Characterization algorithm of failure distribution for lsi yield improvement

This paper describes about an algorithm, which can efficiently characterize a process-induced random failure distribution and a design-induced systematic failure distribution from unknown-induced failure distributions of a memory LSI to predict a reason for yield degradation in it. Our developed algorithm analyzes a function "T(f) isn't greater than 1 or not" related kind and the content of the mathematics measure "f" comes from a histogram of all pair of failure distances to distinguish between a random distribution and a systematic distribution. The cause specification using this algorithm is based on the fact that process induced failure caused by particles or contamination results in a random distribution and design-induced failures have some kind of distribution shape or period depending on circuit layout. Memory LSI failure analysis with our algorithm is able to distinguish between an almost random (process-induced) distribution in one direction, and a systematic (design induced) distribution in the other. This algorithm is effective for a device containing an array of cells.