Next-generation plasmonic hybrid graphene biosensor: Machine learning-assisted development for efficient detection of chikungunya virus

Abstract

Early detection of Chikungunya virus (CHIV) is vital for effective disease management, as the infection has been reported in nearly sixty countries worldwide. CHIV infection typically causes fatigue, muscular pain, joint swelling, and severe headaches, necessitating reliable and highly sensitive diagnostic approaches. This research presents the design and optimization of a novel double square bracket-shaped resonator refractive index biosensor (DSBRRIB) to enhance the efficiency and precision of chikungunya detection in virus-infected blood plasma and red blood cells (RBCs). The silicon substrate used is 8000 nm by 8000 nm in dimension with a conventional thickness of 500 nm, and it measures within the wavelength range of 1950 to 2100 nm. The proposed biosensor leverages graphene-based plasmonic structures to improve sensitivity, stability, and overall performance. Targeting specific blood components adds diagnostic specificity, ensuring higher accuracy in distinguishing natural and affected samples. A comparative analysis of graphene materials with varying layers and resonator configurations was conducted to determine optimal detection efficiency. The optimized biosensor achieved quality factor value of 3196.33 nm/RIU, with corresponding figures of merit value of 487.64 for chikungunya detection. Detection limits were as low as 0.000436 RIU for plasma and 0.000379 RIU for RBCs. Machine learning-assisted optimization further enhanced performance, with the best predictive testing accuracy reaching 0.099319 for bracket gap variations.