Effect of processing parameters and thermal cycling on the microstructure and strength of eutectic tin-silver solder on Cu
Abstract
Detailed studies to characterize the microstructural development and mechanical properties of eutectic Sn-Ag solder joints were carried out on samples reflow soldered on copper under various processing conditions. Some samples were reflow soldered at 30$\sp\circ$C above the melting point and then solidified at different cooling rates. Analysis of these samples showed that increasing the cooling rate increased the volume fraction of primary Sn-dendrites, decreased the amount of $\rm\eta(Cu\sb6Sn\sb5)$ intermetallics in the bulk solder, and resulted in finer microstructures with higher hardnesses. Subsequent isothermal annealing of some of these reflow soldered joints at 125$\sp\circ$C resulted in an initially fairly rapid decrease in hardness to a given level for each cooling rate studied. However, the cooling rate had little or no effect on the shear strength of the solder joints. Studies of the effect of Cu substrate dissolution on the microstructure of solder joints were carried out using samples reflow soldered isothermally at various temperatures and times. Analysis of some of these samples showed that the Cu concentration, the volume fraction of primary Sn-dendrites and $\rm\eta(Cu\sb6Sn\sb5$) intermetallics in the solder increased with reflow temperature and time as a result of Cu dissolution. The isothermal growth kinetics were analyzed using the Nernst-Brunner equation and a numerical method based on it was proposed to predict the amount of Cu dissolution into molten solder for non-isothermal reflow conditions. Isothermal growth of the intermetallic layer between the solder and Cu substrate was investigated, and it was found that layer thickening with time could be described by a power law. A numerical method to predict the layer thickness for non-isothermal reflow was developed which utilized the experimental power law thickening and Nernst-Brunner dissolution kinetic parameters. The method also took into account the additional increase in layer thickness caused by precipitation of dissolved Cu during the cooling part of the reflow process. The intermetallic layer growth resulting from non-isothermal solid state aging was also examined for samples reflow soldered at three different cooling rates and then cycled from ambient to 125$\sp\circ$C. The experimental layer thicknesses were compared to those predicted by a numerical model.
This paper has been withdrawn.