ABSTRACT Long‐wave infrared (LWIR) spectroscopy is indispensable for molecular fingerprinting in chemical analysis and environmental monitoring via thermal radiation. The recent advancements in miniaturized spectrometers in the visible and near‐infrared spectra have witnessed rapid and significant progress. However, fully miniaturized LWIR spectrometers of high performance, high efficiency, and low cost are still rare. Here, a snapshot computational spectrometer is proposed and demonstrated in the LWIR wavelength range. Consisting of a plasmonic metasurface array with only a few elements and a commercial uncooled thermal camera, the spectrometer can reconstruct spectra within a broad LWIR band by leveraging the compressed sensing algorithm together with the dictionary learning technique. The results of narrowband spectral reconstruction show a high wavelength accuracy, with a relative error of less than 1% and a resolution down to 110 nm for most LWIR wavelengths. Broadband spectral reconstruction across the entire LWIR band has been successfully demonstrated for radiation from a globar source and the absorption of some typical chemical analytes, including ethanol and three types of polymer films. The proposed spectrometer may hold profound potential to advance research in mid‐infrared spectroscopy and find wider applications.
Lin et al. (Sat,) studied this question.