Dual-band graphene-assisted metamaterial absorber with machine learning integration for high-sensitivity THz biosensing

In this work, a compact dual-band terahertz (THz) metamaterial absorber integrated with graphene is presented for high-performance biosensing applications. The structure is composed of four concentric octagonal rings, where two are metallic and two are graphene-based, supported on a SiO2 substrate backed by a gold ground plane. The dual-resonant response is achieved through the metallic rings, while the graphene layers enhance absorption and tunability by improving impedance matching with free space. The absorber demonstrates near-unity absorption at 0.9623 THz and 1.4079 THz, with maximum sensitivities of 1.7435 THz/RIU and 1.475 THz/RIU, and corresponding high figures of merit and quality factors, making it highly suitable for precise biosensing. A detailed parametric analysis confirms the influence of structural dimensions on absorption efficiency, while the equivalent circuit model validates the underlying resonance mechanisms. Furthermore, machine learning models are employed to predict absorption characteristics and support biosensing tasks such as virus detection, thereby reducing the need for repeated full-wave simulations. The proposed design effectively combines compact geometry, high sensitivity, excellent selectivity, and ML-assisted optimization, positioning it as a versatile candidate for next-generation THz biosensing and biomedical diagnostics.