A fast transferable method for predicting the glass transition temperature of polymers from chemical structure

TL;DR

This paper introduces a rapid, transferable method combining Group Additive Properties and QSPR to accurately predict the glass transition temperature of polymers from their chemical structure, even outside the training data.

Contribution

The novel QSPR-GAP approach enables fast, accurate predictions of polymer T_g using minimal molecular descriptors, overcoming limitations of traditional GAP methods.

Findings

Only two descriptors needed for PAEK polymers.

Method successfully predicts T_g for polymers outside the training set.

Transferable to other properties and polymer classes.

Abstract

We present a new method that successfully predicts the glass transition temperature $T_{\! \textrm{g}}$ of polymers based on their monomer structure. The model combines ideas from Group Additive Properties (GAP) and Quantitative Structure Property Relationship (QSPR) methods, where GAP (or Group Contributions) assumes that sub-monomer motifs contribute additively to $T_{\! \textrm{g}}$, and QSPR links $T_{\! \textrm{g}}$ to the physico-chemical properties of the structure through a set of molecular descriptors. This method yields fast and accurate predictions of $T_{\! \textrm{g}}$ for polymers based on chemical motifs outside the data sample, which resolves the main limitation of the GAP approach. Using a genetic algorithm, we show that only two molecular descriptors are necessary to predict $T_{\! \textrm{g}}$ for PAEK polymers. Our QSPR-GAP method is readily transferred to other physical properties, to measures of activity (QSAR), or to different classes of polymers such as conjugated or bio-polymers.