Machine Learning-Enhanced Modelling of Dual-Core Silver-Coated PCF-SPR Sensor for Analyte Sensing

Photonic crystal fiber (PCF)-based surface plasmon resonance (SPR) sensors have garnered significant attention for their remarkable sensitivity in detecting molecular interactions and bio-molecular binding events. This study has proposed an approach to enhance the sensitivity of dual-core silver-coated PCF-SPR sensors and modelling by integrating machine learning (ML) algorithms. This sensor stands out with its dual-core design, boasting impressive capabilities such as the highest wavelength sensitivity of 30000 nm/RIU and a 3.33×10-6 resolution. It also covers a wide refractive index (RI) range from 1.34 to 1.41, using silver as the plasmonic material. Additionally, it utilizes ML techniques, including support vector regression, random forests, and multiple-variable linear regression, to enhance modelling accuracy, particularly in predicting confinement loss and the real part of effective RI. The high sensitivity of the proposed PCF-SPR sensor makes it a strong contender for use in bio-sensing applications.