Towards Excursion Detection for Implant Layers based on Virtual Overlay Metrology

Abstract

Virtual overlay metrology has been developed for a series of nine implant layers using a hybrid approach that combines physical modeling with machine learning. The prediction model is evaluated on production data. A high prediction capability is achieved and the model is able to follow variations in the implant-layer overlay and to identify outliers. We will use the prediction model to link excursions to a possible root cause. Furthermore, a KPI based on scanner metrology is defined that can be monitored continuously, and for every wafer, for detecting excursions with a similar root cause.