Hybrid Ag/Graphene/Cu/Al/WS₂ terahertz metasurface biosensing platform for high-sensitivity refractive-index sensing toward wastewater monitoring using machine learning

Abstract Rapid and reliable monitoring of organic contaminants in wastewater requires sensitive sensing platforms capable of resolving small changes in the local dielectric environment. A hybrid Ag/Graphene/Cu/Al/WS₂ asymmetric metasurface is proposed, exploiting plasmonic–excitonic coupling and Fano-type resonance to enhance near-field confinement and refractive-index sensitivity. The 2 × 2 resonator array was numerically investigated using COMSOL Multiphysics over the 0.4–1.4 THz range, with graphene chemical potential, incident angle, resonator geometry, and analyte refractive index systematically varied. The optimized sensor exhibits resonance behavior around 0.74–0.76 THz and achieves a maximum local sensitivity of 909 GHz/RIU and a figure of merit of 17.483 RIU⁻¹ over an RI range of 1.3256–1.542 RIU. Peak transmission remains above 96.7% up to an incidence angle of 80°, while the resonance frequency shows a linear dependence on refractive index with R² = 0.90245. The Random Forest model achieves R² ≥ 0.999735 for transmission prediction. These results establish the proposed hybrid metasurface as a tunable and broad-angle platform for terahertz refractive-index sensing toward wastewater monitoring applications.