Amorphous Metal-Organic Framework Database for Amorphization Prediction and CO₂ Direct Air Capture Screening.

Area of Science:

• Materials Science
• Computational Chemistry
• Chemical Engineering

Background:

• Amorphous metal-organic frameworks (AMOFs) represent a largely unexplored area with significant application potential.
• Efficiently navigating the chemical space of AMOFs requires data-driven methods to understand structure-property relationships.

Purpose of the Study:

• To develop a novel data-driven approach for exploring the amorphous phase of metal-organic frameworks (MOFs).
• To create a comprehensive database of AMOF structures and identify promising candidates for CO2 direct air capture (DAC).

Main Methods:

• Utilized chemically accurate reactive force field (ReaxFF) simulations for melt-quenching to generate a database of 487 AMOF structures.
• Developed and trained machine learning models on the AMOF database to predict thermal amorphization propensity based on structural features.
• Employed Widom insertion calculations to screen AMOF structures for CO2 direct air capture (DAC) applications.

Main Results:

• Successfully constructed a database of 487 amorphous metal-organic framework structures.
• Machine learning models accurately predicted the likelihood of MOFs undergoing thermal amorphization.
• Identified three novel AMOF structures with favorable characteristics for CO2 direct air capture.

Conclusions:

• The developed data-driven approach using ReaxFF simulations and machine learning is effective for exploring the amorphous phase of MOFs.
• This methodology expands the searchable chemical space beyond existing crystalline MOF databases.
• The identified AMOFs show promise for advancing CO2 capture technologies.