Rapid inverse design of metasurfaces with an asymmetric transfer function for all-optical image processing using a mode matching model

Abstract

Metasurfaces have recently emerged as an ultra-compact solution to perform all-optical image processing, including phase contrast imaging. Most metasurfaces used in imaging processing applications operate over a restricted numerical aperture. This limitation imposes constraints on the discernible features that can be effectively visualized and consequently leads to the appearance of undesirable artifacts. Engineering a metasurface that exhibits an asymmetric linear optical transfer function over a relatively large numerical aperture, while maintaining a strong contrast, has proven to be a challenge. In this study, we present a novel approach to designing relatively high numerical aperture and contrast nonlocal metasurfaces (up to a numerical aperture of around 0.5 and an intensity contrast of approximately 50%) with unit cells consisting of several plasmonic nanorods through the use of a rapid, quasi-analytic mode-matching technique, coupled with an optimization algorithm. The combination of these methods facilitates the rapid conceptualization of nonintuitive arrangements of metallic nanoparticles, specifically tailored to perform phase contrast imaging. These designs hold substantial promise in the development of ultra-compact imaging systems.