Batteries are crucial for energy storage applications on both large scales, such as the Notrees Wind Storage Demonstration Project in Texas , and small scales, such as smart phones. Lithium-ion batteries have become widely used due to their high-energy density and variety of cell types [2, 3]. Most lithium-ion batteries use organic solvents as liquid electrolytes, and battery performance depends critically on the properties of the electrolyte. Due to the wide range of possible organic solvents, there is a need for accurate and efficient methods of predicting key properties such as density, dielectric constant, viscosity, and enthalpy of vaporization. Molecular dynamics simulations are one such method, but the quality of the predictions depends on the accuracy of the intermolecular force field used. The performance of the General AMBER force field (GAFF)  was evaluated by computing density, dielectric constant, viscosity, and enthalpy of vaporization for nineteen different organic solvents. Densities were in reasonable agreement with experiment, but there were significant errors in the other properties. To refine the force field, two different approaches were taken. In the first approach, partial charges q and the Lennard-Jones energy parameters Îµ were scaled to optimize density and dielectric constant. In the second approach, partial charges q and the Lennard-Jones atomic size parameters Ï were scaled to match density and dielectric constant. The first approach yielded better agreement for densities, dielectric constants, and viscosities than the original GAFF force field, and gave comparable results for enthalpy of vaporization. The second approach resulted in better agreement for densities and dielectric constants but worse agreement for viscosities and enthalpy of vaporization than the original GAFF force field. In addition, we scaled q, Îµ, and Ï together to optimize all four properties at the same time on Acetonitrile using Latin Hypercube Sampling. We then presented Pareto frontier and feasible region for this multi-objective optimization problem. The q, Îµ, and Ï scaling parameters were picked from the feasible region and this modified force field was evaluated by computing phase equilibria properties.
 I. Gyuk, âSmoothing Renewable Wind Energy in Texasâ, Energy.gov., 2013, https://www.energy.gov/articles/smoothing-renewable-wind-energy-texas.
 A. Manthiram, X. Yu and S. Wang, âLithium battery chemistries enabled by solid-state electrolytesâ, Nature Reviews Materials, 2017, 2, 1-16.
. B. Dunn, H. Kamath and J. Tarascon, âElectrical energy storage for the grid: a battery of choicesâ, Science, 2011, 334, 928â936.
. J.M. Wang, R.M. Wolf, J. W. Caldwell, P. A. Kollman, D. A. Case, âDevelopment and testing of a general amber force fieldâ, Journal of Computational Chemistry, 2004, 25, 1157â1174.