3D-Printed Frequency-Diverse Metasurface for Camera-Based Terahertz Frequency Analyzer
Date
2025
Authors
Quader, S.
Abdullah, M.
He, E.
Fumeaux, C.
Withayachumnankul, W.
Editors
Advisors
Journal Title
Journal ISSN
Volume Title
Type:
Journal article
Citation
IEEE MICROWAVE AND WIRELESS TECHNOLOGY LETTERS, 2025; 35(12):2015-2108
Statement of Responsibility
Sakib Quader, Mariam Abdullah, Estrid He, Christophe Fumeaux, Withawat Withayachumnankul
Conference Name
Abstract
In wireless communications, accurate frequency estimation is essential to support reliable demodulation, channel management, and spectrum surveillance. At terahertz frequencies, the higher carrier values and narrow channel spacing demand much finer frequency estimation, which requires high-resolution frequency discrimination. Conventional grating-based approaches are bulky and limited in resolution, making them unsuitable for compact terahertz systems. In this letter, we present a 3D-printed pseudo-random dielectric metasurface that enables subgigahertz frequency discrimination in the 220–330-GHz band through spatially diverse near-field patterns. The design leverages frequency-dependent scattering to create unique intensity distributions captured by a terahertz camera. We demonstrate that these spatial signatures can be used for frequency discrimination by training a convolutional neural network (CNN) to identify the frequency from a single image. To substantiate the reliability of the metasurface response, we employ multiple beam configurations with varying incidence angles in our experimental setup. The results demonstrate high classification accuracy over the operational range, thus corroborating the metasurface as a viable passive frequency analyzer for terahertz communications.
School/Discipline
Dissertation Note
Provenance
Description
Access Status
Rights
© 2025 IEEE. All rights reserved, including rights for text and data mining, and training of artificial intelligence and similar technologies. Personal use is permitted, but republication/redistribution requires IEEE permission.