Gas phase multicomponent detection and analysis combining broadband dual-frequency comb absorption spectroscopy and deep learning Academic Article uri icon


  • AbstractIn absorption spectroscopy, analysis of multicomponent gas mixtures becomes challenging when absorption features overlap (blended spectra). Here we propose a gas sensor which can accurately identify the species and retrieve the concentrations of components in a gaseous mixture in a broad spectrum. The sensor integrates a mid-infrared dual-frequency comb laser source for spectrum acquisition and a deep learning algorithm for spectral analysis. The sensor was tested on gas phase mixtures of methane, acetone and water vapor. A prototype sensor was assessed in realistic scenarios in real time. We also systematically analyzed and presented explicit visualizations to explain the underlying working mechanism of the algorithms.

published proceedings

  • Communications Engineering

author list (cited authors)

  • Tian, L., Xia, J., Kolomenskii, A. A., Schuessler, H. A., Zhu, F., Li, Y., ... Zhang, S.

complete list of authors

  • Tian, Linbo||Xia, Jinbao||Kolomenskii, Alexandre A||Schuessler, Hans A||Zhu, Feng||Li, Yanfeng||He, Jingliang||Dong, Qian||Zhang, Sasa