Coarse-Grained Molecular Dynamics Study of the Curing and Properties of Highly Cross-Linked Epoxy Polymers

polyolefins, from which the volumetric expansive coefficients and glass transition temperatures (T g ) could be determined. Up to now, this method has been extensively applied to complex polymers, [4] including dendrimers [5,6] and crosslinked networks, [7,8] to examine effects of various factors (i.e., molecular weight, [9] tacticity, [10] confinement [11] ) on the volumetric properties.However, in order to achieve repeatable results, adequate thermodynamics equilibrium should be usually attained in these AA MD simulations. Although a great variety of schemes have been proposed to improve the sampling efficiency, restrained by the current computational power, the productive AA MD simulations are only generally suitable for investigating short-term behaviors of small short-chain polymer (namely, oligomer) systems. To study specific polymer systems at the wide timeand space-scales, one of quite promising schemes is multiscale modeling, which groups every AA segment into one super-atom or coarse-grained (CG) bead. [12,13] In this scheme, it is critical to obtain the "potential energy" (referred to free energy or potential of mean force in most cases) that can accurately describe chain structure and interbead interactions. [14,15] Numerous CG methods have been developed, [16] among which most typical ones include iterative Boltzmann inversion (IBI), [17] force match, [18][19][20] inverse Monte Carlo, [21] conditional reversible work (CRW), [22,23] and MARTINI, [24] etc. Mainly, these methods differ in the target functions that are used to parameterize the nonbonded CG potentials. For example, the IBI method [17] adopts the structural distributions, the CRW method [22] uses the interaction free energy, and the MARTINI method [24] aims to reproduce the partitioning free energies.In general, the CG potentials parameterized until now suffer from representability and transferability. [16,25,26] Namely, even at the same state as the parameterized one with some properties, simulations with these CG potentials cannot reproduce other properties; and CG potentials that are parameterized at some state cannot reliably be used at other states. In order to improve the predictive power of CG simulations, numerous studies have been carried out during the past decade. Here only some of typical ones are mentioned. As one of early works, Qian et al. [27] modified the IBI potentials by a temperature factor to reproduce the thermal expansion coefficients over a broad range of temperature. This method was called one-point extrapolation Multiscale SimulationsThe poly(ethylene oxide) (PEO) is employed as one typical example to demonstrate a new multiscale modeling scheme for simulating high-molecularweight polymeric melts. In this scheme, the structural distributions and the densities at five elevated temperatures at 1 atm, which are obtained from molecular dynamics (MD) simulations of all-atomistic oligomeric melt, are employed as the target functions to parameterize the coarse-grained (CG) potentials. The extensive CG MD simulations rep...

Section: Resultsmentioning

confidence: 99%

Section: Discussionmentioning

confidence: 99%

“…Furthermore, the bonded CG potentials are summations of the functions of the bond length l , the bond angle θ, and the dihedral angle φ, as follows

U_{bonded}^{CG} = \sum U_{normalb} + \sum U_{normala} + \sum U_{normald}

Section: Computational Detailsmentioning

confidence: 99%

Section: Computational Detailsmentioning

confidence: 99%

See 2 more Smart Citations

Multiscale Modeling Scheme for Simulating Polymeric Melts: Application to Poly(Ethylene Oxide)

2017

“…In such cases, for studying cross‐linked phenolic and other thermosetting resins that hold experimental constraints for structural analysis, a theoretical approach based on molecular dynamics (MD) simulations is expected to be effective . Fueled by the evolution of computational sciences in polymer physics and the computational performance of desktop computers over the past two decades, many studies of epoxy resins using atomistic and coarse‐grained MD simulations have been reported and the modeling methodology of cross‐linked epoxy resins has been refined through these studies, revealing numerous molecular‐level findings that cannot be experimentally obtained . Progress in the computational studies of epoxy resins over the last two decades has mainly been the result of rapid industrial growth of carbon‐fiber reinforced plastics that use epoxy resins as matrix resins.…”

Section: Introductionmentioning

confidence: 99%

Molecular Dynamics Simulations of Cross‐Linked Phenolic Resins Using a United‐Atom Model

Izumi

Shudo

Hagita

et al. 2018

because of their excellent physical properties such as mechanical strength, electrical insulation, heat resistance, and solvent resistance. [1] These properties result from their highly cross-linked network structures, which comprise phenolic and methy lene units. In these structures, three methylenes can connect to a phenolic ring such that one can be at the para (p) and two can be at the ortho (o) positions that are adjacent to the hydroxyl group of the phenolic ring, as shown in Figure 1. To further improve these properties, characterization of the relation between the network structure and the physical properties is one of the most important issues. However, this relation has not been completely understood experimentally because a fully cured resin is insoluble and infusible, which makes structural analysis difficult.In such cases, for studying cross-linked phenolic and other thermosetting resins that hold experimental constraints for structural analysis, a theoretical approach based on molecular dynamics (MD) simulations is expected to be effective. [2] Fueled by the evolution of computational sciences in polymer physics and the computational performance of desktop computers over the past two decades, many studies of epoxy resins using atomistic and coarse-grained MD simulations have been reported and the modeling methodology of cross-linked epoxy resins has been refined through these studies, revealing numerous molecularlevel findings that cannot be experimentally obtained. [2][3][4][5][6][7][8][9][10][11][12] Progress in the computational studies of epoxy resins over the last two decades has mainly been the result of rapid industrial growth of carbon-fiber reinforced plastics that use epoxy resins as matrix resins. However, detailed reports on atomistic simulations of cross-linked phenolic resins have been limited compared with the studies of epoxy resins. [13][14][15][16][17] We have investigated the application of atomistic MD simulation methods of epoxy resins to phenolic resins and were the first to propose a simple modeling method to generate highly cross-linked structures from a relaxed structure of phenolic oligomers. In the proposed method, a pair of nearest reactive carbons on the phenolic rings was connected via a newly inserted methylene unit, and this cross-linking reaction was repeated until a target degree of cross-linking was reached. [13] The structure modeling was successful, and a highly cross-linked structure (cross-linking degree of 0.92) was obtained. However, Molecular Dynamics SimulationsA new molecular modeling algorithm for conducting large-scale molecular dynamics simulation studies of cross-linked phenolic resins is developed using a united-atom model. A phenol-formaldehyde polycondensation system is simulated by a pseudoreaction algorithm taking into consideration (i) the difference in the experimental reaction rate constants at ortho and para positions of phenolic units and (ii) the geometry of the reactants. To avoid formation of locally strained cross-linked structures that c...

Complex Curing Pathways and Their Influence on the Phthalonitrile Resin Hardening and Elasticity

Rudyak

Гаврилов

Guseva

et al. 2017

Phthalonitrile compounds with siloxane bridges are recently suggested for producing thermosetting polymer composites with expanded processing range, excellent mechanical properties, and increased thermo‐oxidative stability. Mesoscale chemistry concept is adopted to describe multipathway curing process in phthalonitriles by adding triple link (triazine) formation into chemical reactions scheme. Coarse‐grained dissipative particle dynamics simulations have shown significant changes in matrix topology and mechanical properties with increase of the triazine reaction rate. In contrast to typical changes induced by high functionality crosslinkers, matrices with high content of triazine reach higher conversion degree, are less topologically interconnected, and are softer than ones without triazine.