Construction of Sabatier Volcanoes for CO₂ Hydrogenation to C1-2 Oxygenates Using Data-Efficient Machine Learning

Area of Science:

• Catalysis
• Chemical Engineering
• Materials Science

Background:

• Global priority on CO2 hydrogenation for abatement and chemical production.
• Lack of understanding of complex reaction networks hinders catalyst design.

Purpose of the Study:

• Establish a data- and computation-efficient framework for CO2 hydrogenation to C1-2 oxygenates.
• Identify optimal catalyst properties for efficient CO2 conversion.

Main Methods:

• Computational framework applied to ≈1.2 nm Au, fcc-Co, Cu, Ni, Pd, Pt, and Rh nanoparticles.
• Nonlinear neural network (NN) model for activation energy inference.
• Pathfinder algorithm and energetic span model for network analysis and Sabatier volcano plots.

Main Results:

• Reaction networks not accurately described by Brønsted-Evans-Polanyi relationships.
• Optimal C2 oxygenate formation requires simultaneous CHx formation, C-C coupling, protonation, and suppression of methanation/poisoning.
• Mutually exclusive requirements on monometallic nanoparticles indicate need for multi-component/multimetallic catalysts.

Conclusions:

• The developed framework aids computational discovery of CO2 hydrogenation catalysts.
• Multicomponent and multimetallic catalysts are promising targets for efficient CO2 conversion.
• Addresses mechanistic complexity and data limitations in catalyst design.