Multi-modality analysis of glucose aqueous solution using photoacoustic and dielectric spectroscopy for non- invasive glucose monitoring

Quantitative analysis of glucose using conventional optical spectroscopy suffers from a lack of repeatability due to high optical scattering in skin tissue. Here we present a multi-modality analysis of glucose aqueous solution using photoacoustic spectroscopy (PAS) and broadband dielectric spectroscopy (BDS). These techniques involve the direct detection of the acoustic and electromagnetic waves propagating through or reflecting from tissue without their being scattered. They therefore have potential for better tolerance to the variation of scattering. For PAS, to differentiate signals induced by water absorption, we select another laser wavelength (1.38 mu m) that exhibits the same absorbance for water at 1.61 mu m. Furthermore, one of the two photoacoustic signals is used to normalize the variations of acoustic properties in differential signal. Measured results for glucose solutions (0.2 g/dL) showed that the differential signal has a sensitivity of 1.61%/g. dL(-1) and a detection limit of 120 mg/dL. We also tested glucose detection with BDS (500 MHz to 50 GHz) by detecting glucose hydration bonding at around 10-20 GHz. Using a partial least square analysis and first derivation on broadband spectra, we obtained an RMS error 19 mg/dL and a detection limit of 59 mg/dL. Using both the low-scattering ultrasonic and microwave detection techniques, we successfully captured the glucose footprint in the physiological range.