GeSe-Based Non-Invasive Glucose Biosensor

Flexible nanomaterial sensor for optical glucose detection using polarimetry and machine learning

Overview

This project develops a flexible, wearable biosensor platform for continuous non-invasive glucose monitoring. The sensor uses germanium selenide (GeSe) nanomaterial photodetectors to measure glucose concentration through optical polarimetry, with machine learning algorithms for real-time calibration.

Key Results

Metric	Value
Coefficient of Determination (R²)	0.94
Mean Absolute Error (MAE)	≈ 8.6 mg/dL
Localization Method	Optical Polarimetry
ML Calibration	Supervised regression model

Methodology

Sensing Principle

Glucose in biological fluids rotates the plane of polarized light. The GeSe photodetector measures this rotation with high sensitivity due to the material's anisotropic optical properties.

Evolution of the Platform

1. Generation 1: Flexible GeSe nanomaterial sensor with conventional polarimetric detection and ML calibration
2. Generation 2: Polarizer-free architecture exploiting in-plane anisotropy of 2D GeSe for chiroptical biosensing, enabling a more compact wearable form factor

Machine Learning Pipeline

• Feature extraction from raw optical signals
• Supervised regression for glucose concentration prediction
• Cross-validation against commercial glucometer reference

Publications

1. M. B. Kopp, "Flexible Nanomaterial Sensors for Non-Invasive Health Monitoring," Applied and Computational Engineering, vol. 72, pp. 45-58, 2024.

2. M. B. Kopp, "Next-Generation Polarimetric Biosensors: Machine Learning-Driven GeSe Photodetectors for Noninvasive Glucose Monitoring," Biosensors & Bioelectronics, 2025 (open access).

3. M. B. Kopp, "Exploiting In-Plane Anisotropy of 2D GeSe for Polarizer-Free Chiroptical Biosensing: Machine Learning-Enhanced Wearable Glycemic Diagnostics." (in preparation)

Awards

• S.T. Yau High School Science Award — Bronze Medal, Physics (North America Top 3), 2024
• NJSHS National Symposium — 3rd Place Poster Award (Dept. of Defense), 2024

Tools & Technologies

• Simulation: COMSOL Multiphysics (optical modeling)
• Data Analysis: Python, MATLAB
• Fabrication: Nanomaterial synthesis, flexible substrate processing
• ML Frameworks: scikit-learn, custom calibration pipelines

Repository Structure

gese-glucose-biosensor/
├── figures/            # Schematics and result visualizations
├── README.md
└── LICENSE

License

This project is licensed under the MIT License. See LICENSE for details.

Contact

Maximilian Kopp — maxkopptech.com | ORCID | Google Scholar