Statistical Metrology: Tools for Understanding Variation

As parametric variation increases in importance with shrinking dimensions and increasing integration, the need to understand and manage such variation is becoming critical. Statistical metrology is a collection of tools and techniques for the systematic characterization and study of variation in semiconductor manufacturing. In addition to methods for collecting large volumes of data, important analytic approaches are being developed (1) to decompose parameter distributions into wafer-level, die-level, and wafer-die interaction contributions; and (2) to model the spatial effect of layout, process, or other factors on observed variation. Statistical metrology has been used to study interlevel dielectric thickness and polysilicon critical dimension variation, and new applications to yield improvement, design rule generation, and variation impact analysis will make statistical metrology an important part of future manufacturing and design practice.