Machine learning-driven terahertz graphene-MXene-gold metasurface biosensor for dual COVID-19 and cervical cancer biomarker detection

作者：Jacob Wekalao, Hussein A. Elsayed, Haifa A. Alqhtani, Mostafa R. Abukhadra, Ahmed Mehaney, Amuthakkannan Rajakannu · 发表于：Sensing and Bio-Sensing Research · 年份：2025 · DOI：10.1016/j.sbsr.2025.100920 · 被引用次数：2 · 研究领域：Terahertz technology and applications、Millimeter-Wave Propagation and Modeling、Metamaterials and Metasurfaces Applications

This study introduces a terahertz (THz) biosensor designed for rapid and precise detection of COVID-19 viral proteins and cervical cancer biomarkers. The device uses a metasurface structure composed of U-shaped MXene-coated resonators and L-shaped gold-coated elements on a graphene-functionalized circular platform built on a silicon dioxide substrate. It operates across 0.308–0.312 THz for COVID-19 detection and 0.799–0.809 THz for cervical cancer screening, achieving sensitivities of 200 GHz/RIU and 300 GHz/RIU. The corresponding figures of merit range from 2.222 to 2.667 RIU −1 for COVID-19 and 4.848–5.455 RIU −1 for cervical cancer, with quality factors above 14 for the latter. A one-dimensional convolutional neural network model optimized the sensor's response, reaching up to 95 % prediction accuracy and showing a strong linear correlation between resonance frequency and refractive index (R 2 > 0.99). The architecture can be produced using standard lithography and thin-film deposition methods compatible with existing microelectronic processes. The design allows biofunctionalization with target proteins such as SARS-CoV-2 spike antigens and HPV-related biomarkers. Simulation results align with recent experimental findings in THz biosensing, supporting the feasibility of laboratory testing and clinical application.