Machine learning accelerated nitrogen electrofixation on dual-atom catalysts

Area of Science:

• Materials Science
• Catalysis
• Computational Chemistry
• Machine Learning

Background:

• Atomically dispersed catalysts, including single-atom catalysts, offer high atomic utilization and tunable electronic properties.
• Dual-atom catalysts, created by introducing a second single-atom site, present new opportunities for enhanced catalytic applications.
• Optimizing dual-atom catalysts is complex due to numerous influencing factors like elemental composition and atomic arrangement.

Purpose of the Study:

• To employ machine learning (ML) to accelerate the screening and design of dual-atom catalysts.
• To identify novel dual-atom catalyst systems for efficient electrochemical nitrogen reduction reactions (NRR).
• To establish a robust framework for exploring new catalysts in energy conversion systems.

Main Methods:

• Machine learning (ML) models were utilized for high-throughput prediction of potential dual-atom catalyst candidates.
• Density Functional Theory (DFT) computations were performed to elucidate the reaction mechanisms of promising catalysts.
• Experimental synthesis and electrochemical testing were conducted to validate the predicted catalyst performance.

Main Results:

• ML-driven screening identified a CrNi/MoSe2 dual-atom catalyst with a predicted ultralow limiting potential of -0.45 eV for nitrogen fixation.
• DFT calculations revealed the underlying catalytic mechanisms contributing to the enhanced performance.
• Experimental verification confirmed that the synthesized CrNi/MoSe2 catalyst exhibits high efficiency in electrochemical nitrogen reduction reactions.

Conclusions:

• The study demonstrates the efficacy of ML in accelerating the discovery of advanced dual-atom catalysts.
• The identified CrNi/MoSe2 catalyst shows significant potential for efficient electrochemical nitrogen reduction.
• This work provides a scalable framework for designing next-generation catalysts for energy conversion applications.