Nanophotonic inverse design with SPINS: Software architecture and practical considerations

Abstract

A computational nanophotonic design library for gradient-based optimization called SPINS is presented. Borrowing the concept of computational graphs, SPINS is a design framework that emphasizes flexibility and reproducible results. The mathematical and architectural details to achieve these goals are presented, and practical considerations and heuristics for using inverse design are discussed, including the choice of initial condition and the landscape of local minima. arXiv:1910.04829v2 [physics.app-ph] 31 Oct 2019 design. These include the design of objective functions and the choice of initial condition based on an analysis of the local minima reached by the optimization process.The paper outline is as follows. Section 2 provides a mathematical overview of inverse design, and Section 3 provides an overview of the framework and how the inverse design formulation is implemented. Section 4 presents an example of designing wavelength demultiplexers and discusses salient points in the overall design process. Finally, Section 5 discusses practical considerations and analyzes key properties of gradient-based nanophotonic optimization.

Inverse Design for Silicon Photonics: From Iterative Optimization Algorithms to Deep Neural Networks

“…Other merits such as performance, compactness or fabrication feasibility may also be considered for different application scenarios. Based on the number of DOF, three types of structures are widely used for inverse design: empirical structure [21][22][23][24][25][26][27][28][29][30][31][32], QR-code like structure [33][34][35][36][37][38][39][40][41][42][43][44][45] and irregular structure [46][47][48][49][50][51][52][53][54][55][56][57][58]. An empirical structure making use of existing classical structures usually has only a few DOF.…”

MetaNet: a new paradigm for data sharing in photonics research

“…Further code refactorization, such as that featured in Ref. [44], can provide added modularity and even interchangability of electromagnetic solvers. Code refinement for diffractive optics optimization, including refractorization, will be the subject of future study.…”

Mechanically reconfigurable multi-functional meta-optics studied at microwave frequencies

“…To expand the functionality of metaoptics, adjoint-based topology optimization is capable of designing high-performing dielectric structures with nonintuitive index distributions and objective functions that are challenging to achieve with traditional methods 9,10 . While much of the work has been on 2D platforms such as silicon photonic waveguides 11,12 and at optics 13,14 , free-space 2.5D and 3D devices have also been explored recently [15][16][17][18][19] . In a previous work, a passive 3D device was designed that functions as an red-green-blue (RGB) color splitter and polarizer, scaled up to microwave frequencies 16 .…”