Insights into Lithium Sulfide Glass Electrolyte Structures and Ionic Conductivity via Machine Learning Force Field...

Area of Science:

• Materials Science and Engineering
• Computational Materials Science
• Electrochemistry

Background:

• Sulfide-based solid electrolytes (SEs) are crucial for all-solid-state batteries (ASSBs) due to high ionic conductivity and mechanical stability.
• Glassy SEs (GSEs) with mixed Si and P formers offer promising synthesis routes and dendrite suppression, but their complex structures impede property understanding.
• Accurate modeling of SE structures and properties is essential for designing next-generation ASSBs.

Purpose of the Study:

• To develop a machine learning force field (ML-FF) for accurate simulation of lithium sulfide-based GSEs.
• To investigate the structure-property relationships, including mechanical properties and ionic conductivity, of various GSE compositions.
• To elucidate the mechanisms governing lithium ion diffusion in these electrolytes.

Main Methods:

• Development of a novel machine learning force field (ML-FF) specifically for lithium sulfide-based GSEs.
• Molecular dynamics (MD) simulations utilizing the ML-FF to explore binary (Li2S-SiS2, Li2S-P2S5) and ternary (Li2S-SiS2-P2S5) compositions.
• Validation of simulation results against experimental data and density functional theory (DFT) calculations for density, elastic modulus, and structure factors.

Main Results:

• The ML-FF accurately reproduced key material properties (density, elastic modulus, radial distribution functions, neutron structure factors).
• Distinct local environments for Si and P were identified, with Si favoring edge-sharing in Li2S-SiS2 and mixed sharing in ternary glasses.
• Lithium ionic conductivity varied with composition, with 75Li2S-25P2S5 showing the highest (3.6 mS/cm) and 50Li2S-50SiS2 the lowest (2.1 mS/cm) at 300 K.
• Lithium ion diffusion in ternary glasses correlated with the rotational dynamics of SiS4 and PS4 tetrahedra.

Conclusions:

• The developed ML-FF is a powerful tool for simulating and understanding sulfide-based GSEs.
• The study provides insights into the structural basis for ionic conductivity in these materials.
• This work facilitates the rational design and optimization of SEs for ASSBs.