Investigating the Stability of Organic Electrolytes for Lithium-Air Batteries with High-Throughput Simulations

Lithium-air batteries have gained significant attention as a solution for applications requiring high gravimetric energy density. Currently, however, electrolyte instability remains a key challenge limiting the practical use of aprotic Li-air batteries. Common organic electrolytes are largely unstable against the oxygen electrode, experiencing degradation due to hydrogen removal, proton removal, nucleophilic attack, oxidations and reductions. Previous works have computationally screened electrolyte candidates for stability based on ground state electrolyte properties. However, the chemical processes that cause electrolyte instability are often kinetically driven and thus thermodynamics and kinetics must both be investigated in order to fully evaluate candidate stability. Computational screening for kinetic stability requires automating the determination of reaction barriers for the numerous potential reaction mechanisms.<br/><br/>High-throughput calculation of reaction barriers is challenging due to the difficulty of locating transition state structures. As such, transition state calculations often rely heavily on user input and chemical intuition and existing automated approaches such as double-ended string methods frequently fail for complex molecules. In this work, we present an automated pipeline for high-throughput determination of reaction free energy barriers for kinetically-controlled electrolyte decomposition processes. By integrating cheminformatics-based reaction encoding, relaxed potential energy scans, and nudged elastic band calculations, our approach enables determination of free-energy barriers for reactions of interest of candidate electrolytes at scale. In particular, cheminformatics tools are used to programmatically generate products from reactants and track reactive atoms, relaxed potential energy scans to obtain initial reaction paths, and climbing-image nudged elastic band calculations to obtain transition state guesses from these starting paths.<br/><br/>We apply this approach to determine reaction barriers for nucleophilic attack by superoxide, which is formed during electrochemical cycling in Li-air batteries. We screen 100+ electrolytes with 400+ potential reactions, validating expected trends in the stability of carbonate and sulfone candidates. We study the effects of chemical functionalization with electron-donating and withdrawing groups and their interplay with steric factors. Finally, we identify functional groups and other chemical modifications that could increase the stability of organic electrolytes in aprotic Li-air batteries.