Deep Learning Assisted Design of a Plasmonic Refractive Index Sensor Based on Photonic Crystal Fiber

A sensor for identification of liquid Analytes exhibiting a detection range spanning from refractive index of 1.33 −1.36 has been proposed in this work. The sensor has been designed utilizing a circular coating of gold layer over a Photonic Crystal Fiber (PCF) with symmetrical structure for facilitating the Surface Plasmon Resonance (SPR) phenomenon. The design was numerically explored using the finite element method, while the deep learning model utilized the generated dataset to forecast the sensor's operational characteristics. The prediction technique provided a low mean bias error of less than 2% and was substantially faster than traditional techniques. It achieved a maximum amplitude sensitivity of 357.05 RIU⁻¹ and a maximum wavelength sensitivity of 15000 nm/RIU. The sensor's wavelength resolution was determined to be 7.143×10⁻⁶ RIU. It demonstrates remarkable sensitivity in detecting a wide range of chemical and biological compounds, making it suitable for diverse sensing applications across multiple domains. Due to the enhanced sensing performance and application of deep learning approach, the proposed sensor can serve as a cost-effective and time-saving alternative to existing sensors, eliminating the need for costly and time-consuming simulations.