Inverse design of broadband, strongly-coupled plexcitonic nonlinear metasurfaces

Abstract

Hybrid photonic structures of plasmonic metasurfaces coupled to atomically thin semiconductors have emerged as a versatile platform for strong light-matter interaction, supporting both strong coupling and parametric nonlinearities. However, designing optimized nonlinear hybrid metasurfaces is a complex task, as the multiple parameters' contribution to the nonlinear response is elusive. Here we present a simple yet powerful strategy for maximizing the nonlinear response of the hybrid structures based on evolutionary inverse design of the metasurface’s near-field enhancement around the excitonic frequency. We show that the strong coupling greatly enhances the nonlinear signal, and that its magnitude is mainly determined by the Rabi splitting, making it robust to geometrical variations of the metasurface. Furthermore, the large Rabi splitting attained by these hybrid structures enables broadband operation over the frequencies of the hybridized modes. Our results constitute a significant step towards achieving flexible nonlinear control, which can benefit applications in nonlinear frequency conversion, all-optical switching, and phase-controlled nonlinear metasurfaces.

Expanding the Horizons of Machine Learning in Nanomaterials to Chiral Nanostructures

“…For example, the strong coupling between plasmonic nanocavities and monolayer semiconductors was optimized by a genetic algorithm with a novel figure-of-merit, with the evaluation of each configuration being based on the finite-difference time-domain method. [ 102 ] Furthermore, genetic algorithms can be accelerated further by ML, as was shown for the discovery of stable, compositionally variant, geometrically similar NP alloys for catalysis, resulting in a 50-fold reduction in the number of required energy calculations compared to a traditional “brute force” genetic algorithm. [ 103 ]…”

Strong Coupling at Room Temperature Achieved by Nanoscale Inverse Design

“…Furthermore, our optimization function can be combined with additional plasmonic phenomena derived from the near-field response, such as directionality, incoming polarization selectivity, focusing, plasmonic laser, enhanced nonlinearity, and more, leading to new functionalities. We have recently shown that the structures obtained via the near-field inverse design method can be used to greatly enhance parametric nonlinear processes, allowing for broadband nonlinear control over the entire range of frequencies of the hybridized modes …”

“…We have recently shown that the structures obtained via the near-field inverse design method can be used to greatly enhance parametric nonlinear processes, allowing for broadband nonlinear control over the entire range of frequencies of the hybridized modes. 29…”