Acceleration of yield enhancement activity by utilizing real-time fail bitmap analysis

Abstract

A methodology to enhance baseline yield by utilizing fail bitmap (FBM) analysis is presented in this paper. Yield impact of each process step is estimated from FBM-defect correlation and FBM classification. Statistical accuracy of the data is evaluated by using actual wafer-to-wafer variation of our fab, and is significantly improved by increasing FBM sample size. In order to analyze a large amount of data in real-time manner, throughput of the data analysis is also enhanced by an automated system for timely data feedback to yield enhancement activities.