Journal of Molecular Graphics and Modelling 117 (2022) 108311

Contents lists available at ScienceDirect

Journal of Molecular Graphics and Modelling

journal homepage: www.elsevier.com/locate/jmgm

Effect of adsorption, hardener, and temperature on mechanical properties

of epoxy nanocomposites with functionalized graphene: A molecular
dynamics study
Arman Salehi, Samrand Rash-Ahmadi *
Department of Mechanical Engineering, Urmia University, Urmia, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: Investigating Epoxy/hardener ratio and adsorption rate in epoxy/graphene oxide nanocomposites is of great
Nanocomposites importance, since these values can affect on the mechanical properties of the nanocomposite. In this study,
Mechanical properties molecular dynamics simulation was used to investigate and compare the mechanical properties of epoxy/gra­
Epoxy thermoset resins
phene and graphene oxide nanocomposites (EPON 828, EPON 862, Epoxy Novolac, and Cycloaliphatic Epoxy
Graphene oxide
Molecular dynamics
with GNs and GO). Also, the effect of different weight percentages of graphene oxide (0,1,3 and 5 wt %), different
weight percentages of epoxy compared to hardener, adsorption rate, and different temperatures were studied.
The results showed that increasing the weight percentage of graphene oxide in epoxy matrices improved the
adsorption rate between Epoxy/GO and the strength of nanocomposites. In addition, the amount of Young’s
modulus slightly decreased with increasing the temperature. Besides, the highest amount of Young modulus was
obtained by increasing the weight percentage of epoxy to hardener at 63:37 wt %. Moreover, by comparing the
mechanical properties of epoxy nanocomposites at 5 wt % graphene oxide, the highest Young modulus were
found to be related to Novolac/GO 4.27 Gpa and EPON 862/GO 4.24 Gpa. This study contributes to a more
comprehensive understanding on the behavior of the mechanical properties of epoxy/graphene oxide
nanocomposites.

1. Introduction reinforcer in polymer nanocomposites [8]. Epoxy nanocomposites

reinforced by graphene have wide applications and roles in various in­
According to statistics in recent years, polymer nanocomposites have dustries. Epoxy/graphene nanocomposites (Epoxy/GNs) include a wide
found a spreading of applications in various industries for their low range of industries: medicine [9], piezoresistive coatings [10], elec­
weight and decent mechanical strength compared to metals [1]. As a tronic component encapsulations [11], and thermal management de­
matrix in nanocomposites, epoxy resins are one of the most important vices [12].
and widely used polymers, utilized as matrix in polymer nano­ According to previous studies, functionalization of graphene nano­
composites on account of their excellent adhesion, toughness, thermal sheets with oxide functional groups can further improve the mechanical
properties, and proper electrical conductivity [2]. In addition, after its properties of epoxy/graphene nanocomposites [13–15]. In an experi­
discovery as a valuable material [3], graphene significantly improves mental study on the properties of epoxy/graphene oxide composites,
the mechanical properties compared to the pure matrix of these polymer Galpaya et al. [16] showed that adding 0.1 wt % graphene oxide to the
nanocomposites [4–7], based on its excellent thermal conductivity, epoxy matrix increases the toughness by 50%. Also, adding 0.5 wt %
mechanical resistance, and high flexibility as a reinforcer in polymer graphene oxide increases 35% of the Young modulus of the nano­
nanocomposites. Also, by functionalizing graphene nanosheets (GNs) composite. The results of the study by Wan et al. [17] indicated that by
with oxide groups, graphene can considerably affect the mechanical adding 0.25 wt% graphene oxide to the epoxy matrix, modulus and
properties of polymer nanocomposites compared to pristine graphene tensile strength of the composite increase 3.15 ± 0.11 to 3.56 ± 0.08
and ameliorate the mechanical properties of polymer nanocomposites. Gpa and 52.98 ± 5.82 to 92.94 ± 5.03 Gpa, respectively.
As a result, graphene oxide (GO) has become an effective material as a In the same vein, in an experimental study with an increase of

* Corresponding author.15311-57561, Iran.

E-mail address: [email protected] (S. Rash-Ahmadi).

https://doi.org/10.1016/j.jmgm.2022.108311
Received 12 June 2022; Received in revised form 4 August 2022; Accepted 16 August 2022
Available online 2 September 2022
1093-3263/© 2022 Elsevier Inc. All rights reserved.
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

Fig. 1. Molecular dynamic structure of hardener and epoxy thermoset resins.

graphene oxide (0.25–1 wt %), Huskić et al. [18] improved Young density on mechanical properties of DETDA + EPON828 (DGEBA).
module of Epoxy/GO composite for GO130, 2.40 ± 0.69 to 2.56 ± 0.32 Throughout their study, it was observed that as a result of the increase in
Gpa and for GO1200, 2.17 ± 0.14 to 2.46 ± 0.19 Gpa. Besides, the covalent bonds between the epoxy resin and the hardener, increasing the
experimental results of Zhao et al. [19] indicated that addition of 0.5 wt crosslink density in the DETDA + EPON828 matrix (DGEBA) enhanced
% graphene oxide to epoxy composite improves 80% tensile strength the mechanical properties of the epoxy resin.
and 49% flexural strength of epoxy/graphene oxide composite. Like­ Not all available experimental and laboratory tools can accurately
wise, the thermal and mechanical properties of the composite are describe the behavior properties of nanocomposites due to various fac­
improved as a result of adding 0.8 wt % graphene oxide to epoxy/­ tors such as the impact of environmental and laboratory conditions on
graphene oxide composite [20]. the behavior of materials and also limitations in describing the proper­
In a molecular dynamics (MD) study, Shokuhfar and Arab [21] used ties of materials at the molecular level. In fact, using computational
molecular dynamics simulations to investigate the effect of crosslink simulations and comparing them with laboratory and experimental

Fig. 2. The process of Crosslinking epoxy resins with hardener (DETDA) in two steps.

2
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

Fig. 3. Graphic output of simulation of epoxy/graphene oxide nanocomposite and Covalent bond formation between graphene carbon atoms (a)Epoxide group (b)
Hydroxyl group.

methods can provide us with a more describing and comprehensive structures of epoxy/graphene and epoxy/graphene oxide nano­
understanding on the behavior of materials. One of these computational composites were modeled. After modeling the nanocomposite, the force
tools is molecular dynamics simulation which is able to predict the potentials between the atoms of thermoset resins, the atoms of graphene
behavior of materials and nanocomposites at molecular scales. In this oxide, and graphene were determined.
study, molecular dynamics simulation was used to investigate the me­
chanical properties of various types of epoxy thermoset resins nano­ 2.1. Modeling of epoxy resins and hardener structures
composites EPON 828 (diglycidyl ether of bisphenol A, DGEBA), EPON
862 (bisphenol F diglycidyl ether, DGEBF), CE (Cycloaliphatic Epoxy), In this section, four types of epoxy thermoset resins were modeled
and Novolac (Epoxy Novolac) reinforced with graphene and graphene including: EPON 828 (diglycidyl ether of bisphenol A, DGEBA), EPON
oxide under the different weight percentages of oxide groups (0,1,3 and 862 (bisphenol F diglycidyl ether, DGEBF), CE (Cycloaliphatic Epoxy),
5 wt %). Also, all nanocomposites were subjected to axial tension. In and Novolac (Epoxy Novolac). The molecular dynamic structures of
addition, the mechanical properties of nanocomposites were investi­ epoxy resins and hardener of DETDA (diethyltoluenediamine) are
gated under different weight percentages of diethyltoluenediamine shown in Fig. 1. These four types of resins are almost the most widely
hardener (DETDA) relative to the epoxy matrix. Moreover, the effect of used resins in various industries, which were also utilized in the current
different temperatures on the mechanical properties of nanocomposites study [22]. In general, in order to create crosslinks between epoxy
was studied. The results of the Young module and tensile strength thermoset molecules and hardener (DETDA) molecules, each nitrogen
simulation of nanocomposites were compared to the experimental re­ atom (N) of hardener (DETDA) can form a covalent bond to two carbon
sults of the previous literature with a high level of overlap. The study atoms (C) of epoxide groups in the epoxy resins [23]. Fig. 2 shows the
and comparison of mechanical properties (Young’s modulus and tensile formation of epoxy resins as well as the crosslinking between epoxy
strength) of nanocomposites provide us with a broader understanding resins and hardeners (DETDA).
on the behavior of epoxy/graphene nanocomposites for the widespread
use and role of these nanocomposites in various industries as well as 2.2. Modeling of graphene nanosheet and graphene oxide structures
different environmental conditions. In addition, this study was neces­ (GNs, GO)
sary due to the lack of satisfactory study on the comparison of me­
chanical properties of various epoxy/graphene oxide nanocomposites. The dimensions of graphene and graphene oxide at 50 × 50 Å were
modeled by material studio software [24]. The number of carbon atoms
2. Simulation and method in graphene is 1008 atoms. Two types of epoxide (-O-) and hydroxyl
(-OH) oxide functional groups were used to functionalize graphene
Prior to starting the simulation, the dimensions and molecular nanosheets with oxide functional groups. On the other hand, two

3
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

Table 1 simulation curing process was performed by Kallivokas et al. [25]. Fig. 3
Total number of atoms of epoxy/graphene, graphene oxide nanocomposites. shows the graphical output of the simulation of epoxy/graphene oxide
Nanocomposites Number of atoms of nanocomposites in different oxide functional nanocomposites under the different weight percentages of graphene
groups oxide.
0 wt% (GNs) 1 wt% 3 wt% 5 wt%
2.4. Force field
Novolac/GO 5210 5919 6063 6096
EPON 862/GO 5072 5204 5933 5957
EPON 828/GO 5201 5913 6153 6178 To calculate the potential function of nanoparticles in epoxy/gra­
CE/GO 5312 5902 6390 6408 phene oxide nanocomposites, the sum of two potentials of Tersoff and
12-6 Lenard-Jones was used by hybrid/overlay command. Tersoff po­
tential function [26] was employed for the interactions between the
covalent bonds were used to functionalize the graphene nanosheets by
carbon atoms in graphene nanosheets. Also, the 12-6 Lenard-Jones po­
the epoxide functional group, which included the bonding of one oxygen
tential function [27] was recruited for non-bonded or van der Waals
atom with two graphene carbon atoms. Also, the hydroxyl functional
interactions between epoxy resins, hardener (DETDA) and Oxide func­
group consists of one covalent bond between this functional group and a
tional groups attached to graphene nanosheets. Moreover, the cutoff
graphene carbon. Fig. 3 shows how to create a covalent bond between
radius was considered 30 Å.
the carbon atoms of a graphene nanosheet with the epoxide and hy­
Harmonic bond energy is used for interactions between bond length
droxyl functional groups. These oxide functional groups were randomly
atoms, whose equation is:
distributed under different weight percentages (1, 3 and 5 wt %) on
graphene nanosheets. In addition, the method of calculating is given in Ebonds = Kr (r − r0 )2 (2–1)
equation (1) for the weight percentage of oxide functional groups in
graphene nanosheets. When graphene nanosheets are modeled by Ma­ r is the force constant, r is the distance between two atomic bonds, r0 is
terial Studio software, the number of graphene carbon atoms as well as the bond equilibrium distance for two atomic bonds. The harmonic
their atomic mass is also known. Also, by adding atoms of oxide func­ angle energy equation was used for the interactions between angle bond
tional group (hydroxyl, epoxide) to graphene nanosheets, the weight atoms, and its equation is as follows:
percentage of functional groups The oxide was determined using equa­
tion (1). Weight percentages of oxide groups in graphene nanosheets Eangle = Kθ (θ − θ0 )2 (2-2)
were determined through the total mass of atoms of oxide functional Kθ is the force constant, θ0 is the angle between bonds, θ0 is the
groups, divided by the total mass of graphene oxide atoms, and multi­ balance of the angle bond. The Opls dihedral energy equation was uti­
plied by 100 weight percent of graphene oxide nanosheets. lized to create interactions between angle dihedrals, and the dihedral
mass of oxide functional groups energy equation is:
wt. % = × 100 (1)
mass of Graphen oxide nanosheets 1 1 1
Edihedrals = K1 [1 + cos( ∅ )] + K2 [1 − cos(2 ∅ )] + K3 [1 + cos(3 ∅ )]
2 2 2
1
2.3. Modeling the structure of epoxy/graphene oxide nanocomposites + K4 [1 − cos(4 ∅ )] (2–3)
2
∅ is the dihedral angle, K is the force constant. Also, 12-6 Leonard
At first, after modeling graphene nanosheets and adding specific
Jones equation was used to show non-bonded or van der Waals in­
weight percentages of oxide functional groups, types of epoxy resins and
teractions with its equation as below:
hardener (DETDA) were modeled using Material Studio software. Then
[( ) ]
in a specific dimension (50x50x26), 70% of epoxy resin and 30% of σ 12 (σ)6
Enon bond = 4ε − , r < rc (3)
hardener were added to the graphene oxide nanosheets in these specific r r
dimensions. Next, the data file was taken from the modeled
σ is the equilibrium distance parameter obtained from the arithmetic
nanocomposites.
mean between two atoms, ε is the well depth parameter and is obtained
Four types of resins were used to model the structure of epoxy/gra­
from the geometric mean between atomic surfaces, r is the distance
phene oxide nanocomposites for composite matrix including: EPON 828
between non-bonded atoms, rc is the cut-off radius. Tersoff potential was
(diglycidyl ether of bisphenol A, DGEBA), EPON 862 (bisphenol F
also used to apply covalent bonds between carbon atoms of graphene
diglycidyl ether, DGEBF), CE (Cycloaliphatic Epoxy), and Novolac
nanosheets. Energy equation is as follows:
(Epoxy Novolac). A type of hardener DETDA (diethyltoluenediamine)
was employed as a hardener (DETDA) in these epoxy/graphene oxide 1 ∑∑
E= Vij (4)
nanocomposites. The weight percentage of epoxy to hardener for all 2 i i∕
=j
modeled nanocomposites was 70:30 wt %, including 70 wt % epoxy
matrix and 30 wt % hardener. In addition, to strengthen epoxy matrix, E is the energy of the whole system, Vij is the bonding energy.
graphene and graphene oxide were added to epoxy nanocomposites
under different percentages of oxide functional groups. Also, function­ 2.5. Details of MD simulation
alized graphene nanosheets were utilized to oxide functional groups
(hydroxyl and epoxide) (-OH) and (-O-) in different weight percentages The entire epoxy/graphene oxide nanocomposites simulations were
(1,3 and 5 wt %). Table 1 shows the total number of atoms which performed using open source code Large-scale Atomic/Molecular
separately create epoxy/graphene and graphene oxide nanocomposites Massively Parallel Simulator (LAMMPS) [28]. After obtaining the data
in different weight percentages. Dimensions of epoxy/graphene oxide files of nanocomposites, these data files were executed by Lammps
nanocomposites were modeled as cubic with dimensions of 50 x 50 × 26 software. Next, the mentioned potential forces were applied. In addition,
Å. The density value of epoxy matrix is also 1.3 g/cm3. The process of the type of material atoms, coefficients, bond lengths, angles, and di­
curing and crosslinking epoxy resins with a hardener in nanocomposites hedrals between different atoms were determined. Besides, periodic
is such that for each epoxide group of epoxy resins, one of the bonds of boundary conditions were used for all nanocomposite simulations.
the oxygen atom with carbon is broken, and also a hydrogen atom is Table 1 shows the number of atoms of the simulated nanocomposites.
added to the oxygen atom. In the molecule hardener (DETDA), two The system energy was minimized before starting the simulation. Also,
hydrogen atoms separate from each nitrogen atom of hardener. This (isothermal-isobaric) ensemble (NPT) was used to minimize stress and

4
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

Fig. 4. Stress-strain diagrams of Epoxy/GO nanocomposites at different weight percentages of graphene oxide at 300 K.

relaxation between the atoms of the simulation system at a pressure of 1

atm and a temperature of 300 K. The simulation system was brought to 2.8. The radial distribution function (RDF)
temperature and volume equilibrium using (constant volume and con­
stant temperature) ensemble (NVT) at 300 K. Also, time step of 0.0001 The radial distribution function (RDF) in equations is represented by
ps was considered. g (r) indicating the probability of finding a particle at a distance r from
another particle, which is calculated by the following equation [29]:
2.6. Mechanical deformation V nB
g(r) = . (6)
NB 4πr2 dr
To peruse the mechanical properties of epoxy/graphene and gra­
NB is the number of atoms B, nB is the number of atoms of type B
phene oxide nanocomposites, a total of 32 tensile strength simulations
which are at distance)r(from the atoms A, V is the total volume of the
were performed to obtain the maximum strain stress and Young’s
nanocomposites.
modulus of nanocomposites. For this purpose, the simulation box
(modeled nanocomposites) was first compressed with the change box
3. Results and discussion
command. At the next step, it was subjected to axial tension along the X,
Y and Z directions by the deformation command. The strain rate, used
3.1. The effect of oxide functional groups on the mechanical properties of
for the whole simulation of epoxy/graphene nanocomposites, was
Epoxy/GO nanocomposites
considered 0.7 Ps− 1 .

The effect of different weight percentages of oxide functional groups

2.7. Interaction energy
on the mechanical properties of Epoxy/GO nanocomposites is shown in
the diagrams of Fig. 4 and Fig. 5. According to Fig. 4, the stress-strain
The interaction energy parameter is shown for calculating the
diagram of Epoxy/GO nanocomposite is in different weight percent­
adhesion rate between matrix and reinforcement. According to equation
ages of graphene oxide. Based on the diagrams in this Figure, by
(5), the interaction energy is calculated by the total potential energy of
increasing the weight percentage of oxide groups to graphene oxide in
the nanocomposite subtracted from the potential energy of epoxy matrix
epoxy nanocomposites (Epoxy/GO), the strain-stress diagrams in
and graphene oxide. As the amount of negative value of interaction
Epoxy/GO nanocomposites increased and also improved the strength of
energy escalates, the amount of adhesion increases.
Epoxy/GO nanocomposites. This can be due to the increased weight
( )
Einteraction = EEpoxy/GO − EEpoxy + EGO (5) percentages of graphene oxide, enhancing the interaction and adsorp­
tion between the epoxy matrix and graphene oxide, which in turn

5
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

Fig. 5. Comparison Young’s modulus of Epoxy/GO nanocomposites at Different Weight Percentages of graphene oxide at 300 K in three directions (X, Y, Z).

Table 2
Comparison of Young’s modulus of Epoxy/GO nanocomposites in different weight percentages of graphene oxide.
System GO (wt. %) Tensile strength (MPa) young modulus (Gpa) Method Ref.

EPON862/GNs 0 – 2.59 MD [32]

EPON828(DGEBA)/GO 0.5 64.79 3.36 ± 0.17 Experimental [17]
EPON828(DGEBA)/GO 0.7 50.8 2 MD [31]
EPON828(DERAKANE)/GO 0.8–1.2 75.93–58.12 2.89–2.81 Experimental [20]

EPON862/GNs 0 45.57 2.67 MD This study

EPON862/GO 1 64.13 3.52 MD This study
EPON862/GO 3 64.55 4.20 MD This study
EPON862/GO 5 65.13 4.24 MD This study

EPON828/GNs 0 41.50 2.20 MD This study

EPON828/GO 1 59.45 3.46 MD This study
EPON828/GO 3 63.06 3.49 MD This study
EPON828/GO 5 66.18 3.95 MD This study

CE/GNs 0 55.57 3.25 MD This study

CE/GO 1 58.38 3.54 MD This study
CE/GO 3 66.99 3.68 MD This study
CE/GO 5 69.65 3.75 MD This study

Novolac/GNs 0 65.12 3.55 MD This study

Novolac/GO 1 66.15 3.87 MD This study
Novolac/GO 3 72.10 4.17 MD This study
Novolac/GO 5 64.36 4.27 MD This study

increases the adhesion between the epoxy matrix and graphene oxide. functional groups in graphene oxide nanosheets significantly affects the
These factors intensify the strength of nanocomposites by raising the mechanical properties and strength of epoxy/graphene oxide nano­
oxide group weight percentages of graphene nanosheets. Also, these composites. In addition, increasing the weight percentage of oxide
findings are agreement with the experimental results of Huskić et al. groups to graphene oxide in Epoxy/GO nanocomposites raises the sur­
[18], which were obtained from increasing the surface interaction and face interaction between epoxy resins and oxidized graphene nano­
adhesion of epoxy matrix to graphene oxide with a rising weight per­ sheets, thus improving the strength of nanocomposites. Besides,
centage of graphene oxide. Escalating the weight percentage of oxide according to the diagrams in Fig. 4, functionalizing graphene nanosheets

6
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

Fig. 6. Interaction energy of epoxy/graphene oxide nanocomposites a) Novolac/GO b) EPON862/GO c) EPON828/GO d) CE/GO in different weight percentage
at 300K

with oxide groups (0–1 wt %) significantly escalated the strength of comparing the Young’s modulus of nanocomposites in all three di­
nanocomposites. As shown in Fig. 4 in (b) and (c), such rise in strength rections, the amount of Young’s modulus in the Z direction is much
for EPON862/GNs and EPON828/GNs nanocomposites had a greater lower than the X, Y. The reason is the reduction of van der Waals and
increase in strength than the other two nanocomposites. The reason for electrostatic forces between the epoxy matrix and graphene oxide in the
this rise in strength might be related to the surface interaction between Z direction. Moreover, by comparing Young’s modulus of these nano­
the properties of the epoxy resin matrix with the oxide functional groups composites in Table 2 in different weight percentages of graphene oxide,
in graphene nanosheets. the highest amount of Young’s modulus is related to Novolac/GO, fol­
Also, the diagram of Fig. 5 shows the changes in Young’s modulus (in lowed by EPON862/G O, EPON828/GO, and CE/GO.
Three directions) of Epoxy/GO nanocomposites in terms of different In an MD study on graphene oxide-reinforced polyvinyl alcohol
weight percentages of graphene oxide (0,1,3 and 5 wt%). According to (PVA) nanocomposites functionalized with oxide groups (epoxide,
Table 2 and the diagrams in Fig. 5, by comparing the Young modulus (in carboxyl, hydroxyl, and carbonyl), Wang et al. [29] stated that when the
Y direction) of Epoxy/GO nanocomposites in different weight percent­ number of oxygen atoms of the functional groups in the graphene oxide
ages of graphene oxide (0,1,3 and 5 wt%), it is observed that with raising of the polyvinyl alcohol is greater, the surface interaction between the
weight percentage of oxide groups in Graphene oxide nanosheets of matrix (PVA) and the graphene oxide is subsequently stronger. This
(0–5 wt%), Young modulus of nanocomposites increase for each as factor can considerably influence the mechanical properties of PVA/GO
follow: EPON862/GO, 59%; EPON828/GO, 80%; Novolac/GO, 21%; nanocomposites. In this study, by raising the weight percentages of
and CE/GO, 16%. The prediction of the Young modulus results obtained graphene oxide, epoxy/graphene oxide nanocomposites also increased
from the MD simulation is in agreement with the experimental results of the number of oxygen atoms present in graphene oxide and thus effec­
the study by Li et al. [30] on epoxy/graphene oxide composites. In tively improved the Young modulus of epoxy/graphene oxide nano­
addition, the results of predicting the Young’s modulus value of nano­ composites. These findings are in line with the results obtained by Wang
composites in MD method were slightly higher than the results of Li et al.
et al., since the simulation environment was more ideal than the labo­ Also, the molecular dynamics studies of Shukla et al. [31] were on
ratory environment, and the environmental conditions might affect the epoxy (DGEBA, EPON828) nanocomposites with hardener (TETA)
results of the experiment. Also, according to Fig. 5, with the increase in reinforced with graphene oxide nanosheets under different weight per­
the weight percentage of graphene oxide in all three directions, the centages. They used two types of columbic and 6–9 Leonard Jones po­
amount of Young’s modulus of nanocomposites increases. Besides, by tential function for the interactions between epoxy resin with hardener

7
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

Fig. 7. (a) Changes stress-strain of epoxy/graphene oxide composite in different weight percentages of hardener at 300 K(b) changes of Young’s modulus at different
weight percentages of hardener at 300 K.

and graphene oxide nanosheets. The results of Young’s modulus and epoxy matrix. Based on this part, by increasing the weight percentage of
tensile strength of Shukla et al.‘s nanocomposites are shown in Table 2, hardener (DETDA) relative to the epoxy matrix, the amount of cross­
which overlap well with the results of Young’s modulus and tensile linking between the epoxy molecules and the hardener molecules in­
strength of the current study. In addition, by comparing tensile strength creases and results in an improvement of the strength of the
of these nanocomposites in Table 2 in different weight percentages of nanocomposite. Also, it can be said that with raising the weight per­
graphene oxide, the highest amount of Tensile strength is related to centage of the epoxy matrix compared to hardener (DETDA), the num­
Novolac/GO nanocomposites. The prediction of the tensile strength re­ ber of hardener molecules (DETDA) increased, and this rise in the
sults obtained from the MD simulation is in agreement with the exper­ number of hardener molecules expanded the possibility of forming more
imental results of previous studies on epoxy/graphene oxide covalent bonds with epoxy resin molecules. As a result, the strength of
nanocomposites. Novolac/graphene oxide nanocomposites were enhanced.
On the other hand, according to Fig. 7, part (b) shows Young’s
3.2. Investigation of adsorption rate between epoxy matrices with modulus of epoxy/graphene oxide nanocomposite in relation to
graphene and graphene oxide nanosheets in nanocomposites different weight percentages of hardener. Based on the diagrams in this
Fig. 7 (b), by increasing the weight percentage of epoxy matrix
The effect of different weight percentages of graphene oxide on the compared to the hardener (DETDA) 90:10 to 74:26 wt % Young’s
amount of interaction energy between different Epoxy/GO nano­ modulus increases from 3.53 to 4.02 Gpa. Also, with increasing the
composites is shown in Fig. 6. As can be seen, in all nanocomposites, the weight percentage of epoxy matrix compared to the hardener from
amount of interaction energy increases with the increasing weight per­ 74:26 to 68:32 wt%, the amount of Young modulus decreases to 4.02
centages of graphene oxide. Besides, the highest amount of interaction Gpa, and its value is fixed. However, with the increase of hardener
energy (Adsorption) was related to Novolac/GO and EPON862/GO weight percentage in 63:37 wt%, the Young Modulus nanocomposite
nanocomposites. The amount of interaction energy is directly related to reaches its maximum value of 4.26 Gpa. Moreover, Young Modulus is
the adhesion and adsorption of the matrix with reinforcement in nano­ reduced and fixed at 4.06 Gpa with the increase of hardener weight
composites. As a result, increased interaction energy can improve the percentage from 63:37 to 48:52 wt%. Based on the results of the dia­
mechanical properties of nanocomposites. One of the important issues of grams in Fig. 7 (b), it can be said that at 63:37 wt% the hardener weight
the study was the effect of graphene oxide in different weight percent­ percentage has the highest Young modulus in epoxy/graphene oxide
ages of oxide on the mechanical properties of epoxy/graphene oxide nanocomposite relative to the weight percentage of epoxy matrix or
nanocomposites. This effect was critical for the improvement of the approximate amount of epoxy to hardener between 65:35 to 60:40 wt%.
mechanical properties of epoxy/graphene oxide nanocomposites as a Most of the mechanical properties of the nanocomposite (Young’s
result of adding oxide functional groups to graphene oxide in epoxy modulus) occur at these weight percentages. This rise in Young’s
nanocomposites. As the other significance, such effect created nano­ modulus and the strength of the nanocomposite from 74:26 to 63:37 wt
composites with more efficient and ideal mechanical properties. % is due to the increase in crosslinking between the epoxy resin mole­
cules and the hardener (DETDA) molecules, which leads to the forma­
tion of a covalent bond between the molecules.
3.3. Effect of weight percentage of hardener (DETDA) on mechanical
There is also the point that hardener molecules (DETDA) have the
properties of Epoxy/GO nanocomposite
capacity to form covalent bonds with epoxy molecules [21], which is
due to the greater increase in the weight of epoxy resin than the hard­
In this section, among the epoxy matrices, epoxy Novolac was
ener 56:44 to 48:52 wt %. The covalent bond formation capacity is
selected as the candidate among the epoxy resins by the matrix, in order
incomplete, and there is no possibility of covalent bond formation.
to pursue the effect of the weight percentages of hardener (DETDA) in
Therefore, it keeps the strength and young modulus of nanocomposites
this matrix on the mechanical properties of this nanocomposite. The
constant during increasing the weight percentage of hardener compared
diagrams in Fig. 7 show the effect of the weight percentage of hardener
to epoxy resin. This is because the weight percentage of the epoxy matrix
(DETDA) in the epoxy matrix on the mechanical properties of the epoxy/
than hardener is important. In other words, in case the amount of
graphene oxide nanocomposite. According to these diagrams, part (a)
hardener molecules is slightly higher than epoxy resin molecules, the
shows the stress-strain diagram of epoxy/graphene oxide nano­
capacity of hardener molecules will not be filled to form bonds with
composites at different weight percentages of hardener relative to the

8
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

Table 3 nanocomposites in the Z direction is much lower than in the X and Y

Young’s modulus of Epoxy/GO nanocomposites in different weight percentages directions, which indicates the anisotropic of nanocomposites. The
of graphene oxide in three directions (X, Y, Z). reason for this difference in Young’s modulus can be due to the decrease
System GO (wt. %) Young modulus (Gpa) of electrostatic and van der Waals forces between graphene oxide
X direction Y direction Z direction
nanosheets and epoxy matrix in that direction (Z).

EPON862/GNs 0 2.60 2.67 1.31

EPON862/GO 1 3.41 3.52 1.78 3.4. The radial distribution function (RDF) of epoxy/graphene and
EPON862/GO 3 4.07 4.20 2.13 graphene oxide nanocomposites
EPON862/GO 5 4.16 4.24 2.15

EPON828/GNs 0 2.18 2.20 1.1 Fig. 8 shows the diagrams of the radial distribution function (RDF)
EPON828/GO 1 3.34 3.46 1.74 between carbon and hydrogen atoms in Epoxy/GO and Epoxy/GNs
EPON828/GO 3 3.37 3.49 1.76 nanocomposites under different graphene oxide weight percentages.
EPON828/GO 5 3.83 3.95 1.96
According to the diagrams, in all cases, the distance (r) of the first peak is
CE/GNs 0 3.14 3.25 1.63 approximately equal to 1.1 Å. This distance indicates the bond length
CE/GO 1 3.43 3.54 1.78 between the carbon and hydrogen atoms, which is completely consistent
CE/GO 3 3.56 3.68 1.84
CE/GO 5 3.64 3.75 1.89
with the previous literature, and the results of the radial distribution
function (RDF) diagrams of Shokuhfar and Arab [21]. According to the
Novolac/GNs 0 3.44 3.55 1.79
diagrams in Fig. 8, by increasing weight percentages of graphene oxide,
Novolac/GO 1 3.75 3.87 1.95
Novolac/GO 3 4.05 4.17 2.12 the peak heights decrease slightly, which indicates an increase in the
Novolac/GO 5 4.15 4.27 2.15 higher density of atoms in Epoxy/GO nanocomposites. According to
these diagrams, with increasing distance, the amount of peaks decreases,
balances, and approaches the g(r) = 1.
epoxy resin molecules. In addition, if the amount of hardener molecules
is much less than epoxy resin molecules, the hardener capacity is full to
3.5. The effect of different temperatures on Young’s modulus of epoxy/
form bonds with epoxy resin molecules and requires more hardener
graphene oxide nanocomposites
molecules. The reason to this issue is that the weight percentage of the
epoxy than hardener is an important matrix and can considerably in­
Fig. 9 shows the effect of different temperatures on the young’s
fluence the mechanical properties of nanocomposites.
modules of epoxy novolac/graphene oxide nanocomposite at 3 wt %.
Table 3 shows the value of Young’s modulus of nanocomposites in
The Novolac/graphene oxide nanocomposite was stretched under
three directions (X, Y, Z) under different weight percentages of graphene
different temperatures in the Y direction. Fig. 9 shows the changes in
oxide. According to Table 3, the value of Young’s modulus of
Young’s modulus diagram of epoxy/graphene oxide nanocomposites in

Fig. 8. The radial distribution function (RDF) of C–H (a)Novolac/GO (b)EPON862/GO (c)EPON828/GO (d)CE/GO nanocomposites at different weight percentages
of graphene oxide at 300K

9
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

in Young’s modulus. The results of this study contribute to a more

comprehensive understanding on the mechanical properties of epoxy/
graphene and graphene oxide composites. In addition, it can lead to the
development and further application of graphene oxide reinforced
epoxy composites.

Declaration of competing interest

The authors declare that they have no known competing financial

interests or personal relationships that could have appeared to influence
the work reported in this paper.

Data availability

Data will be made available on request.

References
Fig. 9. Changes in Young’s modulus of novolac/graphene oxide nano­
composites at different temperatures. [1] M. El Rhazi, S. Majid, M. Elbasri, F.E. Salih, L. Oularbi, K. Lafdi, Recent progress in
nanocomposites based on conducting polymer: application as electrochemical
sensors, Int. Nano Lett. 8 (2018) 79–99.
terms of different temperatures. As shown in this Figure, Young’s
[2] P. Mohan, A critical review: the modification, properties, and applications of epoxy
modulus decreases from 4.180 to 4.158 Gpa with an increase in tem­ resins, Polym.-Plast. Technol. Eng. 52 (2013) 107–125.
perature from 300 to 700 K. The reasons to the decrease in Young [3] K.S. Novoselov, A.K. Geim, S.V. Morozov, D.-e. Jiang, Y. Zhang, S.V. Dubonos, et
modulus and strength of epoxy/graphene oxide nanocomposite with al., Electric field effect in atomically thin carbon films, Science 306 (2004)
666–669.
increasing temperature can be as follow: decrease in viscosity, increase [4] K. Bilisik, M. Akter, Graphene nanoplatelets/epoxy nanocomposites: a review on
in kinetic energy, and rise in the distance between epoxy resin molecules functionalization, characterization techniques, properties, and applications,
and graphene oxide molecules. Escalating the temperature also reduces J. Reinforc. Plast. Compos. 41 (2022) 99–129.
[5] M. Koziol, M. Jesionek, P. Szperlich, Addition of a small amount of multiwalled
the van der Waals forces and the interaction between the epoxy novolac carbon nanotubes and flaked graphene to epoxy resin, J. Reinforc. Plast. Compos.
matrix with graphene oxide. It is important to study the influence of 36 (2017) 640–654.
temperature on the mechanical properties of epoxy/graphene oxide [6] S. Stankovich, D.A. Dikin, G.H. Dommett, K.M. Kohlhaas, E.J. Zimney, E.A. Stach,
et al., Graphene-based composite materials, Nature 442 (2006) 282–286.
nanocomposites, since the tensile strength against different tempera­ [7] R.L. Whitby, Chemical control of graphene architecture: tailoring shape and
tures is required depending on the use of these nanocomposites in properties, ACS Nano 8 (2014) 9733–9754.
different industries and environmental conditions. [8] T. Li, Z. Meng, S. Keten, Interfacial mechanics and viscoelastic properties of patchy
graphene oxide reinforced nanocomposites, Carbon 158 (2020) 303–313.
[9] T. Hammer, W. Netkueakul, P. Zolliker, C. Schreiner, R. Figi, O. Braun, et al.,
4. Conclusion Composites of epoxy and graphene-related materials: nanostructure
characterization and release quantification, NanoImpact 20 (2020), 100266.
[10] M. Bragaglia, L. Paleari, F.R. Lamastra, D. Puglia, F. Fabbrocino, F. Nanni,
In this study, using molecular dynamics simulations, the mechanical
Graphene nanoplatelet, multiwall carbon nanotube, and hybrid multiwall carbon
properties of various epoxy/graphene oxide nanocomposites were nanotube–graphene nanoplatelet epoxy nanocomposites as strain sensing coatings,
studied under different graphene oxide weight percentages, different J. Reinforc. Plast. Compos. 40 (2021) 632–643.
temperatures, and also different hardener weight percentages relative to [11] N. Saba, M. Jawaid, O.Y. Alothman, M. Paridah, A. Hassan, Recent advances in
epoxy resin, natural fiber-reinforced epoxy composites and their applications,
the epoxy matrix. The simulation results showed that the Young J. Reinforc. Plast. Compos. 35 (2016) 447–470.
modulus of nanocomposites increases in all epoxy/graphene oxide [12] Y. Zhang, S.-J. Park, Imidazolium-optimized conductive interfaces in multilayer
nanocomposites as a result of increasing the amount of adsorption and graphene nanoplatelet/epoxy composites for thermal management applications
and electroactive devices, Polymer 168 (2019) 53–60.
interaction energy between epoxy matrix and graphene oxide by adding [13] H. Al Mahmud, M.S. Radue, S. Chinkanjanarot, G.M. Odegard, Multiscale modeling
graphene oxide weight percentages (0, 1, 3, and 5 wt %). The highest of epoxy-based nanocomposites reinforced with functionalized and non-
amounts of Young modulus between nanocomposites in 5 Wt.% Gra­ functionalized graphene nanoplatelets, Polymers 13 (2021) 1958.
[14] Y. Oya, K. Inuyama, T. Okabe, J. Koyanagi, R. Matsuzaki, Analysis of structure
phene oxide belong to Novolac/GO and EPON862/GO, which were 4.27 characteristics in laminated graphene oxide nanocomposites using molecular
and 4.24 Gpa, respectively. On the other hand, the lowest Young dynamics simulation, Adv. Compos. Mater. 27 (2018) 427–438.
modulus at 5 Wt. % belong to EPON828/GO (3.95 Gpa) and CE/GO [15] C. Zhao, Y. Zhao, X. Pei, S. Liu, Y. Xia, C. Min, et al., Probing mechanical properties
of graphene oxide/epoxy composites based on nano indentation technique,
(3.75 Gpa). In addition, with increasing the temperature of the nano­ J. Reinforc. Plast. Compos. (2022), 07316844221075706.
composites was 300–700 K, the Young modulus decreases from 4.180 to [16] D. Galpaya, M. Wang, G. George, N. Motta, E. Waclawik, C. Yan, Preparation of
4.158 Gpa due to the reduction of the interactions between the epoxy graphene oxide/epoxy nanocomposites with significantly improved mechanical
properties, J. Appl. Phys. 116 (2014), 053518.
matrix and graphene oxide. Moreover, by raising the ratio weight per­
[17] Y.-J. Wan, L.-C. Tang, L.-X. Gong, D. Yan, Y.-B. Li, L.-B. Wu, et al., Grafting of
centages of epoxy to hardener 90:10 to 68:32 wt %, the number of epoxy chains onto graphene oxide for epoxy composites with improved mechanical
crosslinking increases, and Young’s modulus of nanocomposites im­ and thermal properties, Carbon 69 (2014) 467–480.
proves from 3.53 to 4.01 Gpa. Besides, the weight percentage of epoxy to [18] M. Huskić, S. Bolka, A. Vesel, M. Mozetič, A. Anžlovar, A. Vizintin, et al., One-step
surface modification of graphene oxide and influence of its particle size on the
hardener in the amount of 63:37 wt % reaches the highest value of properties of graphene oxide/epoxy resin nanocomposites, Eur. Polym. J. 101
Young modulus nanocomposite was 4.26 Gpa. After 63:37 wt%, with (2018) 211–217.
increasing the weight percentage of epoxy matrix to hardener, the [19] H. Zhao, J. Ding, H. Yu, Variation of mechanical and thermal properties in
sustainable graphene oxide/epoxy composites, Sci. Rep. 8 (2018) 1–8.
amount of Young’s modulus decreases to 4.02 Gpa, and its value is fixed. [20] Z. Yu, Z. Wang, H. Li, J. Teng, L. Xu, Shape memory epoxy polymer (SMEP)
This decrease in Young’s modulus is due to the reduction in the capacity composite mechanical properties enhanced by introducing graphene oxide (GO)
of cross-link between epoxy resin and hardener atoms, since each into the matrix, Mater. Today Proc. 12 (2019) 1107.
[21] A. Shokuhfar, B. Arab, The effect of cross linking density on the mechanical
hardener (DETDA) has the capacity to form cross-link with four epoxy properties and structure of the epoxy polymers: molecular dynamics simulation,
resins, which decreases with the rise in the ratio of epoxy resin to J. Mol. Model. 19 (2013) 3719–3731.
hardener. This reduction in the capacity of cross-link leads to a decrease [22] C. Acebo, X. Ramis, A. Serra, Improved epoxy thermosets by the use of poly
(ethyleneimine) derivatives, Phy. Sci. Rev. 2 (2017).

10
A. Salehi and S. Rash-Ahmadi Journal of Molecular Graphics and Modelling 117 (2022) 108311

[23] V. Varshney, S.S. Patnaik, A.K. Roy, B.L. Farmer, A molecular dynamics study of [29] Y. Wang, G. Yang, W. Wang, S. Zhu, L. Guo, Z. Zhang, et al., Effects of different
epoxy-based networks: cross-linking procedure and prediction of molecular and functional groups in graphene nanofiber on the mechanical property of polyvinyl
material properties, Macromolecules 41 (2008) 6837–6842. alcohol composites by the molecular dynamic simulations, J. Mol. Liq. 277 (2019)
[24] D.S.J.D.S. Biovia, San Diego, Materials Studio Modeling Environment, 2015. 261–268.
[25] S.V. Kallivokas, A.P. Sgouros, D.N.J.S.m. Theodorou, Molecular dynamics [30] Z. Li, R.J. Young, R. Wang, F. Yang, L. Hao, W. Jiao, et al., The role of functional
simulations of EPON-862/DETDA epoxy networks: structure, topology, elastic groups on graphene oxide in epoxy nanocomposites, Polymer 54 (2013)
constants, and local dynamics, Soft Matter 15 (2019) 721–733. 5821–5829.
[26] J. Tersoff, Empirical interatomic potential for silicon with improved elastic [31] M.K. Shukla, A. Kumar, A. Yadav, K. Sharma, Improved mechanical properties of
properties, Phys. Rev. B 38 (1988) 9902. graphene oxide reinforced cross-linked epoxy nanocomposites: a molecular
[27] W.L. Jorgensen, J. Chandrasekhar, J.D. Madura, R.W. Impey, M.L. Klein, dynamics approach, Mater. Today Proc. 11 (2019) 679–685.
Comparison of simple potential functions for simulating liquid water, J. Chem. [32] C.M. Hadden, D.R. Klimek-McDonald, E.J. Pineda, J.A. King, A.M. Reichanadter,
Phys. 79 (1983) 926–935. I. Miskioglu, et al., Mechanical properties of graphene nanoplatelet/carbon fiber/
[28] S. Plimpton, Fast parallel algorithms for short-range molecular dynamics, epoxy hybrid composites: multiscale modeling and experiments, Carbon 95 (2015)
J. Comput. Phys. 117 (1995) 1–19. 100–112.

11