$1.5M NSF Grant to Develop New Photonic Materials
MIE/ECE Associate Professor Yongmin Liu and ECE Assistant Professor Sunil Mittal, in collaboration with Professor Wenshan Cai from Georgia Institute of Technology and Professor Alexandra Boltasseva from Purdue University, are leading a $1,500,000 NSF DMREF grant for “Accelerating the Design and Development of Engineered Photonic Materials based on a Data-Driven Deep Learning Approach.”
DMREF is the principal NSF program responsive to the National Science and Technology Council’s (NSTC’s) Office of Science and Technology Policy (OSTP) Subcommittee on the Materials Genome Initiative (MGI). DMREF seeks to foster the design, discovery, and development of materials to accelerate their path to deployment by harnessing the power of data and computational tools in concert with experiment and theory. DMREF emphasizes a deep integration of experiments, computation, and theory; the use of accessible digital data across the materials development continuum; and strengthening connections among theorists, computational scientists (including data scientists), and experimentalists as well as those from academia, industry, and government.
Abstract Source: NSF
Advances in photonic materials that generate, process, or detect light can transform diverse areas of science and engineering, including lasers, optical fiber communications, augmented and virtual reality displays, solar energy harvesting, as well as quantum computing and sensing. By rationally engineering the composition and/or structure of materials at various length scales, it is possible to dramatically enhance their optical responses and performance. However, the traditional approach for the discovery and development of new photonic materials relies on trial-and-error and case-by-case explorations, which are often time consuming and ineffective. This project will use advanced artificial intelligence techniques to develop new artificial photonic materials that can be engineered to have prescribed properties and surpass naturally occurring materials. The research seamlessly integrates materials science, photonics, engineering, physics and artificial intelligence. In tandem with research, the team will develop a multi-channel education program to enhance the learning experience of a broad spectrum of the society, and prepare the next-generation workforce and technology leaders.
Technical Description:
The project aims to accelerate the pace of the discovery, design, and implementation of new engineered photonic materials, particularly photonic metamaterials, with user-defined spatial, spectral, linear, non-linear and quantum properties through a data-driven approach. This approach will consolidate properties of constituent material compositions, their geometric structures spanning atomic length scales to micrometers, and their underlying symmetries and topology. The project consists of three research thrusts, including (1) establishing deep learning frameworks to construct photonic metamaterials with high efficiency and accuracy; (2) integrating information on the tailorable optical properties of the constituent material platform into deep learning models, to benefit the design and development of reconfigurable metamaterials; and (3) investigating hybrid material systems that couple topological photonic structures designed by deep learning with quantum emitters and optical nonlinearities. The team will accomplish the interdisciplinary research by fusing theory, computation, deep learning, materials engineering, fabrication and experimentation in a closed-loop manner. Through the project, new fundamental knowledge and insights about the interdependent relationships among structure, properties, performance, and processing across different scales will be gained. In alignment with the Materials Genome Initiative (MGI), the project will create a comprehensive library of different artificial meta-atoms and meta-molecules and their optical responses, and eventually drive transformative applications of photonic metamaterials for classical and quantum information processing.