Abstract

The effects of package temperature on failure mechanisms and lifetimes under mechanical shock loading were studied with the help of five different types of high-density packages (a WL-CSP and four CSP-BGAs) assembled on both double-layer and multi-layer FR4 boards. The localized heating of the packages by means of integrated heating elements was utilized in order to produce similar hot spots to those occurring in products in service. The results showed that the temperature can have a significant effect on the lifetimes of component boards under mechanical shock loading but that the effect varied according to the structures of the component boards. The average number of drops to failure of the WL-CSP component boards increased significantly with an increase in the temperature of the package, while the average number of drops to failure of the CSP-BGA component boards generally decreased. On the other hand, the drop reliability of one out of four CSP-BGA component board types was insensitive to temperature. The failure modes and mechanisms were clarified with the help of physical failure analyses that revealed different failure modes in the component boards. Furthermore, depending on the component board type, the primary failure mode may change with temperature from that identified at room temperature. Particular attention was paid to the nucleation and propagation of cracks at different test temperatures. Computational case studies were designed in order to identify the significance of a change in temperature on three factors: (a) the stiffness of the PWB; (b) the strength and elastic modulus of the solder, and (c) the thermomechanical loads. The influences of each factor on the strains and stresses in the proximity of the solder interconnections were evaluated by means of the finite element method. The results of the statistical and physical failure analyses were rationalized with the help of the results from the finite element analyses. They showed that the effects of a change in temperature on the lifetimes of the component boards under mechanical shock loading can be explained by changes in the nucleation site and/or the propagation of cracks. The results presented in this paper point out that single-load reliability tests can form an incomplete understanding of the failure mechanisms in real service environments and modifications to the currently employed reliability test standards that are needed.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call