Using Data Science and Experimental Tools to Determine Molecular Design Principles for Commercial Quinone Posolytes

Quinones are promising electrolyte candidates for aqueous redox flow batteries (ARFBs) and are widely commercially available, but their application is limited by challenges balancing redox potential, chemical stability, and solubility. While redox potential is relatively straightforward to measure or compute, chemical stability, particularly resistance to degradation, remains complex and experimentally demanding. In this study, we sought to quantify quinone degradation across diverse substitution patterns and to identify molecular features that contribute to improved performance using machine learning regression algorithms. Experimental results confirmed that redox potential and stability are not linearly correlated, complicating simultaneous optimization. We therefore sought to optimize each metric independently and construct a Pareto front based on predicted experimental metrics. This work demonstrates the power of descriptor-driven virtual screening, supported by experimental validation, for the rational design of next-generation quinones for ARFBs and assesses the current landscape of commercially available quinones for their application.