A novel molecular dynamics (MD) simulation methodology to capture brittle fracture in epoxybased thermoset polymer under mechanical loading is presented. The ductile behavior of amorphous polymers has been captured through traditional MD simulation methods by estimating the stress-strain response beyond the yield point; however, brittle fracture in highly crosslinked polymer materials such as epoxy thermoset has not been addressed appropriately and is the primary objective of this work. In this study, a numerically cured epoxy system comprising molecules of epoxy resin and hardener is generated. During the virtual deformation test, it is observed that the inherent molecular vibration due to temperature re-equilibrates the elongated covalent bonds and this molecular vibration impedes further stretching the covalent bond leading to scission. In order to overcome the influence of thermal vibration, an approach that employs deformation tests at absolute zero temperature conditiona concept borrowed from the quasi-continuum method, is developed. Bond dissociation energy is measured to quantify the extent of failure in the system by calculating the bond potentials during the deformation tests. Applying zero temperature condition to the deformation test, however, requires a large amount of computational time due to intermediate energy minimization processes. To improve the computational efficiency, an ultrahigh strain rate (UHSR ≈ 10 13 s -1 ) approach is developed by which the thermal vibration is decoupled from the deformation test using a strain rate higher than molecular vibration frequency. Note that the deformation test performed by traditional MD simulation methods used strain rates ranging 10 8 -10 10 s -1 . Simulation results show that the UHSR approach successfully captures brittle fracture in epoxy polymer due to covalent bond dissociation with high computational efficiency.

No abstract

The carbon fiber/polymer matrix interphase region plays an important role in the behavior and failure initiation of polymer matrix composites and accurate modeling techniques are needed to study the effects of this complex region on the composite response. This paper presents a high fidelity multiscale modeling framework integrating a novel molecular interphase model for the analysis of polymer matrix composites. The interphase model, consisting of voids in multiple graphene layers, enables the physical entanglement between the polymer matrix and the carbon fiber surface. The voids in the graphene layers are generated by intentionally removing carbon atoms, which better represents the irregularity of the carbon fiber surface. The molecular dynamics method calculates the interphase mechanical properties at the nanoscale, which are integrated within a high fidelity micromechanics theory. Additionally, progressive damage and failure theories are used at different scales in the modeling framework to capture scale-dependent failure of the composite. Comparisons between the current molecular interphase model and existing interphase models and experiments demonstrate that the current model captures larger stress gradients across the material interphase. These large stress gradients increase the viscoplasticity and damage effects at the interphase which are necessary for improved prediction of the nonlinear response and multiscale damage in composite materials.

A study of novel direct covalent grafting of photoactive mechanophore units into an epoxy matrix to create self-sensing thermoset network nanocomposites.

This study investigates the glass transition temperature (T g) of novel stresssensitive composites capable of detecting a damage precursor using molecular dynamics (MD) simulations. The molecular structures of a cross-linked epoxy network (which consist of epoxy resin, hardener and stress-sensitive material) have been simulated and experimentally validated. The chemical constituents of the molecular structures are di-glycidyl ether of bisphenol F (DGEBF: epoxy resin), di-ethylene tri-amine (DETA: hardener) and tris-(cinnamoyloxymethyl)-ethane (TCE: stress-sensitive material). The crosslinking degree is varied by manipulating the number of covalent bonds through tuning a cutoff distance between activated DGEBF and DETA during the non-equilibrium MD simulation. A relationship between the cross-linking degree and T g s has been studied numerically. In order to validate a proposed MD simulation framework, MD-predicted T g s of materials used in this study have been compared to the experimental results obtained by the differential scanning calorimetry (DSC). Two molecular models have been constructed for comparative study: (i) neat epoxy (epoxy resin with hardener) and (ii) smart polymer (neat epoxy with stress-sensitive material). The predicted T g s show close agreement with the DSC results.

This article presents a novel approach to model the mechanical response of smart polymeric materials. A cyclobutanebased mechanophore, named ''smart particle'' in this article, is embedded in an epoxy polymer matrix to form the selfsensing smart material. A spring-bead model is developed based on the results from molecular dynamics simulation at the nanoscale to represent bond clusters of a smart polymer. The spring-bead network model is developed through parametric studies and mechanical equivalence optimization to represent the microstructure of the material. A statistical network model is introduced, which is capable of bridging the high-accuracy molecular dynamics model at the nanoscale and the computationally efficient finite element model at the macroscale. A comparison between experimental and simulation results shows that the multiscale model can capture global mechanical response and local material properties.

This paper presents a multiscale modeling approach to simulate the self-sensing behavior of a load sensitive smart polymer material. A statistical spring-bead based network model is developed to bridge the molecular dynamics simulations at the nanoscale and the finite element model at the macroscale. Parametric studies are conducted on the developed network model to investigate the effects of the thermoset crosslinking degree on mechanical response of the self-sensing material. The comparison between experimental and simulation results shows that the multiscale framework is able to capture the global mechanical response with adequate accuracy and the network model is also capable of simulating the self-sensing phenomenon of the smart polymer. Finally, the molecular dynamics simulation and network model based simulation are implemented to evaluate damage initiation in the self-sensing material under monotonic loading. Nomenclature I= fluorescence intensity density = original length of the ith spring = variation of the spring length N = number of springs in this subzone = number of actual potential crosslinked covalent bonds = number of all potential crosslinked covalent bonds = loading strain = equivalent strain in the ith spring ̅ = average equivalent strain in the bead-centered subzone = crosslinking degree Keywords: spring-bead model, network model, statistical distribution of crosslinking degree, self-sensing smart material

No abstract