Large-range two-dimensional sub-nano-misalignment sensing for lithography with a piecewise frequency regression
Optics Letters
|March 15, 2024
Summary
This study introduces a novel convolutional regression filter for precise two-dimensional (2D) misalignment measurements. The new method achieves 0.82 nm accuracy over a 3 mm range, overcoming limitations of traditional circular gratings.
Area of Science:
- Optics and Photonics
- Metrology
- Machine Learning Applications
Background:
- Circular gratings are typically used for coarse alignment, not high-precision 2D misalignment measurements.
- Complex phase distribution in gratings hinders accurate data extraction with conventional methods.
Purpose of the Study:
- To develop a novel method for high-precision 2D misalignment measurement.
- To overcome the limitations of traditional frequency filtering techniques for circular gratings.
Main Methods:
- Introduction of a novel convolutional regression filter.
- Application of the filter for two-dimensional (2D) misalignment measurement.
Main Results:
- Achieved an impressive accuracy of 0.82 nanometers (nm) over a 3 millimeter (mm) range.
- Demonstrated robustness against system errors and noise.
Conclusions:
- The proposed convolutional regression filter offers an effective solution for 2D misalignment sensing.
- This strategy shows promise for future advancements in precision measurement applications.
