Machine Learning-Enabled Framework for High-Throughput Screening of MOFs: Application in Radon/Indoor Air Separation.

Area of Science:

• Materials Science
• Environmental Science
• Computational Chemistry

Background:

• Radon exposure poses significant health risks, particularly in underground settings.
• Effective methods for radon separation from indoor air are crucial for public health.
• Metal-organic frameworks (MOFs) show potential as adsorbents for gas separation.

Purpose of the Study:

• To develop and validate a hybrid artificial intelligence (AI) machine learning framework for high-throughput screening of MOFs.
• To identify MOFs with high adsorbent performance score (APS) and regenerability for radon separation.
• To provide a flexible and transferable framework for designing novel MOFs for radon capture.

Main Methods:

• Initial screening of MOFs using a pore-limiting diameter filter.
• Application of random forest classification and grand canonical Monte Carlo simulations.
• Variable importance analysis for interpretability and Materials Studio for mechanism elucidation.

Main Results:

• Discovery of two MOF candidates with superior APS in radon/N2 and radon/O2 systems compared to existing benchmarks.
• Achieved regenerability (R %) exceeding 85% for the identified MOF candidates.
• Demonstrated framework's model transferability across multiple datasets.

Conclusions:

• The proposed AI-driven framework effectively screens MOFs for radon separation.
• Identified MOFs offer enhanced performance for radon capture and regeneration.
• The framework provides a strategic approach for designing advanced MOFs for indoor air purification.