TY - GEN
T1 - Deep Learning for Novel Infrared, Millimeter wave, and Terahertz Metamaterials
AU - Padilla, Willie
AU - Rozman, Natalie
AU - Deng, Yang
AU - Peng, Rixi
AU - Malof, Jordan
N1 - Publisher Copyright:
© 2024 IEEE.
PY - 2024
Y1 - 2024
N2 - Electromagnetic metamaterials and metasurfaces possess a property termed electromagnetic similitude, permitting them to operate over a large swath of the electromagnetic spectrum from radio frequencies to visible with a simple scaling of the geometry. Additionally, metamaterials are composites meaning they can be fashioned so that their metallic or dielectric constituents can serve alternative functions, i.e. metamaterials are multi-functional. These two key metamaterial features make them ideal candidates for novel devices operating across the electromagnetic spectrum. Although design of metamaterials may often follow intuition, more complex designs or unit-cells which are are not sub-wavelength present more complex challenges. We present new deep learning designed metamaterials operating as broadband detectors of W-band radiation in the millimeter wave regime, as high quality factor (Q-factor) resonators achieving the highest Q-factor published to-date, and as engineered diffusers of infrared radiation for imaging applications. Our exploration of metamaterials across diverse regimes emphasizes the great potential and significance of the metamaterial deep learning design concept for the future of photonics.
AB - Electromagnetic metamaterials and metasurfaces possess a property termed electromagnetic similitude, permitting them to operate over a large swath of the electromagnetic spectrum from radio frequencies to visible with a simple scaling of the geometry. Additionally, metamaterials are composites meaning they can be fashioned so that their metallic or dielectric constituents can serve alternative functions, i.e. metamaterials are multi-functional. These two key metamaterial features make them ideal candidates for novel devices operating across the electromagnetic spectrum. Although design of metamaterials may often follow intuition, more complex designs or unit-cells which are are not sub-wavelength present more complex challenges. We present new deep learning designed metamaterials operating as broadband detectors of W-band radiation in the millimeter wave regime, as high quality factor (Q-factor) resonators achieving the highest Q-factor published to-date, and as engineered diffusers of infrared radiation for imaging applications. Our exploration of metamaterials across diverse regimes emphasizes the great potential and significance of the metamaterial deep learning design concept for the future of photonics.
KW - artificial intelligence
KW - deep learning
KW - metamaterial
KW - metasurface
KW - terahertz
UR - http://www.scopus.com/inward/record.url?scp=85207166410&partnerID=8YFLogxK
U2 - 10.1109/IRMMW-THz60956.2024.10697668
DO - 10.1109/IRMMW-THz60956.2024.10697668
M3 - Conference contribution
AN - SCOPUS:85207166410
T3 - International Conference on Infrared, Millimeter, and Terahertz Waves, IRMMW-THz
BT - 2024 49th International Conference on Infrared, Millimeter, and Terahertz Waves, IRMMW-THz 2024
PB - IEEE Computer Society
T2 - 49th International Conference on Infrared, Millimeter, and Terahertz Waves, IRMMW-THz 2024
Y2 - 1 September 2024 through 6 September 2024
ER -