Machine Learning-Accelerated First-Principles Study of Atomic Configuration and Ionic Diffusion in Li10GeP2S12 Solid Electrolyte

Changlin Qi,Yuwei Zhou,Xiaoze Yuan,Qing Peng,Yong Yang,Yongwang Li,Xiaodong Wen

DOI: https://doi.org/10.3390/ma17081810

IF: 3.4

2024-04-16

Materials

Abstract:The solid electrolyte Li10GeP2S12 (LGPS) plays a crucial role in the development of all-solid-state batteries and has been widely studied both experimentally and theoretically. The properties of solid electrolytes, such as thermodynamic stability, conductivity, band gap, and more, are closely related to their ground-state structures. However, the presence of site-disordered co-occupancy of Ge/P and defective fractional occupancy of lithium ions results in an exceptionally large number of possible atomic configurations (structures). Currently, the electrostatic energy criterion is widely used to screen favorable candidates and reduce computational costs in first-principles calculations. In this study, we employ the machine learning- and active-learning-based LAsou method, in combination with first-principles calculations, to efficiently predict the most stable configuration of LGPS as reported in the literature. Then, we investigate the diffusion properties of Li ions within the temperature range of 500–900 K using ab initio molecular dynamics. The results demonstrate that the atomic configurations with different skeletons and Li ion distributions significantly affect the Li ions' diffusion. Moreover, the results also suggest that the LAsou method is valuable for refining experimental crystal structures, accelerating theoretical calculations, and facilitating the design of new solid electrolyte materials in the future.

materials science, multidisciplinary,chemistry, physical,physics, applied, condensed matter,metallurgy & metallurgical engineering

What problem does this paper attempt to address?

This paper focuses on the atomic configurations and ion diffusion of the lithium-ion solid electrolyte Li10GeP2S12 (LGPS) through efficient first-principle calculations. LGPS plays a critical role in the development of solid-state batteries, but its complex structure, such as the co-occupation of disordered Ge/P sites and the fractional occupancy of lithium ion defects, results in numerous potential atomic configurations, making the computational cost prohibitively high. Currently, the method of electrostatic energy standard screening is widely used to select favorable candidate configurations.

In this study, the authors employed the LAsou method, which combines machine learning and active learning, along with first-principle calculations, to predict the most stable configuration of LGPS and investigate the diffusion properties of lithium ions in the temperature range of 500-900K. The results indicate that different frameworks and lithium ion distributions significantly affect lithium ion diffusion. Furthermore, the LAsou method holds value in optimizing experimental crystal structures, accelerating theoretical calculations, and designing future novel solid electrolyte materials.

The study also reviews the research history of LGPS materials, including the impact of different configurations on properties such as conductivity and thermodynamic stability, as well as various methods previously utilized, such as molecular dynamics simulations and electrostatic energy standards. The paper points out that although electrostatic energy standards are commonly used for screening, their consistency with the stability ranking from high-precision first-principle calculations is not always reliable. Therefore, the introduction of the LAsou method aims to improve the efficiency and accuracy of predicting stable configurations.