Single-atom alloy catalysts designed by first-principles calculations and artificial intelligence

Area of Science:

• Catalysis Science
• Materials Science
• Computational Chemistry

Background:

• Single-atom-alloy catalysts (SAACs) are crucial for efficient industrial reactions due to optimized reactant dissociation and intermediate binding.
• Discovering new SAACs is challenging due to the vast number of potential candidates and the need for fast, reliable property prediction.

Purpose of the Study:

• To develop a rapid and reliable method for predicting the catalytic properties of numerous potential single-atom-alloy catalysts.
• To identify novel, highly efficient, and stable SAAC candidates for various catalytic applications.

Main Methods:

• Applied a compressed-sensing data-analytics approach using density-functional theory inputs.
• Utilized subgroup discovery, a data-mining method, for qualitative analysis of symbolic regression models.

Main Results:

• Successfully predicted the efficiency of experimentally validated SAACs.
• Identified over 200 novel SAAC candidates, some exhibiting superior stability and efficiency compared to existing ones.
• Introduced a new qualitative analysis technique for complex regression models.

Conclusions:

• Data analytics is vital for unbiased catalyst design, offering a systematic approach to discover optimal SAACs.
• The developed methodology provides a practical framework for identifying promising SAACs for diverse catalytic processes.