Development of integrated process control system utilizing neuralnetwork for plasma etching

摘要

The purpose of this study is to provide the integrated processcontrol system utilizing neural network modeling, to search for theappropriate choice of control input, and to keep the process outputwithin the desired range in the real etch process. Variations in theprocess output are classified as the drift and the shift. The drift iscaused by a natural noise that changes over a period of time slowly andsteadily partly due to the aging of equipment. Although the drift movesthe process output away from the target value, its variation isinfinitesimal. On the other hand, the shift results in a largervariation due to the various causes and its width is normally greaterthan that of the drift. Without appropriate procedures, the processoutput will move away from the target value greatly. Therefore thecontrol strategy is to minimize the process shifts, in which processoutputs are measured by monitoring wafers periodically