High-Throughput Screening and Prediction of High Modulus of Resilience Polymers Using Explainable Machine Learning

Abstract

The ability to store and release elastic strain energy, as well as mechanical strength, are crucial factors in both natural and man-made mechanical systems. The modulus of resilience (R) indicates a material's capacity to absorb and release elastic strain energy, with the yield strength (σ y ) and Young's modulus (E) as R = σ y 2 /(2E) for linear elastic solids. To improve the R in linear elastic solids, a high σ y and low E combination in materials is sought after. However, achieving this combination is a significant challenge as both properties typically increase together. To address this challenge, we propose a computational method to quickly identify polymers with a high modulus of resilience using machine learning (ML) and validate the predictions through high-fidelity molecular dynamics (MD) simulations. Our approach commences by training single-task ML models, multitask ML models, and Evidential Deep Learning models to forecast the mechanical properties of polymers based on experimentally reported values. Utilizing explainable ML models, we were able to determine the critical substructures that significantly impact the mechanical properties of polymers, such as E and σ y . This information can be utilized to create and develop new polymers with improved mechanical characteristics. Our single-task and multitask ML models can predict the properties of 12 854 real polymers and 8 million hypothetical polyimides and uncover 10 new real polymers and 10 hypothetical polyimides with exceptional modulus of resilience. The improved modulus of resilience of these novel polymers was validated through MD simulations. Our method efficiently speeds up the discovery of high-performing polymers using ML predictions and MD validation and can be applied to other polymer material discovery challenges, such as polymer membranes, dielectric polymers, and more.

Tuning the Mechanical and Dynamic Properties of Elastic Vitrimers by Tailoring the Substituents of Boronic Ester

“…Simulation studies have been conducted to reveal the dependence of binding energy of boronic ester cross-links on the mechanical performance of final materials . Therefore, the hybrid Monte Carlo (MC)/Molecular Dynamics (MD) model has been successfully applied to investigate associative polymer networks (see the Supporting Information).…”

Accelerating Polymer Discovery with Uncertainty-Guided PGCNN: Explainable AI for Predicting Properties and Mechanistic Insights

“…There is a rising trend in utilizing machine learning (ML) to achieve significant time and cost savings in the development of new polymer materials. − However, predicting multiple properties of polymers based on their monomer composition has remained a longstanding challenge in the field of material informatics and cheminformatics. − One of the primary obstacles has been the lack of open databases with easily accessible data. For instance, existing databases like PoLyInfo and Polymer Genome do not allow for automatic batch downloading of data.…”

“…Previous attempts at utilizing machine learning techniques for polymer structure–property prediction have employed Gaussian process regression, , support vector regression, and artificial neural networks. ,, These methods predominantly employ hand-crafted descriptors as input features. To improve predictive accuracy, researchers have increasingly utilized deep learning (DL) models, notably graph neural networks (GNNs), which exhibit promising capabilities in predicting polymer properties. , Within these models, the message passing graph neural network (MPNN) framework has gained significant traction due to its ability to effectively capture the intricate interplay among atoms within a molecule. , …”