Disorder-Dependent Li Diffusion in Li₆PS₅Cl Investigated by Machine-Learning Potential

Area of Science:

• Materials Science
• Solid-state Chemistry
• Computational Materials Science

Background:

• Argyrodite solid-state electrolytes, such as Li6PS5Cl, show promise for safer and more conductive batteries than liquid electrolytes.
• Understanding Li-ion diffusion mechanisms in these materials is crucial for optimizing their performance, but current theoretical models have limitations.
• Controlling disorder is a key strategy to enhance ionic conductivity, yet the precise relationship remains unclear.

Purpose of the Study:

• To investigate the Li-ion diffusion mechanism in Li6PS5Cl at room temperature (300 K).
• To elucidate the role of chloride (Cl) ion disorder on ionic conductivity.
• To provide a theoretical basis for optimizing ionic conductivity in argyrodite solid-state electrolytes.

Main Methods:

• Large-scale, long-term molecular dynamics (MD) simulations utilizing machine-learning potentials (MLPs).
• Simulations were performed on a 5 × 5 × 5 supercell (6500 atoms) for 25 ns to ensure conductivity convergence.
• Analysis of Li-ion conductivity, activation energies, and equilibrium site occupancies.

Main Results:

• Computed Li-ion conductivity, activation energies, and site occupancies agree with experimental data.
• Maximum Li-ion conductivity was observed when 25% of 4c sites were occupied by Cl ions, not at maximum disorder (50%).
• Non-Arrhenius diffusion behavior was identified, with varying activation energies at higher temperatures (>400 K), explained by inter- and intra-cage jumps.

Conclusions:

• The study clarifies the complex relationship between Cl disorder and Li-ion diffusion in Li6PS5Cl.
• Inter- and intra-cage jumps are key factors governing Li-ion transport and conductivity.
• Findings provide a pathway for designing and optimizing argyrodite materials for enhanced ionic conductivity in solid-state batteries.