A Control & Diagnosis Scheme for Semiconductor Manufacturing

In this paper we discuss work seeking to integrate control and diagnostic applications. Two such applications will be discussed, both in the context of semiconductor manufacturing operations. The first application deals with the analysis of a multistep control system deployed around a photolithographic workcell. The specific effort is directed towards integrating the photolithography controller with a physically based simulation system capable of diagnosis. The second application deals with a multivariate, real-time statistical process control system applied on a plasma etcher. This system is also being extended to perform fault discrimination. The intent is to connect both control systems to a diagnostic system based on evidential reasoning.