Abstract:Gas separation using polymer membranes promises to dramatically drive down the energy, carbon, and water intensity of traditional thermally driven separation, but developing the membrane materials is challenging. Here, we demonstrate a novel graph machine learning (ML) strategy to guide the experimental discovery of synthesizable polymer membranes with performances simultaneously exceeding the empirical upper bounds in multiple industrially important gas separation tasks. Two predicted candidates are synthesized and experimentally validated to perform beyond the upper bounds for multiple gas pairs (O2/N2, H2/CH4, and H2/N2). Notably, the O2/N2 separation selectivity is 1.6-6.7 times higher than existing polymer membranes. The molecular origin of the high performance is revealed by combining the inherent interpretability of our ML model, experimental characterization, and molecule-level simulation. Our study presents a unique explainable ML-experiment combination to tackle challenging energy material design problems in general, and the discovered polymers are beneficial for industrial gas separation.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to develop high - performance polymeric gas - separation membrane materials to overcome the high - cost problems of traditional thermally - driven separation techniques in terms of energy consumption, carbon emissions and water resource consumption. Specifically, through a novel graph - based machine learning (ML) strategy, the paper guides experimental discovery of synthetic polymeric membrane materials that can simultaneously exceed the empirical upper limits in multiple industrially important gas - separation tasks. The separation performance of gas pairs such as oxygen/nitrogen (O2/N2), hydrogen/methane (H2/CH4) and hydrogen/nitrogen (H2/N2) are of particular concern in the study. The paper points out that, despite the progress in theoretical and modelling tools, the development of new polymeric materials with high gas - separation performance is still very time - consuming, mainly because the traditional trial - and - error method is inefficient. In addition, existing polymeric membrane materials usually face a trade - off between permeability and selectivity, which is known as the "upper - bound line". To break through this limitation, the paper proposes a method that combines interpretable data - augmented graph neural networks (GREA) and semi - supervised graph - imbalanced regression (SGIR) for designing and screening polymeric candidate materials with potentially upper - bound - exceeding performance. Through this method, the researchers not only successfully predicted and synthesized two new polymers, but also these polymers showed significantly better performance than existing polymeric membranes in experimental verification, especially in terms of the selectivity of O2/N2 separation, reaching a level 1.6 to 6.7 times higher than that of existing polymeric membranes. This achievement is of great significance for industrial gas - separation applications, especially for key separation processes in the clean - energy industry and climate - change mitigation measures.

Superior Polymeric Gas Separation Membrane Designed by Explainable Graph Machine Learning

Recent Progress on Submicron Gas-Selective Polymeric Membranes

Machine learning enables interpretable discovery of innovative polymers for gas separation membranes

Molecular‐Sieving Membrane by Partitioning the Channels in Ultrafiltration Membrane by in Situ Polymerization

Machine learning in gas separation membrane developing: Ready for prime time

Accelerating Discovery of Polyimides with Intrinsic Microporosity for Membrane‐Based Gas Separation: Synergizing Physics‐Informed Performance Metrics and Active Learning

A perspective on data-driven screening and discovery of polymer membranes for gas separation, from the molecular structure to the industrial performance

Machine Learning‐Guided Design and Synthesis of Eco‐Friendly Poly(ethylene oxide) Membranes for High‐Efficacy CO2/N2 Separation

Design of Polymeric Membranes for Air Separation by Combining Machine Learning Tools with Computer Aided Molecular Design

Creation of Polymer Datasets with Targeted Backbones for Screening of High-Performance Membranes for Gas Separation

Combining Machine Learning and Molecular Simulations to Unlock Gas Separation Potentials of MOF Membranes and MOF/Polymer MMMs

Gas Separation Via Hybrid Metal–Organic Framework/Polymer Membranes

AI powered, automated discovery of polymer membranes for carbon capture

Integrating Molecular Simulations with Machine Learning Guides in the Design and Synthesis of [BMIM][BF4]/MOF Composites for CO2/N2 Separation

Ultra-selective molecular-sieving gas separation membranes enabled by multi-covalent-crosslinking of microporous polymer blends

A Review on the Morphology and Material Properties of the Gas Separation Membrane: Molecular Simulation

Imputation of missing gas permeability data for polymer membranes using machine learning

Rational Matching Between MOFs and Polymers in Mixed Matrix Membranes for Propylene/propane Separation

Machine Learning Guided Polyamide Membrane with Exceptional Solute-Solute Selectivity and Permeance

Polymeric Membranes for CO2 Separation and Capture

Machine Learning for Polymer Design to Enhance Pervaporation-Based Organic Recovery