|
|
Effect of Temperature and Current Stressing on Low Temperature Solder BGA Drop Performance
Analysis Lab |
|
Authored By:Alyssa Yaeger, Ph.D., Michael Meilunas Universal Instruments Corporation NY, USA Sufyan Tahat Auburn University AL, USA SummaryBGA components constructed with low temperature solder (LTS) SnBiSbNi spheres were assembled using a matching alloy paste and then subjected to simultaneous direct current and temperature stressing to generate electromigration of the bismuth atoms. The samples were then drop tested to evaluate solder joint lifetime performance and the results were compared to non-stressed control samples. Solder joint failure analysis was performed by cross-sectional examination using scanning electron microscopy. The experiment described indicates that the current densities, joint temperatures, and stress durations examined could result in a significant reduction in drop reliability - even when minimal bismuth electromigration was observed in cross-section. The extent of the reliability reduction was highly dependent upon the solder joint temperature with joint temperatures of 85°C and 100°C resulting in rapid joint degradation while joint temperatures of 75°C showed little or no reduction in reliability relative to the control samples – even when the 75°C samples were subjected to relatively high current densities. ConclusionsSimultaneous temperature and current stressing affect the performance of LTS BGA solder joints in drop testing, with both increased temperature and current level showing larger impacts for the same stress duration. Populations which were temperature / current stressed at a mild service-like conditions of 0.50A did not show obvious bismuth accumulation through 90 days but did show some degree of bismuth accumulation at day 120 which did not appear to significantly impact drop test performance relative to the benchmark criterion. These samples failed at corner joints due to solder cracking near the PCB pad regardless of the anode and cathode position. Similarly, the slightly harsher 75°C 0.80A condition did not result in bismuth accumulation at the anodes by day 90, but did result in some accumulation by day 120. Importantly, the accumulation at day 120 did not appear to significantly impact the drop test performance relative to similarly aged Control samples and the benchmark criterion, and the primary failure mode was solder joint fracture near the PCB side of the joints regardless of anode and cathode position. The mild conditions performed reasonably well in drop performance, with only the 75oC 0.50A 30-day population failing the benchmark test, as all three populations ultimately showed improvement over time up to 120 days. It is possible that the mild temperature / current stress levels may have inhibited but did not reverse the improvement in drop performance that had been documented with the Control population through 120 days. The current-accelerated populations with 75oC joint temperature at 2.27A produced some degree of bismuth accumulation through 30 days of temperature / current stressing. These test specimens failed due to corner joint solder fracture along the PCB pad during drop testing and performed similarly to the comparable Control samples. It is believed that additional stress time would eventually have resulted in reduced drop performance under this stress condition. 85oC populations showed a decrease in drop reliability over time, with the 85oC 0.80A population performing slightly better than the 85oC 2.84A samples, but with both populations failing the performance benchmark by day 30. The 0.50A population fared better, failing the benchmark requirement only after 120 days of temperature / current stress. Bismuth accumulation was observed after 90 days at 85oC 0.50A and in all populations at both 0.80A and 2.84A, with failure occurring in “anode down” solder joints in most samples. These results indicate that 85oC is an “accelerated condition” as this temperature is only 10oC above peak service condition and yet drives relatively rapid joint degradation with modest current application. The 100oC populations performed similarly to each other in drop testing, with poor performance after only 5 days of temperature / current stress and decreasing performance over time. Most samples showed bismuth accumulation at the anode, and cracking occurred at the corner-most “anode down” solder joints. It is important to note that the bismuth layer thickness was relatively thin, on the order of 1 to 5µm and that all samples survived at least 12 drops prior to failure, indicating that there may be a limit on the impact of additional bismuth accumulation on drop performance (i.e., drop performance will not degrade significantly with increasing bismuth thickness once a thin layer of bismuth has formed). Attempting to derive acceleration factors between the various test conditions is difficult due to the constantly evolving nature of the LTS solder joints which showed improvement over time at room temperature conditions and at 75°C stress conditions but fast degradation at 85°C and 100°C stress conditions. Further complicating the analysis is that the time between stress conditioning and drop testing was limited to 24 to 96 hours for the described experiment, and shorter or longer durations between stress conditioning and drop testing may potentially impact the results in an unknown manner. Similarly, the time between reflow soldering and the application of the stress conditioning may also affect the drop test results, and these “time factors” have not yet been evaluated. Initially Published in the SMTA Proceedings |
|
Comments
|
|
|
|
|