Decoding Lithium Metal Battery Degradation with Symmetric-Cell Artificial Intelligence Diagnostics (SAID)

Area of Science:

• Materials Science
• Electrochemistry
• Artificial Intelligence

Background:

• Lithium metal anode degradation is a critical challenge in battery technology.
• Crosstalk effects in asymmetric cells complicate the analysis of electrode-specific degradation.
• Symmetric cell configurations allow for the isolation of individual electrode contributions.

Purpose of the Study:

• To develop an AI-driven diagnostic tool for deciphering lithium metal anode degradation.
• To leverage early-cycle lithium-lithium symmetric cell data for predictive analysis.
• To identify key degradation fingerprints and their impact on long-term battery performance.

Main Methods:

• Development of a Symmetric-cell Artificial Intelligence Diagnostics (SAID) tool.
• Utilizing early-cycle lithium-lithium symmetric cell data as input.
• Validation of SAID predictions through experimental analysis and application to full cells.

Main Results:

• SAID accurately predicts polarization acceleration (elbow points) with a 13.3% mean absolute percentage error.
• The study identified a persistent, initial-nucleation-related fingerprint influencing long-term cell polarization.
• The findings were experimentally validated and shown to be applicable across different electrolytes in full cells.

Conclusions:

• SAID provides valuable insights into lithium metal anode degradation mechanisms.
• The AI approach offers a powerful method for battery design and understanding hidden chemical correlations.
• This work advances the field of energy storage by enabling more accurate degradation prediction and analysis.