M. A. N.

Dewapriya
School of Engineering Science,
Simon Fraser University,
Burnaby, BC V5A 1S6, Canada
e-mail: mandewapriya@sfu.ca

R. K. N. D. Rajapakse1
Faculty of Applied Sciences,
Simon Fraser University,
Burnaby, BC V5A 1S6, Canada
e-mail: rajapakse@sfu.ca

Molecular Dynamics Simulations

and Continuum Modeling of
Temperature and Strain Rate
Dependent Fracture Strength of
Graphene With Vacancy Defects
We investigated the temperature and strain rate dependent fracture strength of defective
graphene using molecular dynamics and an atomistic model. This atomistic model was
developed by introducing the influence of strain rate and vacancy defects into the kinetics
of graphene. We also proposed a novel continuum based fracture mechanics framework
to characterize the temperature and strain rate dependent strength of defective sheets.
The strength of graphene highly depends on vacancy concentration, temperature, and
strain rate. Molecular dynamics simulations, which are generally performed under high
strain rates, exceedingly overpredict the strength of graphene at elevated temperatures.
Graphene sheets with random vacancies demonstrate a singular stress field as in continuum fracture mechanics. Molecular dynamics simulations on the crack propagation
reveal that the energy dissipation rate indicates proportionality with the strength. These
findings provide a remarkable insight into the fracture strength of defective graphene,
which is critical in designing experimental and instrumental applications.
[DOI: 10.1115/1.4027681]

Introduction

The extraordinary electromechanical [13] and magnetic [4]

properties of graphene have drawn remarkable attention from scientists and engineers. Graphene based nanoelectromechanical systems (NEMS), such as resonators [5], have demonstrated
intriguing applications in various engineering disciplines from telecommunication [6] to biomedicine [7]. However, as in many
crystalline materials, defects are unavoidable during synthesizing
and fabrication of graphene based NEMS [79]. Defects, such as
vacancies (missing atoms), drastically reduce the strength and
stiffness of graphene [10] that critically influence the performance
of NEMS. On the other hand, vacancy defects have been used to
control the electromechanical properties of graphene. Irradiation
induced vacancies turn graphene into two-dimensional amorphous
membrane, which opens new possibilities for engineering graphene based NEMS [11]. Defective graphene has also been used
as a high-capacity anode material in Na-/Ca-ion batteries [12].
Graphene reinforcements remarkably improve the fracture toughness of composite materials, where the defects ensure proper
bonding between the sheets and the polymer matrix [13]. A thorough understanding on the strength and stability of defective graphene are essential to design these advanced NEMS and
composite materials.
Experiments on the mechanical behavior of graphene are limited due to practical problems in designing experiments at the
nanoscale [79]. Therefore, atomistic methods such as molecular
dynamics (MD) have been extensively used to study the mechanics of graphene [1424]. Wrinkles in graphene can be manipulated
using substrate-supported nanoparticles, which allows controlling
the electromechanical properties [14]. Graphene shows a brittleto-ductile fracture transition at higher vacancy concentrations,
1
Corresponding author
Contributed by the Applied Mechanics Division of ASME for publication in the
JOURNAL OF APPLIED MECHANICS. Manuscript received April 6, 2014; final manuscript
received May 11, 2014; accepted manuscript posted May 15, 2014; published online
June 2, 2014. Assoc. Editor: Pradeep Sharma.

Journal of Applied Mechanics

which opens a new branch of advanced NEMS [15,16]. Even with

cracks (a row of vacancies), armchair graphene sheets are significantly stronger than other ultrastrong nanomaterials, such as silicon carbide [17]. The strength of graphene heavily depends on the
orientation of the sheet [18], and the deformation is completely
elastic up to the fracture, which occurs 18% of strain [19].
Cracks in graphene predominantly propagate along the armchair
or zigzag directions [20]. The free zigzag edges are energetically
and kinetically stable than the free armchair edges [21]; fracture
of an armchair sheet creates free zigzag edges. The strength of
graphene with grain boundaries significantly decreases with the
increase of temperature, but the strength slightly increases as the
strain rate increases from 108 s1 to 1010 s1 [22].
The main disadvantage of MD is the computational cost that
can only be overcome by using very high strain rates, which are
generally 109 s1 [1524]. The experimental strain rates, however, are around 103 s1 [1,10]. Molecular dynamics simulations
at 300 K revealed that as the strain rate decreases from 1011 s1
to 108 s1, the strength of pristine armchair graphene reduces by
10%, and this strength reduction is temperature dependent [17].
Strain rates below 107 s1 are impractical in MD simulations
due to high computational cost. Graphene sheets are subjected to
high temperatures (1300 K) when synthesizing by chemical
vapor deposition [7], and also during the fabrication of graphene
based composite materials [13]. Several studies have focused on
the temperature dependent mechanical properties of pristine graphene. However, the effects of temperature on defective sheets
have not been well understood. We recently developed a temperature dependent atomistic model to investigate the strength of pristine graphene with a crack [23]. This model is computationally
very efficient and highly accurate compared to the MD. However,
the model does not predict the strength of defective sheets, and
also the strain rate dependent strength was not investigated.
The classical continuum mechanics is computationally very efficient and provides significant in sights into the mechanical
behavior of graphene [2325]. Griffiths criterion and the stress
intensity factor can be used to characterize the fracture strength of

C 2014 by ASME
Copyright V

AUGUST 2014, Vol. 81 / 081010-1

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms

graphene [23]. The concept of surface stress in three-dimensional

crystals accurately predicts the edge energy induced strain of graphene [24]. EulerBernoulli beam theory with von Karman geometric nonlinearity has been used to investigate the mechanics of
multilayer sheets [25]. Therefore, the continuum concepts can be
quite effective when used to perform the preliminary analysis of
graphene based systems, which will save an extensive amount of
computational cost compared to the MD.
In this paper, we present a comprehensive MD study to assess
the effects of vacancies on the strength of graphene at various
temperatures; the crack propagation is also investigated. We present an atomistic model to evaluate the strain rate and temperature
dependent strength of defective graphene. This model is an efficient computational tool to assess the strength of graphene at practical strain rates as opposed to the very high strain rates in MD.
We further develop a novel continuum based fracture mechanics
framework to characterize the strength of defective graphene.
The paper is organized as follows. Section 2 describes the MD
simulations of defective graphene at various temperatures. Section 3
presents an atomistic model to evaluate the fracture strength of defective graphene. The strain rate and temperature dependent fracture
strength of graphene is obtained in Sec. 4. Section 5 presents a novel
continuum based fracture mechanics framework to evaluate the
strength of defective graphene. Conclusions are drawn in Sec. 6.

Molecular Dynamics Simulations

2.1 Potential Field. We performed molecular dynamics simulations using LAMMPS package [26] with the adaptive intermolecular reactive empirical bond order (AIREBO) potential [27].
The AIREBO potential consists of three subpotentials, which are
the REBO, LennardJones, and torsional potentials. The REBO
potential gives the energy stored in atomic bonds; the Lennard
Jones potential considers the nonbonded interactions between
atoms, and the torsional potential includes the energy from torsional interactions between atoms.
According to the REBO potential [28], the energy stored in a
bond between atom i and atom j can be expressed as
 

(1)
Eij REBO f rij Vij R bij Vij A
where VijR and VijA are the repulsive and the attractive potentials,
respectively; bij is the bond order term, which modifies the attractive potential depending on the local bonding environment; rij is
the distance between the atoms i and j; f(rij) is the cut-off function.
The cut-off function in REBO potential [28], given in Eq. (2), limits the interatomic interactions to the nearest neighbors
8
>
>
>
<

" 1;
#
p rij  R1
;
f rij 1 cos
>
R2  R1
>
>
:
0;

rij < R1
R1 < rij < R2

(2)

R2 < rij

and 2 A
,
where R(1) and R(2) are the cut-off radii, which are 1.7 A
respectively. The values of cut-off radii are defined based on the
first and the second nearest neighboring distances of the relevant
hydrocarbon. The cut-off function, however, causes nonphysical
strain hardening in carbon nanostructures [29]. Therefore, modi to 2.2 A
, have been used to
fied cut-off radii, ranging from 1.9 A
eliminate this nonphysical strain hardening [17,18,21]. In this
study, we used a truncated cut-off function ft(rij), given in Eq. (3)
[30], to eliminate this strain hardening

ft rij

1;
0;

rij < R
rij > R

(3)

. Similar cut-off functions have been

where the value of R is 2 A
used in Refs. [19] and [21] to simulate the fracture of graphene.
081010-2 / Vol. 81, AUGUST 2014

Fig. 1 A graphene sheet, with a single vacancy, is subjected to

a strain e0 along the armchair direction. The zigzag direction is
perpendicular to the armchair direction. The applied strain, e0,
induces stress concentrations at the carboncarbon bonds
circled in red.

2.2 MD Parameters. The size of the simulated graphene

sheet was 5 nm  5 nm with 1008 carbon atoms. Periodic boundary conditions were used along the in-plane directions to eliminate
the effects of free edges. The sheet was allowed to relax over 30
ps before applying strain; the time step was 0.5 fs. During the
relaxation period, the pressure component along in-plane directions were kept at zero using NPT ensemble implemented in
LAMMPS. The NPT ensemble controls the temperature using
NoseHoover thermostat [31,32], which induces a nonphysical
thermal expansion in graphene [24]. An initial random out-of ) was imposed on the
plane displacement perturbation (0.05 A
carbon atoms to eliminate this nonphysical thermal expansion.
Strain was applied by pulling the sheet along the armchair or zigzag direction at a strain rate of 109 s1. The stress component perpendicular to the pulling direction was kept at zero by controlling
the relevant pressure component, using the NPT ensemble, to simulate uniaxial tensile test. Figure 1 shows a graphene sheet subjected to a strain e0 along the armchair direction; in this case, the
stress component along zigzag direction, which is perpendicular
to the armchair direction, was kept at zero. We used VMD package [33] to visualize the MD simulation trajectories.
2.3 Calculation of Stress. Stress in molecular dynamics simulations has been interpreted using either the Cauchy stress
[19,24] or the virial stress [15,23]. The Cauchy stress is computationally efficient than the virial stress. However, the Cauchy stress
induces a nonphysical initial stress (at zero strain) at higher temperatures, whereas the virial stress gives the initial stress as zero
[24]. The Cauchy stress is the gradient of the potential energy per
unit volume versus strain curve; the virial stress [34], rij, is
defined as
"
#
N 

1 X 1X
a a a
Rbi  Rai Fab

m
v
v
(4)
rij
j
i j
V a 2 b1
where i and j are the directional indices (x, y, and z); a is a number
assigned to an atom; b is a number assigned to neighboring
atoms of atom a which varies from 1 to N; Rib is the position of
atom b along the direction i; Fjab is the force along the direction j
on atom a due to atom b; ma and va are the mass and the velocity
of atom a, respectively; V is the total volume. It should be noted
that stress of a single atom could be obtained by ignoring the summation over a, then V is the volume of a single atom.
The definition of volume in the virial stress, however, is ambiguous [15]; the virial stress is quite similar to the Cauchy stress
when instantaneous volume is used in the virial calculation [23].
In this work, we used the instantaneous volume to calculate the
Transactions of the ASME

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms

virial stress. When the applied strain is e, the instantaneous volume can be approximated as V0(1 e), where V0 is the volume at
zero strain. However, we observed a minor lateral contraction of
graphene sheets at higher strains. Therefore, in this work, we
obtained the instantaneous volume by considering the instantaneous positions of atoms given in MD simulations, which is more
accurate since this volume takes into account the observed lateral
, which is
contraction. Thickness of graphene was assumed 3.4 A
the interlayer spacing of graphene in graphite. Five MD simulations, with different randomly distributed vacancies, were performed for each vacancy concentration at a given temperature.
The strength is less sensitive (<5%) to the distribution of vacancies in the sheet. Therefore, the average strength of these five simulations was used for the analysis.
2.4 Effects of Temperature and Vacancies. As shown in
Fig. 2, the temperature and vacancies profoundly change the
stressstrain behavior of graphene. A vacancy concentration of
2% reduces the strength of graphene by 50%, and the toughness
is reduced by 75%. The pristine graphene sheets fail by homogeneous nucleation of defects. In defective sheets, vacancies act
as the fracture nucleation sites. The stress concentration at vacancies leads to the failure of defective sheets at much lower strains
compared to the pristine sheets. In the studied temperature range
(i.e., 300 K to 1500 K), one missing carbon atom reduces the
strength of armchair and zigzag graphene by 15% and 20%,
respectively; this indicates that the nucleation of defects requires a
significant amount of energy compared to the energy stored up to
the fracture.
The strength of pristine zigzag sheets reduces with increasing
temperature by 0.032 GPa K1, which is 37% higher than the
reduction rate of armchair graphene. As shown in Fig. 2, however,
zigzag sheets are stronger than armchair sheets even at higher
temperatures. The strength of defective zigzag sheets is less temperature dependent (19%) compared to the pristine ones. This
high temperature dependency of pristine sheets indicates that the
homogeneous nucleation of defects heavily depends on the temperature. However, defective and pristine armchair sheets show
similar rates of reduction in the strength with increasing temperature. This can be due to higher stress concentration of armchair
sheets (17% at 300 K) compared to zigzag sheets [23] that arise
from the bond arrangement at the crack tip (see Fig. 1).

of a defect (i.e., breaking a carboncarbon bond) at a strain of

13.99%. This defect nucleation could occur at any place since
the sheet is infinitely large as the periodic boundary conditions
are implemented along the in-plane directions. Figure 3(b) shows
that the defective sheet experiences significant stress concentration at the vacancies even at relatively lower strains 3.5%. The
crack propagation initiates at a strain of 7.7%. The nearneighbor vacancies coalesce as the crack propagates; thereby, it
propagates at a much lower speed compared to the pristine sheet.
The fracture is almost complete around a strain of 7.77%, but
five carbon chains hold the two pieces together, and these chains
carry high stresses, which can also be seen in Fig. 2(a). However, as the strain increases to 8.85%, two of these cross-links
have broken and lengths of the other three links have increased
so they do not carry any more stresses. These cross-links can be
a main reason for the recently observed [15,16] brittle to ductile
fracture transition of graphene at higher vacancy concentrations
(8% to 10%).
Figure 4(a) compares the variation in potential energy as a
crack propagates in defective armchair sheets at 300 K. The sheets
with lower vacancy concentrations store a higher amount of strain
energy before the fracture, and this energy is released during the
fracture. The figure shows that the energy dissipation rate
decreases as the vacancy concentration increases. The number of
near-neighbor vacancies is higher at larger vacancy concentration.
Therefore, more vacancies coalesce as a crack propagates, which
results a much lower crack propagation speed. In Fig. 4(b), the
energy dissipation rate of defective graphene is compared with its
strength, and it clearly indicates proportionality. The energy dissipation rate (G) in Fig. 4(b) is defined as G DU/Dt, where DU is
the change in potential energy and Dt is the time taken for the initial failure.
As observed in this section, the MD simulations provide a remarkable insight into the fracture behavior of graphene. However,
the main limitation of MD is the unavoidable high strain rates,
which is around 1012 times the experimental strain rates. In the
ensuing section, we propose an atomistic model, based on the Baileys criterion and the Arrhenius equation, to assess the strain rate
and temperature dependent fracture strength of defective graphene. The model provides an efficient computational tool to
characterize the fracture strength of graphene.

3
2.5 Fracture and Crack Propagation. Our molecular dynamics simulations show that even pristine graphene sheets do not
completely break at the fracture; Fig. 3 shows that a few carbon
chains connect the two nearly separated pieces; as shown in Fig.
3(a), the fracture of pristine graphene is initiated by the nucleation

Atomistic Modeling of Fracture Strength

In our recent study [23], we combined the Arrhenius equation

and the Baileys criterion with the quantized fracture mechanics
to obtain temperature dependent strength of pristine graphene
with a central crack. This model, however, cannot be used to
evaluate the strength of defective sheets, and the strain rate

Fig. 2 Stressstrain curves of pristine and defective (a) armchair and (b) zigzag sheets at various temperatures. Defective sheets have 2% of randomly distributed vacancies.

Journal of Applied Mechanics

AUGUST 2014, Vol. 81 / 081010-3

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms

Fig. 3 The fracture of (a) pristine and (b) defective armchair sheets with 2% vacancy concentration. The simulations were done at 300 K.

dependent strength was also not investigated. In this work, we

use the Arrhenius equation and the Baileys criterion to model
the temperature and strain rate dependent fracture strength of defective graphene.
The Baileys criterion of durability [35] provides a basis to estimate the lifetime of materials at various temperatures [36]. The
criterion is expressed as
tf
0

dt
1
sT; t

(5)

where tf is the time (t) taken to the fracture; s(T,t) is the durability
function at temperature T, which is generally determined by
081010-4 / Vol. 81, AUGUST 2014

experiments [36]. The Arrhenius equation, however, is a good

approximation to the durability function [17].
The Arrhenius equation [37] expresses the temperature dependent rate of a chemical reaction (k) as k A  exp[DE/(kBT)],
where A is a constant that depends on the chemical bonding; DE is
the activation energy barrier; kB is the Boltzmann constant. When
a mechanical force F is applied to a molecule, the activation
energy barrier reduces by an amount of FDx, where Dx is the
change in the atomic coordinates due to F [38]. We defined a durability function for graphene in the form of Arrhenius equation as

s0
U0 =b  crt
sT; t exp
kB T
n

(6)

Transactions of the ASME

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms

Fig. 5 Graphical representation of the solutions of Eq. (9) for

armchair graphene at various temperatures

for zigzag sheets, where 91.7 and 108.9 are the tensile strengths,
in GPa, of armchair and zigzag sheets at 300 K, respectively. r(t)
is the stress at time t, which we expressed in terms of the strain
rate e_ as
_ bet
_ 2
rt aet

Fig. 4 (a) Change in the potential energy (DPE) during fracture

of armchair graphene with various vacancy concentrations
from 0.1% to 4%. Down and up arrows indicate the fracture initiation and completion points, respectively. (b) Correlation
between the energy dissipation rate and the strength of defective armchair (ac) and zigzag (zz) sheets with various vacancy
concentrations used in Fig. 4(a). The linear regression lines are
also shown.

where s0 is the vibration period of atoms that is 5 fs for carbon in

graphene [23]; n is the number of bonds in the sheet; U0 is the
interatomic bond dissociation energy that is 4.95 eV for a carbon
carbon bond [27]; b represents the reduction of average bond dissociation energy due to the presence of vacancies; we defined b,
using MD simulations at 300 K, as

b

1;
0:165a k;

a0
a>0

(7)

where a is the vacancy percentage. Even presence of a single vacancy reduces the strength drastically; this strength reduction is
considered by the constant k. The values of k are 1.13 and 1.21 for
armchair and zigzag sheets, respectively. c vq, where v is the
3; the value of v is close to the
activation volume, which is 8.25 A
representative volume of a carbon atom in graphene, which is
3. q is a directional constant that takes into account the dif8.6 A
ferent bond orientation along the armchair and zigzag directions
(see Fig. 1); q is 1 for armchair sheets and it is 91.7/108.9 (0.82)
Journal of Applied Mechanics

(8)

where a and b are the second and the third order elastic moduli,
respectively; the values of a and b were obtained from regression
analysis of the stressstrain curves given by MD simulations at
300 K, where a and b are 1.11 TPa and 3.20 TPa for armchair
sheet, the corresponding values for zigzag sheet are 0.91 TPa and
1.90 TPa.
We calculated the failure time tf by numerically solving Eq.
(5), which can be simplified as
tf
0

crt  U0 =b
s0
exp
dt
kB T
n

(9)

The solutions of Eq. (9) for pristine armchair sheets, at a strain

rate of 109 s1, are graphically represented in Fig. 5; the figure
shows that the time taken to the fracture, tf, gradually decreases as
the temperature increases. We obtained the fracture stress r(tf) by
substituting the tf into Eq. (8). The ensuing section compares the
fracture strength given by the proposed model with the MD simulations results.

4 Temperature and Strain Rate Dependent

Fracture Strength
The atomistic model, proposed in Sec. 3, has three key features:
(1) The model is temperature dependent; (2) the model is strain
rate dependent; and (3) it is able model both pristine and defective
graphene. Figure 6 shows that the proposed model accurately predicts the strength of pristine and defective armchair and zigzag
graphene in practical ranges of temperature (300 K to 1500 K) and
vacancies (up to 2%). A strain rate of 109 s1 is used in the MD
and the numerical model. Figure 7 shows that the model accurately predicts the strain rate dependent strength of pristine armchair and zigzag graphene; the predicted strengths are within 4%
of the MD results. Strain rates below 107 s1 are impractical in
MD due to very high computational cost. Therefore, the proposed
AUGUST 2014, Vol. 81 / 081010-5

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms

Fig. 6 Comparison of the strength of (a) armchair and (b) zigzag sheets given by the proposed model and the MD
simulations

numerical model is very useful to explore the strength of graphene

at practical strain rates, which are 103 s1.
Figure 8 shows that the strain rate has a very high impact on the
strength, especially at elevated temperatures. A strain rate

Fig. 8 Strain rate and temperature dependent strength of (a)

pristine armchair and (b) pristine zigzag sheets

109 s1 is commonly used in MD simulations. As the strain rate

decreases from 109 s1 to 103 s1, the strength reduction at
1500 K is about five times higher than the strength reduction at
300 K. This result indicates that the MD simulations at elevated
temperatures highly overpredict the strength of graphene. This
finding is very useful since graphene is subjected to various temperatures and strain rates during fabrication of graphene based
composite materials [13]; higher mixing rates at elevated temperatures could tear graphene sheets into pieces.
Temperature and strain rate dependent fracture strength of pristine graphene (rpf ), given in Figs. 8(a) and 8(b), can be condensed
into a simple equation as
rpf rp0  ue_T

Fig. 7 Strain rate dependent fracture strength of graphene. MD

simulations ware performed at 300 K.

081010-6 / Vol. 81, AUGUST 2014

(10)

where rp0 is the projected strength of pristine sheets at 0 K; the values are 97 GPa and 114 GPa for armchair and zigzag sheets,
_ is the gradient of the curves in Fig. 8, which is
respectively. u(e)
3
_ [58.8 3.7log(e)]/10
_
GPa K1 for
strain rate dependent; u(e)
3
_
GPa K1 for zigzag
armchair sheets and it is [70.4 4.5log(e)]/10
_ at 109 s1 are 0.025 GPa K1 and
sheets. The values of u(e)
0.030 GPa K1 for armchair and zigzag graphene sheets, respectively. The corresponding values given by the MD simulations
are 0.023 GPa K1 and 0.032 GPa K1 for armchair and zigzag
Transactions of the ASME

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms

Fig. 9 The strain rate dependent strength of armchair and zigzag sheets with higher vacancy percentages at 300 K

Fig. 10 Variation of the strength of graphene with the square

root of vacancy percentage

graphene sheets, which are very close to the values given by the
proposed equation. Researchers are currently making an excellent
progress toward synthesizing defects free graphene sheets [9]. A
simple analytical approach such as Eq. (10) is quite useful to
determine the strength of pristine graphene under various processing conditions.
We evaluate the strain rate dependent strength of highly defective graphene at 300 K. The results, at a strain rate of 109 s1,
were compared with the MD simulations. Figure 9 shows that the
model is quite accurate up to 8% of vacancy concentration, and
then it slightly underpredicts the strength. A recent study [15]
showed that the fracture of graphene transforms from brittle to
ductile around 9% of vacancy concentration. The proposed model
assumes a brittle fracture of graphene, irrespective to the vacancy
concentration; thereby it marginally underpredicts the strength.
However, the agreement between the model and the MD is quite
reasonable even up to a vacancy concentration of 16%. Figure 9
shows that MD simulation at a commonly used strain rate
(109 s1) gives a finite strength (25 GPa) around 16% of vacancy
concentration; the model shows that graphene almost completely
loses its strength (5 GPa) under a practical strain rate (103 s1).
The figure also shows that armchair and zigzag sheets have identical strength at higher vacancy concentrations (>10%) and lower
strain rate (103 s1).

interestingly, K IC , given in Table 1, is strain rate and temperature independent. The average values of K IC are 115 GPa and
129 GPa for armchair and zigzag sheets, respectively. The constant d, however, is temperature and strain rate dependent and
_
_ is
where u(e)
it takes the form of Eq. (10); d d0 u(e)T,
defined in Eq. (10); d0 is 1 GPa for both armchair and zigzag
sheets.
The observed formal similarity in the fracture strength of the
sheets with a single crack and the sheets with random vacancies
occurs since the failure in the both cases occurs due to the stress
concentration. Figure 11 shows the stress concentration, just
before fracture, of armchair sheets with random vacancies and the
sheets with a single crack. Length and width of the graphene
sheets, with a single crack, are selected to be ten times the crack
length in order to avoid the finite-size effects in MD simulations
[30]. According to Fig. 11, the sheet with a crack length (2 a) of
and the sheet with a vacancy concentration (a) of 1% have
7.3 A

similar fracture strength. The sheets with crack lengths of 9.7 A

have comparable strengths with the sheets that have
and 14.5 A
2% and 4% vacancy concentrations, respectively. In the defective
sheets, the failure strain (ef) almost remains constant even though
the strength significantly reduces with the increase of vacancy
concentration. This indicates that vacancies reduce the stiffness of
graphene, which can also be seen in Fig. 2(a). In the sheets with a
crack, however, the ef gradually decreases with the strength,
which indicates a relatively less reduction in the stiffness.

5 Continuum Based Fracture Mechanics Framework

for Defective Graphene
In continuum fracture mechanics, the fracture strength rf is
expressed in terms of the critical stress
intensity
factor (KIC) and
p

the crack length (2 a) as rf KIC = 2pa [39]. Our recent study

[23] showed that graphene slightly deviates from this continuum
concept due to discontinuity
p of the sheets, and the above equation
becomes rf KIC = 2pa c, where c is a constant that depends
on the orientation of the crack (i.e., armchair or zigzag). Figure 10
shows
p that the strength of graphene indicates proportionality with
1= a, where a is the vacancy percentage. This clearly demonstrates that the sheets with random vacancies and the sheets with a
single crack show similar strength characteristics. Therefore, we
express the fracture strength of defective sheets as
K IC
rf p  d
a
Journal of Applied Mechanics

(11)

Table 1 The values of K IC (GPa) and d (GPa) at various temperatures and strain rates
Armchair

Zigzag

9 1

3 1

10

10 s

9 1

10 s

103 s1

Temperature (K)

KIC

KIC

KIC

KIC

250
300
350
600
900

115
116
116
115
114

7.4
8.6
10.0
15.5
22.3

114
116
114

18.3
22.9
25.2

130
127
128
130
128

8.5
8.8
10.6
18.3
26.3

131
129
129

22.3
24.9
28.6

AUGUST 2014, Vol. 81 / 081010-7

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms

Fig. 11 Stress concentration of armchair graphene (a)(c) with a single crack of length 2a and (d)(e) with various vacancy
concentrations (a). The simulation temperature is 300 K.

Conclusions

In summary, we used MD simulations to study the influence of

vacancy concentration on the strength of graphene at various temperatures. Results reveal that vacancy defects and temperature
have a profound impact on the strength. We also developed an atomistic model to assess the temperature and strain rate dependent
fracture strength of defective graphene. The model reveals that
MD simulations highly overpredict the strength of graphene at
elevated temperatures. The proposed model provides an efficient
computational tool to evaluate the design strength of graphene
under various processing conditions. The model can also be used
to characterize the fracture strength of other nanostructures such
as silicon nanotubes and gold nanowires. Graphene with random
vacancies showed singular stress field as in continuum fracture
mechanics, and the energy dissipation rate shows proportionality
with the strength. We also developed a continuum based fracture
mechanics framework to characterize the strength of defective
graphene. These findings provide valuable insights into the
strength of defective graphene which is useful in experimental
and instrumental applications where defects are unavoidable.

Acknowledgment
This work was financially supported by Natural Sciences and
Engineering Research Council (NSERC) of Canada. Computing
resources were provided by WestGrid and Compute/Calcul
Canada.
081010-8 / Vol. 81, AUGUST 2014

References
[1] Lee, C., Wei, X., Kysar, J. W., and Hone, J., 2008, Measurement of the Elastic
Properties and Intrinsic Strength of Monolayer Graphene, Science, 321(5887),
pp. 385388.
[2] Singh, V., Irfan, B., Subramanian, G., Solanki, H. S., Sengupta, S., Dubey, S.,
Kumar, A., Ramakrishnan, S., and Deshmukh, M. M., 2012, Coupling
Between Quantum Hall State and Electromechanics in Suspended Graphene
Resonator, Appl. Phys. Lett., 100(23), p. 233103.
[3] Klimov, N. N., Jung, S., Zhu, S. Z., Li, T., Wright, C. A., Solares, S. D., Newell, D. B., Zhitenev, N. B., and Stroscio, J. A., 2012, Electromechanical Properties of Graphene Drumheads, Science, 336(6088), pp. 15571561.
[4] Espinosa-Ortega, T., Lukyanchuk, I. A., and Rubo, Y. G., 2013, Magnetic
Properties of Graphene Quantum Dots, Phys. Rev. B, 87(20), p. 205434.
[5] Chen, C., Rosenblatt, S., Bolotin, K. I., Kalb, W., Kim, P., Kymissis, I.,
Stormer, H. L., Heinz, T. F., and Hone, J., 2009, Performance of Monolayer
Graphene Nanomechanical Resonators With Electrical Readout, Nat. Nanotechnol., 4(12), pp. 861867.
[6] Chen, C. Y., Lee, S., Deshpande, V. V., Lee, G. H., Lekas, M., Shepard, K., and
Hone, J., 2013, Graphene Mechanical Oscillators With Tunable Frequency,
Nat. Nanotechnol., 8(12), pp. 923927.
[7] Novoselov, K. S., Falko, V. I., Colombo, L., Gellert, P. R., Schwab, M. G., and
Kim, K., 2012, A Roadmap for Graphene, Nature, 490(7419), pp. 192200.
[8] Banhart, F., Kotakoski, J., and Krasheninnikov, A. V., 2011, Structural Defects
in Graphene, ACS Nano, 5(1), pp. 2641.
[9] Gao, L. B., Ni, G. X., Liu, Y. P., Liu, B., Neto, A. H. C., and Loh, K. P., 2014,
Face-To-Face Transfer of Wafer-Scale Graphene Films, Nature, 505(7482),
pp. 190194.
[10] Zandiatashbar, A., Lee, G. H., An, S. J., Lee, S., Mathew, N., Terrones, M.,
Hayashi, T., Picu, C. R., Hone, J., and Koratkar, N., 2014, Effect of Defects on
the Intrinsic Strength and Stiffness of Graphene, Nat. Commun., 5, p. 3186.
[11] Kotakoski, J., Krasheninnikov, A. V., Kaiser, U., and Meyer, J. C., 2011, From
Point Defects in Graphene to Two-Dimensional Amorphous Carbon, Phys.
Rev. Lett., 106(10), p. 105055.

Transactions of the ASME

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms

[12] Datta, D., Li, J. W., and Shenoy, V. B., 2014, Defective Graphene as a HighCapacity Anode Material for Na- and Ca-Ion Batteries, ACS Appl. Mater.
Interfaces, 6(3), pp. 17881795.
[13] Rafiee, M. A., Rafiee, J., Srivastava, I., Wang, Z., Song, H. H., Yu, Z. Z., and
Koratkar, N., 2010, Fracture and Fatigue in Graphene Nanocomposites,
Small, 6(2), pp. 179183.
[14] Zhu, S., and Li, T., 2014, Wrinkling Instability of Graphene on SubstrateSupported Nanoparticles, ASME J. Appl. Mech., 81(6), p. 061008.
[15] Carpenter, C., Maroudas, D., and Ramasubramaniam, A., 2013, Mechanical
Properties of Irradiated Single-Layer Graphene, Appl. Phys. Lett., 103(1), p.
013102.
[16] Xu, L. Q., Wei, N., and Zheng, Y. P., 2013, Mechanical Properties of Highly
Defective Graphene: From Brittle Rupture to Ductile Fracture, Nanotechnology, 24(50), p. 505703.
[17] Zhao, H., and Aluru, N. R., 2010, Temperature and Strain-Rate Dependent
Fracture Strength of Graphene, J. Appl. Phys., 108(6), p. 064321.
[18] Jhon, Y. I., Jhon, Y. M., Yeom, G. Y., and Jhon, M. S., 2014, Orientation
Dependence of the Fracture Behavior of Graphene, Carbon, 66, pp.
619628.
[19] Cao, A. J., and Qu, J. M., 2013, Atomistic Simulation Study of Brittle Failure
in Nanocrystalline Graphene Under Uniaxial Tension, Appl. Phys. Lett.,
102(7), p. 071902.
[20] Kim, K., Artyukhov, V. I., Regan, W., Liu, Y. Y., Crommie, M. F., Yakobson,
B. I., and Zettl, A., 2012, Ripping Graphene: Preferred Directions, Nano
Lett., 12(1), pp. 293297.
[21] Zhang, B., Mei, L., and Xiao, H. F., 2012, Nanofracture in Graphene
Under Complex Mechanical Stresses, Appl. Phys. Lett., 101(12), p.
121915.
[22] Yi, L. J., Yin, Z. N., Zhang, Y. Y., and Chang, T. C., 2013, A Theoretical
Evaluation of the Temperature and Strain-Rate Dependent Fracture Strength of
Tilt Grain Boundaries in Graphene, Carbon, 51, pp. 373380.
[23] Dewapriya, M. A. N., Rajapakse, R. K. N. D., and Phani, A. S., 2014,
Atomistic and Continuum Modelling of Temperature-Dependent Fracture of
Graphene, Int. J. Fract., 187(2), pp. 199212.
[24] Dewapriya, M. A. N., Phani, A. S., and Rajapakse, R. K. N. D., 2013,
Influence of Temperature and Free Edges on the Mechanical Properties of Graphene, Modell. Simul. Mater. Sci. Eng., 21(6), p. 065017.

Journal of Applied Mechanics

[25] Rokni, H., and Lu, W., 2013, Surface and Thermal Effects on the Pull-In
Behavior of Doubly-Clamped Graphene Nanoribbons Under Electrostatic and
Casimir Loads, ASME J. Appl. Mech., 80(6), p. 061014.
[26] Plimpton, S., 1995, Fast Parallel Algorithms for Short-Range MolecularDynamics, J. Comp. Phys., 117(1), pp. 119.
[27] Stuart, S. J., Tutein, A. B., and Harrison, J. A., 2000, A Reactive Potential for
Hydrocarbons With Intermolecular Interactions, J. Chem. Phys., 112(14), pp.
64726486.
[28] Brenner, D. W., 1990, Empirical Potential for Hydrocarbons for Use in Simulating the Chemical Vapor-Deposition of Diamond Films, Phys. Rev. B,
42(15), pp. 94589471.
[29] Shenderova, O. A., Brenner, D. W., Omeltchenko, A., Su, X., and Yang, L. H.,
2000, Atomistic Modeling of the Fracture of Polycrystalline Diamond, Phys.
Rev. B, 61(6), pp. 38773888.
[30] Dewapriya, M. A. N., 2012, Molecular Dynamics Study of Effects of Geometric Defects on the Mechanical Properties of Graphene, M.A.Sc. thesis, The
University of British Columbia, Vancouver, Canada.
[31] Nose, S., 1984, A Molecular-Dynamics Method for Simulations in the Canonical Ensemble, Mol. Phys., 52(2), pp. 255268.
[32] Hoover, W. G., 1985, Canonical DynamicsEquilibrium Phase-Space Distributions, Phys. Rev. A, 31(3), pp. 16951697.
[33] Humphrey, W., Dalke, A., and Schulten, K., 1996, VMD: Visual Molecular
Dynamics, J. Mol. Graphics Modell., 14(1), pp. 3338.
[34] Tsai, D. H., 1979, The Virial Theorem and Stress Calculation in Molecular
Dynamics, J. Chem. Phys., 70(3), pp. 13751382.
[35] Bailey, J., 1939, An Attempt to Correlate Some Tensile Strength Measurements on Glass: III, Glass Ind., 20, pp. 9599.
[36] Freed, A. D., and Leonov, A. I., 2002, The Bailey Criterion: Statistical Derivation and Applications to Interpretations of Durability Tests and Chemical Kinetics, Z. Angew. Math. Phys., 53(1), pp.160166.
[37] Arrhenius, S., 1889, On the Reaction Rate of the Inversion of the Non-Refined
Sugar Upon Souring, Z. Phys. Chem., 4, pp. 226248.
[38] Kuo, T. L., Garcia-Manyes, S., Li, J. Y., Barel, I., Lu, H., Berne, B. J., Urbakh, M.,
Klafter, J., and Fernandez, J. M., 2010, Probing Static Disorder in Arrhenius Kinetics
by Single-Molecule Force Spectroscopy, PNAS, 107(25), pp. 1133611340.
[39] Anderson, T. L., 1991, Fracture Mechanics: Fundamentals and Applications,
CRC Press, Boca Raton, FL, Chap. II.

AUGUST 2014, Vol. 81 / 081010-9

Downloaded From: http://appliedmechanics.asmedigitalcollection.asme.org/ on 07/17/2014 Terms of Use: http://asme.org/terms