To speed the search for well-matched solar materials, scientists applied random forest (RF) screening, a machine learning technique, for the design, synthesis, and characterization of a polymer that could facilitate rapid development of optoelectronic materials for organic photovoltaics (OPVs). Scientists from Osaka University gathered data on 1200 OPVs from around 500 studies. Using RF machine learning (artificial intelligence), they built a model combining the bandgap, molecular weight, and chemical structure of these OPVs, together with their power conversion efficiency (PCE), to predict the efficiency of potential devices. Exploring new polymers for polymer solar cells using materials informatics. Example of a polymer structure composed of electron donor, electron acceptor, and alkyl chains (upper panel). Classification by random forest method (middle panel). Synergetic combination of materials informatics, practical experiments, and human intelligence (lower area). Courtesy of Osaka University. RF uncovered a correlation between the properties of the materials and their actual performance in OPVs. To exploit this, the model was used to automatically “screen” prospective polymers for their theoretical PCE. The list of top candidates was then whittled down based on chemical intuition about what could be synthesized in practice. OPVs, a class of solar cells based on a light-absorbing organic molecule combined with a semiconductor polymer, are inexpensive, lightweight, safe, and easy to produce. However, the PCEs of OPVs are still too low for full-scale commercialization. “The choice of polymer affects several properties, like short-circuit current, that directly determine the PCE,” researcher Shinji Nagasawa said. “However, there’s no easy way to design polymers with improved properties. Traditional chemical knowledge isn’t enough. Instead, we used artificial intelligence to guide the design process.” Photoelectric conversion in polymer solar cell and chemical structures of the active materials. Courtesy of Osaka University. The information obtained through RF screening led the team to create a previously untested polymer. Although an OPV based on this polymer proved less efficient than expected, the model provided useful insights into the structure-property relationship. Researchers believe that the model’s usefulness could be improved by including more data, such as the polymers’ solubility in water. “Machine learning could hugely accelerate solar cell development, since it instantaneously predicts results that would take months in the lab," said professor Akinori Saeki. "It’s not a straightforward replacement for the human factor — but it could provide crucial support when molecular designers have to choose which pathways to explore.” The research was published in The Journal of Physical Chemistry Letters (doi:10.1021/acs.jpclett.8b00635).
