Open Collections

Abstract

This thesis focuses on Silicon Photonics, specifically the design and optimization of basic on-chip devices using advanced methodologies. It is structured into five chapters, exploring theoretical concepts, practical design processes, and experimental results to improve the performance of Silicon Photonics. The first chapter introduces key concepts in Silicon Photonics and discusses two main devices: the Y-splitter, and the wavelength division multiplexer. These devices are central to the following chapters, with a focus on using Inverse Design. The second chapter provides the theoretical background, explaining Inverse Design principles and machine learning methods, such as gradient descent and optimization algorithms like L-BFGS-B and Adam. The chapter also covers loss functions, the Adjoint method. The third chapter outlines the design process, applying Inverse Design to devices like the 3dB Y-splitter, and combined wavelength demultiplexer/3dB splitter. It highlights how Inverse Design improves device performance and make new devices. It also present a python GUI for inversely design Y-splitters which is developed in this thesis. The fourth chapter presents experimental results and evaluates the performance of the designed devices, offering insights into their effectiveness. The final chapter summarizes the key findings, contributions, and potential future research directions in Silicon Photonics. This thesis demonstrates how inverse design can enhance silicon photonic devices and enable the creation of structures beyond manual design capabilities.