Artificial intelligence-augmented, triboelectric-induced ion mobility for mid-infrared gas spectroscopy (preprint)

ABSTRACT

As the outbreak of the Covid-19 pandemic, isopropyl alcohol (IPA) molecules played significant role as a biomarker for anti-virus diagnosis. However, conventional gas molecules detection exhibited dramatic drawbacks, like the strict working conditions of ion mobility methodology and weak light-matter interaction of mid-infrared spectroscopy, yielding a limited response of targeted molecules. We propose a synergistic methodology of artificial intelligence (AI)-augmented ion mobility and mid-infrared spectroscopy(IMMS), leveraging the complementary features from the sensing signal in different dimensions to reach superior accuracy for IPA identification. We pull in “cold” plasma discharge from triboelectric generator which improves the mid-infrared spectroscopic response of IPA with good linear prediction. Moreover, even with interferences of more than three different carbon-based gases, this synergistic methodology achieved ～99.08% accuracy for a precise gas concentration prediction. The synergistic methodology of AI-augmented IMMS creates a mechanism of gas sensing for accurate gas mixture and regression prediction in healthcare.