Machine Learning-Accelerated Development of Li-Rich Layered Oxides with High Reversible Capacity and Coulombic...

Area of Science:

• Materials Science
• Electrochemistry
• Data Science

Background:

• Li-rich layered oxides offer high energy density for Li-ion batteries.
• Structural instability and oxygen redox activity limit their practical application, causing a capacity-efficiency trade-off.

Purpose of the Study:

• To utilize machine learning for identifying critical synthesis factors influencing initial Coulombic efficiency in Li_1.2Ni_0.13Co_0.13Mn_0.54O₂.
• To guide experimental synthesis for optimizing Li-rich layered oxide cathodes.

Main Methods:

• Trained four machine learning models on a dataset of 203 samples.
• Employed a gradient boosting decision tree model, achieving R² = 0.802 on the test set.
• Identified lithium-to-transition metal ratio and presintering atmosphere as key parameters.

Main Results:

• Machine learning identified the lithium-to-transition metal ratio and presintering atmosphere as critical synthesis parameters.
• Optimized synthesis via machine learning guidance (low-temperature air presintering) reduced Li₂MnO₃ phase, promoted {010} plane exposure, and mitigated rock-salt formation.
• Achieved a high discharge capacity of 301.02 mAh g^-1 and an initial Coulombic efficiency of 81.05%.

Conclusions:

• Machine learning effectively identifies critical synthesis parameters for Li-rich layered oxides.
• Data-driven synthesis optimization significantly enhances electrochemical performance, particularly initial Coulombic efficiency.
• This approach demonstrates the power of integrating machine learning with experimental synthesis for advanced battery material development.