High-Throughput Screening of Molecules for Flow Batteries

Abstract

Development of next generation stationary energy storage technologies is of crucial importance for increasing the share of the intermitted renewable electricity generation in the world. Traditional trial and error based approach for discovery of novel materials is very slow, so we have been thinking of how to accelerate the process. Some crucial parameters such as redox potentials can be readily evaluated with computational tools, such as those developed in our CompBat project [1], [2], but other parameters such as stability are more difficult to be evaluated. In this work we will demonstrate our web-based application for evaluating redox potentials of molecules from the SMILES structure, illustrating very impressive or problematic performance depending on the training data used for the model.For the question of stability, we believe that sufficient quantity of experimental data is required for training of specific machine learning models for stability prediction. For this purpose, we will illustrate our experimental approach for high-throughput electrochemistry, to evaluate the stability of molecules in the time-scale of voltammetry. Moreover, we will discuss the order of magnitude of the kinetics of the decomposition reactions that can be evaluated by voltammetry, and illustrate that voltammetry is a good tool for roughly separating more stable species from the less stable, but other techniques are required for evaluating stability in the time scale of battery operation.References https://compbat.eu/A. Hamza, F.B. Németh, Á. Madarász, A. Nechaev, P. Pihko, P. Peljo, I. Pápai, N-alkylated pyridoxal derivatives as negative electrolyte materials for aqueous organic flow batteries: Computational screening, 10.26434/chemrxiv-2023-bzc6w