Computing Confinement Loss of Open-Channels Based PCF-SPR Sensor with ANN Approach

Surface plasmon resonance (SPR) is a sensitive spectroscopic technique for measuring changes in the refractive index (RI) of a medium in contact with a sensor surface. Confinement loss computation is crucial for designing an SPR-based photonic crystal fiber (PCF) sensor as it directly impacts its performance and sensitivity. The study highlights the simple and fast-training feed-forward Artificial Neural Network (ANN) as a technique for calculating the confinement loss of an open-channels-based PCF-SPR sensor. Moreover, ANN demonstrates proficiency in generating precise predictions for the aforementioned optical properties within the conventional parameter range, encompassing wavelengths between 0.60 μm and 0.90 μm, as well as an analyte RI range spanning from 1.33 to 1.40. Additionally, this work demonstrates a higher speed in predicting output values for varying sensor structural parameters than direct numerical simulation techniques. This study signifies a noteworthy progression in the field through the utilization of ANN-driven optimization methods for integrated silicon photonics devices. This advancement leads to improved speed and accuracy in nredicting sensor outputs.