A framework for extracting defect density information for yield modeling from in-line defect inspection for real-time prediction of random defect limited yields

This paper describes a technique for filtering optically measured in-line defect densities (DD/sub meas/) in order to arrive at defect densities (DD/sub real/) which can be combined with critical areas to predict single layer random defect limited yield. The technique involves correlating in-line defects with bitmap failures to establish the relationship between the optically measured defect density and the defect density required for critical area based yield modeling. That relationship, once established, applies to logic devices as well as memories. This procedure can predict single layer random defect limited yields on a real-time basis as lots are inspected.