Deep learning enabled moiré chiral metasurfaces.

Area of Science:

• Optics and Photonics
• Materials Science
• Computational Physics

Background:

• Moiré metasurfaces with twisted angles demonstrate significant optical chirality.
• The quasi-periodic nature of moiré superlattices leads to computational challenges in design due to large supercells and complex parameter spaces.
• Existing methods struggle with the nonlinear relationship between geometry and optical response, complicating forward modeling and inverse design.

Purpose of the Study:

• To propose a novel deep learning-based strategy to overcome computational limitations in designing moiré chiral metamaterials.
• To accelerate the prediction of optical spectra and enable efficient inverse design of chiroptical responses.
• To facilitate customized metasurface design through digital fitting, reducing development time and resources.

Main Methods:

• Development of an innovative deep learning model for predicting spectra of moiré chiral metamaterials.
• Implementation of the deep learning model for inverse design of chiroptical responses.
• Utilizing digital fitting for metasurface customization.

Main Results:

• The proposed deep learning model accurately and efficiently performs both forward spectral prediction and inverse design.
• Customization of metasurfaces is achieved through digital fitting, significantly reducing design time and computational cost.
• Demonstrated application of moiré chiral metasurfaces in differentiating biomolecular enantiomers for advanced biological analysis.

Conclusions:

• Deep learning offers a powerful solution to the computational challenges associated with moiré chiral metasurface design.
• The developed strategy enables rapid, accurate, and customized design of these advanced optical materials.
• Moiré chiral metasurfaces hold significant promise for enantiomer-specific biological sensing and other chiral applications.