Computational Materials Science 152 (2018) 381–392

Contents lists available at ScienceDirect

Computational Materials Science

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

MD-based characterization of plastic deformation in Cu/Ag nanocomposites T

via dislocation extraction analysis: Effects of nanosized surface porosities
and voids
⁎
A. Kardani, A. Montazeri
Faculty of Materials Science and Engineering, K.N. Toosi University of Technology, Tehran, Iran

A R T I C LE I N FO A B S T R A C T

Keywords: Recently, copper-silver nanocomposites (NCs) have been utilized in medical instruments owing to their ability in
Cu/Ag nanocomposite destroying the bacterial cell wall, which prevents inflammation of the body tissue. It has been revealed that
Porosity introducing nanosized porosities in their structure can lead to an increase in the interfacial surface area with the
MD simulation tissue promoting the quality of treatment. However, since yielding and occurrence of plastic deformation are not
Plastic deformation
acceptable in medicine, analyzing the mechanical behavior of these NCs having nanopores is an important
Dislocation
challenge. Therefore, the focus of this study is to assess the role of porosities on the deformation mechanism of
Cu/Ag NCs under uniaxial tensile loading conditions. Accordingly, several perfect and defected samples are
systematically studied through molecular dynamics simulation. It is observed that plastic deformation of perfect
sample occurs through twinning. For samples with surface voids, this is happened as a result of perfect dis-
locations gliding. Meanwhile, for volumetric porosities, the deteriorating effect is stopped passing the critical
void content. This is ascribed to the formation of many motionless dislocations such as stair-rod, Hirth and Frank
as confirmed via the dislocation extraction analysis. Consequently, it is demonstrated that presence of surface
voids can be more destructive.

1. Introduction their mechanical properties. Whereas, the presence of defects in sub-

In recent years, the use of copper and silver as two antibacterial
agents has been extensively interested in medical instruments.
According to studies conducted by researchers in the field of micro-
biology, copper and silver prevent inflammation of the body tissue by
destroying the bacterial cell wall [1,2]. On the other hand, metal matrix
nanocomposites (MMNCs) have been extensively implemented due to
their excellent mechanical properties and commercialization of their
manufacturing processes [3]. Accordingly, Cu-Ag nanocomposites
(NCs) can be widely utilized in medical implants leading to in-
tensification of the antimicrobial activity. Meanwhile, owing to the
increased interfacial area, the use of nanoporous implants can accel-
erate the process of ion exchange at the NC surface, which reduces time
of the treatment period [4]. Since release and absorption of drugs are
increased by reducing the dimensions of the voids to the nanoscale,
creation of nanopores can also be considered in terms of the drug de-
livery issue [5]. However, the presence of voids is not always useful.
Various defects can occur in the structure of a NC over its formation
process. Surface voids with uncontrollable dimensions or the presence
of them in the volume of materials can lead to substantial challenges in
stances with medical applications, causes a high sensitivity. Any yield
and plastic deformation in the implants are not acceptable [6]. There-
fore, characterization and analysis of the mechanical behavior of im-
plants are very important. Although failure of NCs is a macroscopic
phenomenon, this is resulted by atomic-scale structural variations of the
sample. Therefore, to understand the failure mechanism and to in-
vestigate the structural changes during deformation, it is necessary to
use a tool that can detect the displacement of atoms in the crystalline
structure of material.
A suitable option for simulating mechanical properties and in-
vestigating atomic structural changes of nanostructured materials is
utilizing atomic-scale modeling techniques such as molecular dynamics
(MD) simulation [7]. The plasticity of crystal materials is usually con-
trolled by dislocation slipping. According to the Hall-Petch relation,
grain boundary prevents the activity of dislocations, which in turn,
results in the different mechanisms affecting the plastic deformation of
nano-sized grains [8]. By reduction of the grain size to below 10 nm,
plastic deformation occurs through grain boundary sliding, while in-
creasing this parameter enhances nucleation of partial dislocations and,
as a result, the concentration of stacking faults and twins are increased

⁎
Corresponding author.
E-mail address: a_montazeri@kntu.ac.ir (A. Montazeri).

https://doi.org/10.1016/j.commatsci.2018.06.018
Received 11 April 2018; Received in revised form 6 June 2018; Accepted 8 June 2018
0927-0256/ © 2018 Elsevier B.V. All rights reserved.
A. Kardani, A. Montazeri Computational Materials Science 152 (2018) 381–392

[9,10]. According to the previous studies, temperature, strain rate and

sample geometry are the most important factors affecting the nuclea-
tion and slipping of dislocations [11–13]. It has been well established
that by increasing the temperature, decreasing the strain rate and re-
ducing the dimensions of the specimen under tension, the stress re-
quired for nucleation of dislocations is decreased. Subsequently, the
yield strength and strain of the samples are reduced. According to the
study of Zhang et al. [14] on the effect of sample geometry, nucleation
of dislocations in a cubic sample of single crystal copper occurred easier Fig. 1. Schematic illustration of the uniaxial tensile test in LAMMPS.
and faster than that of its cylindrical specimen. This was attributed to
the much greater stresses in the cube edges compared to the stress
open-source LAMMPS code developed by Sandia National Laboratory
distributions in the cylinder wall. Study of stacking faults is another
[33]. In order to determine the interatomic interactions between cop-
subject that has been considered by the researchers [15–20]. According
per–copper, silver-silver and copper-silver atoms in the force and en-
to these investigations, stacking faults can change the deformation
ergy fields, Embedded Atom Model (EAM) function was implemented.
mechanism by creating twins. Also, these areas can change the slip
This potential function has been successfully used in many previous
direction of the dislocations. Consequently, formation of stacking faults
studies to simulate copper-silver systems [34–37]. Simulations were
can prevent dislocation movements by changing the crystal lattice
carried out using the canonical ensemble (NVT) at a constant tem-
structure from FCC to HCP [21].
perature of 300 K. The Nose-Hoover thermostat was used to maintain
In addition to the external variables such as temperature, strain rate
the simulation temperature at 300 K. The use of this thermostat causes
and sample geometry, internal factors have also been studied by re-
fewer fluctuations than other methods [38]. The initial velocities were
searchers. They include presence of crystalline defects and voids within
determined based on the Maxwell-Boltzmann distribution at the desired
the structure, which can influence the deformation mechanism. For
temperature. In addition, velocity-Verlet integration algorithm [39]
example, Xu et al. [22] demonstrated that presence of voids within the
with a time-step of 2 fs was used for solving equations of motion
single crystal copper can accelerate the propagation of dislocations
through time. Since the initial configuration of the samples may not be
from their surroundings. In addition, Traiviratana et al. [23] showed
in the equilibrium state, the system was relaxed to 100,000 steps (i.e.,
that most of the dislocations nucleated from the surrounding of a void
200 ps) before loading. After relaxation, according to the quasi-statics
are the partially dislocation type. Based on the previous studies,
method introduced by authors [40], the left side of the representative
stacking faults can also be extended from the voids into the material
volume element (RVE) was kept fixed during imposing the uniaxial
[24–26]. Asari et al. [27] showed that by increasing the dimensions of
tensile loading. This was followed by applying the incremental dis-
the void, internal dislocations need more stress to pass through it.
placement on the other side to obtain the desired strain. Fig. 1 presents
Therefore, these imperfections can also appear as an obstacle for slip-
schematic illustration of a typical RVE under uniaxial tension. At each
ping of dislocations. In general, position and amount of defects can be
loading step, a displacement of 0.1 Å was applied along the sample
considered as the most important factors affecting the mechanical
length. After each displacement, the system was relaxed for 30,000
properties of the samples. So far, significant parts of studies have been
steps. Free boundary conditions (BC’s) were considered in direction of
focused on the developments of single phase nanostructures. However,
the axial stretch. Also, periodic BC’s were imposed in the lateral di-
the NCs deformation has some remarkable differences compared to
rections.
other materials. Most of the researches in this field focus on the effect of
As shown in Fig. 2a, rectangular RVE with the dimensions of
the second phase on the matrix deformation. For example, in the re-
10 × 4 × 4 nm in the x, y, and z directions, respectively, was con-
search of Sun et al. [28], influence of the second phase geometry and its
structed to resemble the copper-reinforced with Ag and the porous
distribution was investigated. It was found that the presence of particles
samples. The basic MD cells were created in two steps. First, the NC
with spherical shape and random distribution can lead to formation and
matrix consists of Cu atoms were created using the built-in tools in
expansion of the stacking faults. According to the conducted studies,
LAMMPS guided by the specific metal lattice parameters. At a later
the interface of nanoparticles (NPs) with matrix has a significant con-
stage, a central hole with the diameter of 2 nm was included to ac-
tribution in determining the mechanical properties of NCs. In addition,
commodate the Ag atoms (Fig. 2b and c). In all of the simulations, there
difference between the atomic radius of the second phase and matrix, as
was no overlapping between silver inclusion and copper atoms as de-
well as the increase of temperature, can weaken the strength of the
picted in Fig. 3. It is noted that all graphical illustrations have been
interface [23,29,30]. Mathiazhagan et al. [31] showed that dislocations
made using OVITO software [41]. The geometrical characteristics of
can be emitted from the interface. The amount of reinforcement phase
silver nanoparticle and the RVE were chosen so that a sample con-
is also very effective in the nucleation and movement of dislocations
taining 2.6% mass fraction of Ag was achieved in each case.
[32]. In this study, to investigate the deformation mechanism of Cu-Ag
To model porous samples, some voids with the diameter of 1 nm
NCs, first, the perfect nanocomposite sample is examined through
were distributed at the surface or in the volume of the copper matrix
uniaxial tensile test using MD simulations. Then, the tensile procedure
(See Fig. 4). The volume fraction (VF) for a void with the mentioned
is repeated under the same conditions in the presence of surface and
diameter was about 0.32%. Therefore, the VF for a surface void was half
volume voids at different amounts. This is followed by comparing the
of this amount for the volumetric pore. Accordingly, to produce dif-
results with those of the perfect sample. Accordingly in the second part,
ferent cases having various VF of these porosities, the corresponding
details of MD simulation and sampling techniques are introduced. Then,
amount of them could be introduced in the RVE as shown in Table 1.
in the third section, the results of the tensile test are presented for all
The centro-symmetry (CS) parameter [42] was used to identify the
samples. Also, the governing deformation mechanisms are system-
partial dislocations, twinning and stacking faults. For the specified atom
atically described in this part. Finally, results of the research are sum-
i , this parameter is defined as follows:
marized and the conclusions are made.
N /2
⎯→
⎯ ⎯→
⎯
2. Details of MD simulation CS = ∑ | Ri + Ri + N /2 |2
i=1 (1)
In this research, MD simulation method was used for analyzing the
effect of surface and volumetric porosities on the deformation me- where N is the number of nearest neighbors of atom i equal to 12 for the
⎯→
⎯ ⎯→
⎯
chanism of Cu-Ag NCs. All simulations were conducted utilizing the FCC metals. Also, Ri and Ri + N /2 are vectors from the central atom to a

382
A. Kardani, A. Montazeri Computational Materials Science 152 (2018) 381–392

Fig. 2. Construction steps of the perfect NC sample: (a) Building the rectangular copper matrix, (b) Creation of the central void, and (c) Accommodation of the Ag
nanoparticle.

Fig. 3. Perfect Cu/Ag nanocomposite sample: (a) initial configuration, (b) after relaxation.

Fig. 4. Implantation of voids into samples with: (a) surface and (b) volumetric porosities.

383
A. Kardani, A. Montazeri Computational Materials Science 152 (2018) 381–392

Table 1 particular pair of nearest neighbors in the crystal lattice. The CS value is
Characteristics of nanoporous samples having different volume and surface 0 for a system without any defect. Meanwhile, positive values demon-
voids content. strate presence of defects and free surfaces within the crystalline
Voids content (%) Number of voids in the Number of voids at structure. Dislocations were identified implementing a dislocation ex-
volume the surface traction analysis (DXA) developed by Stukowski and Albe [43]. Ad-
ditionally, the color classification based on a pattern matching algo-
1.3 4 8
rithm was employed to distinguish between defected and perfect
2.5 8 16
5.0 16 32 crystalline lattice. This algorithm acts on the basis of common neighbor
analysis (CNA) approach [44].

3. Results and discussion

3.1. Perfect NC under tension: Mechanical behavior and deformation

mechanism

In order to examine the mechanical properties and the underlying

Fig. 5. Various stages of the uniaxial tensile loading: (a) ε = 0, (b) ε = 0.2, (c)
ε = 0.8, and (d) ε = 1.1.
Fig. 6. Stress–strain curve of the perfect NC sample showing some stress fluc-
tuations after the yield point.
deformation mechanism, the perfect sample was first put under the
uniaxial tension as illustrated in Fig. 5. Fig. 6 depicts the corresponding
stress–strain curve of this sample from which, different mechanical
properties including could be achieved. As seen, the stress–strain re-
lationship in the elastic region is almost linear. This is followed by the
saw-tooth fluctuations in the plastic deformation zone. For the sake of
simplicity, the reader may refer to [45] that documents in detail the
physical mechanism governing this stress fluctuation. Elastic modulus
of this sample at the atomistic scale was obtained from slope of the
stress–strain curve at low strains (i.e., < 3%) as 64.8 GPa. Additionally,
the yield strength defined as the stress at which the first plastic de-
formations occurred was found to be 11.1 GPa. These values were in
accordance with the results of similar studies presented in Table 2 de-
monstrating the validation of the tensile testing procedure implemented
in the present study.
To capture the mechanism affecting deformation of the NC sample
after the yield point, we further proceeded to extract all types of dis-
locations within the crystalline structure of the model. This was
achieved through the DXA algorithm provided by the OVITO software
[43] through the atomistic snapshots. Based on the DXA results, when
the material is in the elastic region, there are no dislocations observed
in the structure. Meanwhile, passing the yield point, the first partial
dislocations are nucleated as illustrated in Fig. 7. Deng et al. [47]
showed that the first dislocations of a single crystal material are nu-
cleated from the specimen surface. This was not the case happened in
the present sample in which, dislocations were nucleated from the NP/
matrix interface (See Fig. 7). It arises because Ag and Cu atoms have
12% difference in terms of the lattice misfit. As this parameter is larger
than 5%, interface of the Ag/Cu sample seems to be semi-coherent
rather than the coherent one [48]. Accordingly, passing the yield point,
the interfacial energy of the semi-coherent interface would be in-
creased. Therefore, this mismatch in the interface region resulted in the
local stress concentration in this area, which in turn, reduced the stress
required for dislocation nucleation. As revealed in Fig. 8, placement of
Ag NP in the Cu matrix increased the local stress value up to 13 GPa.
Similarly, in the research conducted by Amigo et al. [35], it was shown
that presence of four Ag atoms in the Cu crystalline lattice enhances the
local stress up to 7 GPa in the interfacial zone.
According to Fig. 7, a significant proportion of dislocations are
Shockley partial ones. These dislocations alter the atomic order in the

Table 2
Yield strength and Young's modulus of the perfect sample obtained in the present work in comparison with the ones given in the literature.
Study Case Yield strength (GPa) Young's modulus (GPa)

Amigo [35] Cu single crystal and Ag impurities 10.7 61

Xu [22] Cu single crystal with volumetric voids 6.5 86.6
Xie [12] Cu single crystal 11 64.5
Zhan [46] Cu single crystal 8 63
Present work Cu single crystal and Ag NP 11.1 64.8

384
A. Kardani, A. Montazeri Computational Materials Science 152 (2018) 381–392

Fig. 7. Dislocations emission in the perfect NC sample: (a) Nucleation of partial dislocations surrounding the Ag nanoparticle at ε = 0.15 (corresponds to the yield
point in Fig. 6), (b) Spreading of dislocations into the neighbor regions at ε = 0.2, and (c) Propagation of the plastic deformation at ε = 0.3 (configurations have been
characterized by DXA).

Fig. 8. von Mises stress distribution in the perfect nanocomposite sample.

FCC lattice creating stacking faults having the HCP structure [49,50].
To further investigate the issue, CNA results of the perfect NC sample at
the various stages of the deformation have been presented in Fig. 9.
They include the yield point, the deformations occurred at the selective
strains of 0.2 and 0.4, and also at the start of the necking stage. In the
yield strain, stacking faults were formed around the silver nanoparticle.
Then, their percent in the sample were increased up to the strain of 0.4.
After the necking stage, the size of these areas decreased significantly.
This was ascribed to the formation of twins from the stacking faults as
demonstrated in Fig. 10. As seen, stacking faults observed at 4058 ps
corresponding to the strain of 0.51 are ready to creation of twins. The
corresponding snapshot at 4123 ps (in accordance with ε = 0.53) dis-
tinguishes the twins atoms from the other crystalline regions as pre-
sented in Fig. 10c. This phenomenon was also studied by Jiang et al.
[51] in the case of homogeneous single crystal copper. It should be
emphasized that there is a fundamental difference between formation
of twins in the single crystal metal nanowires compared to that for the
case of NC samples. In the former, the deformation occurs through the
surface-nucleated plasticity in which, stacking faults needed to form
twins are initially created from the free surface of the samples.
Meanwhile, in the case of NCs as previously discussed, due to the large
lattice misfit in the interfacial region, this area is prone to initiate the
first stacking faults in the sample. In other words, in the competition
between the free surfaces and interface area, stacking faults are initially
nucleated at the interface. Accordingly, the twinning formation occurs
in this case via the interface-dependent plasticity. This issue has been
illustrated in Fig. 11. It should be pointed out that the formation of
twins from the stacking faults can be a better way to reduce the energy
imposed on the system (Tadmor et al. [52]). This resulted in the re-
duction of stacking faults as a consequence of converting them to twins.
Formation of twins would also manifest itself in the zigzag behavior
observed in the plastic deformation region of the sample as presented in
Fig. 6. This special behavior was also reported by Zhan et al. [46] for
tension of the copper single crystal nanowire.
3.2. Porous nanocomposite under tensile: Role of volumetric porosities on
the results
The stress–strain curve of the volumetric porous samples has been
illustrated in Fig. 12. Results show that, as expected, the yield strength

385
A. Kardani, A. Montazeri Computational Materials Science 152 (2018) 381–392

Fig. 9. Evolution of partial dislocations and stacking fault areas (HCP atoms) in the perfect sample: (a) at the yield point, (b) ε = 0.2, (c) ε = 0.4, and (d) at the initial
of necking stage.

Fig. 10. Steps of twin formation from stacking faults: (a) HCP atoms in the stacking fault regions, (b) Creation of a twin nucleus by overlapping of stacking faults, and
(c) Twined vs. untwined regions.

Fig. 11. Formation of a twinning fault in the nanocomposite sample through the interface-dependent plasticity mechanism (Stacking faults, as the main sources of
twinning formation, are initially nucleated in the interface region).

386
A. Kardani, A. Montazeri Computational Materials Science 152 (2018) 381–392

containing 2.5% volumetric porosities. To explore the mechanism

governing this behavior, we further studied the dislocation length for
the samples introduced in Table 2 during their total deformation (See
Fig. 13). This analysis included various types of dislocations, namely of
Shockley, stair-rod, Hirth, and Frank. As seen, the Shockley dislocation
of these two samples is close to each other. Meanwhile, the other dis-
locations show a significant enhancement for the sample having 5%
voids. According to a research conducted by Martinez et al. [53], pas-
sing the yield point and before significant deformation of the sample,
two dislocations combine to each other creating a stair-rod dislocation
as stated in Eq. (2):
1 1 1
〈2 1 1〉 + 〈1 2 1 〉 → 〈1 1 0〉
6 6 6 (2)

This type of dislocations is usually seen as tetrahedral in the crys-

Fig. 12. Stress–strain curve of NC samples having different VFs of volumetric
porosities.
and Young's modulus were decreased with increasing the voids in the
bulk material. This was attributed to the local stress concentrations in
the vicinity of porosities which in turn, led to easier nucleation of
dislocations. It is noted that the reduction was much more noticeable
for samples having more than 2.5% VF of volumetric porosities. To
facilitate better comparisons, Table 3 lists mechanical properties of Cu/
Ag nanocomposite samples in the presence of different amounts of these
voids along with the results of the perfect specimen.
Surprisingly, it was found that the amount of mechanical properties
reduction in the case with 5% VF of voids is very similar to the sample
talline structure. It is noted that slipping of the stair-rod dislocations
1
with the Burgers vector of 6 〈1 1 0〉 is not possible in the tetrahedron
corners. Accordingly, as shown in Fig. 14, they must decompose into
two Shockley dislocations during the imposed deformation requiring
the energy [54]. Therefore, it was concluded that high amount of stair-
rod dislocations is one of the most important reasons for preventing
mechanical properties reduction in the sample with 5% volumetric
voids compared to the less defected case.
To further proceed with the underlying mechanism, the formation
of other dislocations such as Hirth and Frank were also investigated
employing the DXA analysis. The former takes place when two perfect
dislocations with Burgers vectors perpendicular to each other slide
across the slip planes (See Fig. 15). This collision resembles the Hirth
lock as stated in the following equation [25]:
1 1 1 1 1
〈1 1 0〉 + 〈1 1 0〉 → 〈2 1 1〉 + 〈2 1 1〉 + 〈1 0 0〉
2 2 6 6 3 (3)
Since the Hirth sliding direction does not correspond to the close-

Table 3
Effects of the amount of volumetric porosities on the mechanical properties of the NC sample along with the results of the perfect case.
Case Yield strain Yield strength (GPa) Young's modulus (GPa)

Perfect nanocomposite sample 0.15 11.1 64.8

Sample with1.3%VF of volumetric porosities 0.13 9.2 61
Sample with 2.5%VF of volumetric porosities 0.12 8.8 59.6
Sample with 5%VF of volumetric porosities 0.12 8.4 52.1

Fig. 13. DXA results representing the influence of void density on the dislocation length for various types of dislocations.

387
A. Kardani, A. Montazeri Computational Materials Science 152 (2018) 381–392

Fig. 14. Formation of the stair-rod dislocation at 4141 ps and its decomposition into two Shockley partial dislocations at the snapshot of 4299 ps (Configurations
were characterized by DXA).

Fig. 15. Hirth lock and two Shockley partial dislocations obtained from the interaction of two perfect dislocations in the sample with 5%VF of volumetric porosities
(The snapshot was taken at the time-step of 3246 ps corresponds to the strain of 0.4).

packed slip direction, it is unable to slide and so, is called as the Hirth
lock [55]. Occurring this phenomenon in the crystalline structure can
also limit the movement of other dislocations. Accordingly, increasing
the number of Hirth dislocations could be one of the inhibitor factors of
further reduction in the mechanical properties of the sample containing
5% volumetric voids. In addition to the discussed factors, Frank dis-
locations were also studied for the aforementioned sample. As de-
monstrated in Fig. 16, this type of dislocations was created in the form
of a loop, which could be moved only through climbing based on the
atomic diffusion [56,57]. Since diffusion is a thermally activated pro-
cess happened at the elevated temperatures, increasing the number of
Frank dislocations at the ambient temperature could prevent further
reduction in the mechanical characteristics of the sample having 5%
volumetric voids.

3.3. Surface porosities: An in-depth study towards the underlying

mechanism
After examining the effect of volumetric porosities, the next step
involved analysis of mechanical behavior of the nanocomposite sample
in the presence of surface porosities. We focused on answering the
important and vital question - which type of porosities are more de-
structive in terms of mechanical characteristics of Cu/Ag NCs? To an-
swer this, the tensile procedure was repeated under the same conditions
for samples having different amounts of surface voids. The
Fig. 16. Frank dislocation loop observed at the time-step of 4532 ps corre-
sponds to ε = 0.57 in the sample with 5%VF of volumetric porosities.
corresponding stress–strain curves have been shown in Fig. 17. This was
followed by comparing the results with those of the perfect sample as
listed in Table 4. Our results revealed that in comparison with volu-
metric porosities (See Table 3), surface defects play a critical role in
weakening the mechanical properties of NC samples. To further

388
A. Kardani, A. Montazeri
Fig. 17. Stress–strain curve of NC samples having different VF of surface por-
osities.
examine the issue, the approach proposed by Begau et al. [58] was
implemented to quantify the local dislocation density for the introduced
samples. For a visual representation of the results, the reader may refer
to Fig. 18 that shows the variations of this parameter at the yield stage.
As revealed, dislocation density is higher for the cases with surface
porosities. This would manifest itself in the lower values reported for
the yield strain and stress of these samples compared to the corre-
sponding data for the samples having volumetric pores.
Regarding the effects of nanovoids within the metallic samples on
their yield response, there are several publications in the literature. For
example, it has been demonstrated that dislocations begin to nucleate
from the void surface prior to the maximum stress (i.e., the yield point)
[22]. As previously discussed, presence of voids in the sample leads to
increase of the normal stresses around the porosities. This in turn,
promotes the dislocation nucleation from the voids surface and as a
Table 4
Effect of the amount of surface porosities on the mechanical properties of the NC sample along with the results of the perfect case.
Case Yield strain
Perfect nanocomposite sample 0.15
Sample with 1.3% VF of surface porosities 0.13
Sample with 2.5% VF of surface porosities 0.12
Sample with 5% VF of surface porosities 0.11
Fig. 18. Comparison of the dislocation density at the yield point for perfect and defected samples.
389
Computational Materials Science 152 (2018) 381–392
result, advances the yield point. Furthermore, Cui and Chen explored
the yield response of the metallic films containing nanovoids via MD
simulations [59]. It was shown that yield stress of the samples decreases
to much lower values ascribed to the initiation of dislocations nuclea-
tion shortly prior to the yield point. These results are in line with the
data reported in Fig. 18. As seen, increasing the void content of the
samples causes a significant increase in their dislocation density. Con-
sequently, their yield response deteriorates as reported in Tables 3 and
4.
We would also like to draw the reader's attention to the significant
difference between the results reported for these two cases. According
to Table 4, in the surface defected samples, yield strength and Young's
modulus decreased monotonically with increasing the percentage of
surface porosity. Contrary to the case containing volumetric voids, here,
creating 5% VF of surface voids significantly reduced the mechanical
properties of the specimen compared to those of the case having 2.5%
surface porosities. As seen in Fig. 17, for the latter case, passing the
yield point A, there was observed a drop up to point B. After B, the
curve demonstrated an increase towards point C in which, the tensile
stress reached the yield point value. The same phenomenon was also
observed more intensely for the sample with 5% surface voids.
To explore the underlying mechanism affecting this special beha-
vior, let’s take a closer look at the dislocations creation and emission in
the case with 5% VF of surface voids (See Fig. 19). Based on Fig. 19a,
when the sample reached the yield point at the strain of 0.12 (i.e., point
D), in addition to the NP/matrix interface, some types of dislocations
were also nucleated from the surface voids leading to the declined trend
in the curve. This weakening effect was terminated at point E (corre-
sponds to the strain of 0.13) in which, these new dislocations were
stopped. This was also ascribed to the formation of perfect dislocations
as a result of partial dislocations combination happened around the
voids as seen in Fig. 19b. Finally, further loadings led to the formation
of new dislocations. During this process, there was a rise in the curve up
to point F.
To further analysis of the issue, the CNA results have been presented
Yield strength (GPa) Young's modulus (GPa)
11.1 64.8
9.3 61
8.6 56.5
6.1 44.2
A. Kardani, A. Montazeri
Fig. 19. Snapshots of the sample with 5% VF of surface voids: (a) creation of new dislocations around the surface voids, (b) stopping of partial dislocations with the
formation of perfect ones, and (c) nucleation of new dislocations due to further imposed loadings (Configurations were characterized by DXA).
Fig. 20. Variations of the fraction of HCP atoms (ƞ) during deformation for the
perfect sample.
Fig. 21. Variations of the fraction of HCP atoms (ƞ) during deformation for the
sample having 5% VF of surface porosities.
for the perfect NC along with the data corresponds to the sample with
5% VF of surface voids (See Figs. 20–22). In these curves, ƞ denotes the
ratio of atoms in the HCP structure to the total atoms of the sample.
Additionally, figures inset shows the parameter ξ expressing contribu-
tion of perfect dislocations out of all ones in each loading step. As seen
in Fig. 20, in the case of perfect sample, there were not observed
stacking faults in the elastic region due to lack of partial dislocations.
Passing the yield point (i.e. ε = 0.15 in accordance with Fig. 6), the first
stacking faults were formed through nucleation of partial dislocations.
Consequently, increasing the external tensile loading would lead to
enhancement of the dislocations density, which in turn, promoted the
beginning of necking in the sample. These findings were in a good ac-
cordance with the results of Sun et al. [28]. Additionally, due to the
negligible portion of parameter ξ in this case, it was concluded that
partial dislocation was the main factor dominating plastic deformation
of the perfect sample. Meanwhile, for the surface defected sample as
seen in Fig. 21, the ratio of atoms in the stacking faults was significantly
decreased compared to that of the perfect sample. This was accom-
panied by the incremental trend found for ξ illustrating a rise in the
390
Computational Materials Science 152 (2018) 381–392
Fig. 22. Variations of the fraction of HCP atoms (ƞ) during deformation for the
sample having 5% VF of volumetric porosities.
contribution of perfect dislocations. In other words, for this sample,
presence of more sites for nucleation of partial dislocations increased
the density of these dislocations in the materials structure. Conse-
quently, a significant part of these dislocations were combined together
resulting in the formation of perfect dislocations. Glide of these dis-
locations accelerated the failure of the specimen, which in turn, caused
a significant reduction in the mechanical properties of the sample
containing surface voids as discussed before. To facilitate a better
comparison, Fig. 22 depicts the corresponding curves for the case of NC
sample having 5% VF of volumetric pores. As seen, regions with partial
dislocations are more pronounced in this case compared to the sample
with surface porosities. Additionally, the reverse trend is observed for
perfect dislocations as thoroughly inspected in 3.2.
4. Conclusion
In summary, due to the potential usage of Cu/Ag nanocomposites in
medical implants, we documented in details the deformation me-
chanism of these NCs under tensile loading conditions via numerical-
based MD simulations. Moreover, it has been well-established that ex-
istence of nanosized voids within their structure promotes the quality of
treatment through enhancing the interfacial surface area. Accordingly,
the main goal of the current study was to follow variations in the me-
chanical properties of these NCs in the presence of different amounts of
surface and volumetric porosities. The underlying mechanism gov-
erning this issue was also deeply addressed in this paper through mi-
crostructural characterization provided by means of the dislocation
extraction analysis (DXA). The results showed that in the case of perfect
nanocomposite sample, due to the presence of weak adhesion in the
copper/silver interfacial area, the local stresses in this region were in-
creased compared to other areas. Consequently, the interface converted
to a place for nucleation and release of partial dislocations. In the fol-
lowing, stacking faults resulted from the partial dislocations were
A. Kardani, A. Montazeri
converted to twinning as demonstrated by the zigzag behavior observed
in the plastic deformation region. In the next stage, by investigating
stress–strain curve of the samples having different amounts and types of
porosities, the pore content and distribution type showed its effect in-
tensely. Our simulations revealed that presence of volumetric voids
causes a decrease in the mechanical characteristics of the perfect NC
sample. This was ascribed to the local stress concentrations in the vi-
cinity of porosities which in turn, led to easier nucleation of disloca-
tions. Meanwhile, the weakening effect was stopped after the critical
void content. This was attributed to the formation of stair-rod, Hirth
and Frank dislocations as confirmed by DXA results. It was concluded
that these dislocations were locked during plastic deformation of the
samples with volumetric porosities which in turn, led to prevent the
movement of other dislocations. Additionally, our findings revealed
that in samples containing surface voids, the reduction of mechanical
properties was more significant than their counterparts having volu-
metric voids. This was attributed to the activation of new sites for nu-
cleation of dislocations at the sample surface in addition to the matrix/
nanoparticle interface. It was demonstrated that deformation in these
samples was occurred because of the glide of perfect dislocations. It
should be pointed out that despite these deteriorating effects, the pre-
sence of surface voids is beneficial in terms of biological science.
Acknowledgements
The authors would like to acknowledge high-performance com-
puting supports provided by Shahid Chamran University of Ahvaz.
References
[1] V.M. Villapun, L.G. Dover, A. Cross, Antibacterial metallic touch surfaces, Materials
9 (2016) 1–24.
[2] S. Rtimi, R. Sanjines, C. Pulgarin, J. Kiwi, Quasi-instaneous bacterial inactivation on
Cu-Ag nanoparticulate 3D catheters in the dark and under light: mechanism and
dynamics, Appl. Mater. Interfaces 8 (2016) 47–55.
[3] G.G. Sozhamannan, S. Balasivanandha, R. Paskaramoorthy, Failures analysis of
particle reinforced, Mater. Des. 31 (2010) 3785–3790.
[4] E.M. Carl, Dental implant prostthetics, second ed., Mosby, 2015.
[5] G. Ahuja, K. Pathak, Porous carriers for controlled/modulated drug delivery, Ind. J.
Pharma. Sci. 6 (2009) 599–607.
[6] R.L. Sakaguchi, Craig's restorative dental materials, 13th ed., Mosby, 2012.
[7] Ch.R. Weinberger, G.J. Tucker, Multiscale Materials Modeling for Nanomechanics,
First ed., Springer Series in Materials Science, 2016.
[8] A. Lasalmonie, J. Strudel, J. Mater, Influence of grain size on the mechanical be-
haviour of some high strength materials, J. Mater. Sci. 21 (1985) 1837–1852.
[9] R. Christopher, W. Cai, Plasticity of metal nanowires, J. Mater. Chem. 22 (2012)
3277–3292.
[10] F. Yuan, X. Wu, Hydrostatic pressure effects on deformation mechanisms of nano-
crystalline FCC metals, Comput. Mater. Sci. 85 (2014) 8–15.
[11] B. Ma, Q. Rao, Y. He, Molecular dynamics simulation of temperature effect on
tensile mechanical properties of single crystal tungsten nanowire, Comput. Mater.
Sci. 117 (2016) 40–44.
[12] H. Xie, F. Yin, T. Yu, G. Lu, Y. Zhang, A new strain-rate-induced deformation me-
chanism of cu nanowire: Transition from dislocation nucleation to phase transfor-
mation, Acta Mater. 85 (2015) 191–198.
[13] J. Lao, M. Naghdi Tam, D. Pinisetty, N. Gupta, Molecular dynamics simulation of
FCC metallic nanowires: a review, J. Miner. Metal. Mater. Soc. (TMS) 65 (2013)
175–184.
[14] Y. Zhang, H. Huang, S.N. Atluri, Strength asymmetry of twinned copper nanowires
under tension and compression, Comput. Model. Eng. Sci. 35 (2008) 215–225.
[15] A. Ahadi, P. Hansson, S. Melin, Tensile behavior of single-crystal nano-sized Cu
beams – geometric scaling effects, Comput. Mater. Sci. 135 (2017) 127–133.
[16] J.W. Wang, S. Narayanan, J. Huang, Ze. Zhang, T. Zhu, S. Mao, Atomic-scale dy-
namic process of deformation-induced stacking fault tetrahedra in gold nanocrys-
tals, Nature Commun. 4 (2013) 2340.
[17] W. Pang, P. Zhang, G. Zhang, Ai. Xu, X. Zhao, Dislocation creation and void nu-
cleation in FCC ductile metals under tensile loading: a general microscopic picture,
Sci. Rep. 4 (2014) 6981.
[18] A. Hunter, I.J. Beyerlein, Relationship between monolayer stacking faults and twins
in nanocrystals, Acta Mater. 88 (2015) 207–217.
[19] M. An, Q. Deng, Y. Li, H. Song, M. Su, J. Cai, Molecular dynamics study of tension-
compression asymmetry of nanocrystal a-Ti with stacking fault, Mater. Des. 127
391
Computational Materials Science 152 (2018) 381–392
(2017) 204–214.
[20] Q. Feng, X. Song, H. Xie, H. Wang, X. Liu, F. Yin, Deformation and plastic co-
ordination in WC-Co composite-molecular dynamics simulation of nanoindentation,
Mater. Des. 120 (2017) 193–203.
[21] Y. Lin, D. Pen, Analogous mechanical behaviors in < 100 > and < 110 >
directions of Cu nanowires under tension and compression at a high strain rate,
Nanotechnology 18 (2007) 395705.
[22] Sh. Xu, Y. Su, D. Chen, L. Li, Plastic deformation of Cu single crystals containing an
elliptic cylindrical void, Mater. Lett. 193 (2017) 283–287.
[23] S. Traiviratana, E.M. Bringa, D.J. Benson, M.A. Meyers, Void growth in metals:
atomistic calculations, Acta Mater. 56 (2008) 3874–3886.
[24] K.J. Zhao, C.Q. Chen, Y.P. Shen, T.J. Lu, Molecular dunamics study on the nano-
void growth in face-centered cubic single crystal copper, Comput. Mater. Sci. 46
(2009) 749–754.
[25] Y. Zhang, Sh. Jiang, X. Zhu, Y. Zhao, Influence of void density on dislocation me-
chanisms of void shrinkage in nickel single crystal based on molecular dynamics
simulation, Phys. E 90 (2017) 90–97.
[26] D.S. Aidhy, Ch. Lu, K. Jin, H. Bei, Y. Zhang, L. Wang, W. Weber, Formation and
growth of stacking fault tetrahedra in Ni via vacancy aggregation mechanism, Scr.
Mater. 114 (2016) 137–141.
[27] K. Asari, O.S. Hetland, S. Fujita, M. Itakura, T. Okita, The effect of stacking fault
energy on interactions between an edge dislocation and a spherical void by mole-
cular dynamics simulation, J. Nucl. Mater. 442 (2013) 360–364.
[28] X. Sun, Q. Li, Y. Gu, Xi. Feng, Mechanical properties of bioinspired bicontinuous
nanocomposites, Comput. Mater. Sci. 80 (2013) 71–78.
[29] M.J. Demkowicz, L. Thilly, Structure, shear resistance and interaction with point
defects of interfaces in Cu-Nb nanocomposites synthesized by severe plastic de-
formation, Acta Mater. 59 (2011) 7744–7756.
[30] V.V. Pogorelko, A.E. Mayer, Influence of copper inclusions on the strength of alu-
minum matrix at high-rate tension, Mater. Sci. Eng., A 642 (2015) 351–359.
[31] S. Mathiazhagan, S. Anup, Investigation of deformation mechanisms of staggered
nanocomposites using molecular dynamics, Phys. Lett. A 380 (2016) 2849–2853.
[32] H. Ghaffarian, A. Karimi, S. Ryu, K. Kang, Nanoindentation study of cementite size
and temperature effects in nanocomposites pearlite: a molecular dynamics simu-
lation, Curr. Appl. Phys. 16 (2016) 1015–1025.
[33] S. Plimpton, Fast parallel algorithms for short-range molecular dynamics, Comput.
Phys. 117 (1995) 1–19.
[34] P.L. Williams, Y. Mishin, J.C. Hamilton, An embedded-atom potential for the Cu-Ag
system, Modell. Simul. Mater. Sci. Eng. 14 (2006) 817–833.
[35] N. Amigo, G. Gutierrez, M. Ignat, Atomistic simulation of single crystal copper
nanowires under tensile stress: Influence of silver impurities in the emission of
dislocations, Comput. Mater. Sci. 87 (2014) 76–82.
[36] L. Yuan, Zh. Xu, D. Shan, B. Guo, Molecular dynamics study on the equal biaxial
tension of Cu/Ag bilayer films, Appl. Surf. Sci. 282 (2013) 450–455.
[37] S. Li, W. Qi, H. Peng, J. Wu, A comparative study on melting core-shell and janus
Cu-Ag bimetallic nanoparticles, Comput. Mater. Sci. 99 (2015) 125–132.
[38] W.G. Hoover, Canonical dynamics: equilibrium phase-space distributions, Phys.
Rev. A 31 (1985) 1695.
[39] M.P. Allen, D.J. Tildesley, Computer Simulation of Liquids, First ED, Clarendon
Press, 1986.
[40] S. Bashirvand, A. Montazeri, New aspects on the metal reinforcement by carbon
nanofillers: a molecular dynamics study, Mater. Des. 91 (2016) 306–313.
[41] A. Stukowski, Visualization and analysis of atomistic simulation data with OVITO-
the open visualization tool, Modell. Simul. Mater. Sci. Eng. 18 (2009) 015012.
[42] C.L. Kelchner, S.J. Plimpton, J.C. Hamilton, Dislocation nucleation and defect
structure during surface indentation, Phys. Rev. B 58 (1998) 11085–11088.
[43] A. Stukowski, K. Albe, Extracting dislocation and non-dislocation crysta; defects
from atomistic simulation data, Modell. Simul. Mater. Sci. Eng. 18 (2010) 085001.
[44] H. Tsuzuki, P.S. Branicio, J.P. Rino, Structural characterization of deformed crystals
by analysis of common atomic neighborhood, Comput. Phys. Commun. 177 (2007)
518–523.
[45] A. Kardani, A. Montazeri, Temperature-based plastic deformation mechanism of
Cu/Ag nanocomposites: a molecular dynamics study, Comput. Mater. Sci. 144
(2018) 223–231.
[46] H.F. Zhan, Y.T. Gu, Numerical exploration of plastic deformation mechanisms of
copper nanowires with surface defects, Comput. Mater. Sci. 50 (2011) 3425–3430.
[47] C. Deng, F. Sansoz, Fundamental differences in the plasticity of periodically
twinned nanowires in Au, Ag, Al, Cu, Pb and Ni, Acta Mater. 57 (2009) 6090–6101.
[48] D.A. Porter, K.E. Easteling, Phase Transformations in Metals and Alloys, 2nd ed.,
Chapman Hall, London, 1981.
[49] V. Yamakov, D. Wolf, S. Phillpot, A.K. Mukherjee, H. Gleiter, Dislocation processes
in the deformation of nanocrystalline aluminium by molecular-dynamics simula-
tion, Nat. Mater. 1 (2002) 45–49.
[50] Sh. Takeuchi, E. Kuramoto, T. Yamamoto, T. Taoka, The nature of stacking faults
and partial dislocation in deformed Ni3Ga single crystal, Appl. Phys. 12 (1973)
1486–1492.
[51] J. Jiang, A. Leach, K. Gall, H. Park, T. Rabczuk, A surface stacking fault energy
approach to predicting defect nucleation in surface-dominated nanostructures,
Mechamics Phys. Solids 61 (2013) 1915–1934.
[52] E.B. Tadmor, N. Bernstein, A first-principles measure for the twinnability of FCC
metals, Mech. Phys. Solids 52 (2004) 2507–2519.
[53] E. Martinez, J. Marian, A. Arsenlis, M. Victoria, J.M. Perlado, Atomistically
A. Kardani, A. Montazeri
informed dislocation dynamics in FCC crystals, Acta Mater. 56 (2008) 869–895.
[54] L.A. Zepeda-Ruiz, E. Martinez, M. Caro, Deformation mechanisms of irradiated
metallic nanofoams, Appl. Phys. Lett. 103 (2013) 031909.
[55] S.J. Zhou, D.L. Preston, P.S. Lomdahl, D.M. Beazley, Large-scale molecular dy-
namics simulations of dislocation intersection in copper, Science 279 (1998)
1525–1527.
[56] V.P. Swart, S. Kritzinger, Prismatic dislocation loop rotation and self-climb phe-
nomena in Al-0.13 wt.% Mg, Phil. Mag. 27 (1972) 689–695.
[57] A.V. Bakaev, D.A. Terent, E.E. Zhurkin, P.Yu. Grigor, Molecular dynamics
392
Computational Materials Science 152 (2018) 381–392
simulation of the interaction of dislocation with radiation-induced defect in Fe-Ni-
Cr austenitic alloys, Surf. Invest. X-ray Synchrotron Neutron Techniques 7 (2013)
211–217.
[58] C. Begau, J. Hua, A. Hartmaier, A novel approach to study dislocation density
tensors nd lattice rotation patterns in atomistic simulations, J. Mech. Phys. Solids 60
(2012) 711–722.
[59] Y. Cui, Z. Chen, Molecular dynamics modeling on the role of initial void geometry
in a thin aluminum film under uniaxial tension, Modell. Simul. Mater. Sci. Eng. 23
(2015) 085011.