In the domain of high-temperature semiconductor arrays for electronic noses (e-noses), Metal Oxide Sensors (MOS) have a pivotal role despite their non-linear response to chemical vapors. A prevalent approach to enhance the identification algorithm's performance involves implementing mathematical models during the MOS signal processing. However, certain models rely solely on mathematical goodness-of-fit, overlooking crucial features that render practical e-nose applications ineffective. This paper introduces a theoretical model for the qualitative analysis of MOS signals, focusing on two primary diffusion processes: analyte migration to the sensor's surface and the subsequent dispersion of some of these molecules within the MOS bulk. Additionally, this work discusses a model validation using an ad-hoc e-nose, built with SnO2 gas sensors, and six organic chemicals, detailing main data processing steps. Finally, disclosed results showcase a high success rate for Triacetone Triperoxide (TATP) identification, one of the most significant threats among homemade explosives (HME). The presented conclusions underscore the enhanced efficacy of the proposed signal model for e-nose vapors identification and its practical utility in strengthening pre-emptive HME identification to enhance public safety.
