Related Experiment Video
Updated: Nov 15, 2025
06:56
Micro/Nano-scale Strain Distribution Measurement from Sampling Moiré Fringes
Published on: May 23, 2017
12.5K
Deep learning-based moiré-fringe alignment with circular gratings for lithography
Optics Letters
|March 2, 2021
Summary
A new deep learning strategy achieves nanoscale precision for lithography misalignment measurement. This method offers improved accuracy and robustness over traditional moiré fringe techniques for advanced semiconductor manufacturing.
Area of Science:
- Semiconductor Manufacturing
- Metrology
- Deep Learning
Background:
- Accurate overlay misalignment measurement is critical in lithography for high-precision semiconductor fabrication.
- Traditional moiré-based methods using circular gratings lack the analytical framework for nanoscale accuracy, limiting their use to coarse alignment.
Purpose of the Study:
- To develop a novel, high-precision, two-dimensional misalignment measurement strategy for lithography.
- To overcome the limitations of existing moiré-based techniques in terms of accuracy and robustness.
Main Methods:
- A deep learning strategy, inspired by deep convolutional neural networks, was developed for misalignment measurement.
- The proposed method utilizes micron-scale circular alignment marks.
Main Results:
- The deep learning scheme achieved nanoscale accuracy in misalignment measurement.
- The strategy demonstrated significantly higher precision and robustness against fabrication defects and noise compared to existing methods.
Conclusions:
- The proposed deep learning approach enables a one-step, two-dimensional nanoscale alignment for various lithography types.
- This advancement is crucial for next-generation semiconductor manufacturing processes, including extreme ultraviolet (EUV) lithography.
