Automatic evaluation of line-and-space resist patterns with defects using image recognition technology

New resist materials are necessary to achieve higher resolution for the high NA EUV tools. The feature size shrinkage also increases the possibility of defect generation. Therefore, controlling defects remains essential. There are many factors in the lithography process that can contribute to the formation of defects in resist patterns. As a result, when testing the new resist material for patterning, there are more instances of pattern failures than successful ones. However, understanding pattern flaws can gain knowledge about the mechanism of defect generation. Based on the idea that exploiting the information in pattern failures can guide the resist resolution improvement, this study presents a novel method of interpreting patterns with defects based on an image recognition technology named Hough transform. Approximate 2500 SEM images and part of corresponding simulation results were automatically analyzed. These results were then utilized to extract chemical information.