-
EUROPEAN
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
European Polymer Journal 43 (2007) 2836–2847
www.elsevier.com/locate/europolj
POLYMERJOURNAL
Macromolecular Nanotechnology
Rheological and dynamic-mechanical behaviorof carbon
nanotube/vinyl ester–polyester suspensions
and their nanocomposites
A.T. Seyhan a, F.H. Gojny b, M. Tanoğlu a,*, K. Schulte b
a Izmir Institute of Technology (IZTECH), Mechanical Engineering
Department, 35430 Izmir, Turkeyb Polymer Composites, Technical
University Hamburg-Harburg (TUHH), Denickestrasse 15, D-21073
Hamburg, Germany
Received 10 December 2006; received in revised form 16 March
2007; accepted 10 April 2007Available online 1 May 2007
Abstract
Rheological properties of vinyl ester–polyester resin
suspensions containing various amounts (0.05, 0.1 and 0.3 wt.%)
ofmulti walled carbon nanotubes (MWCNT) with and without amine
functional groups (–NH2) were investigated by utili-zation of
oscillatory rheometer with parallel plate geometry. Dispersion of
corresponding carbon nanotubes within theresin blend was
accomplished employing high shear mixing technique (3-roll
milling). Based on the dynamic viscoelasticmeasurements, it was
observed that at 0.3 wt.% of CNT loadings, storage modulus (G 0)
values of suspensions containingMWCNTs and MWCNT–NH2 exhibited
frequency-independent pseudo solid like behavior especially at
lower frequencies.Moreover, the loss modulus (G00) values of the
resin suspensions with respect to frequency were observed to
increase withan increase in contents of CNTs within the resin
blend. In addition, steady shear viscosity measurements implied
that ateach given loading rate, the resin suspensions demonstrated
shear thinning behavior regardless of amine functional groups,while
the neat resin blend was almost the Newtonian fluid. Furthermore,
dynamic mechanical behavior of the nanocom-posites achieved by
polymerizing the resin blend suspensions with MWCNTs and MWCNT–NH2
was investigatedthrough dynamic mechanical thermal analyzer (DMTA).
It was revealed that storage modulus (E 0) and the loss
modulus(E00) values of the resulting nanocomposites increased with
regard to carbon nanotubes incorporated into the resin blend.In
addition, at each given loading rate, nanocomposites containing
MWCNT–NH2 possessed larger loss and storage mod-ulus values as well
as higher glass transition temperatures (Tg) as compared to those
with MWCNTs. These findings wereattributed to evidences for
contribution of amine functional groups to chemical interactions at
the interface between CNTsand the resin blend matrix. Transmission
electron microscopy (TEM) studies performed on the cured resin
samplesapproved that the dispersion state of carbon nanotubes with
and without amine functional groups within the matrix resinblend
was adequate. This implies that 3-roll milling process described
herein is very appropriate technique for blending ofcarbon
nanotubes with a liquid thermoset resin to manufacture
nanocomposites with enhanced final properties.� 2007 Elsevier Ltd.
All rights reserved.
Keywords: Carbon nanotubes; Thermosets; Rheology;
Nanocomposites; Thermal properties; TEM
0014-3057/$ - see front matter � 2007 Elsevier Ltd. All rights
reserved.doi:10.1016/j.eurpolymj.2007.04.022
* Corresponding author. Tel.: +90 232 750 7806/7890.E-mail
address: [email protected] (M. Tanoğlu).
mailto:[email protected]
-
A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847 2837
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
1. Introduction
Addition of nano-scale particles into polymershas been the
recent study of interest in materials sci-ence to develop
functional nanocomposites offeringenhanced properties as compared
to conventionalpolymer based composites that contains
micro-scaleparticulates such as carbon black (CB) or glassmicro
spheres [1–3]. Due to their high aspect ratioas well as excellent
mechanical, thermal and electri-cal properties, CNTs have been
recently consideredas an ideal nano-filler to fabricate various
types ofpolymer based nanocomposites [1–4]. However,the poor
dispersion states of CNTs within the sur-rounding resin matrix due
to their huge surface areaand inert surfaces leads to some
limitations for han-dling and processing [1–5]. Therefore,
achievementof homogeneous dispersion of CNTs within poly-mers is
the key for the realization of the desiredenhancement in the final
properties of nanocompos-ites. The degree of dispersion is commonly
regardedas functions of the size of the dispersed
particles,wettability by disperse medium, and nature of
theattractive forces between the corresponding constit-uents [7].
From that point of view, some physicaland chemical surface
treatments have been appliedfor CNTs to improve their dispersion
state andcompatibility with the surrounding matrix resin[2–4].
Direct mixing and sonication have been themost common techniques to
disperse CNTs withinthermoset polymer resins such as epoxy,
polyesteror vinyl ester [2,3]. However, these methods werefound to
be ineffective to prevent the agglomeratesof CNTs [1–7]. Shear
intensive 3-roll milling processhas been recently applied for the
same purpose.Gojny et al. [4] revealed that double walled
carbonnanotubes (DWCNTs) exhibited adequate disper-sion within the
epoxy resins through 3-roll millingtechnique, based upon
transmission electronmicroscopy (TEM) investigations. In the
samestudy, they also showed that blending of DWCNTswith very low
content (0.1 wt.%) of epoxy resin via3-roll milling process
improved the mechanicalproperties such as Young’s modulus and
strain tofailure ratio of the corresponding epoxy resins.
In addition to microscopic techniques such asTEM, rheological
examination of polymer suspen-sions containing CNTs has been also
recently per-formed to investigate the dispersion state of CNTsin
the corresponding resin media [7–9]. It wasrevealed that the
Brownian motion of CNT particleswith huge aspect ratio results in
more considerable
viscoelastic rheological behavior as compared tothose of
micrometer size short fiber or particles suchas carbon black [7,8].
However, the number ofworks reported in the literature on the
rheologicalbehavior of CNT/polymer systems is quite limited[6–11].
The rheological behavior of concentratedaqueous nanotube
dispersions was studied by Kin-loch et al. [7]. They found that the
dispersion stateof MWCNTs is highly sensitive to the applied
strainin the linear viscoelastic region and the storage andloss
modulus were independent of frequency. Mitch-ell et al. [10]
investigated the linear viscoelasticproperties of polystyrene (PS)
nanocomposites con-taining SWCNTs with and without surface
func-tional groups. They found that the nanotubes withfunctional
groups have better dispersion in PS thanthose without any
functional group. The authorsalso revealed that nanocomposites with
functional-ized nanotubes gave higher storage modulus andcomplex
viscosity values at low frequency level.Kim et al. [8] studied the
rheological behavior ofepoxy resin suspensions containing amine,
acidand plasma treated carbon nanotubes. It was foundthat the
surface modified CNT/epoxy suspensionsexhibited a very strong shear
thinning behaviorand higher shear viscosity than those with
untreatedCNTs due to the enhanced interfacial bondingbetween CNTs
and the corresponding epoxy resin.Song and Youn [9] also performed
a similar studyand claimed that poorly dispersed CNTs withinepoxy
resin leads to higher storage, loss modulusand complex viscosity in
the resulting resin suspen-sions compared to homogenously dispersed
CNTs.
Measuring the temperature dependent materialproperties is also
critical for determination ofappropriate processing conditions to
predict thebehavior of polymeric components during their ser-vice
life [2–12]. The dispersion state of carbon nano-tubes influences
the dynamic mechanical propertiesof the resulting nanocomposites.
Fidelus et al.[12] investigated the dynamic mechanical propertiesof
the epoxy based nanocomposites containingMWCNTs and SWCNTs. They
found an 8%increase of storage modulus in the nanocompositeswith
SWCNTs relative to the neat epoxy, whilethose with MWCNTs showed
almost no substantialchange in the elastic modulus values. They
alsorevealed that glass transition temperature of thenanocomposites
containing SWCNTs or MWCNTsincreases slightly with respect to
nanotube contents.
In this study, the rheological behavior of vinylester/polyester
based suspensions containing
-
2838 A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
MWCNTs and MWCNT–NH2 was determinedmeasuring dynamic
viscoelastic and steady shearproperties of the suspensions via
parallel plate oscil-latory rheometer. Furthermore, the
correspondingCNT/polymer suspensions were cross-linked
usingperoxide (MEKP) initiator. Dynamic mechanicalbehavior of the
resulting nanocomposites was inves-tigated by dynamic mechanical
thermal analyzer(DMTA) to relate the influence of dispersion
stateof CNTs within the blend with final mechanicaland thermal
properties of the corresponding nano-composites. Transmission
electron microscopy(TEM) was employed to investigate the
dispersionstate of CNTs within nanocomposites. The effectof the
amine functional groups on the surfaces ofcarbon nanotubes upon
their dispersion state andchemical interactions at the interface
between CNTsand the resin blend were discussed in details
consid-ering the liquid and solid state of each involved
resinblend, respectively.
2. Experimental
2.1. Materials
Thin multiwalled carbon nanotubes (MWCNTs)and
amino-functionalized multiwalled carbon nano-tubes (MWCNT–NH2) were
obtained from Nano-cyl (Namur Belgium) and used as
nanofillerswithin the corresponding resin blend. CNTs havean
average diameter of 15 nm with a length ofapproximately 50 lm.
Styrene-free polyester (Poliya420) and vinyl ester resin (Polives
701, a bisphenolA epoxy based vinyl ester resin with 35 wt.% of
sty-rene) were obtained from Poliya Polyester, Turkey.Styrene
emission agent BKY 740, purchased fromAlton Chemie, Germany, was
employed to preventthe styrene evaporation from the resin blend
duringthe polymerization reaction. The formulized resinblend was
composed of 25 wt.% of Poliya 420 and75 wt.% of Polives 701. BKY
740 was furtherintroduced to the prepared suspension at a ratioof 1
wt.%. To polymerize each corresponding resinblend suspension,
Cobalt naphtanate (CoNAP)and methyl ethyl kethone peroxide (MEKP)
wereintroduced to the system as accelerator and initia-tor,
respectively.
2.2. Three (3)-roll milling process
Various amounts of MWCNTs and MWCNT–NH2 (0.05, 0.1 and 0.3 wt.%)
were first dispersed
in the styrene free (Poliya 420) polyester resin byshear
intensive blending via 3-roll milling technique.The collected
CNT/polyester resin mixture was sub-sequently blended with vinyl
ester resin at a weightratio of (1/3) by hand mixing for about 10
min fol-lowed by mechanical stirring for about 15 min.After the
addition of the styrene emission agent(BKY 740, 1 wt.%) into the
prepared resin mixture,the CoNAP and MEKP were introduced into
thesystem at a ratio of 0.2 and 1 wt.%, respectively.The resin
suspensions were allowed to cure at roomtemperature followed by
post curing at 75 �C and120 �C for 2 h, respectively. Prior to the
additionof CoNAP and MEKP into the mixture, liquid sam-ples from
each type of the resin suspensions weretaken for the rheological
characterization. Pleasenote that a special resin blend was used to
manufac-ture nanocomposites because of the fact that someserious
problems with commercial unsaturatedpolyester resin. During 3-roll
milling processing,instant styrene evaporation encountered due to
heatevolved on the rolls caused increase of the viscosityof resin.
Therefore, to overcome these difficulties,the CNTs were first
dispersed in specially synthe-sized high viscous styrene-free
polyester resin withthe application of high shear forces between
gapsof the rolls to break up the agglomerates of theCNTs.
2.3. Rheological measurements
A oscillatory rheometer (TA Instruments) withparallel plate
geometry (500 micrometer gap, and50 mm plate diameter) was used to
analyze the vinylester/polyester resin blend suspensions with
differ-ent amounts of carbon nanotubes with and withoutamine
functional groups. Tests were performed inboth dynamic and steady
modes at room tempera-ture in order to avoid extreme styrene
evaporationduring the measurements. All measurements weretaken in
linear viscoelastic region (LVR) in whichthe storage modulus (G 0)
and loss modulus (G00)were independent of strain amplitude. Dynamic
fre-quency sweeps (DFS) were then conducted in theLVR to
investigate the structure of the suspensions.In the DFS, the strain
amplitude was remained con-stant 35% through whole frequency range.
Pleasenote that prior to main experiments, the corre-sponding value
of strain amplitude was proved tobe in the LVR conducting Dynamic
Strain Sweeps(DSS) at a constant frequency. During the DFS,the
frequency varied stepwise from 0.1 to 80 rad/s.
-
A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847 2839
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
Storage modulus (G 0) and loss modulus (G00) valueswere then
measured as a function of frequency. Fur-thermore, steady shear
sweeps (SSS) were employedto investigate the flow properties of the
suspensionsby considering the viscosity as a function of
increas-ing shear rates.
2.4. Dynamic mechanical characterization of the
nanocomposites
Dynamic mechanical properties of the nanocom-posites prepared
after curing of each type of theresin suspension were investigated
by dynamicmechanical thermal analyzer, (DMTA, GABOEPLOXOR 500 N).
For the measurements, rectan-gular specimens of 50 mm in length, 5
mm in widthand 2 mm in thickness were sectioned from largersamples.
The tests were performed in tensile modeat a frequency of 10 Hz
with a static strain of0.6% and dynamic strain of 0.1%, in a
temperaturerange between �50 and 200 �C with a heating rateof 3
�C/min. The storage modulus (E 0), loss modu-lus (E00) and the loss
tangent (tan d) were determinedas a function of temperature.
2.5. Transmission electron microscopy
The transmission electron microscopy (TEM)was conducted to
investigate the dispersion stateof CNTs with and without amine
functional groups
10-2
10-1
100
101
102
10-1 100
Sto
rag
e m
od
ulu
s, G
' [P
a]
Frequency,
Fig. 1. The frequency dependence of the storage modulus (G 0)
for
within the resulting nanocomposites. The TEMimages were taken
using a Philips EM 400 at120 kV. Ultra thin films of each type of
composites(50 nm) were obtained by ultra microtome cutting.
3. Results and discussion
3.1. Rheological behavior of the suspensions
The storage modulus (G 0) of the resin suspen-sions with
different concentrations of MWCNTand MWCNT–NH2 is shown in Figs. 1
and 2,respectively. Consequently, storage modulus valueswere
observed to increase with an increase in theoscillatory frequency.
Moreover, as the concentra-tions of the CNTs were increased in the
correspond-ing resin blend, higher storage modulus values
wereobtained. At 0.3 wt.% CNT contents, the flowregime of the resin
suspensions was significantlyaltered at low frequencies and a
pseudo-solid likebehavior was more visible. We referred this
aspseudo-solid like behavior since in true-solid likebehavior the
storage modulus (G 0) is fully indepen-dent of frequency [11]. In
addition, suspensions with0.1 and 0.3 wt.% of MWCNTs exhibited
slightlyhigher storage modulus values especially at low
fre-quencies as compared to those with MWCNT–NH2.It was also
reported that the CNTs with very smallsizes are very likely to
produce strong particle–poly-mer interactions even at very low
filling rates
101 102
0.1 wt.% MWCNT
Neat resin blend0.05 wt.% MWCNT
0.3 wt.% MWCNT
ω [rad/sec]
the CNT/resin suspensions with various MWCNT contents.
MA
CR
OM
OL
-
10-2
10-1
100
101
102
10-1 100 101 102
Neat resin blend
0.1 wt.% MWCNT-NH20.3 wt.% MWCNT-NH2
0.05 wt.% MWCNT-NH2
Sto
rag
e m
od
ulu
s, G
' [P
a]
Frequency, ω [rad/sec]
Fig. 2. The frequency dependence of the storage modulus (G 0)
for the CNT/resin suspensions with various MWCNT–NH2 contents.
2840 A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
because of the huge interfacial area between the par-ticles and
polymers [8–10]. So, the high aspect ratioand huge surface area of
CNTs raise the storagemodulus of the resin suspensions. Figs. 3 and
4show the loss modulus values of polymer suspen-sions containing
MWCNTs and MWCNT–NH2,respectively. It was observed that the loss
modulusvalues as a function of angular frequency increases
10-1
100
101
102
103
10-1 100
Lo
ss m
od
ulu
s, G
" [M
Pa]
Frequency,
Fig. 3. The frequency dependence of the loss modulus (G00) for
t
with the incorporation of carbon nanotubes. Inaddition, at 0.3
wt.% loading rate, resin suspensioncontaining untreated CNTs
exhibits viscous behav-ior (G00 > G 0), whereas those with amino
functional-ized CNTs shows viscoelasticity (G00 � G 0).
Thisverifies that amine functional groups over CNT sur-faces alter
the flow characteristic of resin suspen-sions as well.
101 102
0.1 wt.% MWCNT
0.05 wt.% MWCNT
Neat resin blend
0.3 wt.% MWCNT
ω [rad/sec]
he CNT/resin suspensions with various MWCNT contents.
-
10-1
100
101
102
103
10-1 100 101 102
0.3 wt.% MWCNT-NH2
0.1 wt.% MWCNT-NH2
0.05 wt.% MWCNT-NH2
Neat resin blend
Lo
ss m
od
ulu
s, G
" [P
a]
Frequency,ω [rad/sec]
Fig. 4. The frequency dependence of the loss modulus (G00) for
the CNT/resin suspensions with various MWCNT–NH2 contents.
A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847 2841
LE
CU
LA
RN
AN
OT
EC
HN
OL
OG
Y
Figs. 5 and 6 give shear viscosity as a function ofshear rate
for the neat resin blend and resin suspen-sions containing MWCNTs
and MWCNT–NH2,respectively. In principle, shear viscosity of a
purepolymer is divided into two distinct regions includ-ing the
Newtonian and shear thinning regions[9–11]. At low shear rates, the
Newtonian regionindependent of shear rate is observed followed
by
10-1
100
101
102
10-1 100 1
Ne
0.0
0.1
0.3
Vis
cosi
ty [
Pa.
s]
Shear r
Fig. 5. Shear rate dependency of viscosity of neat resin b
shear thinning region through which shear viscositylinearly
declines with shear rate. As seen in the fig-ures, shear thinning
behavior was observed for eachresin suspensions with nanotubes such
that the vis-cosity is reduced with an increase in shear
rates.Moreover, the neat resin blend showed almost theNewtonian
fluid behavior. In other words, the New-tonian region disappeared
even if very low content
01 102 103
at resin blend
5 wt.% MWCNT
wt.% MWCNT
wt.% MWCNT
ate [s-1]
lend and CNT/resin suspensions with MWCNTs.
MA
CR
OM
O
-
10-1
100
101
102
10-1 100 101 102 103
Neat resin blend
0.05 wt.% MWCNT-NH2
0.1 wt.% MWCNT-NH2
0.3 wt.% MWCNT-NH2
Vis
cosi
ty [
Pa.
s]
Shear rate [s-1]
Fig. 6. Shear rate dependency of viscosity of resin blend and
CNT/resin suspensions with MWCNT–NH2.
2842 A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
of CNTs (0.05 wt.%) is added into the resin blend.This implies
that strong particle–particle interactionof CNTs is one major
factor that leads to anincrease in shear viscosity of the
corresponding sus-pensions. On the other hand, at 0.1 wt.%
loadingrate, the initial shear viscosity of the resin suspen-sion
with amino-functionalized nanotubes wereslightly lower than those
of the suspensions withnon-functionalized nanotubes. Moreover,
resin sus-pensions containing 0.3 wt.% of MWCNTs andMWCNT–NH2
exhibited similar initial viscosityvalues. So, based on the
experimental findings, itwas observed that amine functional groups
do notsignificantly contribute to enhancement of disper-sion state
of CNTs within the corresponding resinblends. In fact, it is not
easy to claim a single andprecise comment on rheological behavior
of poly-meric suspensions because it is extremely dependon
dispersing state of fillers inside, particle–particleinteractions,
and interaction between particles anddisperse medium (polymer). In
the present case,since CNTs have high aspect ratio, surface areaand
amine functional groups over their surfaces,both particle–polymer
resin blend and particle–par-ticle interactions became more
crucial, as elucidatedin details during analysis of rheological
data above.Please note that the amino-functionalized nano-tubes
used in our experiments were processed inammonia solution via ball
milling process, duringwhich CNTs are broken in length and their
aspect
ratios are partially diminished. According to infor-mation
provided by the manufacturer of CNTs(Nanocyl, Belgium), the aspect
ratio of multi wallednanotubes with amine functional groups is
eventu-ally five times lower than those without any
surfacetreatment. From that point of view, the incorpora-tion of
amino-functionalized nanotubes (MWCNT–NH2) with their reduced
aspect ratio into the resinsystem is expected to result in
relatively lower stor-age and loss modulus values in addition to
reducedshear viscosity in their corresponding resin suspen-sions as
compared to those with MWCNTs.However, in our case, the effect of
the surface func-tionalization on the rheological behavior is
rela-tively insignificant especially at higher loading rateof CNTs
embedded into the resin blend. Thisimplies a conceivable occurrence
of chemical inter-actions between CNTs and disperse medium
(resinblend) through the amine functional groups overthe CNTs
surfaces. So, the effects of interfacialinteractions and
particle–particle interactions viadistinct aspect ratios of CNTs
upon the shear vis-cosity and dynamic rheological properties of
theCNT/resin systems needs further investigations.Therefore, it is
interesting to monitor the thermo-mechanical properties of the
resulting nanocompos-ites (solid state) prepared for each
correspondingresin suspensions containing MWCNTs andMWCNT–NH2 in
order to further follow ourapproach. Furthermore, TEM was also
conducted
-
A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847 2843
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
to characterize the dispersion state of CNTs withand without
amine functional group within thecured resin blends.
3.2. Thermo-mechanical properties of the resulting
nanocomposites
Figs. 7 and 8 give storage modulus (E 0) and lossfactor (tan d)
values as a function of temperatureobtained from dynamic mechanical
measurementsfor the nanocomposites containing MWCNT andMWCNT–NH2,
respectively. The addition of non-functionalized and
amino-functionalized carbonnanotubes into the polymer system has
some con-siderable effects on the storage modulus in boththe glassy
and the rubbery states, depending onthe nanotube contents within
the resin blend. Thisis due to the stiffening effect of CNTs and
interfacialinteractions along a huge interfacial area betweenthe
CNTs and the polymer matrix. Consequently,CNTs reduced the mobility
of the surrounding poly-mer matrix to some extent leading to an
increase inthe modulus values. This effect is more pronouncedin the
glass transition region. As mentioned above,amino-functionalized
nanotubes have shorter length(lower aspect ratios) as compared to
that withoutsurface treatment as a result of the
functionalizationprocess (ball milling process). So, one can expect
rel-atively lower elastic modulus values from the nano-composites
containing MWCNT–NH2 as compared
102
103
104
-50 0 50
Neat resin blend
0.05 wt.% MWCNT
0.1 wt.% MWCNT
0.3 wt.% MWCNT
Sto
rage
mo
du
lus,
E' [
MP
a]
Temperature
Fig. 7. Storage modulus and loss factor of nanocomposites
to those with MWCNTs. However, we obtainedopposing results in
the present case. As an example,the storage modulus values at 20 �C
for the nano-composites containing 0.3 wt.% of MWCNT–NH2and MWCNTs
were found to be 3170 and2930 MPa, respectively, which are also
higher thanthat of the neat resin blend (2430 MPa). Theseresults
revealed that relatively enhanced dispersionstate of the MWCNT–NH2
within the matrix com-pensates the lower aspect ratio of these
tubes andprovides higher modulus values to their
resultingnanocomposites.
In Figs. 9 and 10, the loss modulus (E00) values ofthe
nanocomposites containing MWCNTs andMWCNT–NH2 were given as a
function of temper-ature, respectively. It was found that the loss
mod-ulus values at the peak points gets higher, as thenanotube
contents increases. In addition, nanocom-posites containing
amino-functionalized nanotubesresults in somewhat higher peak
values as comparedto those with non-functionalized nanotubes.
Inbrief, the loss modulus indicates the energy con-verted into heat
and can thus be used as a measure-ment of viscous component or
unrecoverableoscillation energy dissipated per cycle. From
thatpoint of view, we can further conclude that the sat-isfactorily
dispersed nanotubes with and withouttreatment would assist in
dissipating energy undervisco-elastic deformation of the
surrounding resinblend matrix. Moreover, higher loss modulus
values
0.1
1
100 150 200
CFI
Lo
ss f
acto
r ta
nδ
[oC]
containing non-functionalized nanotubes (MWCNT).
MA
CR
OM
OL
-
102
103
104
0.1
1
-50 0 50 100 150 200
Neat resin blend
0.05 wt.% MWCNT-NH2
0.1 wt.% MWCNT-NH2
0.3 wt.% MWCNT-NH2
M
N
O
G
Lo
ss f
acto
r ta
nδ
Temperature [oC]
Sto
rag
e m
od
ulu
s, E
' [M
Pa]
Fig. 8. Storage modulus and loss factor of nanocomposites
containing amino-functionalized nanotubes (MWCNT–NH2).
0
50
100
150
200
250
300
40 60 80 100 120 140
Neat resin blend
0.05 wt.% MWCNT
0.1 wt.% MWCNT
0.3 wt.% MWCNT
Lo
ss m
od
ulu
s, E
" [M
Pa]
Temperature [oC]
Fig. 9. Loss modulus values of nanocomposites containing
non-functionalized nanotubes (MWCNT).
2844 A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
of the nanocomposites with MWCNT–NH2 as com-pared to those with
MWCNTs showed that aminefunctional groups promotes the involved
chemicalinteractions at the interface in some extend, owingto the
same reasons as previously explained.
Fig. 11 shows the glass transition temperature(Tg) values of the
nanocomposites with MWCNTand MWCNT–NH2 obtained from the slope
of
the storage modulus values in the glass transitionzone. As seen
in the figure, the addition of nano-tubes within the resin blends
increased the corre-sponding Tg values significantly. The mobility
ofthe polymer matrix around the nanotubes is reduceddue to the
presence of the nanotubes. In fact, inter-facial strong bonds are
expected to occur betweenthe polymer matrix and the
amino-functionalized
-
0
50
100
150
200
250
300
20 40 60 80 100 120 140
Neat resin blend0.05 wt.% MWCNT-NH2
0.1 wt.% MWCNT-NH20.3 wt.% MWCNT-NH2
Lo
ss m
od
ulu
s, E
" [M
Pa]
Temperature [oC]
Fig. 10. Loss modulus values of nanocomposites containing
amino-functionalized nanotubes (MWCNT–NH2).
86
88
90
92
94
96
98
100
102
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35
MWCNTMWCNT-NH2
Gla
ss t
ran
siti
on t
emp
erat
ure
,Tg
[oC
]
Nanotube content [wt.%]
Fig. 11. Glass transition temperatures of the nanocomposites as
a function of nanotube content with and without any functional
groups.
A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847 2845
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
nanotubes, thus a enhanced bonding at the interfacefurther
increases the Tg. In addition, the reducedaspect ratio of
amino-functionalized nanotubesmay also have some effects on the
degree of poly-merization. Gryschuk et al. [13] stated that
thelower aspect ratio of MWCNTs (achieved by ballmilling process)
would give better results for highlycross-linked thermoset resins
such as vinyl-esters.
3.3. Dispersion state of carbon nanotubes within the
nanocomposites
Figs. 12-a and 12-b are the TEM micrographsshowing achieved
dispersion state of MWCNTand MWCNT–NH2 at 0.3 wt.% loading
withinthe corresponding resin systems, respectively.
Func-tionalized nanotubes exhibited relatively good
-
Fig. 12-a. TEM image showing the dispersion state of MWCNTs (0.3
wt.%) within the vinyl ester/polyester resin matrix. No
condensedagglomerates, but individual nanotubes are noticeable.
Fig. 12-b. TEM image showing the dispersion state of
amino-functionalized multi walled nanotubes (MWCNT–NH2) (0.3 wt.%)
withinthe corresponding vinyl ester/polyester resin matrix. A
better dispersion of functionalized nanotubes within the resin is
visible.
2846 A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
dispersion within the matrix resin blend without anysignificant
agglomerates. The TEM characteriza-tions also support our findings
that amine func-tional groups over CNTs improve their
dispersionstate within the resins enhancing their compatibilitywith
resin blend matrix.
4. Conclusions
In this study, the rheological behavior of thevinyl
ester/polyester resin blend suspensions con-taining different
weight percentage of MWCNTs
and MWCNT–NH2 was investigated. CNTs withand without amine
groups were dispersed withinthe corresponding resin blend using
3-roll millingmethod. These resin suspensions were then
polymer-ized to reveal the effects of amine functional groupsover
the CNTs surfaces on the chemical interactionsat the interface. The
interactions were monitored byconsidering the rheological behavior
(liquid state) incomparison with the properties of their
matchingnanocomposites (solid state). In that manner,dynamic
mechanical behavior of the resulting nano-composites was also
investigated. Linear dynamic
-
A.T. Seyhan et al. / European Polymer Journal 43 (2007)
2836–2847 2847
MA
CR
OM
OL
EC
UL
AR
NA
NO
TE
CH
NO
LO
GY
viscoelastic measurements revealed that storagemodulus (G 0) and
loss modulus (G00) values of resinsuspensions as a function of
angular frequencyincreased with respect to CNTs regardless of
aminefunctional groups. Moreover, as a consequenceof steady shear
viscosity measurements, it wasobserved that each resin suspension
containingnanotubes exhibited shear thinning behavior, whilethe
neat resin blend was almost the Newtonian fluid.On the other hand,
the storage (E 0) and loss modu-lus (E00) values of the resulting
nanocomposites werefound to increase with an increase in contents
ofCNTs. In addition, at 0.3 wt.% loading rate, resinsuspensions
showed pseudo solid like behavior atlower frequency. Furthermore,
the incorporationof nanotubes into the resin blend shifted the
corre-sponding Tg of the nanocomposites to higher valuesas compared
to those of neat resin blend due to thereduced mobility of the
surrounding resin blendmatrix by CNTs. Furthermore, TEM was
con-ducted to highlight the dispersion state of CNTswith and
without amine functional groups withinthe cured resin samples. It
was found that CNTswere dispersed adequately within the resin
blend,which also proved that 3-roll milling techniquewas capable of
providing homogenous dispersionof nano-fillers within thermoset
resins to manufac-ture nanocomposites.
In brief, we found similar rheological propertiesfrom resin
suspensions independent of aminegroups. We also obtained relatively
higher storageand loss modulus values as well as higher glass
tran-sition temperatures from the nanocomposites con-taining
MWCNT–NH2. This is the corroboratingevidence that enhanced
dispersion state of amino-functionalized nanotubes within the resin
blendmatrix compensates their lower aspect ratio pro-moting the
chemical interactions at the interfacebetween CNTs and resin blend
matrix.
Acknowledgements
Authors acknowledge the supports of The Scien-tific and
Technical Research Council of Turkey
(TÜB_ITAK) and JUL_ICH Research Center of Ger-many for the
financial support for Tubitak-Julich 5project.
References
[1] Gojny FH, Schulte K. Functionalization effect on
thethermo-mechanical behavior of multiwalled carbon nano-tube/epoxy
composites. Compos Sci Technol 2004;34:2303–8.
[2] Gojny FH, Wichmann M, Fiedler B, Schulte K. Influence
ofdifferent carbon nanotubes on the mechanical properties ofepoxy
matrix composites-a comparative study. Compos SciTechnol
2005;65:2300–13.
[3] Gojny FH, Nastalczyk J, Roslaniec Z, Schulte K.
Surfacemodified multiwalled carbon nanotubes in
CNT/epoxycomposites. Chem Phys Lett 2003;370:820–4.
[4] Gojny FH, Wichmann M, Köpke U, Fiedler B, Schulte K.Carbon
nanotube-reinforced epoxy composites: enhancedstiffness and
fracture toughness at low nanotube content.Compos Sci Technol
2004;34:2368–71.
[5] Sandler J, Schaffer MSP, Prase T, Bauhofer W, Schulte
K,Windle AH, et al. Polymer 1999;40:5967–71.
[6] Pötsche P, Battachyya AR, Janke A. Melt mixing
ofpolycarbonate with multiwalled carbon nanotubes: micro-scopic
studies on the sate of dispersion. Eur Polym J2004;40:137–48.
[7] Kinloch IA, Roberts SA, Windle AH. A rheological study
ofconcentrated aqueous nanotube dispersions. Polymer
2002;43:7483–91.
[8] Kim JA, Seong DG, Kang TJ, Youn JR. Effects of
surfacemodification on rheological and mechanical properties
ofCNT/epoxy composites. Carbon 2006;44:1898–905.
[9] Song YS, Youn JR. Influence of dispersion sates of
carbonnanotubes on physical properties of epoxy
nanocomposites.Carbon 2005;43:1378–85.
[10] Mitchell CA, Bahr JL, Arepalli S, Tour JM. Dispersion
offunctionalized carbon nanotubes in polystyrene. Macro-molecules
2002;35:8825–30.
[11] Amari T, Uesugi K, Suzuki H. Viscoelastic properties
ofcarbon black suspension as a flocculated percolation system.Prog
Org Coat 1997;31:171–8.
[12] Fidelus JD, Wiesel E, Gojny FH, Schulte K, Wagner
HD.Thermo-mechanical properties of randomly oriented car-bon/epoxy
nanocomposites. Compos Part A 2005;36:1555–1561.
[13] Gryschuck O, Kocsis JK, Thomann R, Kanya Z, Kiricsi
I.Multiwall carbon nanotube modified vinyl ester andvinyl ester –
based hybrid resins. Compos Part: A 2006;37:1252–9.
Rheological and dynamic-mechanical behavior of carbon
nanotube/vinyl ester-polyester suspensions and their
nanocompositesIntroductionExperimentalMaterialsThree (3)-roll
milling processRheological measurementsDynamic mechanical
characterization of the nanocompositesTransmission electron
microscopy
Results and discussionRheological behavior of the
suspensionsThermo-mechanical properties of the resulting
nanocompositesDispersion state of carbon nanotubes within the
nanocomposites
ConclusionsAcknowledgementsReferences