See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/267425435

Load Transfer Analysis in Short Carbon Fibers with Radially-Aligned Carbon

Nanotubes Embedded in a Polymer Matrix Citation "Load Transfer Analysis in
Short Carbon Fibers with Radia...

Article  in  Journal of Advanced Materials -Covina- · July 2012

CITATIONS READS
17 71

2 authors:

Roberto Guzman de Villoria Brian L Wardle

Universidad de Salamanca Massachusetts Institute of Technology
99 PUBLICATIONS   2,544 CITATIONS    270 PUBLICATIONS   7,947 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

VA-CNT composite electrodes for energy storage View project

Electromechanical Systems View project

All content following this page was uploaded by Roberto Guzman de Villoria on 11 August 2015.

The user has requested enhancement of the downloaded file.

 Load Transfer Analysis in Short Carbon Fibers with
Radially-Aligned Carbon Nanotubes Embedded in a
Polymer Matrix

The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.

Citation Ray, M.C., Roberto Guzman de Villoria and brian L. Wardle.

"Load Transfer Analysis in Short Carbon Fibers with Radially-
Aligned Carbon Nanotubes Embedded in a Polymer Matrix."
Journal of Advanced Materials, issue 4, (Jan. 2009).
As Published http://www.sampe.org/news/jam/2009/2009jamissue4.aspx
Publisher Society for the Advancement of Material and Process
Engineering

Version Author's final manuscript

Accessed Thu Feb 21 02:07:00 EST 2013
Citable Link http://hdl.handle.net/1721.1/71852
Terms of Use Creative Commons Attribution-Noncommercial-Share Alike 3.0
Detailed Terms http://creativecommons.org/licenses/by-nc-sa/3.0/
 Load Transfer Analysis in Short Carbon Fibers with Radially-Aligned Carbon

Nanotubes Embedded in a Polymer Matrix

M. C. Ray
Department of Mechanical Engineering, Indian Institute of Technology, Kharagpur
721302, India

Roberto Guzman de Villoria and Brian L. Wardle

Department of Aeronautics and Astronautics, Massachusetts Institute of Technology,
Cambridge, MA 02139, USA

A novel shortfiber composite in which the microscopic advanced fiber reinforcements

are coated with radially aligned carbon nanotubes (CNTs) is analyzed in this study. A

shear-lag model is developed to analyze the load transferred to such coated fibers from

the aligned-CNT reinforced matrix in a hybrid composite application. It is found that if

the carbon fibers are coated with radially aligned CNTs, then the axial load transferred

to the fiber is reduced due to stiffening of the matrix by the CNTs. Importantly, it is

shown that at low loading of CNTs in the polymer matrix, there is a significant reduction

in the maximum interfacial shear stress, e.g., at 1% CNTs, there is an ~25 % reduction

in this maximum stress. Further, the modification in the load sharing between the fiber

and the matrix plateaus at ~2% CNT matrix loading, indicating a small but critical

window for engineering the interface in this manner. Effects of the variation of the

aspect ratio of the fiber, CNT volume fraction and the application of radial load on the

load transferred to such CNT coated fibers are also investigated.

KEYWORDS: Nanocomposites, Composite Structures

 2

1. Introduction

The identification of carbon nanotubes1 (CNTs) has stimulated extensive

research devoted to the prediction of their elastic properties through experiments and

theoretical modeling. Of interest is to determine elastic properties of the CNTs as an

input to models that predict composite behaviour. Early work by Treacy et al.2

experimentally determined that CNTs have Young’s modulus in the terapascal (TPa)

range. Li and Chou3 linked structural and molecular mechanics (MM) approaches to

compute elastic properties of CNTs. Sears and Batra4 used three MM potentials to

simulate axial and torsional deformations of a CNT assuming that the tube can be

regarded as a hollow cylinder of mean diameter equal to that of the CNT and

determined the wall thickness, Young’s modulus and Poisson’s ratio of the CNT. Shen

and Li5 assumed that a CNT should be modeled as a transversely isotropic material

with the axis of transverse isotropy coincident with the centroidal axis of the tube. They

determined values of the five elastic constants by using a MM potential and an energy

equivalence principle. Batra and Sears6 proposed that the axis of transverse isotropy of

a CNT is a radial line rather than the centroidal axis of the tube and found that Young’s

modulus in the radial direction equals about 1/4th of that in the axial direction. Wu et al.7

developed an atomistic based finite deformation shell theory for single-walled CNT and

found its stiffness in tension, bending and torsion.

A great deal of research has also been carried out on the prediction of effective

elastic properties of CNT-reinforced composites. For example, Thostensen and Chou8

have estimated the elastic moduli of CNT-reinforced composite through

micromechanical analysis. Gao and Li9 derived a shear lag model of CNT reinforced
 3

polymer composites by replacing the CNT with an equivalent solid fiber. Song and

Yoon10 numerically estimated the effective elastic properties of CNT-reinforced polymer

based composites. Siedel and Lagoudas11 carried out a micromechanical analysis to

estimate the effective properties of CNT-reinforced composites. Guzman de Villoria and

Miravete devolped a model to estimate the effect of the CNTs dispersion in composites

matrix by micromechanical analysis12. Jiang et al.13 derived a continuum based model

to study the effect of CNT/matrix interface on the macroscopic properties of CNT-

reinforced composites. Odegard et al.14, 15 have modeled CNT-reinforced composite to

estimate effective elastic moduli using an equivalent-continumm modeling method that

connects computational chemistry and solid mechanics models. To avoid the long times

of simulation of materials at nanoscale level, Yamakov and Glaessgen16 have linked

continuum mechanics in with atomic-level simulations, in one case to study the fracture

tip of several metals. Zhang and He17 theoretically investigated the viscoelastic behavior

of CNT-reinforced composites developing a three-phase shear-lag model. Most

recently, Ray and Batra18 carried out a micromechanical analysis to estimate the

effective elastic and piezoelectric properties of CNT and piezoelectric fiber reinforced

hybrid composite. Several review articles have appeared that summarize the various

advances in these two-phase (CNTs plus a matrix) nanocomposites19-21.

Here we analyze a new hybrid composite composed of micron-scale diameter

advanced fibers with in situ grown radially aligned CNTs and a polymer matrix. Growth

of aligned CNTs on advanced fibers (see examples in Figure 1) have been investigated

by several groups22-26 and recently, bulk composites have been realized using aligned

‘fuzzy’ fibers27-29. The objective of this work is to investigate the load transferred to a
 4

carbon fiber from the matrix in the case where the micon-scale fuzzy fiber is

discontinuous (see Figure 2). A closed-form shear lag model is developed for such

investigation, incorporating a micromechanics model for predicting the radially-

orthotropic properties of the aligned-CNT reinforced polymer matrix. Such composites

can be described as a hybrid nano-engineered composite, where the polymer matrix is

reinforced with radially aligned-CNT resulting in an aligned-CNT nanocomposite matrix

that surrounds the micron-scale advanced fiber (see Figure 3). Such nano-engineered

composites can be fabricated using capillarity-driven wetting of the aligned CNTs by

advanced polymers29, 30.

2. Shear lag model

A schematic sketch of the cylindrical representative volume element (RVE) of the

composite analyzed here is shown in Figure 3. The cylindrical coordinate system ( r , θ

and x ) is considered in such a way that the axis of the RVE coincides with the x axis

while the CNTs are aligned along the r -direction. The model is derived by dividing the

RVE into three zones. The portion of the RVE in the zone − L f ≤ x ≤ L f consists of a

discontinuous micro-scale advanced fiber (carbon is considered here) reinforcement on

which radially aligned CNTs have been grown. When this resulting fuzzy fiber is

embedded in a polymer material, the CNT forest is filled with the polymer creating a

nano-reinforced polymer matrix, what many have called a polymer nanocomposite

31, 32
(PNC) . Thus, the radially aligned CNTs reinforce the polymer matrix and the

portion of the RVE in the zone − L f ≤ x ≤ L f can be viewed as a hybrid composite

comprised of the carbon fiber reinforcement embedded in the CNT-reinforced polymer

 5

matrix composite phase. The radius and the length of the carbon fiber are denoted by

a and 2L f , respectively. The inner and outer radii of the CNT-reinforced matrix phase

are a and R , respectively. The portions of the RVE in the zones − L ≤ x ≤ − L f and

L f ≤ x ≤ L are treated in the model as an imaginary fiber and the matrix phase, both

composed of the polymer material. The radius of the imaginary fiber is also denoted by

a while the inner and outer radii of the matrix phase are also represented by a and R ,

respectively. Thus, the shear lag model developed for the zone − L f ≤ x ≤ L f can be

applied to derive the shear lag models for the zones − L ≤ x ≤ − L f and L f ≤ x ≤ L .

In what follows, the shear lag model for the zone − L f ≤ x ≤ L f is first derived. A

tensile stress σ 0 is applied to the RVE along x direction at x = ± L while the RVE is

subjected to a radial normal stress q 0 at r = R . In order to derive this shear lag model,

the effective properties of the aligned CNT-reinforced matrix phase are needed. This

PNC matrix phase has transverse isotropy in a radial coordinate system due to the CNT

alignment and isotropic nature of the polymer. This is a slight approximation because

the grown CNTs have reduced volume fraction as they grow radially, but volume

fraction may be considered constant over the small (microns) CNT lengths considered.

Micromechanics is used to calculate properties in this region as they have not been

determined experimentally to date. A micromechanics model by Ray and Batra18 is used

to calculate the effective elastic constants for a forest of aligned single-walled CNTs

(properties from Ref.5 given in Table 1) embedded in a polymer (results summarized in

Table 2). These results are used as an input to the shear-lag model.

Returning to the shear-lag model, the governing equations for the different
 6

phases of this RVE concerning equilibrium along x direction are given by

∂σ ix 1 ∂ (rσ ixr )
+ = 0 , i = f and m
∂x r ∂r

(1)

while the relevant constitutive relations are

σ ix = C11
i
∈ix + C12
i
∈iθ + C13
i
∈ri and σ ixr = C 55
i
∈ixr ; i = f and m

(2)

In Eqs. (3) and (4), superscripts f and m denote, respectively, the carbon fiber and the

CNT-reinforced PNC matrix. For the i -th constituent phase, σ ix and σ ri represent the

normal stresses in the x and r , directions, respectively; ∈ix , ∈iθ and ∈ri are the normal

strains along x , θ and r , directions, respectively; σ ixr is the transverse shear stress,

∈ixr is the transverse shear strain and C iij are the elastic constants . It should be noted

here that the principal material coordinates 1, 2, 3 axes are also considered to be

coincident with the problem coordinate axes x , θ , r , respectively. Hence, the

conventional subscripts are used to write the elastic constants appearing in Eq. (2). The

strain-displacement relations for an axisymmetric problem relevant to this RVE are

∂u i wi ∂w i ∂u i ∂w i
∈ix = , ∈θ =
i
, ∈r =
i
and ∈xr =
i
+ ; i = f and m
∂x r ∂r ∂r ∂x

(3)
 7

in which u i and w i represent the axial and radial displacements at any point of the i -th

phase along x and r , directions, respectively. The traction boundary conditions are

given by

σm
r = q 0 and σ m
xr =0
r=R r =R

(4)

and the continuity conditions are

σ fr = σm
r , σ fxr = σm
xr = τ i , uf = um and w f = wm
r =a r =a r =a r =a r =a r =a r =a r =a

(5)

Where, τ i is the transverse shear stress at the interface between the carbon fiber and

the PNC matrix phase The average axial stresses in the different phases are defined as

1 a f 1 R
σ xf = σ
2 ∫ x
2 π rdr , and σ m
x = 2 ∫ x
σ m 2πrdr
πa 0 π( R − a ) a
2

(6)

Now, making use of Eqs. (1) and (4) to (6), it can be derived that

∂σ xf 2 ∂σ xm 2a
= − τ i and = 2 τi
∂x a ∂x R − a2

(7)
 8

Since the radial dimension of this RVE is very small, it is reasonable to assume that the

gradient of σ m
x with respect to the axial coordinate ( x ) is independent of the radial

coordinate ( r ). Thus let us assume that

∂σ m
x
= φ( x)
∂x

(8)

Integrating the governing equation (1) for the PNC matrix phase from r to R , it can be

shown that

2r
φ( x ) = − σm
xr
R −r
2 2

(9)

Substituting σ m
xr = τ i in Eq. (9), the transverse shear stress in the PNC matrix can be
r =a

expressed in terms of the interface shear stress τ i as follows:

⎛ R2 ⎞ a
σm = ⎜ − ⎟
xr ⎜ r r ⎟ (R 2 − a 2 ) τ i
⎝ ⎠

(10)

Also, since the RVE is an axisymmetric problem, it may further be assumed that the

∂w i
gradient of radial displacements with respect to x -direction is negligible and so,
∂x

from the constitutive relation (2) between σ m

xr and ∈xr one can write,
m
 9

∂u m 1
= i σm
∂x
rx
C 55

(11)

Solving Eq. (11), the axial displacement of the matrix phase along x direction can be

derived as follows:

u m = u fa + A1τ i

(12)

a ⎧ 2 r 1 2 2 ⎫
in which u fa = u f and A 2 = ⎨ R ln − (r − a )⎬
r =a Cm
55 ( R
2
−a )⎩
2
a 2 ⎭

(13)

The radial displacements in the two phases can be assumed as in33 :

Bm
w f = A f r and w m = A m r +
r

(14)

where A f , A m and B m are unknown constants. Invoking the continuity conditions for

radial displacement at the interface ( r = a ) it can be found that:

a2 a2
wm = A f + (r − ) A m
r r

(15)
 10

Invoking the continuity condition σ fr = σm

r and satisfying the boundary condition
r =a r =a

σm
r = q 0 , the following equations for solving A f and A m are derived :
r=R

⎡ A11 A12 ⎤ ⎧ A f ⎫ ⎧C13

m f ⎫
− C12 σ xf ⎧ 0 ⎫ ⎧ 0 ⎫ ∂τ i
⎢A ⎨ ⎬= ⎨ ⎬ + ⎨ ⎬ + ⎨ m⎬
A 22 ⎥⎦ ⎩ A m ⎭
q
⎩ C13 ⎭ C11 ⎩ − 1⎭ ⎩ A 2 C13 ⎭ ∂x
m f 0
⎣ 21

(16)

where,

f 2 m f
2(C12 ) a 2 2C13 C12
A11 = f
C11 + f
C12 − f
+ m
C 33 − Cm
23 , A12 = −2C m
33 , A 21 = m
(C 33 − Cm
23 ) 2
+ f
,
C11 R C11

a2
A 22 = C m
23 + C 33 + (C 33 − C 23 )
m m m
, A 2 = A1 r = R .
R2

From Eq. (16), the solutions of A f and A m can be expressed as:

∂τ i ∂τ
A f = L11σ xf + L12 q + L13 and A m = L 21σ xf + L 22 q + L 23 i
∂x ∂x

(17)

The expressions of the coefficients L ij are evident from Eq. (16) and are not shown

here for the sake of brevity. The equilibrium of force along the axial ( x ) direction yields
 11

a R
πR 2 σ 0 = ∫ σ fx 2πrdr + ∫ σ m
x 2πrdr
0 a

(18)

Using Eqs. (6), (17) and the constitutive relations, Eq. (18) can be reduced to

∂τ i R 2 σ 0 L 1 L
= + q 0 − 2 σ xf
∂x L3 L3 L3

(19)

where,

m
C11
L1 = B1L12 − B 2L 22 , L 2 = a +
2
f
( R 2 − a 2 ) − B1L11 + B 2 L 21 , L 3 = A 3C11
m
− B1L13 + B 2 L 23 ,
C11

R R ⎡ 2C f C m 2⎤
m a
A 3 = ∫ 2A1rdr , B1 = 2 ∫ ⎢ 12f 11 r − (C12
m
− C13 ) ⎥ dr ,
a a⎣ C11 r ⎦

R⎡
m a ⎤
2
and B 2 = 2 ∫ ⎢(C12
m
+ C13
m
)r − (C12
m
− C13 ) ⎥ dr .
a⎣ r ⎦

Substitution of Eq. (7) 1 into Eq. (19) yields the governing equation for the average axial

stress in the carbon fiber coated with radially aligned CNTs as follows:

∂ 2 σ xf
− α 2 σ xf = − ησ 0 − μq 0
∂x 2
 12

(20)

where,

2L 2 2R 2 2L1
α2 = , η= and μ =
aL 3 aL 3 aL 3

(21)

Following the above procedure, the governing equation for the average axial stress

( σ xpf ) in the imaginary fiber made of the polymer material lying in the zones

− L ≤ x ≤ − L f and L f ≤ x ≤ L can be written as

∂ 2 σ xpf
− α1 σ xpf = − η σ 0 − μq 0
2

∂x 2

(22)

In the above equation, the expressions for α1 , η and μ are similar to those of α 2 , η
2

and μ , respectively. But these are to be derived by considering C fij = C m

ij = C ij .
p

Solutions of Eqs. (21) and (22) are given by:

η μ
σ xf = c1e αx + c 2 e − αx + σ+ 2 q
α 2
α

(23)

η μ
σ xpf = c 3 e αx + c 4 e − αx + σ0 + q0
α1 α1
2 2

(24)
 13

in which c1 , c 2 , c 3 and c 4 are the constants of integrations to be evaluated from the

following end conditions:

σ xpf = σ 0 at x = ± L and σ xf = σ xpf at x = ± L f

(25)

Utilizing the end conditions given by (25) in Eqs. (23) and (24), the final solutions for

σ xpf and σ xf are obtained as follows:

⎧⎪ cosh( α x) ⎛ ⎞ ⎫
σ xpf = ⎨ 1 ⎜ 1 − η − μq 0 ⎟ + η + μ q 0 ⎪⎬σ
⎪⎩ cosh(α 1L ) ⎜⎝ α1 α1 σ 0 ⎟ α 2 α 2 σ0 ⎪ 0
2 2
⎠ 1 1 ⎭

(26)

⎡ cosh(αx ) ⎧⎪ cosh(α L ) ⎛ ⎞ ⎛ ⎞ ⎫ ⎤
σ xf = ⎢ ⎨
1 f ⎜ η μq
1− 2 − 2 0 ⎟ + η − η + ⎜ μ − μ ⎟ q 0 ⎪⎬ + η + μq 0 ⎥ σ 0
⎜
⎢⎣ cosh(αL f ) ⎪⎩ cosh(α1L ) ⎝ α1 α1 σ 0 ⎟ α 2 α 2 ⎜ α 2 α 2 ⎟ σ 0 ⎪ α 2 α 2σ ⎥
⎠ 1 ⎝ 1 ⎠ ⎭ 0⎦

(27)

In the case that the fiber and matrix are isotropic, and with q 0 =0, the above model

reduces to that presented by Gao and Li9 for a CNT reinforced polymer composite.

Finally, substitution of Eq. (27) into Eq. (7) yields the expression for the interface shear

stress as follows:

a ⎡ α sinh( αx) ⎧⎪ cosh(α1L f ) ⎛⎜ η μq ⎞ ⎛ ⎞ ⎫⎤

⎟ + η − η + ⎜ μ − μ ⎟ q 0 ⎪⎬ ⎥ σ 0
τi = − ⎢ ⎨ 1− 2 − 2 0
2 ⎢ cosh(αL f ) ⎪ cosh(α1L ) ⎝ ⎜ α1 α1 σ 0 ⎟ α 2 α 2 ⎜ α 2 α 2 ⎟ σ 0 ⎪⎥
⎣ ⎩ ⎠ 1 ⎝ 1 ⎠ ⎭⎦

(28)
 14

3. Results and Discussion

The elastic coefficients of arm chair type CNTs with respect to the coordinate system

considered here are obtained from Shen and Li5 which are listed in Table 1. The

polymer material and the carbon fiber are elastically isotropic. The isotropic elastic

coefficients ( C pij ) of the polymer materia18 and the elastic constants ( C fij ) of the high

modulus M40 carbon fiber34 needed for computing the numerical results are as given by

p
C11 = 5.3GPa , C12
p
= 3.1GPa , C11
f
= 373.89GPa and C12
f
= 6.5GPa .

A discussion on the effective properties of the PNC matrix is now in order.

Recently, Ray and Batra18 derived a micromechanics model to predict the effective

properties of CNT and piezoelectric fiber reinforced hybrid composite. In the absence of

piezoelectric fibers this micromechanics model is reduced to a model which predicts the

effective elastic properties of the transversely isotropic PNC matrix with radially aligned

CNTs considered here and is given by:

[C m ] = [C1 ][V3 ]−1 + [C 2 ][V4 ]−1

(29)

The various matrices appearing in (29) are presented in the Appendix. At a particular

value of CNT volume fraction ( VCNT =1.0%), the effective values of the elastic constants

Cm
ij of the PNC matrix predicted from Eq. (29) are presented in Table 2 for different

types of armchair CNTs. Also, for VCNT =1.0%, the elastic constants C m
ij of the PNC
 15

matrix with (10, 10) CNTs presented in Table 2 yield the values of the Young’s modulus

r ) in the radial direction and the Poisson’s ratio ( υ xr ) as 14.34 GPa and 0.369,
( Em m

respectively. The values of the same are also predicted identically from simple rule of

mixtures validating the micromechanics model given by Eq. (29). Thus Eq. (29) can be

used to compute the effective elastic constants C m

ij of the PNC matrix for evaluating the

numerical results.

For presenting the results, following nondimensional parameters are adopted:

σ xf 10τ i
σ* = and τ * =
σ0 σ0

(31)

Unless otherwise mentioned, the values of the geometrical parameters of the RVE are

taken as:

a = 3.5μm , R = a + 10μm , L f / a = 10 and L / L f = 1.2

Arm chair type (10, 10) CNTs are used to compute the numerical results for the axial

stress and the interface shear stress, unless specifically varied. In order to validate the

model derived in the previous section, first the normalized average axial stress in the

fiber without coated with CNTs and the interface shear stress are compared with those

obtained by an existing model9 as shown in Figure 4. For this comparison, the fiber is

an arm chair (10, 10) CNT as considered in Ref.9. It may be noted that the good
 16

agreement between the two sets of results have been obtained verifying the present

model. The marginal differences observed may be attributed to the fact that the model

in Ref.9 did not consider the radial deformation, whereas in the present model radial

deformations have been taken into account. Next, results are computed for carbon

fibers coated with radially aligned CNTs.

The variations of the axial normal stress in the short carbon fiber and the

transverse shear stress at the interface between the fiber and the PNC matrix along the

length of the fiber are shown in Figure 5. It may be observed that the carbon fiber

coated with radially aligned-CNTs shares less load than the fiber without coated with

CNTs. This is attributed simply to the radial and axial stiffening of the polymer matrix by

the CNTs. Note that the axial Young’s moduli of the PNC matrix with 1% (10, 10) CNT

x = 3.41GPa and E x = 3.01GPa , respectively while the radial

and the polymer are E m p

r = 14.34GPa and E r = 3.01GPa , respectively. The CNTs create

Young’s moduli are E m p

a radially orthotropic PNC matrix and an increase in the CNT volume fraction increases

both the axial and radial modulii. Importantly, the maximum interfacial shear stress is

reduced in the case of CNTs reinforcing the matrix. A critical parameter in the design of

polymer matrices for composites is the ratio of this maximum stress to the strength of

the matrix. Load sharing improves this ratio with the presence of only 1% CNTs by 23%,

and it is also expected that the strength of the interface should increase due to the

CNTs as well, further improving the effect. The axial load transferred to the carbon fiber

and the interfacial shear stress decreases with the increase in the radial stiffness of the

PNC matrix as shown in Figure 6 and Figure 7, respectively. Also, compared in Figure 6

and Figure 7is the case where the matrix remains isotropic but the value of its Young’s
 17

modulus is increased to that of E m

r of the PNC with VCNT =1.0% (Poisson’s ratio is

assumed as 0.33). It may be observed from these figures that isotropically stiffening the

matrix causes (as expected) a significant increase in load carried by the matrix and a

reduction in the interfacial maximum shear stress beyond what is seen at 1% radially-

aligned PNC.

In Figure 8, it may be seen that if the value of ( R / a ) decreases then the load

transfer from the matrix to the fiber coated with radially aligned CNTs significantly

decreases as expected due to the larger proportion of overall load carried by the

enhanced-stiffness matrix relative to the fiber. Beyond a few percent volume fraction of

CNTs in the matrix, overall load sharing is not signficantly affected as evidenced both in

Figure 8 and Figure 9. In Figure 9 the effect of CNT volume fraction on the critical length

( L crit ) of the fiber is presented. Here, L crit is measured from the center of the fiber and

determined based on the situation when σ * = 98% of the maximum value of σ * . It may

be noted from this figure that as the CNT volume fraction increases, the critical length of

the fiber decreases rapidly in the region VCNT < 1.0% and then decreases

monotonically. For a particular value of volume fraction of the carbon fiber, the critical

length increases if the aspect ratio of the fiber increases, while for a particular value of

aspect ratio of the fiber this critical length decreases with the increase in the volume

fraction of the carbon fiber. In Figure 10 a critical value of VCNT is shown to exist

beyond which the radial orthotropy of the PNC matrix does not appreciably alter load

sharing capability of the fiber. After ~2% VCNT , there is little change in the load sharing

between the PNC matrix and the fiber.

Variations of maximum values of the axial stress in the carbon fiber coated with
 18

radially aligned CNTs ( VCNT =1.0%) and the interface shear stress with the aspect ratio

of the fiber are presented in Figure 11. The maximum value of the axial load shared by

the fiber increases sharply with the increase in the value of the aspect ratio as long as

L f / a < 12 . For L f / a > 20 , the axial load sharing capability of the fiber becomes

independent of the variation of the aspect ratio of the fiber. In case of interface shear

stress, its maximum value also increases rapidly with the increase in the value of the

aspect ratio of the fiber till L f / a < 8 . The maximum value of τ * becomes saturated for

L f / a > 10 .

The effect of application of radial load on the load transferred to the fiber is

presented in Figure 12. If the applied radial load is compressive, then the maximum

values of the axial normal stress in the fiber and the interface shear stress are higher

than those without the application of radial load ( q 0 = 0 ), and vice versa. The variations

of axial normal stress in the carbon fiber and the interface shear stress along its length

are presented in Figure 13 and Figure 14, respectively, for different arm chair type

CNTs in Table 1. It may be observed from these figures that for a particular value of

VCNT , as the diameter of CNT increases, both the axial normal load transferred to the

fiber and the interface shear stress increase. This may be attributed to the fact that as

the diameter of CNT increases, elastic coefficients of CNT decreases (see Table 1)

which results in the decrease in the values of the effective elastic properties of the CNT-

reinforced PNC matrix. Overall, the type of CNT has a small effect on the composite

fiber load sharing relative to varying volume fraction of the CNTs in the PNC matrix.
 19

4. Conclusions

In this paper, load sharing in a shortfiber composite where the matrix is reinforced with

radially-aligned CNTs has been analyzed. The fiber reinforcement of the composite is a

discontinuous carbon fiber coated with radially aligned CNTs. A shear lag model

considering radial and axial deformations of the different phases of the RVE has been

developed to analyze the axial load transferred to this carbon fiber. Since the radially

aligned CNTs grown on the carbon fiber reinforce the polymer matrix, the effective

elastic properties of the resulting CNT-reinforced PNC matrix are modified. Hence, if the

fiber is coated with CNTs, the axial load transferred to the carbon fiber and the shear

stress at the interface between the fiber and the PNC matrix decrease, and the CNT-

reinforced matrix carries more of this load. If the volume fraction of CNTs increases,

both the axial load transferred to the fiber and the interface shear stress decrease,

including importantly the maximum shear stress at the fiber-matrix interface. The critical

length of the carbon fiber varies little with CNT volume fraction beyond a few percent.

For a particular value of CNT volume fraction, compressive radial load applied to the

RVE increases the axial load transferred to the fiber and the interface shear stress.

Future work should consider load-transfer in randomly-oriented shortfiber composites,

and also load transfer around a broken fiber in continuous filament composites, in

addition to load sharing in the presence of applied shear stress.

 20

Appendix

The various matrices appearing in Eq. (29) are given by

⎡ 0 0 0 0 0 0⎤ ⎡ C11
p
C12p
C12p
0 0 0 ⎤
⎢ p ⎥
⎢ 0
⎢ 0 0 0 0 0⎥⎥ ⎢ C12 C11p
C12p
0 0 0 ⎥
⎢C13n
C n23 C n33 0 0 0⎥ ⎢ v Cp p
v p C12 p
v p C11 0 0 0 ⎥
[C1 ] = v n ⎢ ⎥ , [C 2 ] = ⎢ ⎥,
p 12

⎢ 0 0 0 0 0 0⎥ ⎢ 0 0 0 C p44 0 0 ⎥
⎢ 0 0 0 0 0 0⎥ ⎢ 0 0 0 0 C p44 0 ⎥
⎢ ⎥ ⎢ ⎥
⎢⎣ 0 0 0 0 0 0⎥⎦ ⎢⎣ 0 0 0 0 0 C p44 ⎥⎦

R2
vn = VCNT , v p = 1 − v n , [V3 ] = [V1 ] + [V2 ][C 4 ]−1[C 3 ] , [V4 ] = [V2 ] + [V1 ][C 3 ]−1[C 4 ] ,
R 2 − a2

⎡ C11
n
C12n
C13n
0 0 0 ⎤ ⎡C11p p
C12 p
C12 0 0 0 ⎤
⎢ n ⎥ ⎢ p p p ⎥
⎢C12 C n22 C n23 0 0 0 ⎥ ⎢C12 C11 C12 0 0 0 ⎥
⎢ 0 0 1 0 0 0 ⎥ ⎢ 0 0 1 0 0 0 ⎥
[C 3 ] = ⎢ ⎥ , [C 4 ] = ⎢ ⎥,
⎢ 0 0 0 C n44 0 0 ⎥ ⎢ 0 0 0 C p44 0 0 ⎥
⎢ 0 0 0 0 C n55 0 ⎥ ⎢ 0 0 0 0 C p44 0 ⎥
⎢ ⎥ ⎢ ⎥
⎢⎣ 0 0 0 0 0 C n66 ⎥⎦ ⎢⎣ 0 0 0 0 0 C p44 ⎥⎦

⎡vn 0 0 0 0 0⎤ ⎡vp 0 0 0 0 0⎤
⎢0 ⎥ ⎢0 vp 0 0 0 0 ⎥⎥
⎢ vn 0 0 0 0⎥ ⎢
⎢0 0 0 0 0 0⎥ ⎢0 0 1 0 0 0⎥
[V1 ] = ⎢ ⎥ and [V2 ] = ⎢ ⎥
⎢0 0 0 vn 0 0⎥ ⎢0 0 0 vp 0 0⎥
⎢0 0 0 0 vn 0⎥ ⎢0 0 0 0 vp 0⎥
⎢ ⎥ ⎢ ⎥
⎢⎣ 0 0 0 0 0 v n ⎥⎦ ⎢⎣ 0 0 0 0 0 v p ⎥⎦

(2)
 21

In the above matrices, C nij and C pij are the elastic coefficients of the CNT and the

polymer material, respectively. The volume fractions of the polymer and the CNT with

respect to the volume of the PNC are represented by v p and v n while VCNT denotes

the volume fraction of CNTs with respect to the volume of the RVE.
 22

References

1
S. Iijima, Helical microtubules of graphitic carbon, Nature 1991, 354, 56-58.
2
M. M. J. Treacy, T. W. Ebbessen, J. M. Gibson, Exceptionally high Young's modulus

observed for individual carbon nanotubes, Nature 1996, 381, 678 - 680.
3
C. Li, T. W. Chou, A structural mechanics approach for the analysis of carbon

nanotubes, International Journal of Solids Structures 2003, 40, 2487 - 2499.

4
A. Sears, R. C. Batra, Macroscopic properties of carbon nanotubes from molecular

mechanics simulations, Physical Review B 2004, 69, 1-10.

5
L. Shen, J. Li, Transversely isotropic elastic properties of single-walled carbon

nanotubes, Physical Review B 2004, 69, 1-10.

6
R. C. Batra, A. Sears, Uniform radial expansion/contraction of carbon nanotubes and

their transverse elastic moduli, Modelling and Simulation in Materials Science and

Engineering 2007, 15, 835-844.

7
J. Wu, K. C. Hwang, Y. Huang, An atomistic-based finite-deformation shell theory for

single-wall carbon nanotubes, Journal of the Mechanics and Physics of Solids 2008, 56,

279-292.
8
E. T. Thostenson, T. W. Chou, On the elastic properties of carbon nanotube based

composites: modeling and characterization, Journal of Physics D: Applied Physics

2003, 36, 573-582.

9
X. L. Gao, K. Li, A shear-lag model for carbon nanotube reinforced polymer

composites, International Journal of Solids and Structures 2005, 42, 1649-1667.

10
Y. S. Song, J. R. Youn, Modeling of effective elastic properties for polymer based

carbon nanotube composites, Polymer 2006, 47, 1741-1748.

 23

11
G. D. Seidel, D. C. Lagoudas, Micromechanical analysis of the effective elastic

properties of carbon nanotube reinforced composites, Mechanics of materials 2006, 38,

884-907.
12
R. Guzman de Villoria, A. Miravete, Mechanical model to evaluate the effect of the

dispersion in nanocomposites, Acta Materialia 2007, 55, 3025-3031.

13
L. Y. Jiang, H. L. Tan, J. Wu, Y. G. Huang, K. C. Hwang, Continuum modeling of

interfaces in polymer matrix composites reinforced by carbon nanotubes, Nano Letters

2007, 2, 139-148.
14
G. M. Odegard, T. S. Gates, K. E. Wise, C. Park, E. J. Siochi, Constitutive modeling of

nanotube-reinforced polymer composites, Composites Science and Technology 2003,

63, 1671-1687.
15
G. M. Odegard, T. S. Gates, L. M. Nicholson, K. E. Wise, Equivalent-Continuum

Modeling of Nano-Structured Materials, Composites Science and Technology 2002, 62,

1869-1880.
16
V. I. Yamakov, E. H. Glaessegen, To twin or not to twin, Nature Materials 2007, 6,

795-796.
17
J. Zhang, C. He, A three-phase cylindrical shear-lag model for carbon nanotube

composites, Acta Mechanica 2008, 196, 33-54.

18
M. C. Ray, R. C. Batra, Effective properties of carbon nanotube and piezoelectric fiber

reinforced hybrid smart composite, accepted in ASME Journal of Applied mechanics

2008.
19
E. T. Thostenson, C. Li, T.-W. Chou, Nanocomposites in context, Composites Science

and Technology 2005, 65, 491-516.

 24

20
E. T. Thostenson, Z. Ren, T.-W. Chou, Advances in the science and technology of

carbon nanotubes and their composites: a review, Composites Science and Technology

2001, 61, 1899-1912.

21
J. N. Coleman, U. Khan, W. J. Blau, Y. K. Gun'ko, Small but strong: A review of the

mechanical properties of carbon nanotube-polymer composites, Carbon 2006, 44,

1624-1652.
22
J. O. Zhao, L. Liu, Q. G. Guo, J. Shi , G. Zhai , J. Song, Z. Liu, Growth of carbon

nanotubes on the surface of carbon fibers, Carbon 2008, 46, 380-383.

23
Q. J. Gong, H. J. Li, X. Wang, Q. G. Fua, Z. W. Wanga, K. Z. Lia, In situ catalytic

growth of carbon nanotubes on the surface of carbon cloth, Composite Science and

Technology 2007, 67, 2986-2989.

24
L. T. Qu, Y. Zhao, L. M. Dai, Carbon microfibers sheathed with aligned carbon

nanotubes: Towards multidimensional, multicomponent, and multifunctional

nanomaterials, Small 2006, 2, 1052-1059.

25
L. J. Ci, Z. G. Zhao, J. B. Bai, Direct growth of carbon nanotubes on the surface of

ceramic fibers, Carbon 2005, 43, 883-886.

26
Z. G. Zhao, L. J. Ci, H. M. Cheng, J. B. Bai, The growth of multi-walled carbon

nanotubes with different morphologies on carbon fibers, Carbon 2005, 43, 663-665.
27
R. B. Mathur, S. Chatterjee, B. P. Singh, Growth of carbon nanotubes on carbon fibre

substrates to produce hybrid/phenolic composites with improved mechanical properties,

Composite Science and Technology 2008, 68, 1608-1615.

 25

28
V. P. Veedu, A. Cao, X. Li, K. Ma, C. Soldano, S. Kar, P. M. Ajayan, M. N. Ghasemi-

Nejhad, Multifunctional Composites Using Reinforced Laminae with Carbon Nanotube

Forests, Nature Materials 2006, 5, 457-462.

29
E. J. Garcia, B. L. Wardle, A. J. Hart, N. Yamamoto, Fabrication and Multifunctional

Properties of a Hybrid Laminate with Aligned Carbon Nanotubes Grown In Situ,

Composites Science & Technology 2008, 68, 2034-2041.

30
E. J. García, A. J. Hart, B. L. Wardle, A. H. Slocum, Fabrication of composite

microstructures by capillarity-driven wetting of aligned carbon nanotubes with polymers,

Nanotechnology 2007, 18, 165602 (165611 pp).

31
K. I. Winey, R. A. Vaia, Polymer Nanocomposites, MRS Bulletin 2007, 32, 314-322.
32
P. Ajayan, P. Braun, L. Schadler, Nanocomposite Science and Technology, 2003.
33
Z. Hashin, W. Rosen, The elastic moduli of fiber reinforced materials, Journal of

Applied Mechanics 1964, 31, 223-232.

34
J. D. H. Hughes, The carbon fiber/epoxy interface - a review, Composite Science and

Technology 1991, 41, 13-45.

 26

List of tables

Table 1. Material properties of CNTs (Ref. 5). The 3-axis is aligned with the long

axis of the CNT.

CNT n
C11 C n22 Cn33 n
C12 n
C13 C n44 Cn55 C n66

Type (GPa) (GPa) (GPa) (GPa)

(GPa) (GPa) (GPa) (GPa)

(5,5) 668 668 2143 404 184 791 791 132

(10,10) 288 288 1088 254 87.7 442 442 17

(20,20) 138 138 545 134 43.5 227 227 2

(50,50) 55.1 55.1 218 54.9 17.5 92 92 0.1

 27

Table 2. Material properties of PNC with different arm chair type CNTs

( VCNT = 1.0% ).

PNC with m
C11 Cm
22
m
C 33 m
C12 m
C13 Cm
44 Cm
55 Cm
66

CNT Type (GPa) (GPa) (GPa) (GPa)

(GPa) (GPa) (GPa) (GPa)

(5,5) 5.357 5.357 27.565 3.133 3.115 1.112 1.112 1.112

(10,10) 5.356 5.356 16.628 3.134 3.114 1.112 1.112 1.112

(20,20) 5.349 5.349 10.956 3.139 3.114 1.112 1.112 1.112

(50,50) 5.235 5.235 7.546 3.248 3.111 1.112 1.112 1.112

 28

List of figures

Aligned CNTs
Aligned CNTs

Alumina fiber location

Carbon fiber

50 μm
100 μm

Figure 1. Fibers coated with in situ grown radially aligned CNTs. Fuzzy alumina (left)

and carbon fiber (right).

 29

qo

σ0 σ0

x
z

Figure 2. Model nano-engineered composite with representative volume element

(RVE) indicated at the center.

 30

r r
CNTs Carbon Fiber
Carbon Fiber q0
σ0 σ0
θ

2R
2a
x

CNTs Epoxy
q0 Epoxy
Transverse Cross-Section Lf Lf
L L

Longitudinal Cross-Section

Figure 3. RVE of the composite containing a fiber reinforcement coated with radially-

aligned CNTs.
 31

24

20

16
*
σ , Present
∗
σ & τ

*
τ , Present
12 *
σ , Ref. [10]
∗

*
τ , Ref. [10]
8

0
0 1 2 3 4 5 6 7
x/2a

Figure 4. Model validation by comparison to Ref.9 for the case of an isotropic matrix

phase (no CNTs) around the fiber. L f / a =10

 32

20
*
σ , VCNT=0
18
*
σ , VCNT=1.0%
16
*
τ , VCNT=0
14
*
τ , VCNT=1.0%
12
*
σ &τ

10
*

8
6
4
2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/Lf
Figure 5. Variation of normalized axial normal stress in the carbon fiber and

the interface shear stress along the fiber length ( q 0 = 0 ).

 33

14

12

10

8
*
σ

PNC, VCNT=0
4
PNC, VCNT=0.5%
2 PNC, VCNT=1.5%
Isotropic Stiff Matrix
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/Lf
Figure 6. Variation of axial normal stress in the carbon fiber along its length

( q0 = 0 )
 34

20
PNC, VCNT=0
18
PNC, VCNT=0.5%
16
PNC, VCNT=1.5%
14 Isotropic stiff matrix
12
*

10
τ

8
6
4
2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/Lf
Figure 7. Variation of transverse shear stress at the interface between the

matrix and the carbon fiber along the length of the fiber for different CNT

volume fraction and considering isotropic stiffening of the matrix ( q 0 = 0 ).

 35

14

12

10

8
*
σ

6
R/a=3, VCNT=0
4
R/a=3, VCNT=1.0%

2 R/a=4, VCNT=0
R/a=4, VCNT=1.0%
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/Lf
Figure 8. Load transfer in the carbon fiber for different values of R / a
( VCNT = 1.0% , q 0 = 0 )
 36

0.5
R/a=3, Lf/a=10
0.45 R/a=3, Lf/a=15
R/a=4, Lf/a=10
0.4
R/a=4, Lf/a=15

0.35
Lcrit/Lf

0.3

0.25

0.2

0.15
0 0.5 1 1.5 2 2.5 3
VCNT (%)
Figure 9. Effect of CNT volume fraction on the critical length of a fiber for full

load transfer ( q 0 = 0 ).
 37

*
σ
19
*
τ
18
17
max (σ , τ )
*

16
*

15
14
13
12
11
10
0 0.5 1 1.5 2 2.5 3
VCNT (%)

Figure 10. Variation of maximum axial stress and interfacial shear stress on the carbon

fiber ( L _ f / a = 10 ) as a function of CNT volume fraction ( q 0 = 0 ).

 38

20
18
16
14
max (σ , τ )
*

12
*

10

8 *
σ , VCNT=0
6 σ
*
, VCNT=1.0%
4 σ
*
, VCNT=0
2 σ
*
, VCNT=1.0%
0
0 5 10 15 20 25 30 35 40 45 50
Aspect Ratio (Lf/a)
Figure 11. Variation of maximum values of the axial normal stress and the

transverse shear stress at the interface between the matrix and the carbon

fiber with the aspect ratio of the fiber ( q 0 = 0 , VCNT = 1.0% ).

 39

20
*
σ
*
18 τ

16
max (σ , τ )
*
*

14

12

10

8
-0.8 -0.6 -0.4 -0.2 0 0.2
0.4 0.6 0.8
Normalized radial load (q0/σ 0)

Figure 12. Variation of maximum values of the axial normal stress and the transverse

shear stress at the interface between the matrix and the carbon fiber with the applied

radial load ( L f / a = 10 , VCNT = 1.0% ).

 40

14

12

10

8
*
σ

4 VCNT=0
(5, 5) CNT
2 (20, 20) CNT
(50, 50) CNT
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/Lf
Figure 13. Variation of the axial normal stress in the carbon fiber along its

length when the fiber is coated with different arm chair type CNTs ( q 0 = 0 ,

VCNT = 1.0% ).
 41

20
VCNT=0
18
(5, 5) CNT
16 (20, 20) CNT
14 (50, 50) CNT

12
*

10
τ

8
6
4
2
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x/Lf
Figure 14. Variation of the transverse shear stress at the interface between

the matrix and the carbon fiber along its length when the fiber is coated with

different arm chair type CNTs ( q 0 = 0 , VCNT = 1.0% ).

View publication stats