Machine Learning Analysis and Monomer Screening of Polyamide Nanofiltration Membranes for Lithium Separation

Area of Science:

• Materials Science
• Chemical Engineering
• Machine Learning

Background:

• Nanofiltration (NF) membranes are crucial for precise solute-solute separation, particularly for lithium/magnesium separation.
• Interfacial polymerization is a common synthesis method for NF membranes.
• Predicting membrane performance based on fabrication and operational conditions remains a challenge.

Purpose of the Study:

• To develop machine learning models to predict water permeability and lithium/magnesium selectivity of NF membranes.
• To identify key fabrication and membrane properties influencing NF membrane performance.
• To establish a framework for rational screening of novel monomers for NF membrane synthesis.

Main Methods:

• Utilized Morgan fingerprints (MFs) and molecular descriptors (MDs) to represent monomer properties.
• Employed explainable artificial intelligence (XAI) tools like Shapley additive explanations (SHAP) for performance analysis.
• Developed a material screening approach based on SHAP insights and synthetic accessibility.

Main Results:

• Identified critical relationships between synthesis conditions, membrane properties, and NF membrane performance.
• Screened and identified promising candidate monomers for NF membrane synthesis.
• Validated the screening approach by comparing screened monomers with traditional ones.

Conclusions:

• The study provides a novel, strategic framework for rational monomer screening in NF membrane synthesis.
• The developed approach accelerates the discovery of high-performance membranes for specific separation challenges.
• This work advances membrane technology by enabling tailored membrane design.