Li-S batteries are plagued by parasitic reactions driven by soluble reaction intermediates resulting in stored energy loss and short cycle life. Electrolytes with low solubility limits for these lithium polysulfide (LiPS) intermediates reduce parasitic reactions, improving efficiency and life of cells. Sparingly solvating electrolytes drive sulfur reduction through a solid state pathway, potentially eliminating parasitic losses due to LiPS.¹ Realizing the goal of electrolyte design to enhance cell performance requires a quantitative model of LiPS solubility.

Predicting LiPS Solubility with Electrolyte Donor Number (EDN):

Building on the Gutman Donor Number concept, this talk will describe a new way of thinking about donicity—from the perspective of multicomponent systems (i.e. electrolytes). First, it will be shown that ²³Na NMR can be used as a technique for deriving a pseudo-Gutman DN (²³Na-DN) for any given electrolyte composition. This is similar to work described by Schmeisser et al., in which DNs for ionic liquids were derived from ²³Na NMR.² Next, it will be shown that an accurate descriptor for LiPS solubility also requires knowledge of the density of total available donor sites within a given electrolyte. Raman spectroscopy was used to deduce the specific populations of both solvent and anion species (i.e. establishing the coordinating behavior of solvating species to Li⁺ ions in solution as a function of salt concentration). Finally, it will be shown that the complete descriptor for LiPS solubility, as determined experimentally, relies on both the ²³Na-DN and the structural arrangements of Li⁺ coordinating ligands within a given electrolyte. The combination of these two characteristics establishes an “electrolyte donor number” (EDN). Figure 1 shows that LiPS solubility values converge to a single function versus different functions with respect to the EDN and DN, respectively.

Discussion on the Prediction of Reaction Pathways:

The EDN model predicts LiPS solubility well. Thus, it is not surprising that it also acts as a useful tool for predicting the electrochemical reaction pathways taking place in a Li-S cell. Figure 2 shows the discharge curves for tetraglyme-based electrolytes that span a maximum range of EDN values. It is clear from this data that a higher EDN drastically changes the observed voltage profile, and therefore the mechanistic pathway.

References:

1. Cheng, L.; Curtiss, L. A.; Zavadil, K. R.; Gewirth, A. A.; Shao, Y.;

Gallagher, K. G. ACS Energy Lett. 2016, 1, 503−509.

2. Schmeisser, M.; Illner, P.; Puchta, R.; Zahl, A.; Eldik, R. Chem. Eur. J. 2012, 18, 10969–10982.­­