Accelerated Discovery of Potential Organic Dyes for Dye‐Sensitized Solar Cells by Interpretable Machine Learning Models and Virtual Screening. Issue 6 (24th April 2020)

• Record Type: Journal Article
• Authors: Wen, Yaping; Fu, Lulu; Li, Gongqiang; Ma, Jing; Ma, Haibo
• Is Part Of: Solar RRL. Volume 4:Issue 6(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-04-24
• Subjects: dye-sensitized solar cells -- machine learning -- structure–property relationships -- synthetic accessibility -- virtual screening; Solar energy -- Periodicals; Photovoltaic power generation -- Periodicals; Solar energy -- Research -- Periodicals; Photovoltaic power generation -- Research -- Periodicals; Periodicals; 333.7923
• DOI: 10.1002/solr.202000110
• Languages: English
• ISSNs: 2367-198X

Abstract

Abstract : The development of highly efficient dye‐sensitized solar cells (DSSCs) is greatly hindered by the lack of a reliable and understandable quantitative structure–property relationship (QSPR) model. Herein, an accurate, robust, and interpretable QSPR model is established by combining the machine learning technique and computational quantum chemistry, and with this model, virtual screening as well as the assessment of synthetic accessibility is performed to identify new efficient and synthetically accessible organic dyes for DSSCs. Finally, eight promising organic dyes with high power conversion efficiency and synthetic accessibility are screened out from ≈10 000 candidates. Meanwhile, the interpretability of the model is used for deducing reasonable chemical rules for high‐performance organic dyes, which are expected to contribute to further innovations for the practical applications of DSSCs. Abstract : In this study, a reliable and interpretable quantitative structure‐property relationship model is established by combining machine learning technology and computational quantum chemistry, aiming to predict the efficiency of undiscovered organic dyes for dye‐sensitized solar cells. Moreover, eight promising candidates are screened out from ≈10 000 molecules, and reasonable chemical rules for high‐performance organic dyes are deduced.