Chemometric Modeling of Absorption Maxima of Carbazole Dyes Used in Dye-Sensitized Solar Cells

Abstract

The present chapter reports a partial least squares (PLS)-regression-based chemometric model for the carbazole class of dyes used in dye-sensitized solar cells (DSSCs) to predict the absorption maxima (λmax) values. Quantitative prediction of the λmax values can be an important criterion for molecular design of new dye molecules. We have used only 2D descriptors for the model development purpose as the quantum chemical and electrochemical analyses are time consuming. The developed model was validated extensively using internationally acceptable statistical and validation parameters. A seven-descriptor PLS model with one latent variable (LV) was developed. The statistical results suggested that the model was statistically significant. The majority of the descriptors involved in the model are easily interpretable 2D atom pair descriptors. The model suggests that presence of nitrogen and sulfur atoms at the topological distance of 8, presence of nitrogen and oxygen atoms at the topological distance of 4, higher frequency of oxygen and sulfur atoms at the topological distance of 5, higher frequency of two nitrogen atoms at the topological distance of 5, presence of two oxygen atoms at the topological distance of 4, presence of carbon atoms connected with three aromatic bonds, and presence of two sulfur atoms at the topological distance of 4 in the dye molecules shifted the λmax value toward the longer wavelength. From the information obtained from the model, it has also been suggested that highly conjugated π-systems shift the λmax values toward the longer wavelength. Finally, it can be concluded that the identified features from the PLS model for the carbazole derivatives may be employed for the design and development of new carbazole dyes and to predict λmax values before they are synthesized.