processing of poly(propylene)/carbon nanotube composites using scco2-assisted mixing
TRANSCRIPT
Full Paper
566
Processing of Poly(propylene)/Carbon NanotubeComposites using scCO2-Assisted Mixing
Jia Ma, Hua Deng, Ton Peijs*
scCO2 was used to assist in the preparation of PP/CNT composites. Two types of CNTs wereused: MWNTs with and without HDPE coating (cMWNTs). The morphology of the nanocom-posites and their mechanical and thermal properties were investigated and compared withsamples made by traditional melt compounding.The use of cMWNT leads to better dispersion andproperties in melt-compounded nanocomposites.For systems prepared using scCO2-assisted mix-ing, however, better properties were obtainedusing pristine MWNTs, avoiding the additionalcosts of nanotube modification. It was also shownthat observed improvements in the mechanicalproperties for these materials were due to a com-bination of matrix modification and nanotubereinforcement, rather than a reinforcement effectcaused solely by MWNTs.
Introduction
Undoubtedly, polymer nanocomposites have established
themselves now as a new class of composite materials.
Because of their high intrinsic mechanical properties,
nanoscale dimensions and high aspect-ratio, small
amounts of nanofillers such as nanoclays or carbon
J. Ma, Dr. H. Deng, Prof. T. PeijsSchool of Engineering and Materials Science, and Center forMaterials Research, Queen Mary University of London, Mile EndRoad, London E1 4NS, UKE-mail: [email protected]. H. DengCollege of Polymer Science and Engineering, Sichuan University,State Key Laboratory of Polymer Materials Engineering, Chengdu,610065 Sichuan, ChinaProf. T. PeijsEindhoven Polymer Laboratories, Eindhoven University ofTechnology, PO Box 513, 5600 MB Eindhoven, The Netherlands
Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
nanotubes (CNTs) canprovide the resultingnanocomposite
materialwith significantly improved properties. One of the
most promising areas of nanocomposite research involves
the use of CNTs as nanofillers.[1] CNTs are considered to be
ideal candidates for a wide range of applications in
materials science because of their exceptional mechanical,
thermal, and electronic properties.[2] In particular, their
exploitation as conductive fillers for polymer composites
has shownsignificant success.[3–7] The effectiveuse of CNTs
as reinforcements for polymer matrices still presents some
major difficulties.[8–10] The key challenge still remains in
breaking down bundles of aggregated CNTs and reaching a
fine dispersion in the selected polymer matrix. Aggregates
of nanotubes in polymer composites may act as imperfec-
tions and restrain the reinforcing potential of the CNTs.
Undoubtedly, nanofiller dispersion represents the key-step
for producing high-performance polymer nanocompo-
sites.[8] A number of techniques have been employed to
achieve better CNT dispersion in polymer matrices,
including sonication, covalent functionalization, and
DOI: 10.1002/mame.200900405
Processing of Poly(propylene)/Carbon Nanotube . . .
non-covalent functionalization.[3,9,11–13] Often, these mod-
ifications have been successful in combination with
solution or latex-based processes.[3,11] However, difficulty
in achieving good dispersions in polymer melts remains.
For this reason, recently, dissociated CNTs have been
produced by grafting polymer chains directly onto CNTs in
order to achieve a homogeneous polymer coating on the
CNT surface.[14–16] Owing to the de-aggregation of these
CNTs, enhanced dispersion and improved properties, by
comparison with the direct incorporation of CNTs in
polymer melts, have been reported for high-density
polyethylene (HDPE)-coated multi-walled nanotubes
(MWNTs) in thermoplasticmatrices such as ethylene/vinyl
acetate (EVA), polycarbonate (PC), polyamide (PA) and poly
(propylene) (PP).[14–18] Nanocomposites can be prepared
by different methods such as melt compounding, solution,
and dispersion blending and in situ polymerization. Melt
compounding is themost commonmethod used to process
thermoplastic polymer/CNTnanocomposites because it is a
cost-effective technology for polyolefin-based polymer
systems and is most compatible with current industrial
practices.[19]However,melt compounding, especially in the
case of polyolefin matrices, is generally less effective at
dispersing CNTs and is limited to low nanofiller loadings.
This is due to the high viscosity of the composite systems
causedby theaddition of CNTs.[18]Moreover, thehigh shear
rates and high temperatures employed can also cause
thermal instability of the molten polymers.[20] One
approach to promote compatibility includes the use of
polymer-coated CNTs instead of pristine CNTs in polymer
melts or masterbatch processes, in which a pre-mixed
highly loaded CNT composite is dilutedwith fresh polymer
melt.[21]Considering the high viscosities and high tempera-
tures involved in melt compounding processes, super-
critical CO2 (scCO2) processing could be an interesting
alternative as it may overcome some of these issues. scCO2
processing, as one of the new and cleaner processing
methods for polymer nanocomposites, has recently
received increasing attention.[22,23] The use of scCO2 can
induce plasticization of the polymer which decreases the
glass transition temperature (Tg) and viscosity of polymer
melts, because the scCO2 can swell the polymer, causing
an increase in free volume and a reduction in chain
entanglements. Although the role of scCO2 has not been
systematically investigated in the melt compounding of
nanocomposites, earlier studies have indicated clear
improvements in the dispersion of nanoclays in PP
matrices.[24,25] Because of PP’s good balance between
properties and cost and its wide usage in industry, PP/
CNT nanocomposites have been extensively studied in
recentyears forenhancedmechanicalproperties.[4,6,18,19,26–32]
Andrews et al.[19] fabricated PP/MWNT composites by a
shear mixer. They found a modulus increase from 1.0 to
2.4GPawith a relatively high nanotube content of 12.5wt.-
Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
%but at the expense of a reduction in yield stress from30 to
18MPa. Manchado et al.[31] reported a small modulus
increase from 0.85 to 1.19GPa with a nanotube content of
0.75wt.-% and a strength increase from 31 to 36MPwith a
0.5wt.-% nanotube content for PP/single-walled nanotube
(SWNT) composites. Most recently, Deng et al.[18] investi-
gated the effects of a HDPE coating onto MWNTs on the
mechanical properties of PP/MWNT composites produced
by melt compounding. They observed a property increase
from 1.4 to 1.8GPa for Young’s modulus and from 34 to
38MPa for tensile strength at relatively low loadings
(0.5wt.-%) of coatedMWNTs. Some studies such as those by
Ganb et al.[32] did not show any increase in mechanical
properties. Overall, previous research activities on PP/CNT
bulk composites have shown that mechanical property
improvements by CNTs are moderate with the more
effective reinforcing effects observed at lower nanotube
contents. In thiswork, scCO2wasused topreparePP/MWNT
nanocomposites and the potential benefits of using scCO2
were investigated. PristineMWNTs aswell as HDPE-coated
MWNTs were evaluated and the dispersion, physical and
mechanical properties of nanocomposites obtained using
scCO2 processing were compared with those of traditional
melt compounded samples. Micromechanical modeling
was used to analyze the effective properties of the CNTs in
the nanocomposites. The current study also addressed
effects of PP matrix modification by the CNTs and its effect
on mechanical properties of the nanocomposites.
Experimental Part
Materials and Equipment
The PP used was Moplen1 HP500H from Basell [melt flow
index¼ 1.8 g � (10min)�1]. The pristine MWNTs used were Nano-
cyl-7000 (Nanocyl S.A., Belgium), which is a thin MWNT produced
via catalytic chemical vapor deposition (CCVD), achieving a purity
of 90% carbon. Nanocyl-9000 is an HDPE-coated MWNT (cMWNT)
which contains 31.6% of nanotubes pre-dispersed in a polymer
carrierusingaproprietary technology.[16,33]Allmaterialswereused
as received.A liquidwithdrawalCO2cylinderat5MPapressurewas
supplied by BOC Gases. A custom-built stirred high-pressure
autoclave with paddle type stirrer was used to prepare the
nanocomposites.ARondol laboratorybenchtophot-presswasused
to produce the dumbbell-shaped specimens.
Processing of Nanocomposites
MWNTs (or cMWNTs) were directly added to the PP matrix in an
autoclavetoobtainaCNTconcentrationof0.1,0.25,0.5and1.0wt.-%.
CO2waspumped into the autoclavevia an Iscomodel 260Dsyringe
pump after being chilled to �6 8C. The autoclave was filled with
liquid CO2 and held at 15MPa and 200 8C under stirring, using a
pitchedblade turbine impeller for 30min. At the end ofmixing, the
autoclave was cooled in water to room temperature, and then CO2
www.mme-journal.de 567
J. Ma, H. Deng, T. Peijs
Table 1. Composition of PP/CNT nanocomposites.
Sample CNT type Processing method
scCO2 PP/MWNT Nanocyl-7000 scCO2
scCO2.m PP/MWNT Nanocyl-7000 scCO2; masterbatch
scCO2 PP/cMWNT Nanocyl-9000 scCO2
scCO2.m PP/cMWNT Nanocyl-9000 scCO2; masterbatch
melt.m PP/MWNT Nanocyl-7000 melt-compounded; masterbatch
melt.m PP/cMWNT Nanocyl-9000 melt-compounded; masterbatch
568
wasvented. In anotherprocessing route, amasterbatchof 3.0wt.-%
CNT content was prepared using scCO2 assisted mixing under
15MPa and 200 8C for 10min before being diluted into lower
concentrations of 0.1, 0.25, 0.5 and 1.0wt.-% CNT using the same
conditions as described above. The bulk materials obtained from
the autoclave were frozen in liquid nitrogen and then broken into
pieces, which were subsequently used for hot-pressing into
dumbbell-shaped specimens according to ASTM D-638 at 200 8Cunder 40MPa for 5min. The samples were cooled down to room
temperature using the water cooling system of the hot-press.
For comparison, the melt-compounded nanocomposites were
prepared by a two-step blending process in a mini twin-screw
extruder (DSMXplore 15mLmicrocompounder) at 200 8C.Nitrogen
gas flowwas used to avoid degradation of the polymer during the
mixingprocess.Amasterbatchof3wt.-%MWNTs(orcMWNTs)was
prepared at a processing condition of 50 rpm for 5min. This
masterbatchwas subsequently dilutedwith pure PP at 250 rpm for
15min to produce samples with lower nanotube concentrations.
Theextrudedstrandsproducedbytheextruderwerecut intopellets
and then hot-pressed into dumbbell-shaped specimens. The
compositions of the nanocomposite samples are listed in Table 1.
Characterization
Scanning electron microscopy (SEM) was used to investigate the
morphological characteristics using a JEOL 6300 (accelerating
voltage10 kV) innormal secondaryelectron imagingmode. Prior to
the examination, samples were freeze-fractured in liquid nitrogen
and sputter coatedwith a thin layer of gold. Transmission electron
microscopy (TEM) was carried out using a JEOL JEM 2010. MWNTs
were dispersed in ethanol using ultrasonication for 5min and then
drops of suspension were deposited onto a copper grid for TEM
observation.
Differential scanning calorimetry (DSC) was performed using a
Mettler ToledoDSC822e. Samples of about10mgwereheated from
�50 toþ200 8C at a scan rate of 10 8C �min�1 and held for 10min to
eraseanythermalhistory,andthencooledto�50 8Cusingthesame
scan rate of 10 8C �min�1. Thermogravimetric analysis (TGA) was
carried out undernitrogenatmosphere in the temperature rangeof
20–1000 8C by using a TA TGA Q500. The heating rate used was
20 8C �min�1.
Wide angle X-ray scattering (WAXS) spectrawere recordedwith
a Siemens diffractometer D5000, where the X-ray beam was Ni-
filteredCuKa (l¼1.5405 A) and the radiationwasoperatedata rate
of 40 kVwithafilamentcurrentof40mA.Correspondingdatawere
Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
collected from 5 to 308 at a scanning rate of 0.018 �min�1.Tensile
tests were performed using an Instron 5566 machine. Tests were
conducted on dog-bone shaped specimens according to the ASTM
D-638 standard. The Young’s modulus of each specimen was
calculated, and the gradient of the stress versus strain curve was
measured between 0.05 and 0.2% strain. The values presented for
all the properties are the average of five repetitions of each test.
Results and Discussion
Figure 1 displays the SEM micrographs of PP nanocompo-
sites. Nanotubes are observed as bright spots in the images.
In the case of conventional melt compounding, better
dispersions of CNTs were observed for the HDPE coated
MWNTs samples, owing to the de-aggregated state of the
cMWNTs [16] (Figure1aandb).However, in the case of scCO2
assisted mixing, large aggregates of cMWNTs were
observed (Figure 1d) and evenwhen amasterbatchmethod
was applied the cMWNTs were still not well dispersed
(Figure 1f). Interestingly, pristine MWNTs without HDPE
coating provided a better and more homogeneous disper-
sion in the PP matrix. (Figure 1c) Similar improvements in
dispersion, for scCO2 assisted mixing without the use of
compatibilizers were also reported for nanoclays in PP
matrices.[25] A better dispersion of MWNTs was observed
using the masterbatch method (Figure 1e).
The TEM images of MWNTs and cMWNTs, which are
presented in Figure 2 may help to explain why the
dispersion of cMWNTs in PP matrix was unsatisfactory in
scCO2 assisted mixing. Pristine MWNTs could be dispersed
into individual tubes on the grid with no large aggregates
remaining (Figure 2a). However, cMWNTs (Figure 2b)
showed quite a different structure. The polymer coated
MWNTs form large agglomerates and although the HDPE
coating helps to separate the individual nanotubes, the
break up of these agglomerates still requires high shear
forces. Therefore, cMWNTs can be dispersed well in
conventional melt compounding processes in twin-screw
extruders because of the high shear forces employed.
However, a pitched blade turbine impeller, used in scCO2
assisted mixing, generates much lower shear forces, thus
leading to poor dispersion of these agglomerates.
DOI: 10.1002/mame.200900405
Processing of Poly(propylene)/Carbon Nanotube . . .
Figure 1. SEM images of freeze-fractured samples of PP nanocomposites with 0.5% CNTloading (a) melt.m PP/MWNT (b) melt.m PP/cMWNT (c) scCO2 PP/MWNT (d) scCO2 PP/cMWNT(e) scCO2.m PP/MWNT (f) scCO2.mPP/cMWNT.
Figure 2. TEM images of (a) pristine MWNTs (b) HDPE-coated MWNTs.
Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
The values of Young’s modulus
of the nanocomposites with differ-
ent MWNT loadings are presented
in Figure 3. As a result of the
de-aggregation of the MWNT
bundles by the HDPE coating, the
melt-compounded and master-
batch based PP/cMWNT compo-
sites show a continuous modulus
increase up to 1.7GPa from zero to
1.0wt.-% CNT loading. The same
melt-compounded and master-
batch systems based on uncoated
MWNTs only displayed a Young’s
modulus value of 1.5GPa for
1.0wt.-% CNT loading, due to the
poor dispersion of pristineMWNTs
in the PP matrix. Interestingly,
ScCO2 processed nanocomposites
with pristine MWNTs showed an
increase in Young’s modulus,
which is as good as or even better
than for melt-compounded and
masterbatch systems based on
HDPE coated MWNTs. However,
coated MWNTs showed the lowest
reinforcing efficiency in the case of
ScCO2 processed nanocomposites.
Furthermore, the use of master-
batch in the scCO2 assisted mixing
method did not show improve-
ments in terms of Young’s mod-
ulus,as shownforbothscCO2.mPP/
MWNT and scCO2.m PP/cMWNT
composites. Instead, the long pro-
cessing times needed at these high
temperature conditions may even
have caused degradation of the PP
matrix. Similarly, increases in yield
stress were observed for all scCO2
PP/MWNT samples, with a signifi-
cant 6% increase from zero to
0.1wt.-% CNT loading followed by
roughly constant values across the
rest of the CNT loading range up to
1wt.-% (Figure 4). The degree of
reinforcement of pristine MWNTs
is as good as for melt-compounded
coated PP/cMWNT composites. PP/
cMWNTs composites, showing
poor nanofiller dispersion in the
case of scCO2 assisted mixing, give
lower yield stresses. Although the
HDPE coating indeed helps the
www.mme-journal.de 569
J. Ma, H. Deng, T. Peijs
Figure 3. Young’s modulus of PP nanocomposites as a function ofCNT loading, showing that a similar modulus is achieved forscCO2 processed nanocomposites with pristine MWNTs and melt-compounded nanocomposites with coated cMWNTs. Solid linesrepresent results of nanocomposites with pristine MWNTs,whereas dashed lines represent nanocomposites with coatedcMWNTs.
Figure 4. Yield stress of PP nanocomposites as a function of CNTloading, showing that a similar yield stress is achieved for scCO2processed nanocomposites with pristine MWNTs and melt-com-pounded and masterbatch based nanocomposites with HDPEcoated MWNTs. Solid lines represent results of nanocompositeswith pristine MWNTs, whereas dashed lines represent nanocom-posites with coated cMWNTs.
570Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
dispersion of cMWNTs inmelt compounding processes, the
HDPE coating could also act as a weak interface between
matrix and CNTs, leading to poor stress transfer, thereby
weakening the nanocomposites. This potential effect is
however not found here. ScCO2 assisted mixing achieves a
better dispersion of CNTs without the need of a HDPE
coating, therefore eliminating the potential drawbacks of
the HDPE coating, leading to greater thermal and mechan-
ical reinforcing effects fromtheMWNTs.Also, the relatively
low shear forces involved in scCO2 assisted mixing may
cause less nanotube breakage and damage, while at the
same time avoiding degradation of the polymer.
To calculate the reinforcing efficiency of the CNTs in the
nanocomposites, the rule of mixture is applied to quantify
the effective Young’s modulus of the CNTs. For composites
inwhich discontinuous fibers are not perfectly aligned, the
Young’smodulus of the composite (Ec) canbewritten as: [34]
E ¼ hohLVfEf þ 1� Vfð ÞEm (1)
where hL is the length efficiency factor, ho is the orientation
factor, Em and Ef are the Young’smodulus of thematrix and
the fiber, respectively, and Vf is the volume fraction of the
fiber. hL can vary between 0 and 1. ho is equal to 1 for fully
aligned fibers, 3/8 for random (in-plane) 2D orientation
and 1/5 for random 3D orientation. A similar Equation can
be formulated for the strength, where sm, sm and sf are the
tensile strength of the composites, matrix and fiber,
respectively.
sc ¼ hLhoVfþsf þ 1� Vfð Þsm (2)
Therefore, Equation (1) and (2) canbeused toevaluate the
effective mechanical properties of CNTs in the composites.
In the calculation, the random 3D orientation value of 1/5
was chosen for ho, hL equals 1 for high fiber aspect ratios
such as (1/d)> 100. The volume fractions were deduced
from the weight fractions and densities of the materials.
Figure 5 illustrates the back-calculated effective Young’s
modulus and tensile stress of CNTs. Generally, the effective
properties decrease with increasing nanotube content. The
effective properties of pristine MWNTs in nanocomposites
processed using scCO2 are as good as, or even better than,
those for coated cMWNTs in nanocomposites processed
using melt-compouding and masterbatch. From direct
individual CNT testing, reported in literature, Young’s
moduli of the order of 1 TPa and tensile strength values
ranging from 10 to 150GPa have been reported.[35–38]
However, often back-calculated values of effective nano-
tubeproperties are significantly lower thanexperimentally
measured values of isolated CNTs, presumably due to
difficulties in achieving good dispersions of nanotubes,
strong nanotube/matrix interaction and nanotube align-
ment. Our calculated values of strength are reasonable
DOI: 10.1002/mame.200900405
Processing of Poly(propylene)/Carbon Nanotube . . .
Figure 5. (a) Effective Young’s modulus and (b) effective strengthof CNTs as calculated from the rule of mixtures, showing thatscCO2 processed nanocomposites incorporating pristine MWNTsresult in nanotube efficiency similar to that of melt-compoundedand masterbatch based nanocomposites based on HDPE coatedMWNTs. The dotted line represents the upper bound for nano-tube moduli.
Table 2. DSC measurement results.
Sample CNT Tm Xc Tc
wt.-% -C % -C
neat PP 0 166.0 37.9 110.3
0.1 164.7 43.7 118.9
scCO2 PP/MWNT 0.25 164.2 43.1 119.0
1.0 165.0 41.0 119.9
0.1 164.4 42.3 119.0
scCO2.m PP/MWNT 0.25 165.0 42.4 119.5
1.0 164.4 43.4 119.9
0.1 164.9 43.3 116.2
scCO2 PP/cMWNT 0.25 164.4 44.6 115.7
1.0 127.0/166.0 42.9 116.5
0.1 164.2 43.7 115.5
scCO2.m PP/cMWNT 0.25 164.7 43.3 116.2
1.0 127.3/165.7 43.7 116.4
melt.m PP/MWNT 1.0 165.0 40.2 121.9
melt.m PP/cMWNT 1.0 165.4 39.8 117.8
among the experimental data for CNTs, however, the
effective modulus values deduced from nanocomposites
with very low nanotube content (0.06 vol.-%) are signifi-
cantly higher than the theoretical modulus values for
nanotubes indicated by the dotted line in Figure 5a. The
reason for this could be related to a modification of the
matrix through the addition of CNTswhich causes errors in
the selectedvalueofEm.Because of thehighmatrix content,
the calculated results are strongly affected by variations of
Em.Small changesofEmvaluesof1.25 to1.40MPawill result
inaneffectivemodulus reduction from2400to1 200GPa in
the calculations. In a previous study, Young’s modulus of
the composites has been reported to increasewith polymer
Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
matrix crystallinity.[39] Similar results were also shown in
many other studies and changes in crystallinity and
crystalline morphology of the polymer matrix can have
pronounced effects on the mechanical properties of
nanocomposites next to mechanical reinforcing effects
from nanofillers.[10,40,41]
DSC data of themelting temperature (Tm), crystallization
temperature (Tc) and crystallinity (Xc) of the PP nanocom-
posites are presented in Table 2. The addition of nanotubes
only slightly decreases the Tm of PP at 166 by 1 8C. A Tm of
around 165 8C indicates the melting of a-crystals of PP.
Additionally, samples with HDPE coated MWNT content
above 0.5wt.-% show an additional broad peak at around
127 8C corresponding to the melting of the HDPE coating
(Figure 6). Interestingly, the scCO2 processed sample before
hot-pressing shows a broad peak in the 150–160 8Ctemperature range apart from the peak of 166 8C. Thislower Tm around 155 8C suggests the melting of PP b-phase
crystals.[41,42] Our X-ray diffraction (XRD) results also
confirm the presence of a g-phase of PP which is discussed
in the text below. All samples after hot-pressing only show
a Tm, at 166 8C, indicating no change to the PP a-crystals
upon the addition of CNTs,which agreeswellwithprevious
reports in the literature.[19,29] Therefore, it is assumed that
b and g-phase crystals formation of PP during scCO2
processing was erased from the final specimens after hot-
pressing. The crystallization temperature (Tc) and the
degree of crystallinity (Xc) are increased by the addition
of MWNTs or cMWNTs into the PP, suggesting that
nanotubes act as nucleating agents which reduce the
www.mme-journal.de 571
J. Ma, H. Deng, T. Peijs
Figure 6. DSC heating scan for typical PP/CNT nanocomposites,showing an additional broad peak at around 127 8C for sampleswith cMWNTs.
Figure 7. XRD spectra for typical PP/CNT nanocomposites,confirming the presence of a-phase crystals of PP in the nano-composites.
Figure 8. TGA of PP/CNT nanocomposites processed using scCO2assisted mixing, showing a retarded thermal degradation ofnanocomposites by the presence of MWNTs.
572
spherulites size, induce crystals growth at higher tempera-
ture and enhance crystallization. Furthermore, pristine
MWNTsgivehigherTc of PPnanocomposites thancMWNTs,
which indicates a stronger nucleation effect of MWNTs
than for coated MWNTs. Clearly, the surface of any
nucleating agent plays an important role in the crystal-
lization process.[43,44] MWNTs have high surface energies
which can absorb the polymer preferentially along their
crystal structure and act as nuclei for polymer crystals.
HDPEcoatingsdecrease thissurface freeenergy, resulting in
a reduced nucleating efficiency. Results also show a small
variation in Tc with different CNT loadings. Although
nucleating agents significantly increase the number of
nucleation sites, above certain loadings the introduction of
more CNTs may hinder chain mobility and retard crystal
growth. Thismay explain themore pronounced nucleating
effect observed for very low CNT loadings. Hence, the
mechanical reinforcing effects observed are a combination
of the modification of the PP matrix through increased
crystallinity as well as true reinforcing effects from CNT
fillers.[10]
The melting temperature was calculated from the first
heating, the crystallinity was calculated from the second
heating and the crystallization temperaturewasmeasured
from the first cooling.
Figure 7 shows the XRD patterns for pure PP and scCO2
processed nanocomposites before and after hot-pressing.
The pattern of pure PP shows peaks corresponding to PP a-
phase at 2u¼ 14, 17, 18.5, 21.5, 21.9 and 25.48 for six major
reflections: (110), (040), (130), (111), (041) and (060) plane,
respectively.[45] For scCO2 processed pure PP and PP
nanocomposites, an additional peak at 2u¼ 20.18 is
detected, which corresponds to the characteristic (117)
planeof theg-phaseofPP.[46] Theg-phase isonlyobtained in
some special cases such as under high pressure and by the
crystallization of very low molar mass fractions.[46] In our
Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
case, the high pressure used in the scCO2 processing
promoted the formation of g-phase crystals. Apart from
peaks of a and g-phases, a broad small peak observed
at 2u¼ 168 for the scCO2.m PP/cMWNT 0.5% sample before
hot-pressing corresponds to the (300) plane of the b-phase
of PP.[47]Otherb-phasepeaks suchas2u¼ 16.7 (008) and218(301) are too close to the strong a-phase peaks to be
identified. ThePPnanocomposite sampleafter hot-pressing
only showsana-phaseofPPwhichagreeswellwithourDSC
results, suggesting that the thermal history from scCO2
processing was erased after hot-pressing.
Finally, the thermal stability of the nanocomposites was
investigated by TGA (Figure 8). The thermal degradation of
DOI: 10.1002/mame.200900405
Processing of Poly(propylene)/Carbon Nanotube . . .
PP is substantially retarded by the presence of CNTs. The
onsetdecomposition temperatures,Tonset, arehigher forPP/
MWNTnanocompositesbecause theCNTshinder thefluxof
degradation product and improve the heat dissipation
within the composites.[48] However, above 470 8C the
composites seem to degrade at a slightly higher rate than
the neat polymer. A similar Tonsetwas observed for both the
PP/cMWNT and PP/MWNT nanocomposites, indicating
that the HDPE coating of cMWNTs had no significant effect
on the thermal degradation behavior.
Conclusion
PP/MWNT nanocomposites were prepared using scCO2-
assisted mixing and compared with those processed by a
traditional melt compounding method. Results showed
that by using scCO2-assisted mixing, both yield stress and
Young’s modulus of the nanocomposites increase with
MWNT loading. However, unlike in melt compounding
processes, with which the use of HDPE coated MWNTs is
essential for improveddispersionandproperties, in thecase
of scCO2-assisted mixing, similar good dispersions and
mechanical properties are achieved with pristine MWNTs,
which in addition to obvious cost benefitsmay also benefit
from a lower viscosity of the polymer melt under scCO2
conditionsandbetterpreservationofnanotube lengths. It is
interesting that techniques designed to achieve high
quality PP nanocomposites, such as the use of master-
batches or polymer coated MWNTs, are strictly not
necessarywhenusingscCO2-assistedmixing. Inaccordance
with many other nanocomposite studies, the addition of
CNTs enhanced the PP nucleation process. The resulting
increase in crystallinity is an important additional factor
responsible for the enhanced mechanical properties of the
final nanocomposites.
Received: December 5, 2009; Revised: February 12, 2010;DOI: 10.1002/mame.200900405
Keywords: carbon nanotubes; nanocomposites; poly(propylene);processing; scCO2
[1] M. S. Dresselhaus, G. Dresselhaus, P. Avouris, ‘‘Carbon Nano-tubes: Synthesis, Structure, Properties, and Applications’’,Springer, Berlin 2001.
[2] Baughman, A. A. Zakhidov, W. A. deHeer, Science 2002, 297,787.
[3] N. Grossiord, J. Loos, O. Regev, C. E. Koning, Chem.Mater. 2006,18, 1089.
[4] H. Deng, R. Zhang, C. T. Reynolds, E. Bilotti, T. Peijs,Macromol.Mater. Eng. 2009, 294, 749.
[5] R. Zhang, A. Dowden, H. Deng, M. Baxendale, T. Peijs, Compos.Sci. Technol. 2009, 69, 10 1499.
Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[6] H. Deng, T. Skipa, R. Zhang, D. Lillinger, E. Bilotti, I. Alig,T. Peijs, Polymer 2009, 50, 15 3747.
[7] R. Zhang, M. Baxendale, T. Peijs, Phys. Rev. B 2007, 195433.[8] M. Moniruzzaman, K. I. Winey, Macromolecules 2006, 39,
5194.[9] P. Ciselli, Z. Wang, T. Peijs, Mater. Technol.: Adv. Perform.
Mater. 2007, 22, 10.[10] Z. Wang, P. Ciselli, T. Peijs, Nanotechnology 2007, 18, 455709.[11] T. Saito, K. Matsushige, K. Tanaka, Phys. B: Condens. Matter.
2002, 323, 280.[12] Y. P. Sun, K. Fu, Y. Lin, W. Huang, Acc. Chem. Res. 2002, 35,
1096.[13] A. Star, Y. Liu, K. Grant, L. Ridvan, J. F. Stoddart, D. W. Steuer-
man, M. R. Diehl, A. Boukai, J. R. Heath,Macromolecules 2003,36, 553.
[14] D. Bonduel, M. Mainil, M. Alexandre, F. Monteverde,P. Dubois, Chem. Commun. 2005, 781.
[15] S. Peeterbroeck, F. Laoutid, J. M. Taulemesse, F. Monteverde,J. M. Lopez-Cuesta, J. B. Nagy, M. Alexandre, P. Dubois,Materials 2007, 17, 15 2787.
[16] P. Dubois, M. Alexandre, Adv. Eng. Mater. 2006, 8, 147.[17] P. Potschke, S. Pegel, M. Claes, D. A. Bonduel,Macromol. Rapid
Commun. . 2008, 29, 244.[18] H. Deng, E. Bilotti, R. Zhang, T. Peijs, J. Appl. Poly. Sci. 2010,
DOI: 10.1002/app.30783.[19] R. Andrews, D. Jacques, M. Minot, T. Rantell,Mater. Eng. 2002,
287, 395.[20] P. Potschke, A. R. Bhattacharyya, A. Janke, H. Goering, Compos.
Interfaces 2003, 10, 389.[21] K. Prashantha, J. Soulestin, M. F. Lacrampe, P. Krawczak,
G. Dupin, M. Claes, Compos. Sci. Technol. 2009, 69, 1756.[22] A. Garg, E. Gulari, C. W. Manke, Macromolecules. 1994, 27,
5643.[23] S. P. Nalawade, F. Picchioni, L. P. B. M. Janssen, Prog. Polym. Sci.
2006, 311, 19.[24] M. Garcia-Leiner, A. J. Lesser, Polym. Mater. Sci. Eng. 2003, 89,
649.[25] J. Ma, E. Bilotti, T. Peijs, J. A. Darr, Eur. Polym. J. 2007, 43, 12
4931.[26] T. Kashiwagi, E. Grulke, J. Hilding, K. Groth, R. Harris, K. Butler,
Polymer 2004, 45, 4227.[27] E. Assouline, A. Lustiger, A. H. Barber, C. A. Cooper, E. Klein,
E. Wachtel, J. Polym. Sci. B 2003, 41, 520.[28] J. Yang, Y. Lin, J. Wang, M. Lai, J. Li, J. Liu, J. Appl. Polym. Sci.
2005, 98, 1087.[29] S. P. Bao, S. S. Tjong, Mater. Sci. Eng. A 2007, 485, 508.[30] W. Leelapornpisit, M. Ton-That, F. Perrin-Sarazin, K. C. Cole,
J. Denault, B. Simard, J. Polym. Sci. B 2005, 43, 2445.[31] M. A. L. Manchado, L. Valentini, J. Biagiottiand, J. M. Kenny,
Carbon 2005, 43, 1499.[32] M. Ganb, B. K. Satapathy, M. Thunga, R. Weidisch, P. Potschke,
D. Jehnichen, Acta Mater. 2008, 56, 2247.[33] US 0207824 A1 (2008), invs.: M. Alexandre, D. Bonduel,
M. Claes, P. Dubois, S. Peeterbroeck, S. Pegel, P. Potschke.[34] D. Hull, T. Clyne, ‘‘An Introduction to Composite Materials’’,
Cambridge, University Press, Cambridge 1981.[35] M. M. Treacy, T. W. Ebbesen, L. M. Gibson, Nature 1996, 381,
678.[36] M.-F. Yu, O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly, R. S. Ruoff,
Science 2000, 287, 637.[37] A. H. Barber, R. Andrews, L. S. Schadler, H. D. Wagner, Appl.
Phys. Lett. 2005, 87, 203 106.[38] A. H. Barber, S. R. Cohen, A. Eitan, L. S. Schadler, H. D. Wagner,
Adv. Mater. 2006, 18, 83.
www.mme-journal.de 573
J. Ma, H. Deng, T. Peijs
574
[39] K. P. Ryan, M. Cadek, V. Nicolosi, D. Blond, M. Ruether,G. Armstrong, H. Swan, A. Fonseca, J. B. Nagy, W. K. Maser,W. J. Blau, J. N. Coleman, Compos. Sci. Technol. 2007, 67, 1640.
[40] J. N. Coleman, M. Cadek, R. Blake, V. Nicolosi, K. P. Ryan,C. Belton, A. Fonseca, J. B. Nagy, Y. K. Gun’ko, W. J. Blau, Adv.Funct. Mater. 2004, 14, 791.
[41] A. R. Bhattacharyya, T. V. Sreekumar, T. Liu, S. Kumar,L. M. Ericson, H. Hauge, R. E. Smalley, Polymer 2003, 44, 2373.
[42] L. Mandelkern, R. G. Alamo in: Physical Properties of PolymersHandbook J. E. Mark, Ed. American Institute of Physics, 1996,p. 123.
Macromol. Mater. Eng. 2010, 295, 566–574
� 2010 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[43] J. G. Tang, Y. Wang, H. Y. Liu, L. A. Belfiore, Polymer 2004, 45,2081.
[44] J. C. Wittmann, B. Lotz, Prog. Polym. Sci. 1990, 15, 909.[45] G. Natta, P. Corradini, Nuovo Cimento Suppl. 1960, 15, 40.[46] S. Bruckner, P. J. Philips, K. Mezghani, S. V. Meille, Macromol.
Rapid Commun. 1997, 18, 1.[47] S. V. Meille, D. R. Ferro, S. Bruckner, A. J. Lovinger, F. J. Padden,
Macromolecules 1994, 27, 2615.[48] S. T. Huxtable, D. G. Cahill, S. Shenogin, L. P. Xue, R. Ozisik,
P. Barone, M. Usrey, M. S. Strano, G. Siddons, M. Shim,P. Keblinski, Nat. Mater. 2003, 2, 731.
DOI: 10.1002/mame.200900405