Abstract
Reverse iontophoresis (RI) is a promising non-invasive, wearable technology for the transdermal extraction of interstitial fluid (ISF), which contains rich biomarkers relevant to health status. Despite the advancement of wearable sensors, this technology is still restricted for accurate non-invasive biomarkers detection. The main challenge lies in the instability of ISF extraction during RI. We found that this instability is primarily caused by the skin surface pH variations because of the interaction between the RI-induced H(+) movement and the skin recovery ability. Here, we investigated how the skin surface pH affected RI, theoretically and experimentally; and developed a wearable device and a calibration method to enable accurate non-invasive ISF glucose detection, accordingly. The result showed that glucose prediction accuracy was markedly improved, with mean absolute relative difference (MARD) decreased from 34.44% to 14.78% across both healthy and diabetic volunteers.