268
X. Wu and J. A. El-Awady
addition, experimentally, the structure of bulk epoxies has been characterized using
measurements such as material density, degree of cross-linking, averaged length
between cross-linking sites, and distribution of free volume holes (e.g., [8, 17–19]).
However, such measurements provide average quantities and do not give precise
characterization of the local molecular structure, which is sometimes needed to
establish fundamental structure-property relationships [20, 21]. This is particularly
important when characterizing the interphase regions in epoxy-based composites.
These regions are of significant importance since they exhibit strong structural and
chemical variations within a relatively small volume and these variations greatly
influence the performance of the overall composites [8, 22]. Furthermore, while
scanning electron microscopy (SEM) and transmission electron microscopy (TEM)
techniques can be used to analyze the distribution of fibers within the epoxy matrix
and they have an adequate resolution to identify whether there are cracks at the
fiber/matrix interface [23], a direct experimental measurement of the local molecular
structure of the epoxy or the strength at the fiber/epoxy interface is still lacking.
On the other hand, simulations of the epoxy curing process are generally
simplified by only including the material constituents to create a neat cross-linked
network [24–26]. Such simulations are usually based on fully atomistic molecular
dynamics (MD) simulations with inputs from either density functional theory (DFT)
to calculate the chemical reactions during curing [27] or by utilizing reactive force
fields to create a cross-linked epoxy structure [28, 29]. Additionally, MD simulations
can also be utilized to provide an atomic level understanding of the local variations
of the molecular structure and properties in the bulk epoxy or near/at interface
regions as well as provide an avenue to establish structure-property relationships
for epoxies [23]. The main challenge for such atomistic-based simulations is
that the material properties do not directly match the experimentally measured
macroscale properties [30], which can be attributed to the small length and time
scales probed in the simulations [31]. On the opposite end of length scales, the
finite element method (FEM) is often used to predict the mechanical properties and
deformation/failure of epoxy and epoxy-based composites at the macroscale [32–
35]. In conventional FEM, the material parameters and the constitutive rules are
typically based on empirical observations coming from experimental measurements.
Thus, these simulations lack information from small scales, such as the variations
in local molecular structure and properties (e.g., in the interphase region). One of
the promising approaches to overcome some of the limitations of MD and FEM
simulations and to bridge between them is coarse-grained (CG) methods.
All in all, the study of epoxies and epoxy-based composites is motivated by their
advanced properties and applications. In order to achieve desirable properties of
epoxies, the relationships between the curing process (in bulk and at interphase)
and their molecular structure, as well as their thermal and mechanical properties,
need to be established. If these process-structure-property relationships were well
established, one would be able to create different epoxy systems and probe their
properties effectively and efficiently. Current experimental investigations alone
can hardly reveal the structure-property relationships in epoxy systems. Thus,
simulation methods, especially multiscale simulations, are in need of development.
X. Wu and J. A. El-Awady
addition, experimentally, the structure of bulk epoxies has been characterized using
measurements such as material density, degree of cross-linking, averaged length
between cross-linking sites, and distribution of free volume holes (e.g., [8, 17–19]).
However, such measurements provide average quantities and do not give precise
characterization of the local molecular structure, which is sometimes needed to
establish fundamental structure-property relationships [20, 21]. This is particularly
important when characterizing the interphase regions in epoxy-based composites.
These regions are of significant importance since they exhibit strong structural and
chemical variations within a relatively small volume and these variations greatly
influence the performance of the overall composites [8, 22]. Furthermore, while
scanning electron microscopy (SEM) and transmission electron microscopy (TEM)
techniques can be used to analyze the distribution of fibers within the epoxy matrix
and they have an adequate resolution to identify whether there are cracks at the
fiber/matrix interface [23], a direct experimental measurement of the local molecular
structure of the epoxy or the strength at the fiber/epoxy interface is still lacking.
On the other hand, simulations of the epoxy curing process are generally
simplified by only including the material constituents to create a neat cross-linked
network [24–26]. Such simulations are usually based on fully atomistic molecular
dynamics (MD) simulations with inputs from either density functional theory (DFT)
to calculate the chemical reactions during curing [27] or by utilizing reactive force
fields to create a cross-linked epoxy structure [28, 29]. Additionally, MD simulations
can also be utilized to provide an atomic level understanding of the local variations
of the molecular structure and properties in the bulk epoxy or near/at interface
regions as well as provide an avenue to establish structure-property relationships
for epoxies [23]. The main challenge for such atomistic-based simulations is
that the material properties do not directly match the experimentally measured
macroscale properties [30], which can be attributed to the small length and time
scales probed in the simulations [31]. On the opposite end of length scales, the
finite element method (FEM) is often used to predict the mechanical properties and
deformation/failure of epoxy and epoxy-based composites at the macroscale [32–
35]. In conventional FEM, the material parameters and the constitutive rules are
typically based on empirical observations coming from experimental measurements.
Thus, these simulations lack information from small scales, such as the variations
in local molecular structure and properties (e.g., in the interphase region). One of
the promising approaches to overcome some of the limitations of MD and FEM
simulations and to bridge between them is coarse-grained (CG) methods.
All in all, the study of epoxies and epoxy-based composites is motivated by their
advanced properties and applications. In order to achieve desirable properties of
epoxies, the relationships between the curing process (in bulk and at interphase)
and their molecular structure, as well as their thermal and mechanical properties,
need to be established. If these process-structure-property relationships were well
established, one would be able to create different epoxy systems and probe their
properties effectively and efficiently. Current experimental investigations alone
can hardly reveal the structure-property relationships in epoxy systems. Thus,
simulation methods, especially multiscale simulations, are in need of development.