EUV mask modeling based on a wide-angle full-vector beam propagation method
Abstract
Extreme ultraviolet (EUV) lithography imaging models have garnered widespread attention as digital twin simulation engines for the EUV lithography process. However, as the core of EUV lithography imaging simulation, the EUV mask model still faces challenges in balancing computational accuracy, efficiency, memory usage, and simulation area. To address this, we propose an EUV mask model based on the wide-angle full-vector beam propagation method (WA-FVBPM), integrated with the transfer matrix method and plane-wave expansion. By employing an optimized Padé approximation and an acceleration strategy using alternating direction implicit differentiation, this model achieves both acceptable computational accuracy and exceptional efficiency while minimizing memory consumption for near-field calculations of EUV masks with diverse patterns. Embedding the WA-FVBPM-based mask model within the Abbe imaging framework enables accurate calculation of EUV lithography aerial images, yielding a critical dimension (CD) relative error of less than 0.8% for a standard contact array pattern with a CD of 24.4 nm. Actually, the absolute CD error for all the patterns is less than half a discrete grid, with potential for further reduction through fine sampling. Compared to the finite-difference time-domain method, WA-FVBPM improves computational efficiency by approximately 200 times and reduces memory usage by 100 times while maintaining negligible accuracy loss. This facilitates large-scale simulation of EUV masks with complex patterns. The demonstrated advantages underscore the method's significant potential for full-chip EUV lithography simulations.
Fields of Study & Subjects
Advancements in Photolithography Techniques
Advanced optical system design
Electron and X-Ray Spectroscopy Techniques
Publication details
- Journal / source
- Optics Express
- Publisher
- Optica Publishing Group
- Volume
- 34
- Issue
- 9
- Pages
- 17074–17074
- Publication date
- 23 April 2026
- Language
- EN