Deconstruction Analysis and Regeneration Optimization of Solid Electrolyte Components for Energy Storage Based on the...

Area of Science:

• Materials Science
• Electrochemistry
• Computational Chemistry

Background:

• Solid-state electrolytes offer enhanced safety and energy density compared to liquid electrolytes.
• Developing high-performance solid-state electrolyte materials remains a significant challenge in battery technology.
• Machine learning (ML) presents a powerful approach to accelerate materials discovery and optimization.

Purpose of the Study:

• To develop a predictive model for Na superionic conductor (NASICON) solid-state electrolytes using machine learning.
• To identify optimal component ratios and understand the influence of dopants on ionic conductivity.
• To provide a data-driven reference for the design of advanced solid-state electrolyte materials.

Main Methods:

• Construction of a solid-state electrolyte database from existing experimental data.
• Training a random forest (RF) algorithm model to predict electrolyte performance based on component information.
• Analysis of component effects and screening of optimal doping element ratios.

Main Results:

• A predictive model accurately estimated the performance of NASICON solid-state electrolytes.
• The study identified the optimal ratio for 61 doping elements, significantly aiding material optimization.
• Key components and doping elements influencing ionic conductivity were distinguished.

Conclusions:

• Machine learning, specifically RF modeling, effectively predicts and optimizes solid-state electrolyte performance.
• This data-driven approach accelerates the development of high-performance NASICON materials.
• The findings offer valuable guidance for future solid-state electrolyte research and component selection.