This paper presents an innovative use of machine learning (ML) to improve etch modeling by integrating monotonic machine learning methods with ML-based contour metrology. Unlike traditional methods that rely on single gauge-based data, our approach leverages comprehensive contour data extracted from SEM images to predict etching biases. It handles large datasets efficiently and adapts dynamically to new data. A primary element of our strategy involves constructing a retargeting layer with etch bias, derived from features at multiple sites or points of interest (POIs) on a reference layer which is generated with a fuzzy clustering model. These features and their corresponding etch biases serve as training data for our semi-supervised model which will be used for prediction on large scale designs.
In this paper, we introduce a method that employs a deep learning model, built with GPU, to extract contours from a variety of SEM images. The model is trained with images and their corresponding ground truth. Various models are explored, and their predictive results are juxtaposed with the known ground truth. In comparison with CPU, utilizing GPU can augment the speed approximately 20 times.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.