PolyRapid: Automated High-Throughput Screening of Polymers Using a Computational Workflow

Abstract

High-throughput computational screening of polymers offers a powerful way to address the imbalance between the vast number of polymers synthesised for diverse applications and the relatively small subset that can be studied using atomistic simulations. This work presents PolyRapid: an automatic workflow designed to enable the rapid and efficient screening of an extensive polymer library. In this work it is deployed on a library of 103 homopolymers. The workflow integrates an automated annealing protocol with adaptive control, allowing for reproducible simulations with minimal human intervention and minimisation of the computational cost. To this end, we test a number of quantitative conformational and energetic convergence metrics. It achieves equilibration in 95% of systems within four 30 ns annealing cycles, and 100% within nine annealing cycles saving, for our dataset, more than 2500 hours of compute compared to a fixed equilibration protocol. The simulation results are compared with experimental data and compiled into a publicly available repository which reports polymer temperature, tacticity and crystallinity. The availability of a homogenous large set of simulations enables the adoption of machine learning approaches for a variety of tasks. We exemplify this possibility by proposing rapid machine-learning-based method to predict the (computed) polymer density (F1-score = 0.91) and (experimental) glass transition temperature (F1-score = 0.76), using both chemical fingerprint representations and physically meaningful MD-derived descriptors.