40TH EUROPEAN MASK AND LITHOGRAPHY CONFERENCE, EMLC 2025
Abstract
High-resolution images obtained via Scanning Electron Microscopy (SEM) are vital for semiconductor manufacturing and quality control. Nevertheless, SEM images often suffer from noise and charging effects, particularly when imaging non-conductive materials, leading to image distortion and measurement errors.
In this study, we introduce advanced denoising techniques based on non-local and isotropic total variation minimization using the Split-Bregman algorithm to reduce noise in SEM images. Furthermore, we have used a method based on interpolation and filtering to mitigate the charging effect in SEM images.
Our approach yields improved image clarity, enabling precise contour extraction and enhanced critical dimension CD matching, thereby supporting more accurate semiconductor inspection. By addressing the challenges of noise and charging effects in SEM imaging, this work contributes to the advancement of semiconductor manufacturing and quality control processes, ultimately enhancing the reliability and performance of electronic devices.
