Prediction of Anomalous Variation in GaN-based Chemical Sensors

Abstract

We present the result of investigations into anomalous drift behavior in AlGaN/GaN-based chemical sensors; a problem that must be addressed in such sensors to achieve their great promise as a portable sensor technology. Leakage current measurements undertaken on AlGaN/GaN-based sensor devices were fitted with a parallel diode model. Some sensors were then destructively tested using phosphoric acid etchant to decorate the defects, allowing the density of conductive screw dislocations to be quantified. A strong correlation between diode saturation current and defect density was then established. The remaining devices were encapsulated as working sensors and left in an ionic solution to determine the normalized root mean square error as a measure of drift. This also exhibited a strong correlation with the saturation current, thus establishing a link between semiconductor defect density, leakage current, and uncorrelated drift over time for AlGaN/GaN chemical sensors.