Nonaqueous redox flow battery (RFB) is a promising electrical energy storage technology with great potential to finally realize alternative energy sources for the next-generation vehicles and at grid scales. Molecular Dynamics (MD) simulations are promising methods for the computational understanding of physicochemical properties of nonaqueous electrolyte solutions from the molecular level, but the accuracy of their predictions critically depends on the prescribed force fields. Here, we show that several quantum-mechanically refined force fields to model LiTFSI–acetonitrile solutions all yield structures that are in good agreement with the neutron pair distribution function (PDF) experiments, yet produce dramatically different ion diffusivities, several of which are around one order slower than NMR measurements. Such glaring discrepancies indicate that inadequate representation of long-range interactions leads to excessive frustration in the free energy landscape. Better agreement is achieved by proportionally scaling down the atomic charges of the ions. This simplification enabled the accurate simulation of concentration dependences of ionic diffusion for 0.2-2 M LiTFSI solutions without sacrificing fit quality of the PDFs. We argue that not only the structure but also the dynamics constitutes important checkpoints on the road to computationally design of new electrolytes in nonaqueous redox flow battery.

When dynamics matters: Structure-matching alone is insufficient in the development of realistic force fields for nonaqueous electrolyte solutions

2019 Fall National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.