*********************************
There is now a CONTENT FREEZE for Mercury while we switch to a new platform. It began on Friday, March 10 at 6pm and will end on Wednesday, March 15 at noon. No new content can be created during this time, but all material in the system as of the beginning of the freeze will be migrated to the new platform, including users and groups. Functionally the new site is identical to the old one. webteam@gatech.edu
*********************************
Title: High-Efficiency Topology Optimization for Very Large-Scale Integrated-Photonics Inverse Design
Committee:
Dr. Stephen Ralph, ECE, Chair, Advisor
Dr. Anthony Yezzi, ECE
Dr. Wenshan Cai, ECE
Dr. John Cressler, ECE
Dr. Steven Johnson, MIT
Abstract: This work establishes the mathematical and algorithmic framework necessary for a large-scale, photonics inverse-design methodology that is fully compatible with (and experimentally validated on) commercial, semiconductor-foundry platforms. Specifically, this new methodology quickly and efficiently designs high-performing, multi-functional devices and systems that operate despite various sources of fabrication or operating variability. By overcoming typical tradeoffs between design dimensionality, device footprint, functional complexity, computational cost, and realizable performance, this work paves a practical and proven path toward very large-scale integrated photonics (VLSIP), a key step in designing interferometrically stable architectures within fields like quantum computing, machine learning, and even augmented reality. First, this work introduces the field of photonic inverse design within the context of high-yield photonic integration, highlighting fundamental challenges that continue to impede long-term scalability. The algorithmic framework for a high-efficiency, hybrid time-/frequency-domain adjoint solver, along with comprehensive manufacturing constraints, are then presented. The practicality of this new framework is tested by designing numerous compact, broadband, and robust devices, such as polarization splitters and rotators, full-aperture grating couplers, and 90-degree optical hybrids, all of which were fabricated and tested on different commercial foundry platforms. Finally, these individual devices were monolithically integrated to form a Stokes-Vector Receiver (SVR), a high-capacity direct-detect communications system amenable to long-haul signal processing algorithms. The ultra-compact SVR demonstrates reliable performance across the entire C-band, validating the notion that this new photonics design paradigm is not only compatible with large-scale commercial foundries, but yields high-performing systems robust to common forms of fabrication and environmental variability.
