Figure 1. Flow chart of the developed integrated framework of cheminformatics and coarse-grained molecular dynamics (CG-MD) for prediction of the glass-transition temperature Tg of polymers. In this workflow, the molecular descriptors (x1, x2, x3 …) and property of interest (y) are analyzed to develop machine learning-based QSPR model. CG-MD simulations are then carried out to validate the QSPR model and provide physical insights into essential molecular features.
Figure 2. (A) A correlation plot between the observed and predicted values of Tg of polymers by the 7-variable QSPR model. (B) Contour plot of Tg behavior in the plane of grafting density f vs. cohesive interaction strength of the side chains εs.
Figure 3. Illustrations of CG models of polymers: (A) a linear polymer chain with f = 0, (B) an alternatively branched polymer chain with f = 0.5, and (C) a fully branched polymer chain with f = 1. (D) Snapshot of the bulk simulation box consisting of polymer chains.
In this work, the developed Quantitative Structure-Property Relationship (QSPR) model (Fig.2) identifies the most prominent molecular descriptors influencing the Tg of polymers. Coarse-grained molecular dynamics (CG-MD) simulations are performed to delineate mechanistic interpretation and systematic dependence of these influential molecular features on Tg (Fig. 3). This modeling framework not only identifies prominent structural features but also illustrates the mutual and competing influence of them on complex glass forming behavior of polymers. Their work is paving a way to establishing a materials-by-design framework for polymeric materials via molecular engineering.
References
[1] A. Karuth, A. Alesadi, W. Xia, B. Rasulev, Predicting glass transition of amorphous polymers by application of cheminformatics and molecular dynamics simulations, Polymer 218, 2021, 123495.