Compact Chemical Identifier Based on Plasmonic Metasurface Integrated with Microbolometer Array

• Published in: Laser & photonics reviews Vol. 16; no. 4
• Main Authors: Meng, Jiajun, Weston, Luke, Balendhran, Sivacarendran, Wen, Dandan, Cadusch, Jasper J., Rajasekharan Unnithan, Ranjith, Crozier, Kenneth B.
• Format: Journal Article
• Language: English
• Published: Weinheim Wiley Subscription Services, Inc 01.04.2022
• Subjects: Algorithms; chemical sensors; Chemicals; Cyclohexane; Environmental monitoring; Ethyl lactate; Food; Fourier transforms; FTIR spectrometers; Infrared spectra; Infrared spectrometers; infrared spectroscopy; Machine learning; machine learning classification; Metasurfaces; microspectrometers; plasmonics; Weight reduction
• ISSN: 1863-8880; 1863-8899
• DOI: 10.1002/lpor.202100436

The identification of chemicals from their mid‐infrared spectra has applications that include industrial production of chemicals, food production, pharmaceutical manufacturing, and environmental monitoring. This is generally done using laboratory benchtop tools, such as the Fourier transform infrared spectrometer. Although such systems offer high performance, alternative platforms offering reduced size, weight, and cost can enable a host of new applications, e.g. in consumer personal electronics. Here a compact microspectrometer platform for chemical identification, comprising a mid‐infrared metasurface integrated with a lightweight (≈1 g) and very small (≈1 cm3) microbolometer‐based thermal camera is experimentally demonstrated. A machine learning algorithm is trained to analyze the microspectrometer output and classify chemicals based on their mid‐infrared fingerprints. High accuracy identification of four liquid chemicals, concentration quantification of ethyl lactate in cyclohexane down to subpercentage levels, and the classification of food and drug samples is demonstrated. A compact mid‐infrared microspectrometer platform for chemical identification is demonstrated. It consists of a plasmonic metasurface integrated with a thermal camera. Machine learning algorithms are implemented to train classifiers based on the generated thermal images. The high accuracy identification of four liquid analytes, the concentration quantification of an organic solvent, and the classification of food and drug samples are demonstrated.