Reliability of mixed alloy ball grid arrays using impact bend test and drop shock test

Low-silver SnAgCu alloys are quickly becoming the standard Pb-free solders for portable electronic devices. However, the industry´s quick adoption of near-eutectic SnAgCu alloys (SAC305 and SAC405, so called high-Ag Alloys) and its resistance to the adopted process change entails unavoidable situations where the components have a high-Ag SAC alloy while the paste used to assemble these components on the boards has a low-Ag alloy or vice versa. The impact of such mixed alloy assemblies has been and continues to be the subject of many recent studies. In this study, the mechanical reliability of printed circuit boards with a number of Pb-free BGA components assembled using two different Pb-free solder pastes (SACX and SAC105) has been studied using the Dynamic Four Point Bend Test (also called Impact Bend Test) and the Drop Shock Test. In the impact bend test, these boards were subjected to different mechanical shocks resulting in different strains. Results were plotted as the percentage of failed solder joints as a function of the strain experienced at the board and fitted to a Weibull distribution to estimate the usable life of the assembly in such circumstances. The drop shock test was conducted per JEDEC standard JESD22-B111 according to which the boards are subjected to the same mechanical shock repeatedly, and the failure fraction data was plotted as a function of the number of drops. The combinations of the BGA solder sphere alloy and the paste alloy used in the drop shock tests were exactly the same as those used in the Impact Bend Test. Results of both impact bend and drop shock tests show very good correlation. Analysis of the failed samples was done with the Dye and Pry test to identify whether the failure was on the board side or on the component side. Failure analysis of the impact bend test samples showed a slight difference in the distribution of failures on the component side and those on the board side. In drop shock test, most of the solder joints fa- - iled on the component side. Results show that even in the situations where the quantity of the solder coming from the solder paste was small relative to the quantity contributed by the sphere, the impact of the paste alloy was significant. An attempt was made to explain the observations in terms of solder inter-diffusion kinetics and resulting micro-structural changes.