Computational and Machine Learning Methods for CO₂ Capture Using Metal-Organic Frameworks

Area of Science:

• Computational chemistry and materials science
• Application of machine learning in materials discovery
• Adsorbent development for carbon capture

Background:

• Machine learning (ML) and atomic/molecular force fields (FFs) have advanced chemistry and materials science.
• Metal-organic frameworks (MOFs) are promising materials for carbon dioxide (CO2) capture.
• Understanding the relationship between ML-predicted forces and MOF structures is crucial for adsorbent design.

Purpose of the Study:

• To examine the interplay between ML, computational chemistry, and MOF development for CO2 capture.
• To connect ML-predicted atomic forces with MOF structures relevant to CO2 adsorption.
• To review data processing methods for ML algorithms in MOF research.

Main Methods:

• Review of ML applications in atomic and molecular force fields.
• Analysis of quantum ML successes in materials science.
• Examination of data sourcing techniques, including text mining and MOF formula processing.

Main Results:

• ML algorithms can predict atomic forces that correlate with MOF structures for CO2 adsorption.
• Quantum ML shows promise in accelerating materials discovery for CO2 capture.
• Data digitization and processing are key to training effective ML algorithms for MOF synthesis.

Conclusions:

• ML offers significant potential for advancing MOF adsorbent development for CO2 capture.
• Digitizing scientific records is recommended for efficient synthesis of advanced MOFs.
• A future vision for pioneering MOF synthesis routes for CO2 capture is presented.