Optimizing small conjugated molecules for solar-cell applications using an inverse-design method

Small organic conjugated molecules are key elements for low-cost photovoltaic devices. One example is cyanopyridone molecules. By modifying these molecules, for instance through optimally chosen functional groups attached to the backbone, their properties can be improved. However, the very large number of possible modifications makes it difficult to identify the best performing molecules. In the present work, we have used a computational inverse-design approach (PooMa) to identify the positions and types of functional groups attached to a modified cyanopyridone that lead to the best performance in solar-energy harvesting. A QSPR model based on five electronic descriptors has been used to determine the properties of solar cells. Our approach uses a genetic algorithm to search the chemical space containing 18⁴ (104,976) substituted cyanopyridone systems and predicts out of those the best 20 molecules with optimal performance efficiencies (PCE). PooMa uses the Density-Functional Tight-Binding (DFTB) method for calculating the electronic properties. DFTB is a fast method with acceptable accuracy and, therefore, can be used on a normal desktop without expensive hard- or software. In order to get further information about our suggested systems, a DFT method and its derivative TD-DFT are applied.