High-Throughput Screening of Sulfur-Resistant Catalysts for Steam Methane Reforming Using Machine Learning and...

Area of Science:

• Catalysis
• Materials Science
• Computational Chemistry

Background:

• Bimetallic catalysts are crucial for steam methane reforming (SMR).
• Their performance in sulfur-rich environments remains largely unexplored.
• Sulfur poisoning is a major challenge for SMR catalysts.

Purpose of the Study:

• To develop a rapid screening process for sulfur-resistant bimetallic catalysts using machine learning (ML) and microkinetic modeling.
• To identify novel bimetallic materials with enhanced stability against sulfur poisoning in SMR.
• To leverage computational methods for accelerated catalyst discovery.

Main Methods:

• Development of ML models (Ensemble, ANN, SVR) to predict atomic adsorption energies on bimetallic surfaces.
• Utilizing physical and chemical properties of metals and adsorbates as ML input features.
• Construction of a microkinetic model for the SMR reaction and screening of over 500 bimetallic materials.

Main Results:

• ML models achieved high predictive performance (R^2 up to 0.74) for adsorption energies.
• Ensemble learning model was used in conjunction with microkinetic modeling for catalyst screening.
• Four Ge-based alloys (Ge3Cu1, Ge3Ni1, Ge3Co1, Ge3Fe1) and Ni3Cu1 were identified as promising candidates.

Conclusions:

• The combined ML and microkinetic modeling approach enables efficient discovery of sulfur-resistant catalysts.
• Identified Ge-based and Ni3Cu1 alloys offer potential for cost-effective and sulfur-tolerant SMR.
• This methodology accelerates the development of next-generation catalysts for challenging industrial processes.