Related Experiment Videos
Experimental characterization of subwavelength diffraction gratings by an inverse-scattering neural method
Stéphane Robert1, Alain Mure Ravaud, Stéphanie Reynaud
1Laboratoire Traitement du Signal et Instrumentation, Unité Mixte de Recherche, Centre National de la Recherche Scientifique 5516, 23, rue du Docteur Paul Michelon, 42023 Saint-Etienne Cedex 2, France. stephane.robert@univ-st-etienne.fr
Summary
A novel neural network inverse-scattering method accurately characterizes silicon gratings. This non-destructive technique aids fabrication process improvements with minimal measurements.
Area of Science:
- Optics and photonics
- Materials science
- Artificial intelligence in metrology
Background:
- Characterizing gratings with small period-to-wavelength ratios is challenging.
- Accurate grating characterization is crucial for optimizing fabrication processes.
Purpose of the Study:
- To experimentally validate an inverse-scattering method utilizing a neural network for silicon etched gratings.
- To compare the neural network method with traditional microscopic techniques.
Main Methods:
- Experimental testing of a neural network-based inverse-scattering method.
- Characterization using two established microscopic methods.
- Validation by comparing measured and calculated diffracted intensities.
- Assessing accuracy and repeatability through sample scanning.
Main Results:
- The neural network inverse-scattering method demonstrated validity for grating characterization.
- The method proved effective for non-destructive and non-invasive measurements.
- Accuracy and repeatability were estimated under near-usage conditions.
Conclusions:
- The validated inverse-scattering method offers an easy-to-implement solution for grating characterization.
- This technique requires measurement of only a few diffracted intensities.
- It provides a valuable tool for improving grating fabrication processes.