(356c) Data Science Based Electrolyte Design for Organic Redox Flow Batteries

Ability to design materials for targeted functionality rather than traditional intuitive and trial-by-error based methods can bring enormous societal and technological progress. In this context, exploring the large space of potential materials for energy storage device is computationally intractable. In particular, the discovery of organic redox active species - which dictates the energy density, specific capacity and cycle life is critical step in realizing viable redox flow battery (RFB) technology - will be taunting challenge considering billions of organic structural possibilities. Traditional methodology relies on combinatorial approach of additives, functional groups and solvent composition to achieve higher solubility and stability of the specific redox molecule. However, considering the wide range of materials selection, it is imperative to build an artificial intelligence-based design formulation that will help us bottom up design of RFB electrolytes. In recent years, our team has been building database and scientific framework for electrolyte design formulations based on chemistry informed (e.g. solvate structure and inherent properties) machine learning approach. This framework helps us in solvent and functional group selection process, and ultimately enhance the solubility and stability through inverse molecular design. We will discuss our efforts and achievements related to the data science-based framework of electrolyte design formulations applicable for organic molecule based redox flow battery systems.

Ability to design materials for targeted functionality rather than traditional intuitive and trial-by-error based methods can bring enormous societal and technological progress. In this context, exploring the large space of potential materials for energy storage device is computationally intractable. In particular, the discovery of organic redox active species - which dictates the energy density, specific capacity and cycle life is critical step in realizing viable redox flow battery (RFB) technology - will be taunting challenge considering billions of organic structural possibilities. Traditional methodology relies on combinatorial approach of additives, functional groups and solvent composition to achieve higher solubility and stability of the specific redox molecule. However, considering the wide range of materials selection, it is imperative to build an artificial intelligence-based design formulation that will help us bottom up design of RFB electrolytes. In recent years, our team has been building database and scientific framework for electrolyte design formulations based on chemistry informed (e.g. solvate structure and inherent properties) machine learning approach. This framework helps us in solvent and functional group selection process, and ultimately enhance the solubility and stability through inverse molecular design. We will discuss our efforts and recent results related to the data science-based framework of electrolyte design formulations applicable for organic molecule based redox flow battery systems.