Next-generation polarimetric biosensors: machine learning–driven GeSe photodetectors for noninvasive glucose monitoring
Year: 2025
Type: article
Abstract: Accurate and painless glucose monitoring remains a critical need in diabetes management. Here we present a compact, polarization-sensitive biosensing platform that integrates anisotropic Germanium selenide (GeSe) photodetectors with machine-learning calibration for non-invasive glucose detection. Flexible GeSe devices fabricated on polyethylene-terephthalate substrates convert glucose-induced optical rotation directly into electrical signals without external polarizers. In-vitro tests across 50–300 mg dL−1 glucose demonstrated linear phase sensitivity of 1.59 × 10−2 ° (mg dL−1) −1 and a detection limit of ≈12 mg dL−1 (R2 = 0.96). Preliminary in-vivo reflectance measurements from a fingertip confirmed feasibility of needle-free operation with R2 = 0.91 versus glucometer readings. Gradient-boosting regression applied to polarization-photocurrent data improved prediction accuracy by > 20 %, achieving MAE = 8.6 mg dL−1 and RMSE = 11.2 mg dL−1. These findings indicate that GeSe-based polarimetric photodetectors, enhanced by data-driven modeling, offer a practical route toward compact, intelligent, and painless glucose-monitoring technologies.
Source: Biosensors and Bioelectronics X
Author: Maximilian B. Kopp
Language: English
Subfield: Biophysics
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $2,490
OpenAlex ID: W4417266028