Machine-learning-assisted multilayer graphene–silver–ZrN surface plasmon resonance biosensor for high-sensitivity hemoglobin detection

Abstract

This work presents a theoretically optimized multilayer surface plasmon resonance (SPR) biosensor for quantitative hemoglobin detection using the Kretschmann configuration. The sensor integrates a BK-7 prism, silver plasmonic layer, graphene enhancement layer, zirconium nitride (ZrN) protective layer, and aqueous sensing medium. This architecture synergistically combines enhanced electromagnetic confinement with chemical stability, addressing silver's oxidation vulnerability while maintaining superior plasmonic performance. Electromagnetic analysis via transfer matrix method and finite element simulations demonstrates exceptional sensitivity metrics: maximum angular sensitivity of 500°/RIU, figure of merit of 92.25 RIU⁻¹, and detection limit of 0.006 RIU across clinically relevant hemoglobin concentrations (10–40 g/L). Localized electric field enhancement (~10⁶ V/m) at the sensing interface confirms optimal light-matter interaction amplification. Machine learning models predict sensor responses to graphene thickness and refractive index variations with R² > 0.99, enabling rapid optimization. This design advances SPR biosensor technology for sensitive, label-free biochemical detection applications.