Data-Driven Insight into the Reductive Stability of Ion-Solvent Complexes in Lithium Battery Electrolytes

Area of Science:

• Materials Science
• Electrochemistry
• Computational Chemistry

Background:

• Lithium metal batteries (LMBs) offer high energy density but suffer from electrolyte decomposition due to reactive lithium anodes.
• Current trial-and-error methods for designing stable electrolytes are inefficient for practical LMB applications.

Purpose of the Study:

• To develop a data-driven strategy for understanding and predicting the reductive stability of electrolyte solvents for LMBs.
• To accelerate the rational design of advanced electrolyte molecules for next-generation energy storage.

Main Methods:

• Constructed a large database of solvent molecules using graph theory.
• Employed first-principles calculations and machine learning (ML) for comprehensive investigation.
• Utilized Shapley additive explanations (SHAP) for interpretable ML analysis.

Main Results:

• Identified ion-solvent complexes as a primary cause for decreased reductive stability in 99% of electrolytes.
• Established relationships between lowest unoccupied molecular orbital (LUMO) energy and factors like binding energy and bond length.
• Determined dipole moment and molecular radius as critical descriptors for solvent reductive stability.

Conclusions:

• The study provides data-driven insights into ion-solvent chemistry and its impact on electrolyte stability.
• Identified key molecular descriptors that enable the rational design of stable electrolytes for LMBs.
• This approach significantly accelerates the development of advanced electrolytes for high-performance energy storage systems.