Predictive computational modeling of polymer materials is necessary for the efficient design of composite materials and the corresponding processing methods. Molecular dynamics (MD) modeling is especially important for establishing accurate processing-structure-property relationships for neat resins. For MD modeling of amorphous polymer materials, an accurate force field is fundamental to reliable prediction of properties. Reactive force fields, in which chemical bonds can be formed or broken, offer further capability in predicting the mechanical behavior of amorphous polymers subjected to relatively large deformations. To this end, the Reactive Interface Force Field (IFF-R) has been recently developed to provide an efficient means to predict the behavior of materials under these conditions. Although IFF-R has been proven to be accurate for some crystalline organic and inorganic systems, it has not yet been proven to be accurate for amorphous polymer systems. The objective of this study is to use IFF-R to predict the thermomechanical properties of three different epoxy systems and validate with experimental measurements. The results indicate that IFF-R predicts thermo-mechanical properties that agree closely with experiment. Therefore, IFF-R can be used to reliably establish mechanical properties of polymers on the molecular level for future design of new composite materials and processing methods.

No abstract

No abstract

Carbon fiber and graphene-based nanostructures such as carbon nanotubes (CNTs) and defective structures have extraordinary potential as strong and lightweight materials. A longstanding bottleneck has been lack of understanding and implementation of atomic-scale engineering to harness the theoretical limits of modulus and tensile strength, of which only a fraction is routinely reached today. Here we demonstrate accurate and fast predictions of mechanical properties for CNTs and arbitrary 3D graphitic assemblies based on a training set of over 1000 stress-strain curves from cutting-edge reactive MD simulation and machine learning (ML). Several ML methods are compared and show that our newly proposed hierarchically structured graph neural networks with spatial information (HS-GNNs) achieve predictions in modulus and strength for any 3D nanostructure with only 5-10% error across a wide range of possible values. The reliability is sufficient for practical applications and a great improvement over off-the shelf ML methods with up to 50% deviation, as well as over earlier models for specific chemistry with 20% deviation. The algorithms allow more than 10 times faster mechanical property predictions than traditional molecular dynamics simulations, the identification of the role of defects and random 3D morphology, and high-throughput screening of 3D structures for enhanced mechanical properties. The algorithms can potentially be scaled to morphologies up to 100 nm in size, expanded for chemically similar compounds, and trained to predict a broader range of properties.