Automatic Discovery and Optimal Generation of Amorphous High-Entropy Electrocatalysts.

Area of Science:

• Materials Science
• Electrochemistry
• Catalysis

Background:

• Amorphous high-entropy materials offer industrial potential but lack defined structure-activity relationships for optimization.
• Designing and optimizing these materials for applications like catalysis remains a significant challenge.

Purpose of the Study:

• To discover and optimize amorphous high-entropy oxyhydroxide electrocatalysts for the alkaline oxygen evolution reaction (OER) across the entire design space.
• To establish a machine learning-driven composition-activity relationship for efficient catalyst discovery.

Main Methods:

• Utilized synthesis systems to explore over 1,900,000 compositions for amorphous high-entropy oxyhydroxide electrocatalysts.
• Employed machine learning (ML) techniques to establish composition-activity relationships and identify optimal catalyst compositions.
• Derived catalysts from ultrathin 2D coordination polymers of six nonprecious metals, transformed *in situ* into amorphous oxyhydroxides.

Main Results:

• Identified an optimal composition group for amorphous high-entropy electrocatalysts through ML-guided exploration of the design space.
• Validated ML model with high recall (nearly 100%), accurately classifying high- and low-activity regions.
• Achieved an ultralow overpotential (159 mV at 10 mA cm⁻²) and exceptional durability (10,218 h at 1 A cm⁻²) for the alkaline OER.

Conclusions:

• Developed a general strategy for the automatic discovery and optimization of amorphous high-entropy oxyhydroxide electrocatalysts.
• The ML-driven approach significantly impacts the development of advanced amorphous high-entropy materials for various applications.
• Demonstrated the potential of amorphous high-entropy electrocatalysts for efficient and durable alkaline oxygen evolution reactions.