Developing a predictive solubility model for monomeric and oligomeric cyclopropenium-based flow battery catholytes

The implementation of redox active organics in nonaqueous redox flow batteries requires the design of molecules that exhibit high solubility (>1 M) in all battery-relevant redox states. Methods for forecasting nonaqueous solubility would be valuable for streamlining the identification of promising structures. Herein we report the development of a workflow to parametrize and predict the solubility of conformationally flexible tris-(dialkylamino)cyclopropenium (CP) radical dications. A statistical model is developed through training on monomer species. Ultimately, this model is used to predict new monomeric and dimeric CP derivatives with solubilities of >1 M in acetonitrile in all oxidation states. The most soluble CP monomer exhibits high stability to electrochemical cycling at 1 M in acetonitrile without a supporting electrolyte in a symmetrical flow cell.