Characterization and Modeling of Pattern Dependencies and Time Evolution in Plasma Etching

A quantitative model capturing pattern dependent effects and time evolution of the etch rate in Deep Reactive Ion Etching (DRIE) is presented. DRIE is a key process for pattern formation in semiconductor fabrication. Non-uniformities are caused due to microloading and aspect ratio dependencies. The etch rate varies over time and lateral etch consumes some of the etching species. This thesis contributes a physical analysis for capturing and modeling microloading, aspect ratio dependencies, effects of lateral etch and time evolution of the etch rate. This methodology is applied to the study of etching variation on silicon wafers; the integrated model is able to predict pattern density and feature size dependent non-uniformities in trench depth and time evolution of the etch rate. Previous studies of variation in plasma etching have characterized microloading and aspect ratio dependent etching (ARDE) as distinct constant causes for etch non-uniformity. In contrast to these previous works, we present here a time-based methodology for vertical and lateral etch. Thesis Supervisor: Duane S. Boning Title: Professor of Electrical Engineering and Computer Science