Peer-Reviewed · Scientific Reports
A Cyclical Deep Learning Based Framework For Simultaneous Inverse and Forward Design of Nanophotonic Metasurfaces
Authors
Abhishek Mall, Abhijeet Patil, Amit Sethi, and Anshuman Kumar
Abstract
The conventional approach to nanophotonic metasurface design and optimization for a targeted electromagnetic response involves exploring large geometry and material spaces. This is a highly iterative process based on trial and error, which is computationally costly and time-consuming. Moreover, the non-uniqueness of structural designs and the high non-linearity between electromagnetic response and design makes this problem challenging. To model this unintuitive relationship between electromagnetic response and metasurface structural design as a probability distribution in the design space, we introduce a framework for inverse design of nanophotonic metasurfaces based on cyclical deep learning (DL). The proposed framework performs inverse design and optimization for generating meta-atoms and meta-molecules as metasurface units, based on DL models combined with a genetic algorithm. It includes consecutive models that emulate numerical electromagnetic simulation and iterative optimization, generating optimized structural designs while simultaneously performing forward and inverse tasks. The framework also systematically explores unseen metasurface topologies. As an example, we demonstrate inverse design of a gap-plasmon half-wave plate metasurface for user-defined optical response. This method can be generalized for a wide range of targeted optical responses.