Theses Doctoral

Next-Generation Metasurface Applications Powered by Empirical Designs and Machine Learning

Huang, Xiaoyan

Metasurfaces have presented themselves as next generation optical platforms with an unprecedented capability to exert designer amplitude, phase, and polarization control on incoming electromagnetic (EM) waves. Existing works have focused on demonstrating simple optical functionalities (lensing, beam steering, holography) at longer wavelengths (microwave, infrared), and the design methodology has largely been empirical.

In this work, we demonstrate next generation metasurface applications in near infrared and visible wavelengths. The purpose of such applications evolves from imaging to complex machine vision applications, and as such calls for a updated design paradigm that combines traditional, empirical based methods with modern inverse design tools based on machine learning. We expand the potential of metasurfaces by demonstrating their interdisciplinary applications in complex imaging, quantum optics and optical computation. A homebrew fabrication and testing pipeline is developed to support the challenging mission of near infrared and visible frequency usage.

Furthermore, we propose a new design paradigm that combines physics informed intuitions with modern machine learning to simulate and design metasurface in an accurate and time-efficient way. In conclusion, I will discuss the outlook of metasurfaces in real-world applications, whose unique combination of performance and form factor make them ideal candidates for next-generation optical devices.

Files

This item is currently under embargo. It will be available starting 2025-07-18.

More About This Work

Academic Units
Applied Physics and Applied Mathematics
Thesis Advisors
Yu, Nanfang
Degree
Ph.D., Columbia University
Published Here
August 14, 2024