Home How To Use Database Optical Property DeepHL DeepEnergy Scientific Advisory

Deep4Chem: AI Optical Spectrometer

Accurate and reliable prediction of various optical properties of organic compounds is very important for developing new organic compounds with desired properties. Here, we have developed an artificial intelligence (AI) optical spectrometer, Deep4Chem, using deep learning methods and experimental database of optical properties of organic compounds in various solvents, host films, and solid states.

       Deep4Chem Citation

      "Deep learning for development of organic optoelectronic devices: efficient prescreening of hosts
      and emitters in deep-blue fluorescent OLEDs"
      Minseok Jeong, Joonyoung F. Joung, Minhi Han, Chang Woo Koh, Dong Hoon Choi*, and Sungnam Park*
      npj Comput. Mater. 8, 147 (2022) († equally contributed.) (Link)

      "Beyond Woodward–Fieser Rules: Design Principles of Property-Oriented Chromophores Based on
      Explainable Deep Learning Optical Spectroscopy"
      Joonyoung F. Joung, Minhi Han, Minseok Jeong, and Sungnam Park*
      J. Chem. Inf. Model. 62, 12, 2933-2942 (2022) († equally contributed.) (Link)

      "Deep Learning Optical Spectroscopy Based on Experimental Database: Potential Applications to Molecular Design"
      Joonyoung F. Joung, Minhi Han, Jinhyo Hwang, Minseok Jeong, Dong Hoon Choi, and Sungnam Park*
      JACS Au, 1, 4, 427–438 (2021) († equally contributed.) (Link)

      "Experimental database of optical properties of organic compounds"
      Joonyoung F. Joung, Minhi Han, Minseok Jeong, and Sungnam Park*
      Sci. Data, 7, 295 (2020) († equally contributed.) (Link)


Deep4Chem predicts 7 optical properties of a given organic compound in solvent and solid state.
(1) First absorption peak position (λabs,max) and bandwidth (σabs), Extinction coefficient (εmax),

(2) Emission peak (λemi,max) and bandwidth (σemi), Photoluminescence quantum yield (ΦQY),

(3) Fluorescence lifetime (τflu)