Share Email Print

Proceedings Paper

Application of artificial neural networks to real-time control of plasma processes
Author(s): Daniel S. Camporese
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Plasma processes typically involve a number of coupled control parameters which exhibit complex interactions. These parameters jointly effect the plasma process parameters which may or may not be measurable. Ultimately, the process parameters affect the wafer parameters which we would ultimately like to control. The inherent multivariate nature of the problem makes conventional control methodologies difficult to apply. Artificial neural networks (ANNs) offer a promising alternative because of their ability to learn the desired control behavior by direct observation of the process. Once enabled as the controller, the ANN continues to improve its model of the process behavior and thus compensates for slow drifts in the process.

Paper Details

Date Published: 1 January 1992
PDF: 10 pages
Proc. SPIE 1594, Process Module Metrology, Control and Clustering, (1 January 1992); doi: 10.1117/12.56640
Show Author Affiliations
Daniel S. Camporese, Techware Systems Corp. (Canada)

Published in SPIE Proceedings Vol. 1594:
Process Module Metrology, Control and Clustering
Cecil J. Davis; Irving P. Herman; Terry R. Turner, Editor(s)

© SPIE. Terms of Use
Back to Top