Physics-data-driven intelligent optimization for large-aperture metalenses
Yingli Ha, Yu Luo, Mingbo Pu, Fei Zhang, Qiong He, Jinjin Jin, Mingfeng Xu, Yinghui Guo, Xiaogang Li, Xiong Li, Xiaoliang Ma, Xiangang Luo
Abstract
Metalenses have gained significant attention and have been widely utilized in optical systems for focusing and imaging, owing to their lightweight, high-integration, and exceptional-flexibility capabilities. Traditional design methods neglect the coupling effect between adjacent meta-atoms, thus harming the practical performance of meta-devices. The existing physical/data-driven optimization algorithms can solve the above problems, but bring significant time costs or require a large number of data-sets. Here, we propose a physics-data-driven method employing an “intelligent optimizer” that enables us to adaptively modify the sizes of the meta-atom according to the sizes of its surrounding ones. The implementation of such a scheme effectively mitigates the undesired impact of local lattice coupling, and the proposed network model works well on thousands of data-sets with a validation loss of 3×10<sup>−3</sup>. Based on the “intelligent optimizer”, a 1-cm-diameter metalens is designed within 3 hours, and the experimental results show that the 1-mm-diameter metalens has a relative focusing efficiency of 93.4% (compared to the ideal focusing efficiency) and a Strehl ratio of 0.94. Compared to previous inverse design method, our method significantly boosts designing efficiency with five orders of magnitude reduction in time. More generally, it may set a new paradigm for devising large-aperture meta-devices.