Figure 10.

Data‐mining the literature for potential dye candidates for DSSCs; A) A starting database of 9431 dye candidates from the scientific literature was generated using the text‐mining tool, ChemDataExtractor. B) Initial screens remove small molecules, organometallic dyes, and chemicals not absorbing in the solar spectrum. C) Molecules lacking a carboxylic acid or reasonable dipole moment are removed. D) Optical properties of the dyes are assessed. E) HOMO and LUMO energy levels of each dye are checked using DFT to ensure proper integration into a DSSC. F) A final set of five dyes is selected for experimental verification. G) Experimental validation. Reproduced with permission.^{[ 46 ]} Copyright 2018, Wiley‐VCH.