Abstract
This is the second article in a two-part series on the causes of stress voiding and the evidence required to tie it to an IC interconnect failure. The first part, published in the November 1999 issue of EDFA, focuses on the causes and distinguishing characteristics of stress void failures. Here, that information is applied in the analysis of an actual case of stress voiding. The author explains how he developed evidence of stress, nucleation, and diffusion, the three phenomena required to differentiate stress voiding from other failure mechanisms, and how this evidence pointed to specific processing errors and suggested how to fix them. Although the investigation relied heavily on traditional observational tools, they were applied in new ways using a relatively new technique, FIB backside thinning, to gain critical proof of nucleation and confirm suspected processing errors.
