Skip to Main Content
The V-ramp test is an important test for GOI (gate oxide integrity) of MOS devices. The criterion for a product under GOI qualification is to meet the requirement with defect density lower than a target defect density (D0). This paper studies the methods and guidelines in minimum sample size determination provided by JEDEC/FSA joint publication and points out their drawbacks, inconsistency, and misguidance. We provide an exact method and easy-to-use numerical solution by extending JEDEC's formula to any allowed failure number, target defect density, and confidence level. Important guidelines are also provided for reliability practitioners to reduce uncertainty resulted from imperfect sampling procedures and to avoid mistakes commonly happened in defect density evaluation against a target D0. Our proposed exact method can be applied on many other reliability tests that need determining a minimum sample size to save wafers and testing resources as long as the binomial distribution is applicable.