High-Sensitivity Terahertz Gas Sensor Using Graphene-Enhanced Metasurfaces with Machine Learning Optimization
摘要
This paper presents a terahertz gas sensor design incorporating graphene-enhanced metasurfaces with gold and silver resonators. The sensor architecture consists of a central square gold resonator surrounded by a silver square ring resonator on SiO₂ substrate. The sensor demonstrates excellent sensitivity of 200 GHzRIU−1 across a refractive index range of 1.00–1.07 RIU, with consistent quality factors ranging from 9.525 to 9.250. Operating in the 0.1–0.4 THz frequency range, the sensor exhibits tunable transmission characteristics through graphene chemical potential modulation and incident angle variation. Polynomial regression analysis validates the sensor's performance, achieving prediction accuracies of up to 100% and R2 values between 90–100%. The integration of machine learning optimization enhances the sensor's reliability and predictive capabilities, making it particularly suitable for real-time gas detection applications.