Probing Active Sites in CuPd Cluster Catalysts by Machine-Learning-Assisted X-ray Absorption Spectroscopy

Area of Science:

• Catalysis
• Materials Science
• Surface Chemistry

Background:

• Size-selected clusters are crucial model catalysts due to their defined size and composition.
• Understanding structure-activity relationships in these clusters is challenging due to their small size and dynamic nature under reaction conditions.

Purpose of the Study:

• To elucidate the structure-activity relationships of copper-palladium (Cu-Pd) clusters in propane oxidation.
• To accurately identify and characterize active catalytic sites within subnanometric clusters.

Main Methods:

• Utilized machine learning approaches for analyzing operando X-ray absorption near-edge structure (XANES) spectra.
• Applied advanced spectral analysis to determine cluster speciation and structural information.

Main Results:

• Achieved accurate speciation of different Cu-Pd cluster types during the propane oxidation reaction.
• Obtained detailed structural information for each identified cluster type.
• Elucidated the roles of copper and palladium in the catalytic activity.

Conclusions:

• Machine learning analysis of operando XANES spectra provides a powerful method for characterizing catalytic active sites in clusters.
• The study successfully identified active species and clarified the relative contributions of Cu and Pd in Cu-Pd clusters for propane oxidation.