Comparison of a two-stage group-screening design to a standard 2k-p design for a whole-line semiconductor manufacturing simulation model

Abstract

The focus of the paper is on the comparison of results obtained using and not using group screening in an experimental design methodology applied to a semiconductor manufacturing simulation model. A wholeline simulation model of a semiconductor fab is built. The model includes more than 200 tools used in manufacturing 2 products with around 250 steps each. Output analysis results for the equipment utilization and queue sizes have identified the three most critical equipment groups in the fab. Seventeen input factors are set for investigation through a 2-stage group-screening experiment and a fractional factorial using all 17 factors. The result illustrates that the final models can be quite different. While group screening used with simulation can be an appealing, flexible, tractable tool for capacity analysis of a semiconductor manufacturing facility, one must be concerned with the fact that the two techniques can give different answers to the users. Additionally, researchers need to address the proper choice of significance level for group screening.