SPR-based refractive index sensor design with grated Au-ZnS for dengue detection using machine learning

Area of Science:

• Photonics and Optics
• Nanotechnology
• Biomedical Sensing

Background:

• Surface Plasmon Resonance (SPR) sensors are crucial for label-free biosensing.
• Fiber optic SPR sensors offer advantages in miniaturization and remote sensing.
• Developing highly sensitive and specific biosensors is essential for early disease detection.

Purpose of the Study:

• To propose and investigate a novel SPR fiber optic refractive index (RI) sensor.
• To optimize the sensor design for maximum wavelength sensitivity (WS).
• To evaluate the sensor's capability for detecting the dengue virus and its integration with machine learning.

Main Methods:

• Fabrication of a multi-mode fiber (MMF) SPR sensor with a bi-metallic nanostructure (Gold/Zinc Sulphide).
• Systematic evaluation of design parameters to maximize wavelength sensitivity.
• Application of machine learning (ML), specifically a neural network (NN) model, for data analysis and prediction.
• Performance evaluation using metrics like RI sensitivity, Detection Accuracy (DA), Figure of Merit (FOM), and signal-to-noise ratio (SNR).

Main Results:

• The SPR sensor exhibited non-linear RI sensitivity ranging from 9000 to 18,000 nm/RIU.
• Achieved a highest WS of 14285.71 nm/RIU for detecting dengue virus in infected haemoglobin.
• The NN model demonstrated superior performance with a mean square error (MSE) of 0.2828 and R² of 0.9998.
• Key performance metrics included maximum DA of 14.285 μm⁻¹, FOM of 158.73 RIU⁻¹, SNR of 5.55, and QF of 79365.055 nm/RIU.

Conclusions:

• The proposed SPR fiber optic sensor with a bi-metallic nanostructure is highly effective for RI sensing.
• Integration of ML, particularly NN, significantly enhances sensor performance in prediction and classification.
• The developed biosensor shows great potential for sensitive and accurate biological applications, including disease detection.