Enhanced Defect Detection in After Develop Inspection with Machine Learning Disposition

A complementary Machine Learning disposition method was generated and tested for after develop inspections in lithography. For lithography coating defects, this new method showed twice the sensitivity and five times the specificity in a controlled experiment versus the baseline system. Applying the detection method along with process improvements, preventative measures and rework for splatter defects, reduced yield loss from splatters by over 30x. Herein we describe learnings on the use of image enhancement for training and disposition, an Explainable AI system to support understanding, and a process flow to train augmentation based on performance.