Fault diagnosis of plasma etch equipment

The development and implementation of robust methods for fault detection promises to enhance manufacturing by improving our capability to monitor equipment and processes. In order to fully utilize this capability it is important that the machine fault is not only detected, but also diagnosed as belonging to a fault category so that appropriate corrective action can promptly be taken. In this paper we examine the diagnostic performance of two probabilistic modeling techniques in using sensor signals to classify faults. We also discuss how the strengths of these models may be combined in a hierarchical architecture giving rise to a more powerful diagnostic tool.