Deep Generative Modeling and Inverse Design of Manufacturable Free-Form Dielectric Metasurfaces

Generative Design
DOI: 10.1021/acsphotonics.2c01006 Publication Date: 2022-09-22T14:25:17Z
ABSTRACT
Conventional approaches on design and modeling of metasurfaces employ accurate simulation methods. However, these methods require considerable computational power time for every simulation, making them computationally expensive in the long run. To address this high cost learn compact yet expressive representations high-dimensional meta-atoms efficient optimization, deep learning (DL) based have emerged as an alternative solution numerous applications been demonstrated recent years. there are still outstanding challenges DL-assisted that need to be overcome, such limited degrees freedom, insufficient generalizability models, poor fabrication feasibility final designs. Here, concurrently addressing challenges, we propose end-to-end framework generative inverse dielectric free-form metasurfaces. The is generic, it can accommodate a variety physical scenarios including dispersion, incident polarization, operation wavelength using single data set model. We develop shape generation method generate inclusive, free-form, feasible meta-atom library with manufacturability considerations. A forward model exhibits improved terms material spectral window constructed neural networks. In stage, manufacturable realized. As proof-of-concept, meta-lens optimization polarization filter quarter-wave plate demonstrated.
SUPPLEMENTAL MATERIAL
Coming soon ....
REFERENCES (42)
CITATIONS (21)