Inverse design of ZIFs through artificial intelligence methods.

Area of Science:

• Materials Science
• Computational Chemistry
• Chemical Engineering

Background:

• Zeolitic-imidazolate frameworks (ZIFs) are a subclass of metal-organic frameworks (MOFs) with tunable properties.
• Designing ZIFs for specific gas separation applications requires precise control over their diffusion properties.
• Existing design methods often lack the efficiency and specificity needed for industrial applications.

Purpose of the Study:

• To develop and validate a computational tool for designing fine-tuned ZIFs.
• To achieve desired diffusivity (Dᵢ) and selectivity (Dᵢ/Dⱼ) for gas mixtures.
• To demonstrate the tool's capability in meeting industrial performance criteria for gas separations.

Main Methods:

• Integration of a biologically inspired evolutionary algorithm with machine learning.
• Utilizing the tool to design ZIFs with target diffusion properties.
• Testing the designed ZIFs for permeability and selectivity in specific gas mixtures.

Main Results:

• Successful design of ZIFs tailored for specific gas diffusion requirements.
• Demonstrated efficacy in achieving target permeability and selectivity for industrial applications.
• Validation of the computational tool's predictive and design capabilities.

Conclusions:

• The developed computational tool effectively designs ZIFs for targeted gas separation applications.
• This approach offers a powerful strategy for accelerating the discovery of advanced materials for the chemical industry.
• The designed ZIFs meet critical industrial performance benchmarks for gas mixtures.