You are on page 1of 7

Composites Science and Technology 78 (2013) 41–47

Contents lists available at SciVerse ScienceDirect

Composites Science and Technology


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

In situ monitoring of through-thickness strain in glass fiber/epoxy composite


laminates using carbon nanotube sensors
Ali Naghashpour ⇑, Suong Van Hoa
Concordia Centre for Composites (CONCOM), Department of Mechanical and Industrial Engineering, Concordia University, 1455 De Maisonneuve Blvd. W., Montreal, Quebec,
Canada H3G 1M8
Center for Applied Research on Polymers and Composites (CREPEC), Canada

a r t i c l e i n f o a b s t r a c t

Article history: The intention of the work presented in this paper was to find a way to measure the through-thickness
Received 1 August 2012 strain (TTS) in composite laminates. The reason for this was because there did not seem to be any effec-
Received in revised form 8 January 2013 tive technique for the measurement of the TTS, particularly for locations away from the free edge. In the
Accepted 22 January 2013
present work, Multiwalled carbon nanotubes (MWCNTs) were added into epoxy resin to make the resin
Available online 4 February 2013
electrically conductive. The modified resin was then incorporated with long glass fibers to make glass/
epoxy laminates. The laminate was subjected to a uniaxial load, while the electrical resistances across
Keywords:
the thickness of the laminate were measured. Two different types of uniaxial loads were applied. One
A. Carbon nanotubes
A. Polymer–matrix composites (PMCs)
was along the length of the sample and the other across the thickness of the sample. For the case where
B. Electrical properties lengthwise load was applied, classical lamination theory (CLT) was used to calculate the average TTS
B. Mechanical properties (ATTS), and to find stacking sequences that can provide the largest ATTS for a certain load. Strain gage
C. Deformation was mounted on the edge of the laminate in order to provide another means to measure the ATTS. It
C. Laminate theory was found that the electrical resistance across the laminate thickness is sensitive to the axial load along
the length of the sample. The magnitude of the change in electrical resistance across the laminate thick-
ness is proportional to the strain measured by strain gage. However, while the strain measured by strain
gage shows negative strain, there was an increase in the through-thickness electrical resistance (TTER).
For the case of thicknesswise load, the strain is negative and the TTER shows decrease. For the thickness-
wise load, the change in TTER can be related to the ATTS. For the lengthwise load, even though the mag-
nitude of the change in TTER is proportional to the ATTS, the change in TTER may not be completely due
to the ATTS.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Significant amount of work on interlaminar (IL) strains and stres-


ses took place in the 1970 [2] and 1980 [3] where analytical and
Fiber reinforced polymer composites (FRPC) are widely used in numerical analyses were used to calculate the values of the IL
many industrial applications due to their high strength-to-weight strains and stresses. Experimental method to measure IL strain is
and stiffness-to-weight ratios [1]. When composite laminates are required. Small strain gages may be used to measure TTS at free
subjected to loading, apart from in-plane strains and stresses, edges. However it is not possible to determine the TTS at locations
through-thickness strains (TTSs) and stresses also exist. The TTSs away from the free edges. Fiber optics may be used to measure in-
occur due to Poisson coupling effect. Around free edges these give plane strains. Recently attempts have also been made to use fiber
rise to interlaminar stresses which can cause delamination failure. optics to measure TTS [4]. However due to the large diameters and
the small values of the TTS, the interference due to the insertion of
the optical fiber may not give accurate strain values. The advent of
Abbreviations: ATTS, average through-thickness strain; CLT, classical lamination polymer nanocomposites where carbon nanotubes (CNTs) are
theory; MWCNTs, multiwalled carbon nanotubes; TTS, through-thickness strain;
added into the epoxy resin making the matrix electrically conduc-
ERC, electrical resistance change; TTER, through-thickness electrical resistance; IL,
interlaminar; Vf, fiber volume fraction; SEM, scanning electron microscopy. tive gives rise to the potential of using this technique to measure
⇑ Corresponding author at: Concordia Centre for Composites (CONCOM), Depart- the TTSs. CNTs are inherently multifunctional materials, they can
ment of Mechanical and Industrial Engineering, Concordia University, 1455 De be used to make sensors due to their small size, high aspect ratios
Maisonneuve Blvd. W., Montreal, Quebec, Canada H3G 1M8. Tel.: +1 514 848 (length to diameter ratio), exceptional electrical conductivity [5],
2424x7228; fax: +1 514 848 3175.
thermal conductivity [6], physical and mechanical properties [7].
E-mail addresses: al_nag@encs.concordia.ca (A. Naghashpour), hoasuon@alcor.-
concordia.ca (S.V. Hoa). Currently many studies have focused on embedding CNTs in

0266-3538/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.compscitech.2013.01.017
42 A. Naghashpour, S.V. Hoa / Composites Science and Technology 78 (2013) 41–47

polymers to make them behave like strain sensors. Subsequently Table 1


failure in polymer matrix composites (PMCs) can be detected using Measured mechanical properties of both glass fiber/epoxy composites unfilled and
filled with 0.3 wt% MWCNTs with 60% fiber volume fraction.
electrical resistance measurement (ERM) [8]. Fiedler et al. [9] first
reported the ability of CNTs as conductive additive for damage Material type E1 (GPa) E2 (GPa) G1(GPa) #12
sensing in composites. Böger et al. [10] added CNTs and carbon Glass/epoxy 47.2 ± 0.8 11.72 ± 0.75 4.1 ± 0.2 0.28 ± 0.01
blacks in the glass/epoxy composites and found that there is corre- Glass/epoxy/0.3 wt% 49.4 ± 1.1 12.4 ± 1.08 4.22 ± 0.33 0.28 ± 0.01
spondence between the change in in-plane strain and the change MWCNTs

in TTER. Thostenson et al. [11–12] dispersed CNTs into epoxy to


make glass/epoxy/CNTs composite samples and showed that the percolation threshold (0.3 wt%) was used to assure consistency
cracking in tensile samples corresponds with the increase in in- in the electrical resistance. The mixture was processed on three roll
plane electrical resistance. Nofar et al. [13] incorporated MWCNTs mill (EXAKT 80E, EXAKT Technologies Inc.) to disperse the
into epoxy to fabricate glass/epoxy/ MWCNTs composite samples MWCNTs within the epoxy matrix. The neat epoxy matrix and
and found that damage in the samples is detected by significant modified epoxy matrix were heated up to 70 °C for 20 min in a vac-
change in in-plane electrical resistance during tensile and fatigue uum oven to remove air bubbles. Unidirectional glass fibers were
testing. They also reported that electrical resistance change (ERC) wetted with the neat and modified epoxy matrix by hand lay-up
is more sensitive than strain gage measurements for damage mon- on the aluminum mold to make glass/epoxy plates and glass/
itoring. Gao et al. [14] deposited CNT onto glass fiber surfaces and epoxy/0.3 wt% MWCNTs plates. The composite plates were cured
showed that epoxy composites made using this fiber system may using an autoclave.
be used for in situ sensing of strain and damage. Alexopoulos
et al. [15] developed fibers consisting of CNTs in a sheath of poly- 2.3. In-plane strain measurements
mer matrix. These fibers are then embedded inside the glass/poly-
mer structure for detecting mechanical deformation using the Conventional metallic strain gages (Vishay micro measure-
change in electrical resistance of the special fiber. Hena-Zamal ments C2A-06-125LT-350) from Intertechnology were used to
and Hoa [16] embedded MWCNTs in the glass/epoxy composite measure in-plane strain for characterizing mechanical properties.
and indicated that there is correspondence between the change
in TTER and damage accumulation during fatigue testing. The 2.4. Mechanical characterization
above works illustrate very interesting attempts to monitor in-
plane strain and damage in composite materials using CNT sensors. The tensile and shear properties for glass /epoxy composite
The problem of through-thickness stresses and strains and sub- laminates unfilled and filled with 0.3 wt% MWCNTs were experi-
sequent effect of delamination was first investigated rigorously in mentally measured in both longitudinal and transverse directions
the 1980s. Pipes et al. [2,17] carried out numerical calculations for in accordance with ASTM D 3039M-07 and ASTM D3518M respec-
laminates subjected to uniaxial displacement and showed that tively. The mechanical tests were performed with the standard
interlaminar strains and stresses exist at the free edges of the lami- head displacement rate of 1.27 mm/min at room temperature. Five
nate. A review on the interlaminar stresses in composites was given samples were tested to determine each mechanical property and
by Salamon [3]. Murthy and Chamis [18] showed that free edge the average value was made over five measurements. The mea-
delamination depends not only on the lay-up sequence but also on sured properties for both composites (without and with MWCNTs)
the size of the component, and the loading conditions. Moire inter- with 60% fiber volume fraction (Vf) are shown in Table 1. From the
ferometry was used to measure the TTS at the edge of the laminates experimental results, it is observed that the mechanical properties
[19–21]. Even though optical fibers were used to detect the trans- of glass/epoxy composite are enhanced a little (but not signifi-
verse strain in laminates [4], the diameter of the fiber is still large cantly) by adding 0.3 wt% MWCNTs.
(52 micrometers, which is more than 6 times the diameter of a car-
bon fiber, and about half the thickness of ply). Moire interferometry 3. Analysis
can be used to measure the TTS at the edge of the laminate, but it
cannot be used to determine the TTS at locations away from the lam- When a composite laminate is subjected to uniaxial tension, its
inate edge. Here, we attempted to find a new method for in situ thickness changes due to Poisson coupling. This gives rise to the
monitoring of the average through-thickness strain (ATTS). This is TTS. The magnitude of the TTS depends on the stacking sequence.
done by incorporating MWCNTs to glass fibers/epoxy composite. It was desirable to have the largest TTS for a certain uniaxial stress.
For this, classical laminate theory (CLT) [23] was used to determine
the laminate stacking sequence that gives rise to maximum TTS.
2. Experimental characterization and materials

3.1. Formulation of ATTS for symmetric laminate subjected to in-plane


2.1. Materials
loads
MWCNTs with 95% of purity, diameters of 2–20 nm and lengths
When a composite laminate is subjected to in-plane loads,
of 1 to more than 10 lm were purchased from Bayer Material Sci-
plane stress assumption in CLT states that the stresses r3 ; s13 ; s23
ence. Unidirectional S-glass fibers purchased from ACP Composites
in the principal material coordinate system are small and negligi-
Company with thickness of 0.127 mm, Epon 862 and EPIKURE W
ble. Therefore, the three-dimensional stress–strain relation in
purchased from Miller-Stephenson Chemical Company were uti-
plane stress condition for a ply is simplified to the following:
lized as reinforcement, epoxy resin and curing agent respectively.
8
>
>
e1 9>
>
2
S11 S12 S13 0 0 0
38
>
>
r1 9>
>
>
> >
> 6S 7>
> >
2.2. Fabrication of composite laminates >
>
>
e2 >>
> 6 12 S22 S23 0 0 0 7> >
>
r2 >>
>
>
>
< > 6 7> >
e3 = 6 S13 S23 S33 0 0 0 7< 0 =
¼6
6
7 ð1Þ
The epoxy resin and curing agent (26.4 wt%) were first mixed. >
> >
> c23 >
> 6 0 0 0 S44 0 0 7 7>
> 0 >
> >
>
Then 0.3 wt% MWCNTs were added into epoxy matrix (the >
> >
> 6 7>
> >
>
>
> c13
>
>
>
40 0 0 0 S55 0 5>
>
> 0 >>
>
>
percolation threshold of the MWCNTs in epoxy was determined : ; : ;
experimentally to be 0.1883 wt% [22]). A concentration more than
c12 0 0 0 0 0 S66 s12
A. Naghashpour, S.V. Hoa / Composites Science and Technology 78 (2013) 41–47 43

From Eq. (1), through-thickness strains e3 ; c13 ; c23 are given by


e3 ¼ S13 r1 þ S23 r2 ; c13 ¼ 0; c23 ¼ 0 ð2Þ
Eq. (2) shows that the extensional strain is not zero and can be com-
puted directly from the in-plane stresses r1 and r2 and the compli-
ances S13 and S23. Because e3 is not zero, a layer in a state of plane
stress experiences a change in thickness, Dh, which is expressed by:
Dh ¼ e3 h ð3Þ
h is the thickness of the layer. Eqs. (2) and (3) for e3 and Dh are
independent of whether the layer is within a laminate or is isolated
by itself. In fact, in order to indicate that the above expressions for
e3 and Dh are valid for any layer, specifically the kth layer, the
expressions can be rewritten as:
e3k ¼ S13 r1k þ S23 r2k ; Dhk ¼ e3k hk ð4Þ
Hence, by computing the thickness change of each layer, the to-
Fig. 1. Analysis results for [05, hn (n=6,. . .,10), 05] laminate under Nx = 43.307 KN.
tal change in laminate thickness can be calculated by summing the
various layer thickness changes. Therefore, the total thickness 8 9 2 2 38 9
change of an N-layer laminate (DH) is expressed by: < r1 >
> = m n2 2mn < rx >
> =
r2 ¼ 64 n2
7
m2 2mn 5 ry ð12Þ
X
N >
: > > >
DH ¼ Dhk ð5Þ s12 ; mn mn m2  n2
:
sxy ;
k¼1
where m ¼ cos h and n ¼ sin h, finally the ATTS is calculated by
ATTS for an N layer laminate ez can be calculated by summing the substituting the obtained on-axis stresses for each layer from Eq.
change in the thicknesses of the individual layers divided by the ori- (12) into Eq. (6).
ginal thickness of the laminate (H) as:
3.2. Determination of the lay-up with large theoretical ATTS
DH 1 X N
ez ¼ ¼ ðS13 r1k þ S23 r2k Þhk ð6Þ
H H k¼1
In order to select a laminate that exhibits large ATTS when it is
where subjected to in-plane loads for experimentation, using values from
Table 1, the above formulation was used. The thickness of each
X
N
layer of the laminate is assumed to be 0.12 mm for the analysis
H¼ hk ð7Þ
(verified experimentally). A computer program was written to cal-
k¼1
culate ATTS. It should be noted that #23 = 0.428 and G23 = 3.28
From Eq. (4), it is necessary to have the on-axis stresses on each (GPa) [23] are assumed for the analysis. Fig. 1 shows variations
layer to calculate ATTS. The laminate constitutive relations for a of ATTS for the [05, hn(n = 6, . . . , 10), 05] composite laminate with fiber
symmetric laminate subjected to only in-plane loads Nx, Ny, Nxy, in orientations (h) ranging from 0° to 90°. Note that the selected lam-
the absence of thermal and moisture effects can be written as: inate should be sufficiently thick so that strain gage can be placed
8 9 8  9 on its edge to measure TTS. It is shown that the magnitude of ATTS
< Nx >
> = < ex >
> =
increases gradually with increase in the degree of fiber orientations
Ny ¼ ½A ey ð8Þ
>
: >
; >
: c >; (h). Also the magnitude of ATTS decreases as the number of layers
Nxy xy in the composite laminate increases. The largest ATTS is achieved
Off-axis strains are achieved by the following equation: with h = 90°. The [05, 908, 05] composite laminate was selected
8 9 8 9 since it can give large ATTS and at the same time has adequate
>
< ex >
= < Nx >
> = thickness for strain gage application.
ey ¼ ½a Ny ð9Þ
>
: c >; >
: >
;
xy Nxy 4. Experimental methods

where [a]= [A]1 and [A] is extensional stiffness of the laminate and
4.1. Fabrication of the [05, 908, 05] laminate
Aij’s are defined as:
X
N Two plates were made. Fiber volume fractions of plates 1 and 2
Aij ¼  ij ðzk  zk1 Þ
Q ð10Þ are calculated to be about 60% and 66% respectively. ASTM stan-
k¼1
dard D 3039M-07 was used. Average thickness of samples cut from
where Q  ij refers to off-axis reduced stiffnesses. By obtaining the plates 1 and 2 are shown in Table 2. The average thickness of
mid-plane off-axis strains from Eq. (9), the off-axis stresses of each
layer are calculated as:
8 9 Table 2
2 38 9
 11  12  16 > ex >
< rx >
> = Q Q Q < = Average thickness of samples.
6  26 7
5 ey
 22 
ry ¼ 4Q 12 Q Q ð11Þ Plate number Vf (%) Samples Average thickness (mm)
>
: > > >
sxy ;  16
Q  26
Q  66 : cxy ;
Q [05, 908, 05] plate 1 Vf = 60 Sample 1 2.16 ± 0.045
Sample 2 2.09 ± 0.081
The on-axis stresses or stresses in the principal material coordinate Sample 3 2.12 ± 0.064
system can be obtained by a transformation on off-axis stresses Eq. [05, 908, 05] plate 2 Vf = 66 Sample A 2.02 ± 0.041
(11): Sample B 1.98 ± 0.077
Sample C 1.96 ± 0.053
44 A. Naghashpour, S.V. Hoa / Composites Science and Technology 78 (2013) 41–47

samples is determined by making an average over fifteen measure- 4.4. Average through-thickness strain (ATTS) measurement
ments for each sample. Quality controls of the samples were car-
ried out using Hitachi S-4700 Scanning electron microscopy ATTS measurement was performed using conventional metallic
(SEM). Fig. 2 shows a sample SEM of the edge of the laminate. strain gage (L1E-350K-PC06-LE) purchased from MFL Company
The SEM image is taken at the location where the strain gage [24]. The length of the copper wire in the strain gage is 0.74 mm.
was mounted, and after loading. Fig. 2 indicates that no crack exists However there is a backing required for bonding with the sub-
which reveals high quality laminate. strate. The length of the backing is 1.9 mm (the thickness of the
sample was about 2.16 mm, which is adequate). The accuracy of
4.2. Arrangement of electrical connections and strain gage these gages is 3%, as provided from the manufacturer. As such
the strain gage only measures the strain over the 90° layers (which
Tabs were bonded at both ends of the tensile sample to prevent is about 0.96 mm thick), and not over the whole thickness of the
the gripping damage and slipping while performing the tensile composite sample (Fig. 3).
test. Six conductive electrical contact points made from silver-
epoxy paste were mounted on the top (T) and bottom (B) surfaces
of the tensile sample. Then electrical wires were attached with sil- 5. Tensile and compression tests
ver-epoxy paste on both surfaces of the sample to make electrodes
for electrical resistance measurement. For measuring ATTS, a small The tensile sample was mounted in the grips of a 100 KN-MTS
strain gage (1.9 mm long) was bonded on the thickness section of mechanical testing machine and monotonically loaded in tension,
the sample. The dimensions and arrangements of electrical con- to a maximum value of 1100 N. Tensile tests were performed for
nections and bonded strain gage on the tensile sample are shown six samples with a rate of 0.127 mm/min while monitoring TTERs
in Fig. 3. and ATTS simultaneously.

4.3. Electrical resistance measurements 5.1. Results and discussion

TTER measurements were carried out by two-probe method 5.1.1. Tensile load (x direction) while measuring strain and resistance
using Agilent digital Multimeter (34401A). Electrical resistance (z direction)
change (ERC) is defined by the following equation: TTERs and ATTS of the [05, 908, 05] sample with dimensions
(250 mm  25 mm  2.16 mm) and (Vf = 60%) were measured
ERC ¼ DR=Ri ¼ ðRf  Ri Þ=Ri ð13Þ
simultaneously while tensile load was applied. The ATTS over 90°
where Ri and Rf are the TTER before and while loading respectively. layers localized at the center of the laminate was measured using

Fig. 2. SEM of laminate section.

Fig. 3. Dimensions and locations of strain gage and electrical contact points.
A. Naghashpour, S.V. Hoa / Composites Science and Technology 78 (2013) 41–47 45

strain gage and was calculated from analysis based on CLT [23]. L
R¼q ð14Þ
These are plotted in Fig. 4. Good correspondence is obtained. A
Since strain gage is not able to measure ATTS of the whole thick-
where q is electrical resistivity, L is the distance over which the
ness of laminate edge due to its gage length limitation while the
electrical resistance is determined and A is the cross sectional area.
TTER measurement is performed over the whole thickness of lam-
In the case of electrical resistance across the thickness of the
inate, analysis was done to calculate ATTS for the total thickness
laminate, L is thickness t. One can assume that the cross section
based on CLT. The change in electrical resistance was calculated
A does not change during tensile loading in the plane of the sample.
based on Eq. (13). As can be seen in Fig. 5a, the magnitude of ATTS
If the resistivity does not change, then a reduction in t would result
obtained from analysis and ERC measured for the entire laminate
in a reduction in R. The fraction of the change in electrical resis-
thickness increase as load increases. However their signs are differ-
tance is given as:
ent. Strains are negative while electrical resistance changes are po-
sitive. When the laminate is subjected to tensile in-plane loads, the DR R 2  R 1 t 2  t 1
¼ ¼ ¼e ð15Þ
laminate thickness is expected to reduce due to Poisson coupling R R1 t1
effect. This is reflected in the negative strains. Normally, one would
where e is the ATTS.
expect that smaller thickness would result in smaller electrical
Fig. 5 shows the ATTS of about 227 le (2.27  104) at a load
resistance (or negative change in electrical resistance). However
of 1100 N. At the same time, the ratio of the change in electrical
this is not what is observed. A closer examination is necessary.
resistance (DR/R) at 1100 N in Fig. 5 is about +0.007 (7  103).
From Ohm’s law, the electrical resistance R is given by the
The ratio between the TTER change and ATTS is 30.8. This indi-
equation:
cates that TTER is more sensitive than ATTS. However, the sign of
the ratio is negative which means that TTER increases while the
thickness decreases. As such TTER change may not be due only
0 300 600 900 1200
to the reduction in thickness. In order to examine, further com-
Average through-thickness strain (Micro)

0 pression test was done.

-50 5.1.2. Compressive out-of-plane load (z direction) while measuring


strain and electrical resistance (z direction)
In order to check to see if the reduction in sample thickness may
-100
give rise to the change in electrical resistance, an experiment was
carried out to investigate the effect of compressive out-of-plane
-150 load on electrical resistance in thickness direction of the laminate.
The details of the experimental set up are schematically illustrated
-200
in Fig. 6. An L-shaped piece made of electrically non-conducting
glass fiber/epoxy composite was used to support and to insulate
the electrically conductive laminate from metallic grip and also
-250 to prevent bending of the laminate. The laminate was subjected
Load (N)
to compresion in thickness direction while average through-thick-
Measured strain using strain gage over 90 degree layers of the laminate
Calculated strain from analysis (CLT) for 90 degree layers of the laminate
ness strain (ATTS) and through-thickness electrical resistances
(TTERs) at three different locations were measured simultaneously.
Fig. 4. Tensile result, showing ATTS over 90° layers of [05, 908, 05] versus load Fig. 7a and b show the effect of out-of-plane load and strain on
(sample 1). change in TTERs. Fig. 7a indicates relation between TTER versus

0.008
ΔR/Ri)
Electrical Resistance Change (Δ

0.006

0.004

0.002

0
-250 -200 -150 -100 -50 0
Average through-thickness strain (Micro strain)
Experimental Data (ED) Linear fit ED

(a) (b)
Fig. 5. (a) Calculated ATTS and measured ERC versus load for the total thickness. (b) Measured ERC versus calculated ATTS for the total thickness.
46 A. Naghashpour, S.V. Hoa / Composites Science and Technology 78 (2013) 41–47

Fig. 6. Transverse compression in the thickness direction: (a) Sample placed on MTS machine with load cell to apply out-of-plane load. (b) Close-up view of L-shaped support,
electrodes and strain gage bonded to sample and location of loading rod with diameter of 12.7 mm.

Out-of-plane Load (N) Average through-thickness strain (Micro strain)


-1100 -900 -700 -500 -300 -100 -70 -60 -50 -40 -30 -20 -10 0
0 0

Electrical Resistance Change (ΔR/Ri)


Electrical Resistance Change (ΔR/Ri)

-0.004 -0.004

-0.008 -0.008

-0.012 -0.012

-0.016 -0.016

-0.02 -0.02

-0.024 -0.024
Electrodes T1-B1 at edge Electrodes T2-B2 at middle Electrodes T1-B1 at edge Electrodes T2-B2 at middle

Electrodes T3-B3 at middle Electrodes T3-B3 at middle

(a) (b)
Fig. 7. Out-of-plane compression test results. (a) Change in TTERs at edge and middle of the laminate versus comparesive out-of-plane load. (b) Change in TTERs versus
measured ATTS. Initial electrical resistance at the electrodes: T1-B1 (0.169 M-X), T2-B2 (0.167 M-X), T3-B3 (0.168 M-X).

load, while Fig. 7b shows the correlation between TTER and mea-
sured ATTS. It is observed that compression load in the thickness
direction produces negative strains and negative TTERs. Values of
the TTER are different at three different locations. This is due to
the variation in the proximity of the probes to the location of the
loading area. The above indicates that a reduction in thickness of
the sample results in reduction in through-thickness electrical
resistance (TTER). At the same time, a tensile load in the in-plane
direction results in an increase in TTER. Even though a tensile load
in the in-plane direction can produce a reduction in thickness of
the sample due to Poisson coupling effect, the increase in TTER
does not correspond to this reduction in thickness.
Fig. 8a illustrates the hypothesis that the application of a tensile
load along the x direction would produce a reduction in the thick-
ness (z direction) due to Poisson coupling. This reduction in thick-
ness gives rise to negative average through-thickness strain (ATTS).
This reduction in thickness was thought to reduce the distance be-
tween the crossing CNTs. This in turn would reduce the tunneling
resistance between the crossing CNTs and should reduce the
through-thickness electrical resistance (TTER). However an in-
Fig. 8. Poisson coupling reduces the laminate thickness. schematic illustrating (a)
crease in TTER was observed. One explanation for this increase in reduction of the tunneling distance between crossing CNTs (a1 and a2 before and
TTER can be the separation of the CNTs due to lengthwise load as after loading) and (b) one possible explanation for the increase in TTER (b1 and b2
shown in Fig. 8b. before and after loading).
A. Naghashpour, S.V. Hoa / Composites Science and Technology 78 (2013) 41–47 47

6. Conclusions [8] Li C, Thostenson ET, Chou TW. Sensors and actuators based on carbon
nanotubes and their composites: a review. Compos Sci Technol
2008;68:1227–49.
In this work, a new method was attempted for in situ monitor- [9] Fiedler B, Gojny FH, Wichmann MHG, Bauhofer W, Schulte K. Can carbon
ing of ATTS in glass fiber/epoxy composite laminates using nanotubes be used to sense damage in composites? Ann Chim Sci Matér
2004;22:81–94.
MWCNTs. It was observed that the through-thickness electrical
[10] Böger L, Wichmann MHG, Meyer LO, Schulte K. Load and health monitoring in
changes (TTER) are proportional to the change in average glass fibre reinforced composites with an electrically conductive
through-thickness strains (ATTS). As such, the addition of CNTs in nanocomposite epoxy matrix. Compos Sci Technol 2008;68:1886–94.
[11] Thostenson ET, Chou TW. Carbon nanotube networks: sensing of distributed
glass fibers/epoxy composite laminates and measuring the TTER
strain and damage for life prediction and self-healing. Adv Mater
may be used as a new method for the measurement of the 2006;18:2837–41.
through-thickness strains, particularly at locations away from the [12] Thostenson ET, Chou TW. Real-time in situ sensing of damage evaluation in
free edges of the sample. However the reduction in the through- advanced fiber composites using carbon nanotubes networks. Nanotechnology
2008;19:215713.
thickness electrical resistance due to lengthwise loads requires [13] Nofar M, Hoa SV, Pugh MD. Failure detection and monitoring in polymer
more work to explain. matrix composites subjected to static and dynamic loads using carbon
nanotube networks. Compos Sci Technol 2009;69:1599–606.
[14] Gao SL, Zhuang RC, Zhang J, Liu JW, Mäder E. Glass fibers with carbon nanotube
Acknowledgments networks as multifunctional sensors. Adv Funct Mater 2010;20:1885–93.
[15] Alexopoulos ND, Bartholome C, Poulin P, Marioli-Riga Z. Structural health
This research is supported by the Natural Sciences and Engi- monitoring of glass fiber reinforced composites using embedded carbon
nanotube (CNT) fibers. Compos Sci Technol 2010;70(2):260–71.
neering Research Council of Canada (NSERC) and Concordia Center [16] Hena-Zamal H, Hoa SV. Fatigue damage behavior of glass/epoxy composite
for Composites (CONCOM). using carbon nanotubes as sensors. In: 26thICAF Symposium, Montréal; 2011.
[17] Whitney JM, Daniel IM, Pipes RB. Experimental mechanics of fiber reinforced
composite materials. Soc Exp Mech Monogr 1984:4.
References
[18] Murthy PLN, Chamis CC. Free edge delamination: laminate width and loading
conditions effects. NASA Technical memorandum 100238; 1987.
[1] Chou TW. Microstructural design of fiber composites. Cambridge [19] Guo Y, Post D, Han B. Thick composites in compression: an experimental study
(UK): Cambridge University Press; 1992. of micromechanical behavior of smeared engineering properties. J Compos
[2] Pipes RB, Pagano NJ. Interlaminar stresses in composite laminates under Mater 1992;26(13):1930–44.
uniform axial extension. J Compos Mater 1970;4:538–48. [20] Ifju PG, Masters JE, Jackson WC. The use of Moiré interferometry as an aid to
[3] Salamon NJ. An assessment of the interlaminar stress problem in laminated standard test method development for textile composite materials. Compos
composites. J Compos Mater 1980;14:177–94. Sci Technol 1995;53(2):155–63.
[4] Takeda S, Okabe Y, Yamamoto T, Takeda N. Detection of edge delamination in [21] Nailadi CL, Adams DF, Adams DO. An experimental and numerical
CFRP laminates under cyclic loading using small diameter FBG sensors. investigation of the free edge problem in composite laminates. J Reinf Plast
Compos Sci Technol 2003;63:1885–94. Compos 2002;21(1):3–39.
[5] Ebbesen TW, Lezec HJ, Hiura H, Bennett JW, Ghaemi HF, Thio T. Electrical [22] Rosca ID, Hoa SV. Highly conductive multiwall carbon nanotube and epoxy
conductivity of individual carbon nanotubes. Nature 1996;382(6586):54–6. composites produced by three-roll milling. Carbon 2009;47(8):1958–68.
[6] Berber S, Kwon YK, Tomanek D. Unusually high thermal conductivity of carbon [23] Hyer MW. Stress analysis of fiber-reinforced composite materials. Destech
nanotubes. Phys Rev Lett 2000;84(20):4613–6. Publications; 2009.
[7] Ruoff RS, Lorents DC. Mechanical and thermal-properties of carbon nanotubes. [24] http://www.mflstraingauges.com/index_gauge_types.html.
Carbon 1995;33(7):925–30.

You might also like