Effects of defect propagation/growth on in-line defect based yield prediction

The advantages of predicting the yield of the wafers before they reach the end-of-line are manifold. They include: detection and elimination of yield-limiting process excursions and adjustment of wafer starts to meet the demand schedule. In-line defect measurements are the key enablers of the functional yield prediction. Other factors include parametric measurements. The purpose of this paper is to examine the extension of the yield model system to include intra-layer and inter-layer defect propagation statistics. Using real fabline data we present and discuss the results of intra-layer and interlayer defect propagation. The key steps in our yield prediction model are the extraction of critical areas of different layers from the device layout, and analysis of the in-line defect data such as identification of cluster and random defects, separation of `adder´ and `common´ defects, and estimation of defect size distribution