Machine Learning Augmented VOC Identification by Mid-Infrared Nanoantennas with Microfluidics Chambers

Abstract

Infrared vibrational spectroscopy enhanced by plasmonic nanoantenna (PNA) become attractive sensing platforms for molecular recognition and characterization. With the rich information provided by IR fingerprint absorption and wavelength shift by broadband PNA, machine learning serves as a powerful tool to classify the complementary physical (refractive index) and chemical (chemical bond vibration) information and recognize the molecules. We propose an ultrasensitive broadband hook nanoantenna platform integrated with microfluidics channels and perform machine learning analysis to identify chemically similar alcoholic molecules in low concentrations. Our platform paves the way to advanced IR spectroscopic sensing for a wide range of chemical identification.