Abstract
Coefficient of thermal expansion (CTE) mismatch is responsible for most of the failures experienced in electronics packaging. Traditional means of life testing include thermal shock, thermal cycling, and power cycling; however there are several drawbacks to these methods. Thermal shock produces high thermal gradients which can present inaccuracies in time to failure (TTF) calculations through inconsistent fatigue behavior. Power cycling is the fastest of the three methods at cycling rates of up to 6 cycles per hour, which minimizes the time that a test can be observed during a standard workday/week. All of the processes require a large amount of power either through the repetitive heating and cooling of an oven, or the power dissipated by the electronics themselves, leading to unnecessary energy expenditures. Since these tests need to be performed for any given geometry and material, a more efficient test method was required. By equating CTE mismatch to mechanically cycled fatigue testing the lifetime of the sample can be approximated with at least a 10x reduction in TTF, as proposed by [1]. This allows for materials to be decoupled from a system and tested individually at cycle rates up to 1 cycle per second. We present a comparison of different lead-free solders using this accelerated test method. The test involves utilizing a modified tribometer to apply a sinusoidal force with peaks of ±20 N to a soldered sample using spring displacement according to Hooke’s law. Elevated temperature testing was utilized, at a homologous temperature of 0.8, to further accelerate TTF, in increasing the contribution of plasticity where the solders accumulate a significant amount of work per cycle without risk of reflow. As lead-free solder and solder pastes become more popular, a solid characterization of different attachment materials is required to make an informed selection for cost effectiveness versus robustness in a design. In this paper, we examined three different compositions. The control sample of gold-tin (AuSn) solder was used to compare against tin-antimony at compositions of 90:10 (Sn90Sb10) and 95:05 (Sn95Sb5). The footprint geometry was kept constant across all samples. Post failure, the mode of failure was examined via visual inspection under optical microscope and SEM imaging. The microhardness of samples was then captured in different regions confirming variation between hardness experienced by locations of different failure modes. Additionally, due to process, voiding was present in many samples. This voiding was examined and found to be inversely proportional to life, leading to a better understanding of the importance of process control.
