Stepper registration analysis using Monte Carlo modeling

This paper presents an analysis of stepper registration using Monte Carlo modeling. Registration data are collected from an array of many error vectors and assimilated in the form of a multiple linear equation where each coefficient in the equation represents a different geometrical error. Data are collected periodically from several tools. Nested variance analysis was used to study the different families of variation for the geometrical errors. Component coefficient variations are studied for normality and Monte Carlo simulation is used to predict the distributions of errors that will be seen over time. The interaction of intrafield and interfield errors can be easily handled with this approach to generate the total registration error.