Peukert-Informed Prediction and Screening of Manganese-Based High-Rate Potassium Cathode

Area of Science:

• Materials Science
• Electrochemistry
• Computational Chemistry

Background:

• Manganese-based layered oxides are promising cathode materials for potassium-ion batteries (KIBs) due to their low cost and high theoretical capacity.
• However, their practical application is hindered by poor rate performance, which limits their use in high-power devices.

Purpose of the Study:

• To develop an efficient and interpretable machine learning framework for identifying novel manganese-based layered cathode materials for KIBs.
• To overcome the limitations of traditional trial-and-error methods in materials discovery.

Main Methods:

• Employed interpretable machine learning, including Random Forest, high-throughput virtual screening, and Pareto-front optimization.
• Utilized Peukert capacity and a Peukert-derived rate-decay descriptor to analyze capacity and rate-dependent behavior.
• Screened 14,558 candidate compositions derived from 22 dopant elements.

Main Results:

• Identified specific compositional regions that effectively balance practical capacity and rate performance in manganese-based layered cathodes.
• The model-recommended composition achieved approximately 67 mAh g⁻¹ at 500 mA g⁻¹.
• Demonstrated excellent cycling stability, retaining over 70% of initial capacity after 500 cycles.

Conclusions:

• The developed machine learning framework provides an interpretable and applicability-aware methodology for accelerating the discovery of high-rate cathode materials.
• This approach offers generalizable insights for designing advanced energy storage materials, particularly for KIBs.