Defect Density Extraction of Gate Oxide Shorts Incorporating Defect Clustering with Pseudo Transistor Arrays

The performance of gate oxide is a critical factor which influences yield and reliability. Gate oxide short (GOS) has become a dominant failure of CMOS integrated circuits. Thus defect density of GOS is a key parameter for yield prediction of nanometer process. In this paper, a new approach for the defect density extraction of GOS incorporating defect clustering, as well as pseudo transistor arrays as test vehicles to collect mass of testing data are proposed. Experimental results show that it is in a good agreement with inline e-test data when the extracted defect density of GOS is used to predict yield.