Comparative analysis of grating reconstruction: deep learning versus Levenberg-Marquardt methods

L. Fu; X. Wang; K. Frenner; S. Reichelt

doi:10.1117/12.2688328

10 August 2023 Comparative analysis of grating reconstruction: deep learning versus Levenberg-Marquardt methods

L. Fu, X. Wang, K. Frenner, S. Reichelt

Proceedings Volume 12619, Modeling Aspects in Optical Metrology IX; 1261907 (2023) https://doi.org/10.1117/12.2688328
Event: SPIE Optical Metrology, 2023, Munich, Germany

Abstract

Model-based optical scatterometry is a widely utilized non-destructive measuring technique in semiconductor manufacturing for retrieving features on wafers. It offers an attractive solution for quality control and process monitoring. However, the increasing complexity of 3D nanoscale device structures presents significant challenges for optical scatterometry. To address these challenges, it is crucial to integrate different methods and create a hybrid metrology approach that could encompass measurements, modeling, and data analysis techniques. To tackle this objective, we explore in this study two alternative approaches for parameter reconstruction, distinct from the conventional library search method. The first approach utilizes a neural network based on a Resnet architecture, while the second approach employs the Levenberg-Marquardt algorithm, a nonlinear least square fitting technique. By performing a comparative analysis of the two methods, we propose a strategy to combine them for accurate and efficient parameter reconstructions.

Conference Presentation

Citation Download Citation

L. Fu, X. Wang, K. Frenner, and S. Reichelt "Comparative analysis of grating reconstruction: deep learning versus Levenberg-Marquardt methods", Proc. SPIE 12619, Modeling Aspects in Optical Metrology IX, 1261907 (10 August 2023); https://doi.org/10.1117/12.2688328

ACCESS THE FULL ARTICLE

INSTITUTIONAL
Select your institution to access the SPIE Digital Library.

SELECT YOUR INSTITUTION

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.

PERSONAL SIGN IN

No SPIE Account? Create one

PURCHASE THIS CONTENT

SUBSCRIBE TO DIGITAL LIBRARY

50 downloads per 1-year subscription

Members: $195

Non-members: $335 ADD TO CART

25 downloads per 1 - year subscription

Members: $145

Non-members: $250 ADD TO CART

PURCHASE SINGLE ARTICLE

Includes PDF, HTML & Video, when available