MOF-Integrated Ultra-Broadband Nanoantennas for Machine-Learning-Enabled VOC Gas Identification

Current commercial volatile organic compound (VOC) gas sensors can only detect the total amount of gases and cannot identify them, slowing down the development of some emerging applications such as early healthcare diagnosis. Herein, we, for the first time, integrate metal-organic frameworks (MOFs) with ultra-broadband plasmonic nanoantennas, for machine-learning-augmented multiple VOC gas identification. With a large full width at half maximum (FWHM= 4.69 μm), our platform can detect IR fingerprints of all VOC gases located in 6-12 μm. Despite the overlapping IR spectra of molecules, with the help of machine learning, our method can accurately identify different VOC molecules (100% accuracy in our demonstration). Our work provides a methodology for VOC gas identification and toolkits for early healthcare diagnosis.