With the introduction of lead-free solder alloys, the effect of voids on solder joint reliability has rapidly gained importance. In this study, a first analysis of X-rayed CR0805 solder joints shows a significant reduction in void content, from 20% down to 2.5%, after vacuum soldering. The statistical analysis of the void distribution demonstrates that the vacuum option reduces number of voids and median diameter of voids in comparison to the convection soldering process. A subsequent accelerated thermal cycling test of these analysed test vehicles, according to JESD22-A104D, indicates the tendency of a higher characteristic life time for higher void content. In contrast to these findings, the 1% to failure criterion reveals a higher reliability for lower voiding. During the finite element method (FEM) modelling part of this study, two modelling approaches of void implementation into solder joint geometry are investigated: modelling with a constant volume of the standoff for different void contents, and a modelling approach with a random combination of void content and volume of standoff. The modelling approach with the random combination reveals that voids can reduce the lifetime in the “worst case” parameter combination. In particular, the 1% time to failure rate indicates a quantitative correlation with the experimental results. Furthermore, the FEM results suggest a higher impact on reliability for a single void in comparison to a distribution of multiple voids with similar void content. Finally, the FEM study shows a high sensitivity of predicted life time with respect to the standoff height. Based on this finding, the CR0805 solder joint geometry is examined using optical inspection and cross-section polishes with the outcome that the better wetting behaviour during vacuum soldering causes a reduction of the solder alloy volume and consequently further decreases the standoff height.
Read full abstract