Diagnosis of semiconductor manufacturing equipment and processes

This paper describes part of the research performed during the Microelectronic Manufacturing Science and Technology (MMST) program on techniques for the diagnosis of equipment malfunctions and misprocessing during semiconductor manufacturing. The main motivation of this work was to investigate techniques for rapid diagnosis. Towards this goal, a number of equipment-level diagnosis techniques are described. These techniques use equipment models to diagnose equipment malfunctions at a given process step. The results obtained by applying these diagnosis approaches are very encouraging. The various approaches were able to diagnose a number of faults that were deliberately introduced to test the algorithms and the equipment faults that developed during the final demonstration of the MMST program. These techniques were applied to a TI-built Advanced Vacuum Processor (AVP) and an Applied Materials Precision reactor AMT 5000