Machine Learning-Guided Surface Strain Engineering in Connected Platinum-Nickel Nanoparticle Catalysts for Advanced...

Area of Science:

• Materials Science
• Electrochemistry
• Catalysis

Background:

• Catalyst surface structure engineering is crucial for enhancing oxygen reduction reaction (ORR) activity.
• Developing efficient and durable electrocatalysts is vital for energy conversion technologies.

Purpose of the Study:

• To employ a machine learning (ML)-guided strategy for designing support-free, connected nanoparticle catalysts with improved ORR activity.
• To identify key descriptors influencing ORR specific activity (SA) and validate ML predictions through experimental synthesis and characterization.

Main Methods:

• ML analysis of 210 Pt-based ORR catalysts to identify structure-activity relationships.
• H2-annealing-induced surface structuring to create Pt-Ni nanoarchitectures with tunable compressive strain.
• Electrochemical testing to evaluate ORR activity and durability.

Main Results:

• Surface compressive strain (≈-4%) was identified as a critical descriptor for ORR SA.
• An optimized Pt-Ni nanoarchitecture catalyst exhibited a 12-fold increase in ORR SA (5.1 ± 0.5 mA cmPt-2) compared to commercial Pt/C.
• The catalyst demonstrated remarkable durability, retaining compressive strain and showing minimal Ni dissolution after 10,000 potential cycles.

Conclusions:

• ML-guided design is an effective framework for developing high-performance electrocatalysts.
• Strain engineering of catalyst surfaces significantly enhances ORR activity and durability.
• The developed Pt-Ni nanoarchitecture represents a promising next-generation catalyst for energy conversion.