Sampling methodology for SEM-based defect classification: risk, cost, and benefit analysis

Abstract

The current paper analyzes the relative merits of optical and SEM-based defect classification systems, the needs and costs associated with these systems, and the factors limiting the usability of these systems. In particular, we consider the impact of throughput rate and classification accuracy, on excursion detection and the resulting economic benefits. The paper includes a discussion of these models and a comparison is made to obtain the maximum benefits from existing optical and SEM review and classification methodologies. Scenarios for 0.25 micron fabs are used to indicate the procedures and policies that are the most effective from a fab economic perspective.