3D high-NA metalenses enabled by efficient 2D optimization

Abstract

Inverse design through topology and dimension optimization can significantly improve the performance of metasurfaces as compared to the traditional point-by-point designs. However, it is very challenging to optimize large-area 3D gradient metasurfaces due to heavy electromagnetic simulation and slow converging speed. In this work, we propose a quick and efficient design approach of 3D large-area high-NA metalenses by simple rotation and translation of the 2D cylindrical lens, where the later can be optimized using basic searching algorithms. The simulation and experimental results together show the effectiveness of this approach, which has great potential to be generalized for fast optimization of other types of wavefront shaping metasurfaces with spatial symmetry.