Optimization of Fluorinated Ether-Based Quasi-Solid Electrolyte Systems for Lithium–Sulfur Batteries

TL;DR

This paper optimizes a fluorinated ether-based electrolyte for lithium-sulfur batteries using statistical modeling to improve performance and stability.

Contribution

A data-driven Gaussian process regression model is introduced to optimize electrolyte composition in Li–S batteries.

Findings

An optimized electrolyte composition (DOL:OTE:DME = 0.273:0.505:0.222) achieved 861 mAh g–1 initial discharge capacity.

The optimized electrolyte showed only 9.2% capacity loss over 100 cycles at 0.3 C.

Gaussian process regression effectively mapped composition–performance relationships in multicomponent systems.

Abstract

Although quasi-solid-state lithium–sulfur (Li–S)
batteries show great promise for safe and high-energy storage systems,
optimizing electrolyte formulations remains challenging due to the
complex interplay of factors such as ion transport, stability, and
sulfur utilization. In this study, seven quasi-solid electrolyte formulations
were systematically investigated based on a ternary electrolyte component
system of 1,3-dioxolane (DOL), 1H,1H,5H-octafluoropentyl 1,1,2,2-tetrafluoroethyl ether
(OTE), and 1,2-dimethoxyethane (DME). The seven electrolyte formulations
were designed based on a modified mixture design adapted from the
design of experiments (DoE) principles. A Gaussian process regression
(GPR) model was then used to statistically map the relationship between
electrolyte composition and performance responses. Here, GPR is used
as a data-driven approximation to capture composition–performance
trends and guide electrolyte optimization within the ternary design
space. The electrolytes were formed via in situ polymerization
to ensure mechanical stability and maintain favorable interfacial
contact with the electrodes. The model, trained on experimental data,
identified an optimized composition (DOL:OTE:DME = 0.273:0.505:0.222)
with improved predicted performance compared to the initial set. The
optimized electrolyte delivered a high initial discharge capacity
of 861 mAh g–1 at 0.3 C with only 9.2% capacity
loss over 100 cycles showing markedly improved cycling stability compared
to the baseline electrolyte. The statistical modeling provides a powerful
framework for electrolyte development and offers valuable insights
into the composition–performance relationships in multicomponent
electrolyte systems.

Figures