An Sensong, Zheng Bowen, Julian Matthew, Williams Calum, Tang Hong, Gu Tian, Zhang Hualiang, Kim Hyun Jung, Hu Juejun
Department of Materials Science & Engineering, Massachusetts Institute of Technology, Cambridge 02139, MA, USA.
Department of Electrical & Computer Engineering, University of Massachusetts Lowell, Lowell 01854, MA, USA.
Nanophotonics. 2022 Jun 10;11(17):4149-4158. doi: 10.1515/nanoph-2022-0152. eCollection 2022 Sep.
In this paper, we propose a deep learning approach for forward modeling and inverse design of photonic devices containing embedded active metasurface structures. In particular, we demonstrate that combining neural network design of metasurfaces with scattering matrix-based optimization significantly simplifies the computational overhead while facilitating accurate objective-driven design. As an example, we apply our approach to the design of a continuously tunable bandpass filter in the mid-wave infrared, featuring narrow passband (∼10 nm), high quality factors (-factors ∼ 10), and large out-of-band rejection (optical density ≥ 3). The design consists of an optical phase-change material GeSbSeTe (GSST) metasurface atop a silicon heater sandwiched between two distributed Bragg reflectors (DBRs). The proposed design approach can be generalized to the modeling and inverse design of arbitrary response photonic devices incorporating active metasurfaces.
在本文中，我们提出了一种深度学习方法，用于对包含嵌入式有源超表面结构的光子器件进行正向建模和逆向设计。具体而言，我们证明了将超表面的神经网络设计与基于散射矩阵的优化相结合，在简化计算开销的同时，还能促进精确的目标驱动设计。例如，我们将该方法应用于中波红外连续可调带通滤波器的设计，该滤波器具有窄通带（约10纳米）、高品质因数（品质因数约为10）和大的带外抑制（光密度≥3）。该设计由夹在两个分布式布拉格反射器（DBR）之间的硅加热器顶部的光学相变材料GeSbSeTe（GSST）超表面组成。所提出的设计方法可以推广到包含有源超表面的任意响应光子器件的建模和逆向设计。
