Mid-Infrared Photoacoustic Spectroscopy for Non‑Invasive Glucose Monitoring: Design, Skin Phantom Development, and Numerical Modelling of PA-Tissue Interactions

Abstract

Non-invasive glucose monitoring research and development has gained increasing attention over the past two decades, driven by the rising diabetic population and the limitations of conventional glucose monitoring approaches. Photoacoustic spectroscopy has emerged as a promising technique for painless glucose sensing. This thesis explores the use of dual-wavelength (9.25 μm and 10.3 μm) mid-infrared quantum cascade lasers in a photoacoustic spectroscopy system for glucose monitoring. Through improvements to the developed system and the design of a more systematic evaluation using controlled phantom fabrication and measurement conditions, the proposed approach demonstrated more consistent results, with a 75% reduction in frequency shift compared with previous studies. The fabricated skin phantoms exhibited close similarity to human skin in both mechanical and acoustic properties. With the implementation of a machine learning regression model, promising performance was achieved, with 100% of predictions falling within Zone A and B of the Clarke Error Grid. The agreement between the experimental measurements and the COMSOL photoacoustic simulation further supports the validity of both the fabricated skin phantoms and the developed mid-infrared photoacoustic spectroscopy system. Overall, this research demonstrates the potential of a non-invasive photoacoustic glucose monitoring system with an improved evaluation methodology, providing a strong foundation for future in vivo studies.