Artificial neural networks applied in fast-designing ultrabroad bandgap elliptical hole dielectric mode photonic crystal nanobeam cavity

Abstract

Artificial neural networks are employed to predict the band structure of the one-dimensional photonic crystal nanobeam, and to inverse-design the geometry structure with on-demand band edges. The data sets generated by 3D finite-difference time-domain based on elliptical-shaped hole nanobeams are used to train the networks and evaluate the networks’ accuracy. Based on the well-trained forward prediction and inverse-design network, an ultrabroad bandgap elliptical hole dielectric mode nanobeam cavity is designed. The bandgap achieves 77.7 THz for the center segment of the structure, and the whole designing process takes only 0.73 s. The approach can also be expanded to fast-design elliptical hole air mode nanobeam cavities. The present work is of significance for further research on the application of artificial neural networks in photonic crystal cavities and other optical devices design.