Abstract
This study proposes a novel, highly sensitive surface plasmon res-onance (SPR) sensor based on a hexagonal photonic crystal fibre(PCF) sensor, tailored for analytes’ refractive index (RI) detectionfrom 1.30 to 1.43. The sensor’s optical characteristics are analysedusing finite element method (FEM) simulations. The proposed designachieves a peak wavelength sensitivity of 29,000 nm/RIU and anamplitude sensitivity of 2,653.68 RIU⊃1. Additionally, a high reso-lution of 3.45 × 10 RIU underscores its effectiveness in capturingminute variations in refractive index. We conducted a compara-tive evaluation of ten machine learning algorithms for predictingconfinement loss and amplitude sensitivity. Our results show thatensemble methods, particularly Extra Trees, Random Forest, GradientBoosting, and XGBoost, achieve exceptionally high prediction accu-racy for confinement loss (R2 > 0.999), while LightGBM outperformsother models for amplitude sensitivity prediction (R2 = 0.9448). Theproposed sensor is stable and suited for detecting analytes for foodsafety, and adulteration monitoring applications.
| Original language | English |
|---|---|
| Pages (from-to) | 1824-1862 |
| Number of pages | 39 |
| Journal | Taylor & Francis Journal of Electromagnetic Waves and Applications |
| Volume | 40 |
| Issue number | 11 |
| Early online date | 15 Apr 2026 |
| DOIs | |
| Publication status | E-pub ahead of print - 15 Apr 2026 |
UN SDGs
This output contributes to the following UN Sustainable Development Goals (SDGs)
-
SDG 6 Clean Water and Sanitation
Keywords
- Plasmonic sensors
- multi-purpose sensor platform
- Gas sensor
- sulfur dioxide (SO₂) gas
- machinelearning
- foodmonitoring
Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver