Deep neural network enabled active metasurface embedded design

Abstract In this paper, we propose a deep learning approach for forward modeling and inverse design of photonic devices containing embedded active metasurface structures. particular, demonstrate that combining neural network metasurfaces with scattering matrix-based optimization significantly simplifies the computational overhead while facilitating accurate objective-driven design. As an example, apply our to continuously tunable bandpass filter in mid-wave infrared, featuring narrow passband (∼10 nm), high quality factors ( Q -factors ∼ 10 2 ), large out-of-band rejection (optical density ≥ 3). The consists optical phase-change material Ge Sb Se 4 Te (GSST) atop silicon heater sandwiched between two distributed Bragg reflectors (DBRs). proposed can be generalized arbitrary response incorporating metasurfaces.