Professional Documents
Culture Documents
net/publication/227186751
CITATIONS READS
60 326
2 authors, including:
Shailesh I. Kundalwal
University of Toronto
40 PUBLICATIONS 929 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Shailesh I. Kundalwal on 19 October 2015.
Received: 15 November 2010 / Accepted: 21 April 2011 / Published online: 13 May 2011
Ó Springer Science+Business Media, B.V. 2011
Abstract A novel fuzzy fiber reinforced composite extensive work to predict their effective properties.
(FFRC) reinforced with zig-zag single-walled carbon Treacy et al. (1996) experimentally determined that
nanotubes (CNTs) and carbon fibers is proposed. The CNTs have axial Young’s modulus in the terapascal
distinct constructional feature of this composite is (TPa) range. Lu (1997) estimated elastic properties of
that the uniformly aligned CNTs are radially grown CNTs and nanoropes using an empirical force con-
on the surface of carbon fibers. Analytical models stant relation. Li and Chou (2003) linked structural
based on the mechanics of materials approach and the and molecular mechanics (MM) approaches to com-
Mori–Tanaka method are derived to estimate the pute elastic properties of CNTs. Shen and Li (2004)
effective elastic constants of this proposed FFRC. reported that CNTs can be modeled as transversely
The values of the effective elastic properties of this isotropic materials with the axis of transverse isotropy
composite are estimated with and without considering coincident with the centroidal axis of the CNT and
an interphase between the CNT and the polymer developed variational models to determine the values
matrix. It has been found that the transverse effective of the five elastic constants of CNTs. Xiao et al.
properties of this composite are significantly (2005) developed an analytical model based on the
improved due to the radial growing of CNTs on the molecular structural mechanics approach for estimat-
surface of carbon fiber. The effective properties are ing the mechanical properties of carbon nanotubes.
also found to be sensitive to the CNT diameter. Batra and Sears (2007) proposed that the axis of
transverse isotropy of a CNT is a radial line rather
Keywords Fuzzy fiber composites than the centroidal axis of the CNT and found that
Micromechanics Effective properties Young’s modulus in the radial direction equals about
Advanced composites of that in the axial direction. Gupta and Batra
(2008) determined the wall thickness and material
moduli of a CNT based on the frequencies of axial,
1 Introduction torsional, and radial breathing modes. An atomistic-
based continuum model has been developed by Cheng
Since the discovery of carbon nanotubes (CNTs) et al. (2009) for the estimation of the mechanical
(Iijima 1991), researchers have been carrying out properties of single-walled carbon nanotubes.
In order to harness the exceptionally attractive
S. I. Kundalwal M. C. Ray (&)
mechanical properties of CNTs, extensive research is
Department of Mechanical Engineering, Indian Institute
of Technology, Kharagpur 721302, India being carried out for developing CNT-reinforced
e-mail: mcray@mech.iitkgp.ernet.in composites. For example, Thostenson and Chou
123
150 S. I. Kundalwal, M. C. Ray
(2003) have estimated the elastic moduli of CNT- Although a great prospect has been highlighted
reinforced composite through micromechanical analy- through the above mentioned in situ research on
sis. Using the approach of continuum mechanics, CNT-reinforced composites, the manufacturing of
Odegard et al. (2003) predicted the effective elastic such unidirectional continuous CNT-reinforced com-
moduli of CNT-reinforced composite using an equiv- posites in large scale has to encounter some chal-
alent continuum modeling method. Gou et al. (2004) lenging difficulties. Typical among these are the
investigated the interfacial bonding of single-walled agglomeration of CNTs, the misalignment and the
carbon nanotube (SWCNT) reinforced epoxy compos- difficulty in manufacturing very long CNTs. Further
ites using a combination of computational and exper- research on the practical use of CNTs and improving
imental methods. Gao and Li (2005) derived a shear lag the effective properties of existing fiber-reinforced
model of discontinuous CNT-reinforced polymer polymer matrix composites has led to the growth of
composites by considering the CNT as an equivalent CNTs on the surface of the advanced fibers. For
solid fiber. Song and Youn (2006) numerically esti- example, Zhang et al. (2008) produced CNT arrays
mated the effective elastic properties of CNT-rein- on the host aluminum silicate and quartz fibers.
forced polymer based composites. Zhang and He Mathur et al. (2008) experimentally investigated that
(2008) theoretically investigated the viscoelastic the flexural strength and the modulus of the carbon-
behavior of CNT-reinforced composites, developing fiber-reinforced composite can be improved by
a three-phase shear lag model. Jiang et al. (2009) growing CNTs on the surface of the carbon fibers.
determined the maximum volume fraction of CNTs in a Garcia et al. (2008) fabricated a hybrid laminate in
CNT-reinforced composite and investigated its effect which the reinforcements are a woven cloth of
on the effective elastic properties of the composite. alumina fibers with in situ grown CNTs on the
Esteva and Spanos (2009) studied the effect of weak- surface of the fibers. They demonstrated that both
ened interfaces between CNTs and the polymer matrix mechanical and electrical properties of such a lam-
on the effective properties of CNT-reinforced polymer inate are enhanced because of the CNTs grown on
matrix composite. Meguid et al. (2010) developed a the surface of the alumina fibers. Ray et al. (2009)
model of an atomistic-based representative volume carried out a load transfer analysis of short carbon-
element (RVE) which consists of the CNT, the fiber-reinforced composite in which the aligned
surrounding epoxy matrix, and the interface between CNTs are radially grown on the surface of the
CNT and epoxy to find out effective properties of CNT- carbon fibers. Most recently, Ray (2010) proposed a
reinforced epoxies. They homogenized the RVE into a novel hybrid smart composite which exhibits
representative fiber by equating the associated strain improved electro-mechanical properties because of
energies under identical loading conditions. The same the CNTs radially grown on the piezoelectric fibers.
homogenized RVE was then employed in a microme- The long fibers of unidirectional continuous fiber
chanical analysis to predict the effective properties of reinforced composites may be augmented with CNTs
the CNT-reinforced epoxy. Tsai et al. (2010) charac- which are radially grown on their surfaces and the
terized the elastic properties of CNT-reinforced poly- resulting long fiber may influence the effective proper-
mer nanocomposites considering an effective ties of the augmented composites. Such a fiber coupled
interphase between a CNT and the polymer matrix. with radially grown CNTs on its surface is also being
Rio et al. (2010) experimentally as well as theoretically named as ‘‘fuzzy fiber’’ (Mathur et al. 2008; Garcia et al.
investigated the effective properties of CNT-reinforced 2008). However, the polymer matrix composite being
polyester composites. Wernik and Meguid (2011) composed of such long fuzzy fiber reinforcement has not
presented a nonlinear atomistic-based continuum yet been studied. In this paper, a novel continuous
model for predicting the effective mechanical proper- unidirectional fuzzy fiber reinforced composite (FFRC)
ties of CNT-reinforced polymer composite. To exploit has been proposed. The fuzzy fiber reinforcement of the
the attractive elastic properties of CNTs, Ray and Batra FFRC is composed of a long carbon fiber while CNTs
(2009) proposed a hybrid piezoelectric composite are radially grown on the surface of the carbon fiber.
(HPZC) reinforced with CNTs and piezoelectric Analytical models based on the micromechanics para-
fibers. In this HPZC, the CNTs are vertically aligned digm have been derived for predicting the effective
and parallel to the vertical piezoelectric fibers. elastic properties of this proposed FFRC.
123
Micromechanical analysis of FFRC 151
123
152 S. I. Kundalwal, M. C. Ray
123
Micromechanical analysis of FFRC 153
8 nt 9 8 p 9 8 nc 9
imperfect bonding does not affect the effective > 1 > > 1 > > 1 >
>
> >
> >
> p >> >
> >
nc >
longitudinal Young’s modulus of the CNT-reinforced >
> nt >
2 > >
> 2 > > >
> 2 >
>
>
< nt = > < p = < nc >
> > > =
polymer matrix composite and marginally affects the r3 r3 r 3
vnt þ v p p ¼ ð4Þ
transverse properties of the composite for high > nt
> >
23 > > 23 >
> > > > nc
23 >
>
>
> nt >
> >
> p >> >
> >
>
volume fraction ([0.8) of CNTs. In their extensive >
> >
> >
> > > nc >
13
: nt ; : p ; : nc >
13 > > 13 ;
research, Tsai et al. (2010) investigated the effect of 12 12 12
the interphase between the CNT and the polymer In Eq. 4, vnt is the volume fraction of the CNT
matrix formed due to the non-bonded van der Waals with respect to the volume of the RVE of the PMNC
interaction for estimating the effective properties of (Fig. 4) and vp = 1 - vnt. Also, the superscript nc
CNT-reinforced polymer matrix composite. They represents the unwound PMNC material. Substituting
found that the interphase marginally enhances the Eqs. 1 and 2 into Eqs. 3 and 4, the stress and the
effective Young’s modulus of the composite trans- strain vectors in the unwound PMNC material can be
verse to the CNT fiber over that of the composite expressed in terms of the corresponding stress and
without the consideration of the interphase. They also strain vectors of the constituent phases as follows:
estimated the effective properties of the composite
considering perfect bonding between the CNT and frnc g ¼ ½C1 nt þ ½C2 fp g ð5Þ
the surrounding polymer matrix and found them to nc
nt p
f g ¼ ½V1 þ ½V2 f g ð6Þ
closely match with the prediction by the molecular
dynamics simulation. Hence, in the MOM approach Also, the relations among the stresses and strains
being presented here, it is assumed that CNTs and the in the CNT and the polymer phase given by (3) can
polymer matrix are perfectly bonded. In order to be written as
satisfy the perfectly bonding situation between the
½C3 nt ½C4 fp g ¼ 0 ð7Þ
fiber and the matrix, researchers mainly imposed the
iso-field conditions and used rules of mixture (Smith The various matrices appearing in Eqs. 5–7 are
and Auld 1991; Bevensite and Dvorak 1992). given by
According to the iso-field conditions one may assume 2 3
0 0 0 0 0 0
that the normal strains in the homogenized compos- 6 0
ite, the fiber and the matrix are equal along the fiber 6 0 0 0 0 07
7
6 nt 7
direction while transverse stresses in the same phases 6 C13 C23 Cnt
nt
0 0 07
½C1 ¼ vnt 6
6 0
33 7;
are equal along the direction transverse to the fiber 6 0 0 0 0 07
7
length. The rules of mixture allow one to express the 6 7
4 0 0 0 0 0 05
normal stress along the fiber direction and the
0 0 0 0 0 0
transverse strains along the normal to the fiber 2 3
direction of the homogenized composite in terms of Cp11 Cp12 Cp12 0 0 0
6 Cp Cp11 Cp12 0 0 0 7
that in the fiber and the matrix and their volume 6 12 7
6 7
fractions. Such iso-field conditions and rules of 6 vp Cp12 vp Cp12 vp Cp11 0 0 0 7
6
½C2 ¼ 6 7;
mixture (Smith and Auld 1991; Bevensite and
6 0 0 0 Cp44 0 0 7
7
Dvorak 1992) for satisfying the perfect bonding 6 7
4 0 0 0 0 Cp44 0 5
conditions between the CNT fiber and the polymer
matrix can be expressed as 0 0 0 0 0 Cp44
2 3
8 nt 9 8 p 9 8 nc 9 Cnt
11 Cnt
12 Cnt
13 0 0 0
>
> r1 > > >
> r1 >> >
> r1 > > 6 nt 7
> rnt >
> > > > rp > > >
> rnc >
> 6 C12 Cnt Cnt 0 0 0 7
>
> 2 >> >
> 2 >> >
> 2 > 6 22 23 7
< nt = < p = < nc > = 6 0
3 3 3 6 0 1 0 0 0 77
nt ¼ p ¼ nc ð3Þ ½C3 ¼ 6 7;
>
> r > > > r23 > > r > 6 0 Cnt 0 7
> 23
> > >
nt > > p >
> >
>
>
> 23
>
>
>
nc > 6 0 0 44 0 7
> r > > r13 > > r > 6 0
>
: 13 >
; > : p > ; >
: 13 >
; 4 0 0 0 Cnt 0 75
rnt
12 r 12 rnc
12
55
0 0 0 0 0 Cnt
66
and
123
154 S. I. Kundalwal, M. C. Ray
2 3
Cp11 Cp12 Cp12 0 0 0 PMNC with respect to the 1–2–3 coordinate system
6 Cp Cp11 Cp12 0 0 0 7 can be obtained by the following transformations:
6 12 7
6 0 PMNC
6 0 1 0 0 0 77
C ¼ ½TT ½Cnc ½T1 ð12Þ
½C4 ¼ 6 7;
6 0 0 0 Cp44 0 0 7
6 7 where,
4 0 0 0 0 Cp44 0 5 2 3
0 0 0 0 0 Cp44 1 0 0 0 0 0
2 3 60 m 2
n2
mn 0 0 7
vnt 0 0 0 0 0 6 7
60 n2 m2 mn 0 0 7
6 0 vnt 0 0 0 0 7 ½T ¼ 6
6 0 2mn 2mn m2 n2
7
6 7 6 0 0 7
7
6 0 0 7 40
6
½V1 ¼ 6
0 0 0 0 7 0 0 0 m n 5
7 and
6 0 0 0 vnt 0 0 7 0 0 0 0 n m
6 7
4 0 0 0 0 vnt 0 5 with m ¼ cos h and n ¼ sin h:
0 0 0 0 0 vnt
2 3 Therefore, the effective elastic properties of the
vp 0 0 0 0 0
PMNC surrounding the carbon fiber with respect to the
60 vp 0 0 0 07
6 7 principle material coordinate axes of the proposed FFRC
60 0 1 0 0 07
6 7 varies over an annular cross-section of the PMNC phase
½V2 ¼ 6 7 ð8Þ
60 0 0 vp 0 07 of the RVE of the CFF. However, without loss of
6 7
40 0 0 0 vp 0 5 generality, it may be considered that the volume average
PMNC
0 0 0 0 0 vp of these effective elastic properties C over the
volume of the PMNC can be treated as the constant
Using Eqs. 6 and 7, the local strain vectors fnt g
and fp g can be expressed in terms of the composite effective elastic properties CPMNC of the PMNC
strain fnc g and subsequently, using them in Eq. 5, material surrounding the carbon fiber with respect to
the following constitutive relation between the states the 1–2–3 coordinate axes of the FFRC and is given by
of stresses and the states of strains at any point in the Z2p ZR
PMNC 1 PMNC
unwound PMNC material is obtained: C ¼ 2
C r dr dh ð13Þ
p R a 2
0 a
frnc g ¼ ½Cnc fnc g ð9Þ
Thus the effective constitutive relations for the
where the effective elastic coefficient matrix ½Cnc of PMNC material with respect to the principle material
the lamina of the unwound PMNC is given by coordinate axes of the FFRC can be expressed as
PMNC PMNC PMNC
½Cnc ¼ ½C1 ½V3 1 þ½C2 ½V4 1 ð10Þ r ¼ C ð14Þ
and 2.1.2 Effective elastic properties of the CFF
1
½V3 ¼ ½V1 þ ½V2 ½C4 ½C3 ; ½V4
¼ ½V2 þ ½V1 ½C3 1 ½C4 ð11Þ The effective elastic constants of the CFF shown in
Fig. 3 may be predicted by estimating the effective
It may be noted that the matrix ½Cnc directly elastic properties of a lamina of continuous unidirec-
provides the effective elastic properties at a point in the tional fiber-reinforced composite in which the carbon
portion of the PMNC material surrounding the carbon fiber is the reinforcement and the matrix phase is the
fiber where the CNT is aligned with the 3-axis of the PMNC material whose effective elastic properties
FFRC. But it is to be noted that with respect to a local are given by Eq. 13. The cross sections of the RVE of
material coordinate system (10 , 20 , 30 ) as shown in such lamina have been illustrated in Fig. 5a. Here, the
Fig. 3, the matrix ½Cnc also provides the effective length of the carbon fiber aligns with the 1-direction.
elastic properties at a point located in the PMNC where The MOM approach derived in the previous section
the CNT axis (30 -axis) is oriented at an angle h with the is augmented to estimate the effective elastic prop-
3-axis in the 2–3 plane. Thus at any point in the PMNC erties of the CFF. Similar to Eqs. 3 and 4, the iso-
surrounding the carbon fiber, the location dependent field conditions and the rules of mixture for the RVE
PMNC
effective elastic coefficient matrix C of the shown in Fig. 5a can be written as
123
Micromechanical analysis of FFRC 155
8 f 9 8 PMNC 9 8 CFF 9
> 1 > 1 1 > In Eq. 16, vf and vP are the volume fraction of the
>
> > >
> >
>
>
>
>
>
>
> >
>
>
> r f >
> >
> rPMNC >
> >
> rCFF >
> carbon fiber and the PMNC material, respectively,
>
> >
2 > >
> 2 >
> >
> 2 >
>
< rf >
> = > < rPMNC >
= < rCFF >
> = with respect to the volume of the RVE of the CFF.
3 3 3
¼ ¼ ð15Þ Using Eqs. 14–16, and following the procedure for
>
> r23 >
f
> >
>
PMNC
r23 >
> >
> r23 >
CFF
>
>
> >
> >
> >
> >
> > deriving Eq. 10, the constitutive relations of the CFF
> f >
> >
> r13 > >
> r13
PMNC >
> >
> CFF >
>
>
>
> > > >
> >
> r13 > can be obtained as follows:
: f ; : PMNC ; : CFF >;
r12 r12 r12 CFF CFF CFF
r ¼ C ð17Þ
and
8 f 9 8 PMNC 9 8 CFF 9 in which the effective elastic coefficient matrix
r1 > r1 CFF
>
> > >
> >
> > r1 >
> f >
> > >
> > >
PMNC >
>
>
>
>
CFF >
C of the CFF is given by
>
> >
> >
> >
> >
> >
> CFF
>
> 2 >
> >
> 2 > > 2 >
< f = < PMNC = < CFF >
> > = C ¼ ½C5 ½V7 1 þ½C6 ½V8 1 ð18Þ
3 3 3
vf þ v PMNC ¼ ð16Þ
>
> f23 >
> >
> PMNC >
> >
> CFF
23 >
> The various matrices appearing in Eq. 18 are
>
> >
> >
> 23 >
> >
> >
>
> f >> >
> PMNC >
> >
> CFF >
>
>
>
> ; >
> >
> > > >
: PMNC > ; > : CFF >
13 13 13
: f ;
12 12 12
2 3
vPMNC CPMNC
11 vPMNC CPMNC
12 vPMNC CPMNC
13 0 0 0
6 CPMNC CPMNC CPMNC 0 0 0 7
6 12 22 23 7
6 7
6 CPMNC CPMNC CPMNC 0 0 0 7
½C 6 ¼ 6
6
13 23 33 7;
7
6 0 0 0 CPMNC
44 0 0 7
6 7
4 0 0 0 0 CPMNC
55 0 5
0 0 0 0 0 CPMNC
66
2 3 2 3
Cf11 Cf12 Cf13 0 0 0 1 0 0 0 0 0
6 7 6 f
Cf22 Cf23 0 0 0 7
6 0 0 0 0 0 07 6 C12 7
6 7 6 f 7
6 0 0 0 0 0 07 6C Cf23 Cf33 0 0 0 7
½ C 5 ¼ vf 6
6
7; ½C7 ¼ 6 13
7 6 0
7;
6 0 0 0 0 0 07 6 0 0 Cf44 0 0 7
7
6 7 6 7
4 0 0 0 0 0 05 4 0 0 0 0 Cf55 0 5
0 0 0 0 0 0 0 0 0 0 0 Cf66
2 3 2 3
1 0 0 0 0 0 0 0 0 0 0 0 ð19Þ
6 CPMNC CPMNC CPMNC 0 0 0 7 60 vf 0 0 0 07
6 12 22 23 7 6 7
6 PMNC 7 6 7
6 C13 CPMNC CPMNC 0 0 0 7 60 0 vf 0 0 07
½C 8 ¼ 6
6 0
23 33 7; ½V5 ¼ 6
7 60
7;
6 0 0 CPMNC
44 0 0 7 6 0 0 vf 0 077
6 7 6 7
4 0 0 0 0 CPMNC
55 0 5 40 0 0 0 vf 05
PMNC
0 0 0 0 0 C66 0 0 0 0 0 vf
2 3
1 0 0 0 0 0
60 vPMNC 0 0 0 0 7
6 7
6 7
60 0 vPMNC 0 0 0 7
½V 6 ¼ 6
60
7;
6 0 0 vPMNC 0 0 7
7
6 7
40 0 0 0 vPMNC 0 5
0 0 0 0 0 vPMNC
1
½V7 ¼ ½V5 þ ½V6 ½C8 ½C7 and ½V8 ¼ ½V6 þ ½V5 ½C7 1 ½C8
123
156 S. I. Kundalwal, M. C. Ray
123
Micromechanical analysis of FFRC 157
2 3
CCFF
11 CCFF
12 CCFF
13 0 0 0 effective properties of the interphase, the CNT
6 0 0 0 0 0 07 properties, the polymer matrix properties and the
6 7
6 7 procedure of the MT method for multiple inclusions
6 0 0 0 0 0 07
½C9 ¼ vCFF 6
6
7; (Dunn and Ledbetter 1995), a three-phase MT model
6 0 0 0 0 0 077 can be derived to estimate the effective elastic
6 7
4 0 0 0 0 0 05 coefficient matrix of the unwound PMNC. The
0 0 0 0 0 0 explicit formulation of such three-phase MT model
2 3 can be derived as
vP Cp11 vP Cp12 vP Cp12 0 0 0
h
6 Cp Cp11 Cp12 0 0 0 7
6 12 7 ½Cnc ¼ ½Cp þ ðvnt þ vi Þ Ci ½Cp ½AV
6 p 7 i
6 C12 Cp12 Cp11 0 0 0 7
½C10 ¼ 6
6 0
7: ð24Þ þ vnt Cnt Ci ½Ant
6 0 0 Cp44 0 0 77
6 7 1
4 0 0 0 0 Cp44 0 5 vp ½I þ ðvnt þ vi Þ½AV ð25Þ
0 0 0 0 0 Cp44 where ½Cp and vi represent the linear elastic coef-
ficient matrix and the volume fraction of the inter-
phase, respectively, while V denotes the domain
2.2 Mori–Tanaka (MT) method comprising a CNT and the interphase surrounding
this CNT. The concentration tensors ½AV and ½Ant
The previous micromechanics model is based on the appearing in Eq. 25 are given by
assumptions and the rules of mixture delineated by ½AV ¼ ½I þ ½SV ½UV and ½Ant
Eqs. 3 and 4, respectively, which imply the perfect ¼ ½I þ ½DS½Ui þ ½Snt ½Unt ð26Þ
bonding condition between the CNT fiber and the
matrix. However, it is imperative to justify the The various matrices appearing in (26) are
"
validity of these assumption and the rules of mixture
for modeling the perfect bonding conditions of the ½Unt ¼ ð½Snt þ ½C1 Þ þ ½DS
MOM approach. For this purpose, another microm-
1
echanics model based on the Mori–Tanaka (MT) vnt 2
½Snt ½DS þ ½C
method which does not require to satisfy Eqs. 3 and 4 vi
will be presented here. Although it is mentioned in #1
the previous section that the weakened interface or vnt 1
½Snt ½DS þ ½C ;
the interphase between the CNT fiber and the vi
"
polymer matrix does not affect the effective elastic
properties to a considerable extent (Esteva and ½Ui ¼ ½DS þ ð½Snt þ ½C1 Þ
Spanos 2009; Tsai et al. 2010), we intend to
1
investigate the effect of the interphase between the vnt 1
½Snt ½DS þ ½C
CNT and the polymer on the effective properties of vi
this FFRC employing the MT method. The non #1
bonded van der Waals interaction between an atom of vnt 2
½Snt ½DS þ ½C ;
CNT and an atom of the polymer is characterized vi
here by introducing an interphase (Odegard et al. vnt vi
½Uv ¼ ½Unt þ ½Ui ;
2003; Tsai et al. 2010). The effective properties of vnt þ vi vnt þ vi
1 nt
such interphase resembling a solid continuum can be ½DS ¼ ½Snt ½SV ; C ¼ ð C ½Cp Þ1 ½Cp
determined by molecular dynamics simulation and
and C2 ¼ ð Ci ½Cp Þ1 ½Cp : ð27Þ
are readily available in the open literature (Tsai et al.
2010). Thus considering this interphase between a Also, in the above matrices, ½SV and ½Snt indicate
CNT and the polymer matrix, the micromechanical the Eshelby’s tensor for the domains V and nt,
model of the unwound PMNC by the MT method respectively, and ½I is an identity matrix. As assumed
will be a three-phase MT model. Utilizing the in the previous model, the CNT may be treated as a
123
158 S. I. Kundalwal, M. C. Ray
solid circular cylindrical fiber (Gao and Li 2005; Tsai consequently, the Eshelby tensor (Jiang and Martin
et al. 2010). Thus the specific form of the Eshelby 1998) corresponding to transversely isotropic mate-
tensor for cylindrical inclusion given by Qui and rial is utilized for computing ½S1 while the inclusion
Weng (1990) is utilized to compute the matrices is a circular cylindrical fiber.
½SV [and ½Snt . Once ½Cnc is computed from Eq. 25, Finally, considering the CFF as the cylindrical
Eqs. 12 and 13 are used to estimate the average inclusion embedded in the isotropic polymer matrix
effective elastic coefficient matrix CPMNC of the the effective elastic properties ½C of the FFRC can
PMNC material surrounding the carbon fiber. Note be derived by the two-phase MT method (Mori and
that if the perfectly bonding condition between a Tanaka 1973) as follows:
CNT and the polymer is assumed the conventional
½C ¼ ½Cp þ vCFF CCFF ½Cp ½A2 ð29Þ
two-phase MT method (Mori and Tanaka 1973) is
used to predict the effective coefficient matrix ½Cnc in which the matrix of the strain concentration factors
of the unwound PMNC. are given by
The effective elastic properties of the CFF can be 1
½A 2 ¼ A~ 2 vP ½I þ vCFF A
~2 and
predicted by estimating the effective elastic properties
h i1
of a composite in which the carbon fiber is the ~ 2 ¼ ½I þ ½S2 ð½Cp Þ1 CCFF ½Cp
A
reinforcement and the matrix phase is the PMNC
material. Thus according to the two-phase model by the
MT method (Mori and Tanaka 1973), the effective 3 Results and Discussion
elastic coefficient matrix for the CFF is given by
CFF PMNC In this section, numerical values of the effective
C ¼ C þvf Cf CPMNC ½A1
elastic properties of the proposed FFRC are evalu-
ð28Þ ated using the two different models derived in the
in which the matrix of the strain concentration factors preceding sections. Zig-zag single-walled CNTs, the
are as follows carbon fiber and the polyimide polymer matrix are
used for evaluating the numerical results. Their
1
½A 1 ¼ A~ 1 vPMNC ½Iþ ~1
vf A ; material properties are listed in Table 1. The effec-
h PMNC 1 f PMNC i1 tive properties and the thickness of the hollow
~ 1 ¼ ½Iþ ½S1 C
A C C circular cylindrical continuum representing the inter-
phase between a CNT and the polyimide matrix are
where, the Eshelby’s tensor ½S1 is computed based also presented in Table 1. Unless otherwise men-
on the properties of the PMNC matrix and the shape tioned, (10, 0) CNT is used for plotting the results
of the carbon fiber. It is worthwhile to note that and the value of the diameter of the carbon fiber is
the PMNC matrix is transversely isotropic and assumed as 2a = 10 lm. Volume fraction of CNTs
(10, 0) CNT (Tsai et al. 2010) 709.9 172.4 240 240 1513.1 1120 268.7 dn = 0.78
Interphase 29.6 15.2 15.2 15.2 29.6 7.2 7.2 Gap = 0.3333
(14, 0) CNT (Tsai et al. 2010) 557.5 137.5 187.7 187.7 1082.8 779.2 210 dn = 1.1
Interphase 27.63 14.23 14.23 14.23 27.63 6.70 6.70 Gap = 0.3236
(18, 0) CNT (Tsai et al. 2010) 472.9 118.7 159.7 159.7 846.1 596.3 177.1 dn = 1.42
Interphase 27.55 14.19 14.19 14.19 27.55 6.68 6.68 Gap = 0.3158
Carbon (Honjo 2007) 236.4 10.6 10.6 10.7 24.8 7 25 d = 10000
fiber
Polyimide (Odegard et al. 9 6 6 6 9 1.5 1.5 –
2005)
123
Micromechanical analysis of FFRC 159
(vCNT) in the FFRC depends on the CNT diameter, the carbon fiber volume fraction (vf) in the FFRC,
the carbon fiber diameter and the surface to surface the maximum value of vCNT can be determined as
distance of two adjacent radially aligned CNTs at rffiffiffiffiffiffiffiffiffi
pd2n pvf
their roots. If there are no other phases or materials in ðvCNT Þmax ¼ pffiffiffi vf ð30Þ
between CNTs, it is reported (Jiang et al. 2009) that 2ðdn þ 1:7Þ2 2 3
the minimum surface to surface distance between two where dn is the diameter of the CNT. The derivation
adjacent CNTs is the equilibrium van der Waals of Eq. 30 has been presented in the section of
distance, which is about 0.34 nm. Since in the ‘‘Appendix’’. It is evident from Eq. 30 that when
PMNC material, polyimide polymer molecules fill pffiffiffi
vf = 0, vCNT is zero. Also, when vf = p=2 3 i.e.,
the gap between CNTs and the formation of the
there is no PMNC, the value of vCNT is zero. Thus
interphase is also considered, the surface to surface
the maximum value of vCNT given by Eq. 30 will be
distance between the two adjacent CNTs at their
maximum at a particular value of the vf. Figure 7
roots is considered as 1.7 nm. Recall that the FFRC
illustrates the variation of the maximum value of the
lamina can be viewed as being comprised of the
CNT volume fraction in the FFRC (vCNT) with the
CFFs and the polyimide matrix. For fibers with
carbon fiber volume fraction (vf) while the values of
circular cross section, it is well known that hexagonal pffiffiffi
vf varies from 0.1 to p=2 3. It may be observed from
array of packing is the optimal packing of fibers and
this figure that the maximum values of vCNT for
the corresponding maximum fiber volume fraction is
different CNT diameters are almost independent of
0.9069. Hence, in the proposed FFRC, the hexagonal
the CNT diameter and are maximized at vf = 0.24.
packing array of CFFs is considered as shown in
Also, it may be noted that if vf increases then either
Fig. 6 for evaluating the numerical results. It is also
the diameter or the number of the carbon fiber
assumed that the radially grown CNTs are uniformly
increases. This results in the reduction of the length
spaced on the surface of the carbon fiber. Note that
of the CNTs which in turn causes decrease in the
although the volume fraction of CFF is 0.9069, the
volume of the CNTs in the CFF and the FFRC
volume fraction of carbon fiber is much less than
beyond a value of vf as 0.24. In what follows, unless
0.9069.
otherwise mentioned the effective properties of the
The determination of vCNT in the FFRC is an
FFRC are computed for a particular value of vf while
important issue. It is obvious that the constructional
the maximum value of vCNT corresponding to this
feature of the FFRC put a constraint on the
value of vf is considered for the computation. First,
maximum value of vCNT. Based on the surface to
the effective elastic properties of the PMNC sur-
surface distance at the roots of two adjacent CNTs
rounding the carbon fiber computed by the models
and the CNT diameter, the maximum number of
CNTs grown on the surface of a carbon fiber of
particular diameter can be determined. Then based on 0.09
CNT (10, 0), (vCNT)max
0.08
CNT (14, 0), (vCNT)max
0.07 CNT (18, 0), (vCNT)max
0.06
(vCNT)max
0.05
0.04
0.03
0.02
0.01
0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Carbon Fiber Volume Fraction (vf)
123
160 S. I. Kundalwal, M. C. Ray
55 14
50
12
45
40 10
(GPa)
(GPa)
35
8
30
PMNC
PMNC
25
6
C11
C33
20
15 4
MOM Approach, (vCNT)max MOM Approach, (vCNT)max
10 Two-phase MT Method, (vCNT)max Two-phase MT Method, (vCNT)max
2
5 Three-phase MT Method, (vCNT)max Three-phase MT Method, (vCNT)max
0 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Carbon Fiber Volume Fraction (vf) Carbon Fiber Volume Fraction (vf)
55 8
50
7
45
40 6
(GPa)
(GPa)
35 5
30
PMNC
4
PMNC
25
C22
C12
20 3
15
MOM Approach, (vCNT)max 2 MOM Approach, (vCNT)max
10 Two-phase MT Method, (vCNT)max Two-phase MT Method, (vCNT)max
5 1
Three-phase MT Method, (vCNT)max Three-phase MT Method, (vCNT)max
0 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Carbon Fiber Volume Fraction (vf) Carbon Fiber Volume Fraction (vf)
derived in the previous sections are illustrated in overestimates the value of CPMNC 33 . It may be
Figs. 8, 9, 10, 11, 12, and 13. It should be noted that mentioned that the previous researchers (Esteva and
the models by the MOM approach and the two-phase Spanos 2009; Tsai et al. 2010) also reported the
MT method estimate the properties with the assump- similar trend of results for estimating the effective
tion of perfect bonding condition between a CNT and longitudinal properties of CNT-reinforced composite.
the matrix (i.e., without the interphase). On the other The constructional feature of the PMNC reveals that
hand, the model by the three-phase MT method the PMNC would be a transversely isotropic mate-
predicts the properties in the presence of the inter- rial. This is corroborated by the prediction of CPMNC
22
phase. Figure 8 illustrates the variation of the effec- which is identical to that of CPMNC33 as shown in
tive elastic coefficient CPMNC
33 of the PMNC with the Fig. 9. Figure 10 illustrates the variation of the
carbon fiber volume fraction (vf). It can be observed effective transverse elastic coefficient CPMNC
11 of the
that the models by the MOM approach and the two- PMNC with the carbon fiber volume fraction (vf). It
phase MT method, predict identical estimates for may be observed that the three-phase MT method
CPMNC
33 while the three-phase MT method marginally provides slightly enhanced estimate for the value of
123
Micromechanical analysis of FFRC 161
25 275
MOM Approach, (vCNT)max
Two-Phase MT Method, (vCNT)max
20 225
Three-Phase MT Method, (vCNT)max
MOM Approach, (vCNT = 0)
(GPa)
C11 (GPa)
15
150
PMNC
C23
10
75
MOM Approach, (vCNT)max
5
Two-phase MT Method, (vCNT)max
Three-phase MT Method, (vCNT)max
0 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Carbon Fiber Volume Fraction (vf) Carbon Fiber Volume Fraction (vf)
Fig. 12 Variation of the effective elastic coefficient CPMNC of Fig. 14 Variation of the effective elastic coefficient C11 of the
23
the PMNC with the carbon fiber volume fraction FFRC with the carbon fiber volume fraction
123
162 S. I. Kundalwal, M. C. Ray
30 15
25
20 10
C22 (GPa)
C23 (GPa)
15
Fig. 15 Variation of the effective elastic coefficient C22 of the Fig. 16 Variation of the effective elastic coefficient C23 of the
FFRC with the carbon fiber volume fraction FFRC with the carbon fiber volume fraction
6
observed from Fig. 15 that even if the interphase is
not considered, the effective value of C22 of the
proposed FFRC is significantly larger than that of the 4
MOM Approach, (vCNT)max
composite without CNTs. This is attributed to the
Two-Phase MT Method, (vCNT)max
fact that the radially grown CNTs enhance the out-of- 2
Three-Phase MT Method, (vCNT)max
plane stiffness of the matrix surrounding the carbon
MOM Approach, (vCNT = 0)
fibers while the transverse properties are matrix
0
dominant properties. Since the FFRC is transversely 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Carbon Fiber Volume Fraction (vf)
isotropic with 1-axis as the axis of symmetry, the
values of the effective elastic coefficients C33 of the
Fig. 17 Variation of the effective elastic coefficient C12 of the
FFRC are found to be identical to those of C22 and FFRC with the carbon fiber volume fraction
for brevity are not presented here. Radially grown
CNTs also significantly improve the values of the the model by the MOM approach with that by the MT
effective elastic coefficient C23 of the FFRC as shown methods reveals that the MOM approach yields
in Fig. 16. But the interphase does not influence this conservative estimates of the elastic constants of the
constant. It may be observed from Fig. 17 that the proposed FFRC. Hence, predictions by the MOM
growing of CNTs on the surface of the carbon fiber approach have been considered in the subsequent
moderately improves the value of the elastic coefficient results for investigating the effect of the CNT diameter
C12 of the FFRC. Since the FFRC is transversely and the maximum CNT volume fraction on the
isotropic, similar results are also obtained for the effective properties of the FFRC.
effective elastic coefficient C13. Figure 18 depicts the Figures 19, 20, 21 illustrate the comparisons of the
variation of the effective elastic coefficient C55 of the effective elastic constants of the proposed FFRC for
proposed FFRC with the carbon fiber volume fraction. different values of the CNT diameter. It may be
It may be observed that both the MT methods mentioned that as the integer for designating the zig-
overestimates this constant as compared to the MOM zag type CNT increases, the diameter of the CNT
approach for vCNT 6¼ 0. Identical results are also increases. Figure 19 illustrates that the CNT diameter
predicted for C66 but not shown here. Comparison of does not influence the effective elastic coefficient C11
123
Micromechanical analysis of FFRC 163
18 30
MOM Approach, (vCNT)max
16
Two-Phase MT Method, (vCNT)max 25
14 Three-Phase MT Method, (vCNT)max
C22 (GPa)
10
15
8
6 10
Fig. 18 Variation of the effective elastic coefficient C55 of the Fig. 20 Variation of the effective elastic coefficient C22 of the
FFRC with the carbon fiber volume fraction FFRC with the carbon fiber volume fraction
225 15
200
175
150 10
C11 (GPa)
C23 (GPa)
125
100
75 5
50 CNT (10, 0), (vCNT)max CNT (10, 0), (vCNT)max
CNT (14, 0), (vCNT)max CNT (14, 0), (vCNT)max
25
CNT (18, 0), (vCNT)max CNT (18, 0), (vCNT)max
0 0
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Carbon Fiber Volume Fraction (vf) Carbon Fiber Volume Fraction (vf)
Fig. 19 Variation of the effective elastic coefficient C11 of the Fig. 21 Variation of the effective elastic coefficient C23 of the
FFRC with the carbon fiber volume fraction FFRC with the carbon fiber volume fraction
of the proposed FFRC. Figure 20 reveals that the of the CNT volume fraction for a particular value of
increase in the CNT diameter marginally increases vf. However, the variation of the CNT volume
the value of the effective elastic coefficient C22 of the fraction for a particular value of vf would be an
FFRC for low carbon fiber volume fraction. The important study. For this the discrete values of the
same is true for the effective elastic coefficient C23 as carbon fiber volume fraction are considered as 0.2,
depicted in Fig. 21. This may be attributed to the fact 0.3, 0.4 and 0.5. The maximum values of vCNT
that as the diameter of the CNT increases, the value corresponding to these values of vf are 0.0735,
of the CNT volume fraction increases as shown in 0.0721, 0.0658 and 0.0564, respectively. However,
Fig. 7 which results in the increase in the values of for presenting the results, the limiting value of vCNT
the effective elastic constants of the PMNC. Similar has been considered as 0.0564. Once again, Fig. 22
predictions for the other effective elastic coefficients illustrates that the effective value of C11 of the
have also been obtained, but for the sake of brevity proposed FFRC is independent of the variation of the
they are not presented here. So far, the effect of radial CNT volume fraction. Figure 23 depicts the variation
growing of CNTs on the surface of the carbon fiber of the effective elastic coefficient C22 of the FFRC
has been studied by considering the maximum value with the CNT volume fraction. It may be observed
123
164 S. I. Kundalwal, M. C. Ray
140 10
120
8
100
C11 (GPa)
C12 (GPa)
80 6
60
4
vf = 0.2 vf = 0.2
40
vf = 0.3 vf = 0.3
vf = 0.4 2
20 vf = 0.4
vf = 0.5 vf = 0.5
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0 0.01 0.02 0.03 0.04 0.05 0.06
CNT Volume Fraction (vCNT) CNT Volume Fraction (vCNT)
Fig. 22 Variation of the effective elastic coefficient C11 of the Fig. 24 Variation of the effective elastic coefficient C12 of the
FFRC with the CNT volume fraction ((18, 0) CNT) FFRC with the CNT volume fraction ((18, 0) CNT)
30 3.5
25 3
2.5
20
C55 (GPa)
C22 (GPa)
2
15
1.5
10 vf = 0.2 vf = 0.2
1
vf = 0.3 vf = 0.3
5 vf = 0.4 0.5 vf = 0.4
vf = 0.5 vf = 0.5
0 0
0 0.01 0.02 0.03 0.04 0.05 0.06 0 0.01 0.02 0.03 0.04 0.05 0.06
CNT Volume Fraction (vCNT) CNT Volume Fraction (vCNT)
Fig. 23 Variation of the effective elastic coefficient C22 of the Fig. 25 Variation of the effective elastic coefficient C55 of the
FFRC with the CNT volume fraction ((18, 0) CNT) FFRC with the CNT volume fraction ((18, 0) CNT)
that the effective value of C22 is significantly and CNTs are radially grown on the surfaces of these
sensitive to the change in CNT volume fraction. carbon fibers. Two analytical models based on the
Although not shown here, similar trend of the results micromechanics such as the mechanics of materials
are also predicted for the effective elastic constant (MOM) approach and the Mori–Tanaka (MT)
C23. Figure 24 displays that the effective value of C12 method are derived to predict the effective elastic
increases marginally with the CNT volume fraction properties of a lamina made of this proposed FFRC.
and the same is also true for the effective elastic The effective properties of the FFRC are estimated
constant C55 as shown in Fig. 25. with and without the consideration of an interphase
between a CNT and the polymer matrix. The
interphase is a continuum representation of the non
4 Conclusions bonded van der Waals interaction between the atoms
of CNTs and the atoms of the polymer matrix. In the
A novel fuzzy fiber reinforced composite (FFRC) is absence of the interphase, it is assumed that the CNT
proposed which is composed of single-walled zig-zag and the polymer matrix are perfectly bonded and the
CNTs, carbon fibers and polyimide matrix. The models are derived by the MOM approach and the
carbon fiber reinforcements are horizontally aligned two-phase MT method. The three-phase MT method
123
Micromechanical analysis of FFRC 165
is employed to derive the model taking into account Referring to Fig. 6, the RVE of the FFRC can be
the effect of the interphase. Because of the construc- considered as an equilateral triangle. Thus the volume
tional feature of the proposed FFRC, there exists a (VFFRC ) of the RVE of the FFRC is given by
maximum value of the CNT volume fraction for a pffiffiffi
particular value of the carbon fiber volume fraction. FFRC 3 2
V ¼ D L ð31Þ
Such maximum volume fraction of CNT increases 4
with the increase in the diameter of the CNT. The where D = 2R. The volume (Vf) of the carbon fiber
present study reveals that the CNTs being radially is
grown on the surface of the carbon fiber with its axis p
normal to this surface, significantly improve the Vf ¼ d2 L ð32Þ
8
values of the transverse effective elastic constants
C22, C33, C23, C12, C13 and C55 of the proposed where d = 2a. Thus the carbon fiber volume fraction
FFRC over their values without CNTs. This is (vf) in the FFRC can be expressed as
attributed to the fact that the elastic properties of Vf p d2
CNT along its axis are exceptionally larger than vf ¼ FFRC
¼ pffiffiffi 2 ð33Þ
V 2 3D
those transverse to its axis and CNTs are transverse
to the carbon fiber. The in-plane effective elastic Using A (3), the carbon fiber volume fraction (vf )
coefficient C11 is not affected by the radial growing in the CFF can be derived as
of CNTs on the surface of carbon fibers. Although p 2
pffiffiffi
8d L 2 3
the consideration of the interphase between the CNT vf ¼ p 2 ¼ vf ð34Þ
8D L
p
and the polymer matrix was a concern, it does not
affect the in-plane effective elastic coefficient C11 The maximum number ðNCNT Þmax of radially
and its effects on the other effective elastic coeffi- grown aligned CNTs on the surface of the carbon
cients are also not pronounced. The predicted values fiber is given by
of these effective elastic coefficients considering the
pdL
interphase is slightly greater than those when per- ðNCNT Þmax ¼ ð35Þ
fectly bonding condition between the CNT and the 2ðdn þ 1:7Þ2
polymer matrix is considered. Thus for conservative Therefore the volume (VCNT) of the CNTs is
and intuitive estimates, one may neglect the non p
bonded van der Waals interaction between an atom of VCNT ¼ d2n ðR aÞðNCNT Þmax ð36Þ
4
CNT and the polymer matrix and consider the perfect
bond between a CNT and the polymer matrix. The Thus the maximum volume fraction ðvCNT Þmax of
effective values of the transverse elastic coefficients the CNT with respect to the volume of the FFRC can
of the FFRC marginally increase with the increase in be determined as
the CNT diameter as long as the carbon fiber volume rffiffiffiffiffiffiffiffiffi
VCNT pd2n pvf
fraction remains low. Since the transverse properties ðvCNT Þmax ¼ FFRC ¼ p ffiffiffi v f
V 2ðdn þ 1:7Þ2 2 3
of the proposed FFRC are significantly improved
without the cost of the values of the in-plane effective ð37Þ
elastic constants C11, the proposed FFRC will have Finally, the maximum volume fraction (max (vnt))
better strength against the delamination failure. The of the CNTs with respect to the volume of the
present investigation will be useful for the experi- PMNC and with respect to the volume of the CFF
mental verification of the proposed FFRC. (max (vnt )) can be determined in terms of max (vCNT)
as follows:
pffiffiffi 2
VCNT 2 3 D
Appendix ðvnt Þmax ¼ PMNC ¼ ðvCNT Þmax ð38Þ
V p D 2 d2
pffiffiffi
Equation 30 as shown in the results and discussion VCNT 2 3
ðvnt Þmax ¼ CFF ¼ ðvCNT Þmax ð39Þ
section can be derived as follows: V p
123
166 S. I. Kundalwal, M. C. Ray
References Mori, T., Tanaka, K.: Average stress in matrix and average
elastic energy of materials with misfitting inclusions. Acta
Batra, R.C., Sears, A.: Uniform radial expansion/contraction of Metall. 21, 571–574 (1973)
carbon nanotubes and their transverse elastic moduli. Odegard, G.M., Gates, T.S., Wise, K.E., Park, C., Siochi, E.J.:
Model. Simul. Mater. Sci. Eng. 15, 835–844 (2007) Constitutive modeling of nanotube-reinforced polymer
Bevensite, Y., Dvorak, G.J.: Uniform fields and universal composites. Compos. Sci. Technol. 63, 1671–1687 (2003)
relations in piezoelectric composites. J. Mech. Phys. Odegard, G.M., Clancy, T.C., Gates, T.S.: Modeling of the
Solids 40, 1295–1312 (1992) mechanical properties of nanoparticle/polymer compos-
Cheng, H.C., Liu, Y.L., Hsu, Y.C., Chen, W.H.: Atomistic-con- ites. Polymer 46, 553–562 (2005)
tinuum modeling for mechanical properties of single-walled Qui, Y.P., Weng, G.J.: On the application of Mori-Tanaka’s
carbon nanotubes. Int. J. Solids Struct. 46, 1695–1704 (2009) theory involving transversely isotropic spheroidal inclu-
Dunn, M.L., Ledbetter, H.: Elastic moduli of composites sions. Int. J. Eng. Sci. 28(11), 1121–1137 (1990)
reinforced by multiphase particles. ASME J. Appl. Mech. Ray, M.C.: Concept for a novel hybrid smart composite rein-
62, 1023–1028 (1995) forced with radially aligned zigzag carbon nanotubes on
Esteva, M., Spanos, P.D.: Effective elastic properties of piezoelectric fibers. Smart Mater. Struct. 19, 035008
nanotube reinforced composites with slightly weakened (2010)
interfaces. J. Mech. Mater. Struct. 4, 887–900 (2009) Ray, M.C., Batra, R.C.: Effective properties of carbon nano-
Gao, X.L., Li, K.: A shear-lag model for carbon nanotube tube and piezoelectric fiber reinforced hybrid smart
reinforced polymer composites. Int. J. Solids Struct. 42, composite. ASME J. Appl. Mech. 76, 034–503 (2009)
1649–1667 (2005) Ray, M.C., Guzman de Villoria, R., Wardle, B.L.: Load
Garcia, E.J., Wardle, B.L., Hart, A.J., Yamamoto, N.: Fabri- transfer analysis in short carbon fibers with radially-
cation and multifunctional properties of a hybrid laminate aligned carbon nanotubes embedded in a polymer matrix.
with aligned carbon nanotubes grown in situ. Compos. J. Adv. Mater. 41(4), 82–94 (2009)
Sci. Technol. 68, 2034–2041 (2008) Rio, T.G., Poza, P., Rodriguez, J., Garcia, M.C., Hernandez, J.J.,
Gou, J., Minaie, B., Wang, B., Liang, Z., Zhang, C.: Compu- Ezquerra, T.A.: Influence of single-walled carbon nano-
tational and experimental study of interfacial bonding of tubes on the effective elastic constants of poly (ethylene
single-walled nanotube reinforced composites. Comput. terephthalate). Compos. Sci. Technol. 70, 284–290 (2010)
Mater. Sci. 31, 225–236 (2004) Shen, L., Li, J.: Transversely isotropic elastic properties of
Gupta, S.S., Batra, R.C.: Wall thickness and radial breathing single-walled carbon nanotubes. Phys. Rev. B 69,
modes of single-walled carbon nanotubes. ASME J. Appl. 045–414 (2004)
Mech. 75, 061010 (2008) Smith, W.A., Auld, B.A.: Modeling 1–3 composite piezo-
Honjo, K.: Thermal stresses and effective properties calculated electrics: thickness mode oscillations. IEEE Trans.
for fiber composites using actual cylindrically-anisotropic Ultrason. Ferroelectr. Freq. Control 38(1), 40–47 (1991)
properties of interfacial carbon coating. Carbon 45, Song, Y.S., Youn, J.R.: Modeling of effective elastic properties
865–872 (2007) for polymer based carbon nanotube composites. Polymer
Iijima, S.: Helical microtubules of graphitic carbon. Nature 47, 1741–1748 (2006)
354, 56–58 (1991) Thostenson, E.T., Chow, T.W.: On the elastic properties of
Jiang, Y.L., Martin, L.D.: Anisotropic coupled-field inclusion carbon nanotube based composites: modeling and char-
and inhomogeneity problems. Philos. Mag. A 77(5), acterization. J. Phys. D 36, 573–582 (2003)
1341–1350 (1998) Treacy, M.M.J., Ebbessen, T.W., Gibson, J.M.: Exceptionally
Jiang, B., Liu, C., Zhang, C., Liang, R., Wang, B.: Maximum high Young’s modulus observed for individual carbon
nanotube volume fraction and its effect on overall elastic nanotubes. Nature 381, 678–680 (1996)
properties of nanotube-reinforced composites. Composites Tsai, J.L., Tzeng, S.H., Chiu, Y.T.: Characterizing elastic
B 40, 212–217 (2009) properties of carbon nanotube/polyimide nanocomposites
Li, C., Chou, T.W.: A structural mechanics approach for the using multi-scale simulation. Composites B 41, 106–115
analysis of carbon nanotubes. Int. J. Solids Struct. 40, (2010)
2487–2499 (2003) Wernik, J.M., Meguid, S.A.: Multiscale modeling of the non-
Lu, J.P.: Elastic properties of carbon nanotubes and nanoropes. linear response of nano-reinforced polymers. Acta Mech.
Phys. Rev. Lett. 79, 1297–1300 (1997) 217, 1–16 (2011)
Mathur, R.B., Chatterjee, S., Singh, B.P.: Growth of carbon Xiao, J.R., Gama, B.A., Gillespie, J.W.: An analytical molec-
nanotubes on carbon fiber substrates to produce hybrid/ ular structural mechanics model for the mechanical
phenolic composites with improved mechanical proper- properties of carbon nanotubes. Int. J. Solids Struct. 42,
ties. Compos. Sci. Technol. 68, 1608–1615 (2008) 3075–3092 (2005)
Meguid, S.A., Zhu, Z.H.: A novel finite element for treating Zhang, J., He, C.: A three-phase cylindrical shear-lag model for
inhomogeneous solids. Int. J. Numer. Methods Eng. 38, carbon nanotube composites. Acta Mech. 196, 33–54
1579–1592 (1995) (2008)
Meguid, S.A., Wernik, J.M., Cheng, Z.Q.: Atomistic-based Zhang, Q., Qian, W., Xiang, R., Yang, Z., Luo, G., Wang, Y.,
continuum representation of the effective properties of Wei, F.: In situ growth of carbon nanotubes on inorganic
nano-reinforced epoxies. Int. J. Solids Struct. 47, fibers with different surface properties. Mater. Chem.
1723–1736 (2010) Phys. 107, 317–321 (2008)
123