Authored By:
Robert Darveaux, Corey Reichman, Sabira Enayet, Wen-Sung Hsu, and Win Thandar Swe
Amkor Technology, Inc., Chandler, AZ, USA
Summary
Board level reliability testing was used to compare six lead free alloys to tin-lead eutectic using a 98 ball Wafer Level Chip Scale Package (WLCSP). The component had a 0.5mm Ball Grid Array (BGA) pitch, and Al/NiV/Cu pad metallization. Thermal cycling (4 conditions), cyclic bend (2 conditions), cyclic drop (3 conditions), and solder joint array tensile testing (3 conditions) were utilized to compare the alloys. The effects of reflow conditions and aging conditions were quantified.
In drop testing, first failures were in the range of 4 to 1000 drops. Most samples failed by a mixture of bulk solder and interface failure. Drop test life improved with increased Ag content. The effect of mild aging after surface mount was positive for most alloys. The effect of multiple reflows was mixed.
In solder joint array tensile testing, the Ductile-to-brittle transition strain rate (DTBTSR) was in the range of 0.3/sec to 80/sec. DTBTSR improved with decreasing Ag content and with room temperature aging, but it degraded with multiple reflows.
In cyclic bend testing, first failures were in the range of 1000 to 5000 cycles. SAC405 and 63Sn37Pb had the best performance. A 3mm bend deflection had 2x to 3x longer life compared to a 4mm bend deflection.
In temperature cycling, first failures were in the range of 100 to 6000 cycles. Fatigue life increased with Ag content for the SAC alloys. Sn0.7Cu showed good performance under all conditions. 63Sn37Pb showed good performance under 35C<=>110C condition. Sn3.5Ag had poor performance under all conditions due to voiding and some interface failures.
Conclusions
1) Board level reliability testing was used to compare six lead free alloys to tin-lead eutectic using a 98 ball Wafer Level Chip Scale Package (WLCSP). The effects of reflow conditions and aging conditions were quantified.
2) In drop testing, first failures were in the range of 4 to 1000 drops. Most samples failed by a mixture of bulk solder and interface failure. Drop test life improved with increased Ag content. The effect of mild aging after surface mount was positive for most alloys. The effect of multiple reflows was mixed.
3) In solder joint array tensile testing, the ductile-to-brittle transition strain rate (DTBTSR) was in the range of 0.3/sec to 80/sec. DTBTSR improved with decreasing Ag content and with room temperature aging, but it degraded with multiple reflows.
4) In cyclic bend testing, first failures were in the range of 1000 to 5000 cycles. SAC405 and 63Sn37Pb had the best performance. A 3mm bend deflection had 2x to 3x longer life compared to a 4mm bend deflection.
5) In temperature cycling, first failures were in the range of 100 to 6000 cycles. Fatigue life increased with Ag content for the SAC alloys. Sn0.7Cu showed good performance under all conditions. 63Sn37Pb showed good performance under 35C<=>110C condition. Sn3.5Ag had poor performance under all conditions due to voiding and some interface failures.
Initially Published in the SMTA Proceedings
|
