ML‐Assisted Accuracy Prediction of a Silicon Carbide–Based SPR Biosensor for Acetone Concentration Measurement

ABSTRACT An enhanced surface plasmon resonance biosensor is designed for acetone concentration measurement in exhaled breath (EB) and exhaled breath condensate (EBC) using the Kretschmann configuration and the Fresnel multilayer concept. After analyzing different plasmonic metals and substrate materials, it is concluded that for realizing deeper resonance and greater angular sensitivities, for acetone measurement in EB and EBC, a CsF prism is more optimal as a substrate material. Black phosphorus is employed in this design for its potential to act as a selective layer through chemisorption, which fully relies on its extremely active two‐dimensional surface and dipole interaction effect to boost permittivity variation at the metal–dielectric interface. Finally, simulation design is ensured to be accurate through simulation using various methods: optimization of sensor design through the use of Transfer Matrix Method (TMM) and Finite Element Method (FEM), and verification of resonance and electrical field simulation through Finite Difference Time Domain (FDTD) simulation. The designed optimized structure reveals a high sensitivity of 332.53 deg/RIU, a figure of merit of 115.78, a quality factor of 126.47 RIU −1 , and a minimum reflection of 0.0002. In addition, the validity of the accuracy of Artificial Neural Network (ANN) and Adaptive Neuro‐Fuzzy Inference System (ANFIS) models in describing the simulation of sensor response has been verified for a dependable precise measurement of acetone concentration. This designed SPR biosensor unveils a vast improvement over existing designs with immense potential for utilization in noninvasive acetone breath analysis.