Chiral scatterers designed by Bayesian optimization

TL;DR

This paper employs Bayesian optimization to design chiral scatterers that enhance the emission of chiral light, leveraging simulations of Maxwell's equations to navigate complex parameter spaces.

Contribution

It introduces a Bayesian optimization framework for designing chiral scatterers, integrating theoretical insights and rigorous electromagnetic simulations.

Findings

Optimized chiral scatterers with enhanced emission properties

Effective navigation of high-dimensional design spaces

Demonstrated potential for tailored chiral light sources

Abstract

The helicity or chirality of scattered light is strongly linked to the dual symmetry of the scatterer. The latter depends on chiral materials or on scatterers which are not superimposable with their mirror image. This inherently yields asymmetric structures of various shapes with many degrees of freedom. In order to explore these high dimensional parameter spaces, numerical simulations and especially optimization strategies are a valueable tool. Here, we optimize the emission of chiral line sources in two-dimensional dimer setups using Bayesian optimization. We deduce relevant objective functions from recent theoretical findings for chiral electromagnetic fields and employ rigorous simulations of Maxwell's equations.