Terahertz Plasmonic Biosensor Leveraging Ag-Au-Graphene Heterostructures for Quantitative Hemoglobin Analysis with Machine Learning Algorithms for Performance Optimization

Abstract

This investigation presents the design, simulation, and performance analysis of a terahertz-based biosensor for hemoglobin detection. The sensor architecture incorporates a synergistic combination of graphene, gold, and silver metasurfaces in a hierarchical resonator structure. Extensive parametric analysis was conducted to optimize the sensor's performance characteristics. The optimized sensor demonstrates high sensitivity, achieving up to 1000 GHzRIU⁻¹, with a figure of merit of 3.289 RIU⁻¹. Experimental results indicate effective detection of hemoglobin concentrations ranging from 10 to 40 g/L, corresponding to refractive indices between 1.34 and 1.43. Electromagnetic field distribution analysis exemplifies peak absorption at 0.65 THz. Furthermore, the sensor’s potential for binary encoding applications was evaluated with remarkable performance. Machine learning optimization, employing a decision tree regressor, demonstrates an optimal R² score of 100% across various parameter combinations, suggesting potential for the development of accurate sensing systems. The proposed sensor design represents a significant advancement in terahertz biosensing technology, with implications for enhanced medical diagnostics and biomedical research applications.