Authored By:
Sakthi Cibi Kannammal Palaniappan and Martin.K.Anselm, Ph.D.
Center for Electronics Manufacturing and Assembly (CEMA), Rochester Institute of Technology Rochester, NY, USA
Summary
As the traditional eutectic SnPb solder alloy has been outlawed, the electronic industry has almost completely transitioned to the lead-free solder alloys [1] [2]. The conventional SAC305 solder alloy used in lead-free electronic assembly has a high melting and processing temperature with a typical peak reflow temperature of 245 Celsius which is almost 30 Celsius higher than traditional eutectic SnPb reflow profile. Some of the drawbacks of this high melting and processing temperatures are yield loss due to component warpage which has an impact on solder joint formation like bridging, open defects, head on pillow [3], and other drawbacks which include circuit board degradation, economic and environmental factors [4], and brittle failure defects in the circuit board like pad cratering. To overcome this, a detailed study has been carried out on low temperature lead-free solder paste that utilizes Bi bearing alloys.
Three low temperature lead-free solder pastes Sn-58Bi, Sn-57Bi-1Ag and Sn-40Bi-Cu-Ni with the melting temperatures around 138 Celsius (which is 45 Celsius below eutectic SnPb and 79 Celsius below SAC) were printed on Cu-OSP finish test boards. These pastes were assembled with SAC305, Sn99CN and Sn100C solder spheres. The range of Bi concentrations for various resulting mixtures used in this study was calculated to be in the range of 2 to 4 wt%. The mixtures were reflowed under two different low temperatures reflow profiles; (a) a traditional SnPb profile with a peak temperature 217 Celsius and (b) a low temperature SnBi profile with a peak temperature 177 Celsius (recommended by the paste manufacturer). After the assembly process, the mixed solder joints were shear tested to study the failure modes and shear strength at rate of 27.50mils/sec. Cross sectioning was performed to evaluate the possible microstructural changes at room temperature and after aging conditions that may have led to the changes in failure mode observed in shear testing. The isothermal aging condition used in the study is 125 Celsius for 200 hours which mimics 21 years of field storage at 25 Celsius degrees using Arrhenius extrapolation for Cu6Sn5 intermetallic formation. Our study suggests that high temperature reflow profile (217 Celsius peak profile) had better mechanical strength than the low temperature reflow profile (177 Celsius peak profile). A metallurgical explanation for the improvement is presented in this paper. Thus, this paper describes that by generating a robust reflow assembly process for SnBi solder paste, the shear strength can be increased, cost of manufacturing can be reduced and high temperature assembly process (SAC) issues can be minimized which may improve product yield in production.
Conclusions
From the ball shear test, the high temperature reflow profile process resulted in a greater strength than low temperature reflow profile process. With the low temperature reflow process, the peak temperature (177 Celsius) is insufficient to melting and dissolve the lead free ball. The solder ball alloys used in the study have a melting point range from 217 Celsius to 227 Celsius. Incomplete mixing creates a high concentration of Bi precipitates in the high stress region of the shear test resulting in weak solder joint. Whereas with the high temperature reflow profile process, (peak reflow temperature 217 Celsius), the lead-free balls are able to more completely melt and coalesce with the SnBi paste (excluding Sn100C).
From the cross sectional analysis, for the low temperature reflow profile process there is no proper dissolution between the paste and ball, whereas with the high temperature reflow profile process, the paste and the ball are completely dissolved. The reason for this is the higher peak temperature in the high temperature reflow process. With L23+Sn100C process there is partial dissolution between the paste and the ball as the pasty range of L23 is very high and the Ag content in the ball tend to affect the dissolution.
From this study a consistent improvement in high temperature assembly process is observed with the SAC305 solder paste. In addition the results show a tighter distribution of the data for this alloy. Although an improvement in the SN99Cn and SN100C was observed it standard deviation of the results indicate a non-repeatable result and overlap in strength between the two processes.
Micro-alloying improvements in strength for the Sn99Cn and SN100C alloys seems to be limited. Addition of Cu and Ni in the L27 solder paste does not appear to have a comparable increases in strength as compared to SAC305. In fact, L23 (addition of 1wt%Ag) appears to provide the highest strength for the SN99Cn and Sn100C alloys in the high temperature process prior to and following aging.
Initially Published in the SMTA Proceedings
|
