Screening Platform for Promising Na Superionic Conductors for Na-Ion Solid-State Electrolytes.

Area of Science:

• Materials Science
• Electrochemistry
• Computational Chemistry

Background:

• Sodium-ion (Na-ion) batteries offer a cost-effective and abundant alternative to lithium-ion batteries.
• Solid electrolytes in Na-ion batteries enhance safety and energy density but often exhibit lower ionic conductivity than liquid electrolytes.
• Developing high-performance solid electrolytes is crucial for next-generation battery technologies.

Purpose of the Study:

• To develop a machine learning model for identifying novel sodium superionic conductor (NASICON) materials with high ionic conductivity.
• To accelerate the discovery and validation of solid-state electrolytes for Na-ion batteries.

Main Methods:

• Utilized machine learning, specifically gradient boosting algorithms, to classify 3573 NASICON structures.
• Engineered new features based on chemical descriptors like Na content, elemental radii, and electronegativity.
• Validated promising candidates using density functional theory (DFT) and ab initio molecular dynamics (AIMD) simulations.

Main Results:

• Achieved an average prediction accuracy of 84.2% in classifying NASICON materials.
• Identified four novel NASICON compounds (Na3YTaSi2PO12, Na3HfZrSi2PO12, Na3LaTaSi2PO12, and Na3ScTaSi2PO12) with potential for high ionic conductivity.
• Confirmed the thermodynamic stability and favorable ionic conductivity of the predicted materials.

Conclusions:

• The machine learning approach effectively identifies high-performance NASICON materials for solid-state electrolytes.
• The validated compounds represent promising candidates for advancing Na-ion battery technology.
• This work paves the way for accelerated materials discovery in solid-state batteries.