A systematic approach to identify critical yield sensitive parametric parameters

摘要

Identifying and eliminating yield limiting factors is a critical requirement for achieving accelerated yield improvement rates. This paper presents a systematic methodology for identifying parametric yield sensitive parameters in a semiconductor process. The method decomposes the product yield into two major components: a non-random systematic yield Y/sub s/ and a random yield Y/sub r/. The systematic yield Y/sub s/ and all ET (Electrical Test) parameters are then used for the statistical correlation analysis. It divides data into groups and calculates statistics for all ET (Electrical Test) parameters. The critical yield sensitive parameters are identified based on the results of the correlation analysis and the analysis of the statistical significance of the data groups. Applications of the proposed analysis method to many manufacturing lines' data have been very successful in identifying and quantifying parametric yield sensitivity.