Determining Water Solubility of Organic Electrolytes for Aqueous Organic Redox Flow Batteries

A major obstacle to decarbonization of the domestic energy grid is a lack of cost-effective, safe, and nontoxic long-term energy storage technologies. Aqueous organic redox flow batteries provide a cost-effective alternative to other grid-scale technologies such as redox flow batteries that do not require the mining of rare earth metals. However, the development of possible catholyte and anolyte materials for these devices requires a balance of properties including stable and rapid redox reactions as well as sufficient aqueous solubility of redox compounds. Herein, we demonstrate the effectiveness of a machine learning model to predict aqueous solubilities of various organic compounds used in AORFBs, validating our results across a training data set consisting of relevant compounds that were measured for their aqueous solubility by 1H NMR measurements. We have subsequently used the model to determine the aqueous solubility of slightly more than 3000 compounds. This indicated insufficient solubility for 2000 of the molecules, however it did not indicate which compounds would be most suitable for practical testing. Density functional theory and molecular dynamics simulations are used to further interrogate the solubility of select remaining compounds.