Mixed Metals Impact on Reliability
This paper documents the effect of mixing Leaded and Lead Free alloys on BGA devices and how this impacts reliability.
Analysis Lab
DOWNLOAD
|
Authored By:
Rick Gunn
Vice President of Engineering / Chief Technology Officer
Nextek Incorporated, Madison, Alabama USA
Summary
With the adoption of RoHS and implementation of Lead Free solders a major concern is how this will impact reliability. Both commercial and military hardware are impacted by this change even though military hardware is considered exempt from the requirements of RoHS. As the supply chain has moved to the new lead free alloys both markets are being forced to understand these impacts and form risk mitigation strategies to deal with the change.
This paper documents the effect of mixing Leaded and Lead Free alloys on BGA devices and how this impacts reliability. Three of the most common pitch BGA packages are included in the study to determine if the risk is the same as pitches decrease. Metallurgical analysis was performed utilizing cross-sections and SEM to study the alloying of tin/lead and tin/silver/copper both separately and combined.
MIL-STD-883 Method 1010.8 Temperature Cycling was used to accelerate fatigue life of the samples and testing of those samples was performed at regular intervals using a bed of nails tester. Various reflow soldering temperatures were used to assemble the different combinations of alloys. Under-filling of assembled BGA's was also studied as a risk mitigation strategy.
Conclusions
The data from these experiments exposed several facts. First and most importantly we see that mixing of leaded solder and lead free BGA's is not a problem for reliability but in fact improves the reliability of the solder in temperature cycling. The data also shows that mixed alloy groups out performed single alloy systems across all packages. The average failure time for the leaded/leaded 220C group was 3021 cycles and was the lowest of all groups. The second lowest group was the leadfree/leadfree at 245C group with an average failure time of 6561 cycles.
The leaded/leadfree at 220 group had an average failure time of 7799 cycles and the best group was the leaded/leadfree at 235C with and average failure time of 8986 cycles. The data shows that group 3, leaded/leadfree at 235C, is 197% better than group 1, 36% better than group 4 and 15.2% better than group 2. Package pitch only appeared to be a factor in the .5mm devices as all of that package type failed during the test period.
Failures in the .8mm package and 1.0mm package occurred only in the single alloy groups. Underfill of the BGA's improved reliability of the device regardless of pitch. The failure rate for nonunderfilled parts is 55.8% and the failure rate of underfilled parts is 4.1%. Failure location across all groups is 86% of the time on the package side and did not appear to change behavior with pitch.
Initially Published in the IPC Proceedings
|
Comments
|
No comments have been submitted to date.
|
|
|
|
