3) reflow profiles (low temperature reflow of 185 and 200^oC peak, mid temperature reflow of 210 and 220^oC, and traditional SAC reflow of 240^oC peak) with a 40x40mm² PBGA928 component (max concave warpage of -11mil at 100^oC and max convex warpage of 34mil at 250^oC with inversion temperature of 140^oC).  

Defective joints are always seen under low temperature soldering for all three pastes regardless of PBVR.  Increasing reflow temperature to above 200^oC improved the joint shape and greatly reduced the defects.   Under P220 and PBVR of 0.13, all three pastes have achieved the optimal joints without “fatal” defects. DOE analysis had confirmed the reflow peak is the most impact factor to form the optimal joint, followed by PBVR and alloy selection. SAC/BiSnAg joints were dominated by hot-tearing under low temperature reflows, whose wide pasty range, overlapped with the component inversion temperature of 140^oC, is attributed to the higher occurrence.  SAC/DFLT and SAC/DFLT-2 had much narrower pasty range and always solidified above the inversion temperature, showing much less hot tearing. The investigation of the joint reliability performance is still ongoing and will be presented in Phase II.