(500g) A Multiscale Pipeline for Polymer Network Design and Mechanical Property Prediction with Reaction Detection | AIChE

(500g) A Multiscale Pipeline for Polymer Network Design and Mechanical Property Prediction with Reaction Detection

Authors 

Lin, T. S., Massachusetts Institute of Technology
Olsen, B., Massachusetts Institute of Technology
The design of reaction schemes that yield a material with quantitative control over properties is one of the largest challenges in chemistry. The Network Elasticity and Reaction Designer (NERD) is an open source software that enables the prediction and design of mechanical properties of polymer networks, one of the largest categories of materials used widely in engineering, medicine, and consumer products. NERD takes as input from the user a series of BigSMILES or SMILES strings representing polymer and small molecule reactants, aided by the first polymer stochastic graphical editor for users unfamiliar with the grammar of line notations. The reaction chemistry is predicted by publicly available small molecule reaction data from USPTO patent literature and a smaller set of published organic reaction templates. The network-forming polymers are then systematically coarse-grained by first using ab initio quantum chemical calculations of representative oligomers to generate energies for key conformational isomers, followed by application of the rotational isomeric states (RIS) theory to yield appropriate coarse-grained scaling parameters. NERD then applies reaction simulators to generate the topology of the polymer network, providing all needed input to predict the desired mechanical properties. While the initial version of NERD focuses on linear elastic modulus and fracture properties and is restricted to end-linked polymer networks, the computational structure of the platform can easily be expanded to support a broader array of property predictions and to update the accuracy of models as the underlying theory advances.