Visible Achromatic Metalens Design Based on Artificial Neural Network

Abstract

Metasurfaces, known as ultra‐thin and planar structures, are widely used in optical components with their excellent ability to manipulate the wavefront of the light. The key function of the metasurfaces is the spatial phase modulation, originated from the meta‐atoms. Thus, to find the relation between the phase modulation and the parameters of an individual meta‐atom, including the sizes, shapes, and material's optical properties, is the most important but also time‐consuming part in the metasurface design. Here by developing a backpropagation neural network based machine learning tool, the design process of a high performance achromatic metalens can be greatly simplified and accelerated. A library of the phase modulation data from 15 753 meta‐atoms can be generated in less than 1 s by our backpropagation neural network. In the experiment, it is demonstrated that the designed metalens shows an excellent achromatic focusing and imaging ability in the visible wavelengths from 420 to 640 nm without the polarization dependence.

Dielectric metalens for miniaturized imaging systems: progress and challenges

“…The simulation requirement in computation resources can be reduced with the assistance of machine learning algorithms such as neural networks [175][176][177][178] . For instance, a phase library containing 15,753 meta-atoms is generated in less than one second by a backpropagation neural network 179 . However, the massive dataset is expensive and sometimes unrealistic regarding the economic and labor costs.…”

Broadband achromatic metalens design based on machine learning

“…The produced phase will be compared with value from the numerical simulation and then correct the neural network until the MSE was small enough. With the welltrained neural network, we obtained the predicted phase response of 15753 patterns in 0.67 second [8]. 2.…”

“…Fabrication of the designed metalens performance By using the particle swarm optimization method, the final design with 3209 patterns including 3153 cross shapes and 56 hollow circle shapes were determined from the database (15753 meta-atoms) according to the required phase and group delay that can realize the achromatic focusing. Based on this final design, we fabricated our metalens by using the E-beam lithography (EBL) patterning and the atomic layer deposition growth of TiO2 to form the nano-posts array [7,8]. The scanning electron microscope images of the fabricated sample are shown in Fig.…”

Recent advances in strongly resonant and gradient all-dielectric metasurfaces

“…The silicon on insulator material system used in this case is technologically favorable due to the CMOS electronics industry. Similar waveguide cross-sections but with TiO 2 instead of Si have been utilized by Wang et al 160 for metalenses targeted at visible frequencies. In order to improve the process of expanding and selecting within the achievable span of phase and group delay, a backpropagation neural network and particle swarm optimization were utilized.…”