Comparison of Equipment Modeling Methods as Applied to the LPCVD of In-situ P-Doped Polysilicon
Abstract
The orthogonal array design of experiments for process and equipment optimization is combined with a semi-empirical modelling approach based on dimensional analysis. This methodology is applied to In-Situ Phosphorus Doped Polysilicon Process via LPCVD. Taguchi's method of pure empiricism works well for the optimization of four different responses-thickness uniformity across the wafer, uniformity in a batch, growth rate, and the film resistivity for the process under consideration. The average thickness uniformity growth rate of 998 A/min. and a resistivity of 1.25 mohm-cm are achieved. However, the dependence of Taguchi's method on the assumption that the process is free of parameter interactions distorts the prediction of nominal-is-the-best-signal-to-noise ratio for the film thickness uniformity across the wafer. As a solution to such problems of purely experimental approach a thirteen terms semi-empirical model based on dimensional analysis is evolved for the film thickness uniformity in the process under consideration by using stepwise regression.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1989
- Accession Number
- ADA208373
Entities
People
- Parmeet S. Chaddha
Organizations
- Massachusetts Institute of Technology