Data-Efficient Design of High-Entropy Oxygen Carriers for Chemical Looping Using Active Learning.

Area of Science:

• Materials Science
• Chemical Engineering
• Computational Materials Science

Background:

• High-entropy materials offer vast potential across catalysis, energy, and structural applications.
• Their complex compositions hinder systematic exploration, limiting studied material space.

Purpose of the Study:

• To introduce an efficient active learning strategy for navigating complex compositional material spaces.
• To accelerate the discovery of high-entropy materials, specifically oxygen carriers for chemical looping.

Main Methods:

• Integrated predictive modeling, uncertainty estimation, and iterative sampling.
• Employed an active learning approach for continuous learning and focused search.
• Applied the strategy to identify high-entropy oxygen carriers for chemical looping.

Main Results:

• The active learning strategy significantly accelerated the discovery process.
• Identified promising high-entropy oxygen carrier candidates more effectively than traditional methods.
• Demonstrated the general applicability of the strategy to multicomponent materials.

Conclusions:

• Active learning provides an efficient pathway for exploring vast multicomponent material landscapes.
• This approach reduces time and data requirements for materials discovery.
• The methodology is broadly applicable to various high-entropy material systems.