Composites: Part A 103 (2017) 8495

Contents lists available at ScienceDirect

Composites: Part A
journal homepage: www.elsevier.com/locate/compositesa

Electrical, morphological and thermal properties of microinjection

molded polyamide 6/multi-walled carbon nanotubes nanocomposites
Shengtai Zhou a, Andrew N. Hrymak a,, Musa R. Kamal b
a
Department of Chemical and Biochemical Engineering, The University of Western Ontario, London N6A 5B9, Canada
b
Department of Chemical Engineering, McGill University, Montreal H3A 0C5, Canada

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

Article history: A series of polyamide 6/multi-walled carbon nanotubes (PA6/CNT) nanocomposites were prepared using
Received 12 July 2017 a masterbatch dilution process, followed by microinjection molding of a part with a three-step decrease
Received in revised form 19 September in thickness along the flow direction. Morphology observations revealed that there was a preferential
2017
orientation of CNT in the flow direction, which is attributed to the prevailing high shear rates in mIM.
Accepted 23 September 2017
Available online 25 September 2017
The distribution of CNT after melt processing was evaluated by dissolving experiments. Additionally,
the correlation between electrical resistivity and development of microstructure for each section of the
microparts was considered. The thermal behavior of PA6/CNT nanocomposites and corresponding micro-
Keywords:
A. Polymer-matrix composites (PMCs)
parts was evaluated using differential scanning calorimetry. Results indicated that the addition of CNT
B. Microstructures had little effect on the melting behavior of PA6/CNT nanocomposites and corresponding microparts.
D. Thermal analysis However, the crystallization behavior was changed significantly and a double crystallization peak was
E. Injection molding observed for samples incorporating CNT.
2017 Elsevier Ltd. All rights reserved.

1. Introduction However, the direct melt compounding of CNT with a polymer

The pursuit of high performance polymer/carbon nanotubes
(CNT) nanocomposites has continued since the pioneering work
reported by Iijima [1]. A number of CNT-related studies have been
conducted [2] to make full use of their exceptional mechanical
properties [3], excellent electrical and thermal conductivity [4,5],
and self-lubricating nature [6]. Therefore, CNTs are commonly
introduced into metals [7,8], ceramics [9,10], wood [11] and poly-
mer matrices [1115] to achieve high performance multifunctional
nanocomposites. Presently, polymer/CNT nanocomposites have a
wide range of applications in everyday life due to their weight
advantage, good resistance to corrosive conditions and excellent
moldability. However, the expected merits of incorporating CNT
into polymer matrices are always constrained due to the poor
interfacial interaction between CNT and host polymers [16], and
most importantly, the quality of dispersion of CNT in host matrices
[4,17].
Melt blending method is commonly utilized to fabricate CNT-
containing polymer composites because it holds advantages of
posing less threat to the environment (solvent free), higher pro-
duction efficiency and easier to scale to industrial quantities [18].
Corresponding author.
E-mail address: ahrymak@uwo.ca (A.N. Hrymak).
matrix is impractical due to the CNT handling issue [18,19]. More-
over, direct or indirect contact with CNT could potentially pose
immediate or chronic health hazards to on-site workers [20,21].
As a result, melt dilution of commercially available polymer/CNT
masterbatches (polymeric pellets with high CNT loading concen-
trations) with neat polymers emerges as an effective method to
fabricate nanocomposites at specified CNT loading concentrations
[2227].
Currently, there is a high demand for polymeric microparts in
the areas of electronics, automotive, biomedical and microelec-
tromechanical systems (MEMS) [2831]. The advent of microinjec-
tion molding (mIM) technology has made possible mass production
of microparts to meet the increasing needs in the aforementioned
areas. Normally, the definitions of microparts are identified in
terms of part weight (within a few milligrams), part dimensions
and/or the tolerances of part dimensions (within the micrometer
range) [29]. Under such circumstances, the archived molding prac-
tices applied in conventional injection molding (CIM) cannot be
directly transferred to mIM since very fast filling of mold cavities
is required for mIM in order to avoid premature solidification of
polymer melts. Thus, higher molding temperatures, injection pres-
sures and injection speeds are routinely employed in mIM [30].
Recently, a number of studies have been conducted to assess
the products molded from unfilled thermoplastics by mIM

https://doi.org/10.1016/j.compositesa.2017.09.016
1359-835X/ 2017 Elsevier Ltd. All rights reserved.
 S. Zhou et al. / Composites: Part A 103 (2017) 8495
[3136]. However, in the case of applications that require micro-
parts with multifunctional properties, such as electrical, thermal
and mechanical performance, are always accommodated by using
polymeric composites containing multifunctional fillers [37]. As a
result, more and more research attention has been directed to
the mIM of polymer/CNT composites. For instance, Ferreira et al.
[38] studied the state of dispersion of pristine and functionalized
CNT (i.e. p-CNT and f-CNT) in polyamide 6 (PA6) based on the mor-
phology observations and electrical conductivity measurements.
Their results showed improved dispersion of f-CNT in subsequent
microparts when compared with their PA6/p-CNT counterparts.
However, samples which were molded from PA6/p-CNT nanocom-
posites exhibited higher electrical conductivity in comparison with
those microparts containing f-CNT. The authors attributed this
phenomenon to the development of PA6/f-CNT interphase, which
could act as an insulating barrier, thereby impairing the electrical
conductivity. Abbasi et al. [25] reported that, in microinjection
molded samples, CNTs are preferentially aligned in the flow direc-
tion, which was attributed to the severe shearing conditions pre-
sent in mIM. This could also significantly affect the rheological
behavior and electrical conductivity of subsequent moldings [25].
Jiang et al. [39] investigated mIM of polyoxymethylene (POM)/
CNT composites. They found that the orientation of CNT showed
a hierarchical structure in microparts: an obvious orientation of
CNT is observed in the shear layer whereas less orientation of
CNT is detected in the core layer. This typical structure was, how-
ever, absent from CIM counterparts [39]. It is apparent that the dif-
ferences in the thermo-mechanical history experienced by
polymer melts in mIM and CIM play a pivotal role in determining
the microstructure of resultant moldings. In a previous study
[40], we reported that melt temperature has a marked influence
on the electrical conductivity of polystyrene (PS)/CNT microparts.
Higher melt temperature generally contributes to higher electrical
conductivity. Unlike PS/CNT microparts, morphology observations
revealed the existence of CNT aggregates in polypropylene (PP)/
CNT counterparts [41], which is mainly ascribed to the fact that
PP is a semi-crystalline polymer. The crystallization of polymer
chains could affect the state of dispersion of CNT in such polymer
matrices [42]. Additionally, increasing melt and mold tempera-
tures showed a positive effect on the electrical conductivity of
microparts [41]. Thus, it could be concluded that higher molding
temperatures facilitate to the formation of a three-dimensional
(3D) conductive network in subsequent moldings [3941]. The
above shows that the microstructure of polymer/CNT microparts
could be significantly affected by the selection of host polymer
matrices and the applied injection molding conditions which are
associated with various processing parameters, such as melt tem-
perature, mold temperature, injection speed, and so forth. Apart
from the above-mentioned influencing factors, the types of carbon
fillers could also play a role in determining the microstructure in
corresponding microparts [43]. For example, Zhou et al. [43] found
that the utilization of high structure carbon black (CB) and CNT
was advantageous to the enhancement of electrical conductivity
of corresponding microparts, which is closely related to the devel-
opment of microstructure in subsequent moldings.
In this study, a series of PA6/CNT composites were prepared via
masterbatch dilution process. The obtained extrudates were pel-
letized and subjected to mIM under a defined set of processing con-
ditions. PA6 is a typical engineering thermoplastic which exhibits
exceptional mechanical performance, chemical stability, easy flow
and fast processing cycles [44]. A rectangular mold insert which
has a three-step decrease in thickness along the flow direction
[45] was utilized in mIM. The purpose is to evaluate the effect of
abrupt changes in mold geometry on the state of CNT dispersion
within subsequent microparts.
85
2. Experimental
2.1. Materials
A masterbatch of 20 wt% multi-walled carbon nanotubes in
polyamide 6 (PA6/CNT) was kindly supplied by Hyperion Catalysis
International Inc. (Cambridge, MA). The CNT was vapor grown,
with a length over 10 mm and a density of 1.75 g/cm3, coupling
with an outside diameter of approximately 10 nm [46]. The mas-
terbatch was melt diluted with a neat PA6 resin (BASF, Ultramid
B3S) which has a density of 1.13 g/cm3 and melt flow index of
198 g/10 min (275 C @ 5.00 kg load) [47]. The PA6 is an easy flow-
ing polymer which has been adopted as the host matrix for car-
bonaceous fillers in several studies [4749].
2.2. Preparation of samples
Prior to melt blending, the PA6 and PA6/CNT masterbatch pel-
lets were dried under vacuum at 85 C for 8 h. Then, a series of
nanocomposites with CNT loading fractions of 2, 5 and 10 wt%
were prepared via a lab scale co-rotating twin-screw extruder
(Haake Minilab Rheomix CTW5, Thermo Scientific) operating at
60 rpm and 240 C for 10 min. In order to differentiate and express
each system clearly, the obtained blends were named as follows:
CNT2 indicates that the loading concentration of CNT in PA6 is
2% in weight and a similar rule is applicable to other composite
systems. Afterward, the obtained extrudates were pelletized and
subjected to mIM in a Battenfeld Microsystem 50 (Wittmann Bat-
tenfeld GmbH, Austria). The specific machine design and produc-
tion process of microparts were reported in a previous study
[31]. Images of the three-step mold insert and a micropart are dis-
played in Fig. 1a. All sections of the microparts have the same
width of 2.40 mm and the thickness of the thick, middle and thin
sections is 0.85, 0.50, and 0.20 mm, respectively. The thick and
middle sections have a length of 5.00 mm, while the thin section
has a length of 4.80 mm. The mIM process was carried out at a bar-
rel temperature of 260 C and a mold temperature of 80 C. The
injection speed was 300 mm/s. For comparison, round-disc shaped
samples were prepared by a laboratory hot press (Model: 4386,
Carver Inc., Wabash, IN) at 240 C.
2.3. Characterization
2.3.1. DC electrical conductivity
The samples used for volume electrical resistivity measure-
ments were cut, using a scalpel, into three sections, as indicated
by red arrows in Fig. 1b. The direct current (DC) electrical resis-
tance (R, ohm) of each section was measured in the transverse
direction (TD) and parallel to the flow direction (FD), respectively.
The R was determined using a two-probe method [50]. Afterward,
the electrical resistivity (q, ohm cm) for each sample was calcu-
lated with Eq. (1):
A
qR 1
L
where q denotes volume electrical resistivity, L (cm) is the distance
between copper electrodes and A (cm2) is the cross-sectional area.
2.3.2. Scanning electron microscopy
The morphology of obtained extrudates via masterbatch dilu-
tion and corresponding PA6/CNT microparts with respect to sam-
pling positions (Fig. 1c) were recorded using a high resolution
Hitachi S-4500 scanning electron microscope (SEM). The examined
samples were cryogenically fractured in liquid nitrogen, and all
86 S. Zhou et al. / Composites: Part A 103 (2017) 8495
Fig. 1. (a) Images of the mold insert and a micropart; (b) three-step configurations of the microparts, red arrows indicate the boundary of each section; (c) sampling positions
of each section for morphology observations.
fractured surfaces were coated with a thin layer of platinum prior
to observations.
2.3.3. Dissolving experiments and transmission electron microscopy
It has been reported that the dissolving experiments, that show
the state of nanotubes dispersion in a good solvent for the polymer
matrix, are helpful to reflect the state of dispersion that was
achieved in the solid state [51]. Thus, in order to obtain a qualita-
tive evaluation of the CNT dispersion within corresponding sam-
ples, the original masterbatch, CNT5 extrudate and different
sections taken from CNT5 microparts were dissolved in formic acid
for 1 h at 40 C. Small pieces of PA6/CNT masterbatch, CNT5 extru-
date and cut sections from CNT5 microparts were immersed in
about 50 mL formic acid with external stirring. These solutions
were formulated to acquire an equivalent amount of CNT in each
vial. The state of dispersion of CNT in PA6/formic acid solution
was recorded in digital photos by a Canon camera (Japan). Further-
more, the PA6/CNT/formic acid solution was diluted with distilled
water. Afterward, the solution was sonicated (Branson B-220, USA)
for 3 min at 125 W. A drop of the above-mentioned solution was
placed on a copper grid (400-mesh, Formvar Carbon Film, Electron
Microscopy Sciences). To observe the nanostructure of CNT, the
sample was examined by transmission electron microscopy
(TEM, Philips CM 10) at an acceleration voltage of 80 kV, after com-
plete evaporation of the liquid phase. The diameter and length of
CNT were measured manually with the help of an image analysis
software (Image J).
2.3.4. Differential scanning calorimetry
The melting and crystallization behavior of pure PA6 and
diluted PA6/CNT samples as well as each section of corresponding
microparts were determined using a differential scanning
calorimeter (DSC, Q2000, TA Instruments) under nitrogen atmo-
sphere at a constant gas flow rate of 50 mL/min. Thin slices taken
from the middle of each section were cut along the TD. The mea-
surement which consists of a heating phase followed by a cooling
phase was carried out between 40 and 240 C at a ramp rate of
10 C/min. The melting and crystallization temperatures were
detected from the heating and cooling graphs, respectively. The
crystallinity (vc) of each sample was calculated with Eq. (2):
DH m
vc
DHcalc
where DHm is the melting enthalpy of PA6 and DHcalc represents the
expected melting enthalpy of PA6 based on the measured melting
enthalpy for pure polymer and the mass fraction of it in the blend-
ing composite. The melting enthalpy of 100% crystalline PA6 is
230 J/g [47].
3. Results and discussion
3.1. Morphology
The morphology of extrudates obtained from the masterbatch
dilution process was examined using SEM and the results are given
in Fig. 2. The bright dots (the end of CNT) and the fibril-like struc-
ture underneath the polymer surface, which are pointed out by red
arrows, indicate the presence of CNT. In addition, the mean dis-
tance between the individually dispersed CNT decreases with an
increase of CNT loading concentration, which is favorable for the
formation of 3D conductive pathways. In general, CNTs were uni-
formly dispersed within the extrudates, regardless of CNT concen-
tration, which further confirmed that the masterbatch dilution
process is an effective method to fabricate polymeric composites
containing CNT.
The development of internal microstructure within each section
of PA6/CNT microparts, and the effect of CNT concentration on the
morphology of subsequent microparts were evaluated using SEM.
All samples for PA6/CNT microparts were fractured in liquid nitro-
gen along the TD, i.e. perpendicular to flow direction. Fig. 3 shows
the microstructure of the thick section obtained from CNT5 micro-
parts. Fig. 3a and b were taken from the core layer of the thick sec-
tion whereas Fig. 3c and d were taken from the shear layer, which
is close to the mold wall (Fig. 1c). Fig. 3a and c suggest that CNTs
are relatively uniformly distributed within the matrix since no
obvious CNT aggregates were observed in either the core layer or
shear layer. Interestingly, more bright dots could be observed in
Fig. 3d (shear layer) when compared with Fig. 3b (core layer), sug-
gesting that CNTs are preferentially aligned in the flow direction,
which could be attributed to the influence of higher shearing effect
that prevails in the shear layer [40,41]. This is consistent with our
previous observations for PS/CNT microparts [40]. Consequently,
this would increase the probability of CNT-CNT contact along the
2 flow direction, thereby leading to a reduction of electrical resistiv-
ity in the FD when compared with that in the TD.
 S. Zhou et al. / Composites: Part A 103 (2017) 8495 87

Fig. 2. The morphology of cryo-fractured surface of diluted extrudates: (A1A3) 2 wt%, (B1B3) 5 wt%, and (C1C3) 10 wt%, respectively.

Fig. 3. The morphology of thick section of CNT5 microparts from: (a, b) core layer; (c, d) shear layer.

Similarly, it could be deduced from Figs. 4 and 5 that CNT had a detectable bright dots in corresponding layer of the middle
relatively uniform distribution in the middle section and thin sec- (Fig. 4d) and thin (Fig. 5d) sections seemed to be smaller than that
tion of CNT5 microparts, respectively. However, the number of observed for the thick section (Fig. 3d). As a result, it could be
88 S. Zhou et al. / Composites: Part A 103 (2017) 8495

Fig. 4. The morphology of middle section of CNT5 microparts from: (a, b) core layer; (c, d) shear layer.

Fig. 5. The morphology of thin section of CNT5 microparts from: (a, b) core layer; (c, d) shear layer.

expected that this would lead to a higher electrical resistivity for
both the middle and thin sections, indicating the shearing condi-
tions that exsited in different sections of the microparts play a piv-
otal role in determing the microstructure. Studies [40,41] have
shown that there is a significant increase of maximum shear rates
with a simultaneous reduction of the thickness of mold cavities.
Therefore, it seems to be resonable to attribute the development
of microstructure in each section to the variation of shearing condi-
tions. Li and Shimizu [52] found that the CNT could achieve a better
dispersion in a polymer matrix when they were processed at ele-
vated shearing conditions. Chen et al. [53] reported similar findings
in CNT filled poly(vinylidene fluoride) (PVDF) composites. However,
Pegel et al. [19] pointed out that although higher shear rates could
provide suitable CNT dispersion in a host polymer matrix, secondary
agglomeration of CNT was also prevented which is thought to be
detrimental to the enhancement of electrical conductivity of corre-
sponding composites. Moreover, the length reduction of CNT during
the melt processing, especially under severe shearing conditions,
could be another concern which would significantly devalue the
electrical conductivity of subsequent products [54].
 S. Zhou et al. / Composites: Part A 103 (2017) 8495 89

Fig. 6. The morphology of thick section of CNT2 microparts from: (a, b) core layer; (c, d) shear layer.

Fig. 6 displays the internal microstructure of the thick section of
CNT2 microparts. The CNTs were uniformly dispersed in each layer,
as no obvious CNT agglomeration could be observed. Moreover, it
is hard for CNT to form intact 3D conductive pathways within
PA6 due to their lower loading fraction. In addition, Fig. 7 illus-
trates the morphology of the thick section of CNT10 microparts.
As shown in Fig. 7b and d, there is no significant difference in
the quality of dispersion of CNT in each layer of thick section at
such a high filler concentration. Furthermore, a conductive net-
work could be easily observed with CNT particles entangled with
each other, which is advantageous to the enhancement of overall
electrical conductivity.
3.2. Electrical resistivity
The electrical resistivity (q) of compression molded PA6/CNT
samples and each section of corresponding microparts is reported
in Fig. 8. No q could be measured for the thin section of CNT2
microparts due to a lack of conductive pathways, as displayed in
Fig. 8b. However, the q in the TD for both the thick and middle

Fig. 7. The morphology of thick section of CNT10 microparts from: (a, b) core layer; (c, d) shear layer.
90 S. Zhou et al. / Composites: Part A 103 (2017) 8495
Fig. 8. Electrical resistivity of (a) compression molded PA6/CNT samples and sections of (b) CNT2 microparts, (c) CNT5 microparts and (d) CNT10 microparts.
sections is almost two orders of magnitude higher than that mea-
sured in the FD, indicating a preferential orientation of CNT in the
flow direction. In addition, there was a reduction of q in the middle
section when compared with that of the thick section with respect
to the measurement directions. It was illustrated in Section 3.1 that
high shear rates would improve the dispersion of CNT in a polymer
matrix. Thus, there would be a better dispersion of CNT in the mid-
dle section when compared with the thick section. Although the
formation of a conductive network seemed to be unlikely at such
a lower CNT loading concentration (2 wt%), the improvement of
electrical conductivity in the middle section could be explained
by a hopping or tunneling mechanism [55]. In this case, a phys-
ical contact between the incorporated fillers is unnecessary as long
as they are close enough to allow the passage of electrons [56,57].
As the orientation and dispersion of CNT are favored in PA6 with an
increase of shear rates, the average distance between individually
dispersed CNT decreases in the middle section and the transport
of electrons is therefore facilitated through tunneling between
adjacent CNTs [53,55].
Fig. 8c shows that there was an increase of q measured in the
TD and FD for CNT5 microparts from the thick to thin sections.
However, similar to CNT2 microparts, the q measured in the FD
was invariably lower than the value obtained from the compres-
sion molded samples (Fig. 8a), which further confirmed a preferen-
tial alignment of CNT in the flow direction. The increase of q in the
middle section was consistent with the morphology observations
discussed in Section 3.1, which could be attributed to the improved
dispersion of CNT and the disentanglement of CNT aggregates aris-
ing from the increasing shear rates. For example, Ferreira et al. [38]
pointed out that the percolation threshold for PA6/CNT composites
in lIM is 46 wt%. In the case of CNT5, the values of electrical con-
ductivity would be dominated by the formation of conductive net-
work through physical interconnection amid the dispersed CNTs
and small CNT aggregates. Thus, the improved dispersion of CNTs
or disentanglement of CNT aggregates in the middle section of
resulting microparts might be detrimental to the formation of
intact conductive network [19,58], thereby lowering the electrical
conductivity in according section [58].
The q measured in the TD for each section of CNT10 microparts
exhibited a similar trend as reported for their CNT5 counterparts.
However, the q measured in the FD for the middle section was
lower than that for the thick section. It was observed in Fig. 7 that
sufficient conductive pathways were formed within PA6 matrix at
a CNT loading fraction of 10 wt%. The favored orientation of CNT in
the FD tends to increase the probability of CNT-CNT contact, lead-
ing to an enhanced electrical conductivity in the middle section
when compared with that obtained for thick section. Furthermore,
it is clear from Fig. 8c and d that the presence of very high shearing
effect is unfavorable for the construction of conductive pathways
within subsequent microparts since the thin section showed the
highest q, regardless of CNT loading concentrations.
3.3. Dissolving experiments and transmission electron microscopy
To qualitatively observe the dispersion of CNT in PA6/CNT mas-
terbatch, diluted composites and corresponding microparts, the
dissolving experiments were carried out with respect to each
sample. The observation of dissolved samples in formic acid for
 S. Zhou et al. / Composites: Part A 103 (2017) 8495
Fig. 9. Digital photos of PA6/CNT/formic acid solutions after 2 months: 1. PA6/CNT
masterbatch; 2. CNT5 extrudate; from 3 to 5 are the thick, middle and thin sections
of CNT5 microparts, respectively.
PA6/CNT samples is presented in Fig. 9. The solutions of extruded
CNT5 and sections taken from corresponding microparts appear
as homogeneous black solutions, indicating relatively homoge-
neous distribution of individual nanotubes or very small CNT
aggregates [51]. However, a black sediment coupling with nearly
transparent solution is visible for PA6/CNT masterbatch (sample
1), corresponding to the existence or remnants of CNT agglomer-
ates. Therefore, the dispersion of CNT is improved after melt dilu-
tion and subsequent mIM processes.
Characterization of the nanostructure of CNT in respective sam-
ple was carried out by TEM. The CNTs have a hollow structure,
which could be clearly observed in Fig. 10c, d and f. The measured
diameter of CNT ranges from 5 to 10 nm, which is consistent with
the technical data provided by manufacturer. However, Fig. 10f
indicated that the length of CNT is significantly reduced when
compared with their original length (over 10 mm) [46], suggesting
that melt processing has a significant effect on the length reduction
of CNT. Although the CNT5 extrudate has a homogeneous disper-
sion in formic acid (Fig. 9, sample 2), Fig. 10a and b indicate the
Fig. 10. TEM graphs of (a, b) extruded CNT5, (c, d) CNT5 thick section and (e, f) CNT5 middle section.
91
presence of small CNT bundles, implying that the CNT agglomer-
ates that existed in the original masterbatch could be significantly
reduced after subjecting to melt dilution process. Fig. 10c and d
show that there is a thin layer of PA6 matrix surrounding the indi-
vidual CNT, which suggests a good interfacial interaction between
CNT and the host polymer. Therefore, the adhesion of a thin layer
of PA6 to the CNT explained well dispersion of corresponding
materials in formic acid solution.
3.4. Thermal properties
Fig. 11a and b show the DSC heating and cooling behavior,
respectively, of pure PA6 and diluted PA6/CNT extrudates with var-
ious CNT concentrations. It could be deduced from Fig. 11 that the
incorporation of CNT had little effect on the melting behavior of
PA6 matrix. However, the crystallization process was greatly
affected with the presence of CNT. For pure PA6, only a single crys-
tallization peak (Tc1) was detected in the vicinity of 192 C. For
PA6/CNT extrudates, as indicated in Fig. 11b and Table 1, Tc1
shifted to higher values when compared with that obtained from
pure PA6. Moreover, the onset crystallization temperature (Tonset)
increased significantly upon the addition of CNT, indicating that
CNT played a role of nucleation agent. In addition, a second crystal-
lization temperature (Tc2) was observed at higher temperatures for
CNT-containing PA6 composites, and its position gradually shifted
to higher temperatures with an incremental loading fraction of
CNT, which suggested that the appearance of the second crystal-
lization peak in PA6 could be related to the presence of CNT. Sim-
ilar observations have been reported by Logakis et al. [46] and
Krause et al. [59]. The appearance of the double crystallization
peak in PA6/CNT composites could be explained by the formation
of two different types of crystal morphologies or the two-step crys-
tallization due to the presence of one-dimensional filler, i.e. CNT,
which has been detailed by Logakis et al. [46]. In the first case,
the CNT surface may induce the crystallite growth in a manner that
92 S. Zhou et al. / Composites: Part A 103 (2017) 8495

Fig. 11. DSC (a) heating and (b) cooling curves obtained from pure PA6 and PA6/CNT extrudates with various filler concentrations.

Table 1
Melting temperature (Tm), crystallinity (vc), crystallization temperatures (Tc1, Tc2) and onset crystallization temperature (Tonset) of pure PA6, PA6/CNT extrudates and different
sections of corresponding microparts.

Samples Melting process Crystallization process

Tm (C) vc Tc1 (C) Tc2 (C) Tonset (C)
PA6 221.24 0.392 192.60 N/A 198.56
PA6 Thick section 220.30 0.419 191.76 N/A 196.19
PA6 Middle section 220.06 0.499 191.63 N/A 196.56
PA6 Thin section 217.98 0.417 191.31 N/A 196.03
CNT2 extrudate 220.99 0.301 194.25 206.19 215.92
CNT5 extrudate 219.66 0.321 193.13 207.85 218.83
CNT10 extrudate 221.53 0.335 196.00 209.35 218.66
CNT5 Thick section 221.88 0.340 195.99 210.08 220.31
CNT5 Middle section 217.66 0.342 193.15 208.67 219.25
CNT5 Thin section 217.38 0.317 192.68 209.69 217.93
CNT2 Middle section 217.10 0.369 193.69 206.54 214.94
CNT10 Middle section 217.51 0.328 194.56 209.84 220.24

is perpendicular to the nanotube axis, which leads to the develop-
ment of trans-crystalline structure [38]. However, the polymer
chains tend to form the spherulites in the regions where are far
away from the nanotubes [46]. Thus, the co-existence of two differ-
ent types of crystals could possibly contribute to the appearance of
the double crystallization peak. In the other case, it was proposed
that the presence of double crystallization peak could be related to
the two-step crystallization of polymer chains. The incorporation
of fillers in a polymer matrix would greatly accelerate the crystal-
lization process of polymer chains [47]. Phang et al. [60] proposed
that both the high aspect ratio of CNT and the higher crystallization
temperature are favorable for the formation of thermodynamically
stable a-phase crystals. Both mechanisms suggest that the appear-
ance and the intensity of the second crystallization peak could be
related to the number of the incorporated CNTs.
The thermal behavior of pure PA6 and sections taken from cor-
responding microparts is given in Fig. 12. As shown in Fig. 12 and
Table 1, the melting temperature (Tm), Tc1 and Tonset for each sec-
tion of pure PA6 microparts decreased as compared with these val-
ues obtained from pure PA6. This could be attributed to the
influence of thermo-mechanical history experienced by polymer
melts in mIM which determines the microstructure during the
molding process [61]. For example, Kamal et al. [62] found that a
five-layer crystalline structure is observed in a POM micropart in
comparison with commonly observed three-layer structure in
CIM macroparts. Moreover, there would not be sufficient time
allowed for polymer chains to form well-developed crystals during
mIM due to the fast filling rate, high thermal gradients and short
cycle times. Therefore, a slight downshift of Tm observed in sec-
tions of pure PA6 microparts (Table 1) indicates the existence of
imperfections in resulting crystalline structures. However, a sec-
ond melting peak (ca. 222.65 C) is noticeable for the middle sec-
tion, suggesting that the prevailing shearing conditions in the
middle section [41] might partially promote the formation of
well-aligned structures. Table 1 reveals that although the Tm of dif-
ferent sections of PA6 microparts decreased when compared with
that of pure PA6, a remarkable increase of crystallinity (vc) was
achieved for these sections. This could thus be attributed to the
variation of shearing conditions in different sections of microparts,
as given in Ref [40,41]. The higher shear rates facilitated the orien-
tation of polymer chains in flow direction, leading to a higher vc for
sections that experienced relatively high shearing conditions.
Moreover, the cooling effect would be accentuated in different sec-
tions of PA6 microparts [63]. Under such circumstances, however,
a relatively stable crystalline structure was hardly formed by poly-
mer chains, which in turn resulted in a reduction of Tm. Interest-
ingly, the vc of thin section was comparable to that of thick
section. However, the Tm in the thin section was at least 2 C lower
than that of the thick section. Therefore, it appears that a combina-
tion of extremely high shear rates and fast cooling rates in the thin
section is unfavorable for the formation of perfect crystallites,
which contributes to a reduction of Tm.
 S. Zhou et al. / Composites: Part A 103 (2017) 8495 93

Fig. 12. DSC (a) heating and (b) cooling curves obtained from pure PA6 and sections of pure PA6 microparts.

Fig. 13. DSC (a) heating and (b) cooling curves obtained from pure PA6 and sections of CNT5 microparts.

Fig. 13 displays the thermal behavior of pure PA6 and different
sections taken from CNT5 microparts. As shown in Fig. 13a, only a
marginal difference of the Tm was recorded, suggesting that the
mIM process had little effect on the melting behavior of these
microparts. Similar to PA6/CNT extrudates (Fig. 11b), the presence
of CNT in each section of corresponding microparts had a signifi-
cant effect on the crystallization behavior and a double crystalliza-
tion peak was also observed. Compared with the CNT5 extrudate, a
slightly higher Tc2 (Table 1) for each section of corresponding
microparts might suggest a shear induced crystallization effect in
mIM samples. However, the achieved vc for each section of CNT5
microparts was lower than that of corresponding sections of pure
PA6 counterparts, indicating that the presence of CNT might hinder
the formation of ordered crystalline structure or domains, which
resulted from the limited space and confinement effect that
imposed on the polymer chains by CNT, irrespective of the hetero-
geneous nucleation effect induced by the incorporated fillers. The
other possible explanation could be associated with the fact that
the pure PA6 which was used to dilute the masterbatch is different
from the original PA6 in the masterbatch. Thus, the existence of
two different types of PA6 polymer chains might be reflected in
the crystallization process, thereby leading to a relatively lower
vc for the diluted samples such as PA6/CNT extrudates and sections
of corresponding microparts, as given in Table 1.
The effect of CNT loading concentrations on the thermal behav-
ior of the middle section of microparts as compared with pure PA6
is shown in Fig. 14. Except for the middle section of pure PA6
microparts, Fig. 14a and Table 1 reveal that there was an approxi-
mately 4 C reduction of Tm for the middle section of all microparts,
regardless of CNT loading concentrations. This behavior may be
explained by the consideration that the decrease of Tm is associ-
ated with the formation of imperfect crystals. Moreover, there
was a reduction of vc with an incremental loading fraction of
CNT, indicating a steric hindrance effect. Additionally, aside from
CNT10 micropart, the vc for the microparts is invariably higher
than that observed in corresponding extrudates, which further
supports the shear-induced crystallization effect in mIM. Therefore,
it could be concluded that the crystallization process is strongly
94 S. Zhou et al. / Composites: Part A 103 (2017) 8495
Fig. 14. DSC (a) heating and (b) cooling curves obtained from pure PA6 and middle section of PA6/CNT microparts with various filler concentrations.
influenced by the presence of CNT, while the employed molding
conditions or thermo-mechanical history determines the
microstructure of products.
4. Conclusion
A series of polyamide 6/multi-walled carbon nanotubes (PA6/
CNT) nanocomposites were prepared using masterbatch dilution
process via twin-screw extrusion. The obtained extrudates were
subsequently subjected to either compression molding or microin-
jection molding (mIM). Morphology observations indicated that
CNTs were uniformly dispersed in PA6, which reveals that master-
batch dilution is an effective method to fabricate CNT-containing
nanocomposites. As to micromoldings, electrical resistivity mea-
surements and morphology observations indicated that there
was a preferential orientation of CNT in the flow direction and
improved dispersion of CNT was achieved with an increase of
shearing conditions. The improved dispersion of CNT in PA6 after
processing was corroborated by dissolving experiments. Addition-
ally, the enhancement of electrical conductivity of PA6/CNT micro-
parts was closely related to CNT loading concentrations and the
development of internal microstructure. The formation of three-
dimensional (3D) conductive pathways was found to be crucial
to the enhancement of electrical conductivity of as-molded micro-
parts. DSC results indicated that the incorporation of CNT had little
influence on the melting behavior of PA6/CNT nanocomposites and
corresponding microparts. However, the crystallization behavior of
CNT-containing samples was greatly altered with the presence of a
double crystallization peak. It is suggested that the crystallization
process was mainly dominated by the presence of CNT.
Acknowledgement
The authors would like to thank Ian Vinkenvleugel (University
Machine Services, The University of Western Ontario) and Gleb
Meirson (Fraunhofer Project Centre for Composites Research @
Western) for experimental assistance. This work was supported
by Natural Sciences and Engineering Research Council of Canada
(NSERC) and the Network for Innovative Plastic Materials and
Manufacturing Processes (NIPMMP). S. Zhou acknowledges sup-
port from China Scholarship Council. We thank Biotron Integrated
Microscopy (The University of Western Ontario) for providing TEM
analysis.
References
[1] Iijima S. Helical microtubules of graphitic carbon. Nature 1991;35:568.
[2] Van Noorden R. The trials of new carbon. Nature 2011;469(7328):146.
[3] Grobert N. Carbon nanotubes-becoming clean. Mater Today 2007;10(1
2):2835.
[4] Shrivastava NK, Khatua BB. Development of electrical conductivity with
minimum possible percolation threshold in multi-wall carbon nanotube/
polystyrene composites. Carbon 2011;49(13):45719.
[5] Han Z, Fina A. Thermal conductivity of carbon nanotubes and their polymer
nanocomposites: a review. Prog Polym Sci 2011;36(7):91444.
[6] Meng H et al. Friction and wear behavior of carbon nanotubes reinforced
polyamide 6 composites under dry sliding and water lubricated condition.
Compos Sci Technol 2009;69(5):60611.
[7] Moghadam AD et al. Mechanical and tribological properties of self-lubricating
metal matrix nanocomposites reinforced by carbon nanotubes (CNTs) and
graphene - a review. Compos Part B - Eng 2015;77:40220.
[8] Chen B et al. Load transfer strengthening in carbon nanotubes reinforced metal
matrix composites via in-situ tensile tests. Compos Sci Technol 2015;113:18.
[9] Kasperski A, et al. Double-walled carbon nanotube/zirconia composites:
preparation by spark plasma sintering, electrical conductivity and
mechanical properties. Ceram Int 2015; 41(10 Part A): 137318.
[10] Yang Y et al. Quantification and promotion of interfacial interactions between
carbon nanotubes and polymer derived ceramics. Carbon 2015;95:96471.
[11] Baishya P, Maji TK. Functionalization of MWCNT and their application in
properties development of green wood nanocomposite. Carbohyd Polym
2016;149:3329.
[12] Logakis E et al. Structure-property relationships in isotactic
polypropylene/multi-walled carbon nanotubes nanocomposites. Compos Sci
Technol 2010;70(2):32835.
[13] Logakis E et al. Indirect methods for the determination of optimal processing
conditions in conductive polypropylene/carbon nanotubes composites. Chem
Phys Lett 2010;498(1):1258.
[14] Ayatollahi MR et al. Effect of multi-walled carbon nanotube aspect ratio on
mechanical and electrical properties of epoxy-based nanocomposites. Polym
Test 2011;30(5):54856.
[15] Villmow T et al. Influence of injection molding parameters on the electrical
resistivity of polycarbonate filled with multi-walled carbon nanotubes.
Compos Sci Technol 2008;68(3):77789.
[16] Liu T et al. Morphology and mechanical properties of multiwalled carbon
nanotubes reinforced nylon-6 composites. Macromolecules 2004;37
(19):721422.
[17] Wegrzyn M et al. The influence of injection molding parameters on electrical
properties of PC/ABS-MWCNT nanocomposites. J Appl Polym Sci 2013;130
(3):21528.
[18] Tang W, Santare MH, Advani SG. Melt processing and mechanical property
characterization of multi-walled carbon nanotube/high density polyethylene
(MWNT/HDPE) composite films. Carbon 2003;41(14):277985.
[19] Pegel S et al. Dispersion, agglomeration, and network formation of multiwalled
carbon nanotubes in polycarbonate melts. Polymer 2008;49(4):97484.
[20] Helland A et al. Reviewing the environmental and human health knowledge
base of carbon nanotubes. Cinc Sade Colet 2008;13(2):44152.
 S. Zhou et al. / Composites: Part A 103 (2017) 8495
[21] Ryman-Rasmussen JP et al. Inhaled carbon nanotubes reach the subpleural
tissue in mice. Nat Nanotechnol 2009;4(11):74751.
[22] Prashantha K et al. Masterbatch-based multi-walled carbon nanotube filled
polypropylene nanocomposites: assessment of rheological and mechanical
properties. Compos Sci Technol 2009;69(11):175663.
[23] Mahmoodi M et al. The electrical conductivity and electromagnetic
interference shielding of injection molded multi-walled carbon nanotube/
polystyrene composites. Carbon 2012;50(4):145564.
[24] Ptschke P, Fornes TD, Paul DR. Rheological behavior of multiwalled carbon
nanotube/polycarbonate composites. Polymer 2002;43(11):324755.
[25] Abbasi S, Carreau PJ, Derdouri A. Flow induced orientation of multiwalled
carbon nanotubes in polycarbonate nanocomposites: rheology, conductivity
and mechanical properties. Polymer 2010;51(4):92235.
[26] Lpez Gaxiola D et al. Effects of carbon fillers on tensile and flexural properties
in polypropylene-based resins. J Appl Polym Sci 2010;118(3):162033.
[27] Castillo FY et al. Electrical, mechanical, and glass transition behavior of
polycarbonate-based nanocomposites with different multi-walled carbon
nanotubes. Polymer 2011;52(17):383545.
[28] Liu Z et al. Filling behavior, morphology evolution and crystallization behavior
of microinjection molded poly(lactic acid)/hydroxyapatite nanocomposites.
Compos Part A - Appl S 2015;72:8595.
[29] Whiteside BR et al. Micromoulding: process measurements, product
morphology and properties. Plast Rubber Compos 2004;33(1):117.
[30] Giboz J, Copponnex T, Ml P. Microinjection molding of thermoplastic
polymers: a review. J Micromech Microeng 2007;17(6):96109.
[31] Chu J et al. Characterization of the microinjection molding process. Polym Eng
Sci 2010;50(6):121425.
[32] Zhang N et al. Replication of micro/nano-scale features by micro injection
molding with a bulk metallic glass mold insert. J Micromech Microeng
2012;22(6):065019.
[33] Park SH et al. Injection molding micro patterns with high aspect ratio using a
polymeric flexible stamper. Exp Polym Lett 2011;5(11):9508.
[34] Annicchiarico D, Attia UM, Alcock JR. A methodology for shrinkage
measurement in micro-injection moulding. Polym Test 2013;32(4):76977.
[35] Zhao J et al. Effects of process parameters on the micro molding process. Polym
Eng Sci 2003;43(9):154254.
[36] Rhode M, Derdouri A, Kamal MR. Micro replication by injection-compression
molding. Int Polym Proc 2009;24(3):28897.
[37] Ruprecht R et al. Injection molding of microstructured components from
plastics, metals and ceramics. Microsyst Technol 2002;8(45):3518.
[38] Ferreira T, Paiva MC, Pontes AJ. Dispersion of carbon nanotubes in polyamide 6
for microinjection moulding. J Polym Res 2013;20(11):19.
[39] Jiang Z, Chen Y, Liu Z. The morphology, crystallization and conductive
performance of a polyoxymethylene/carbon nanotube nanocomposite
prepared under microinjection molding conditions. J Polym Res 2014;21
(6):115.
[40] Zhou S, Hrymak AN, Kamal MR. Electrical and morphological properties of
microinjection molded polystyrene/multiwalled carbon nanotubes
nanocomposites. Polym Eng Sci 2016;56(10):118290.
[41] Zhou S, Hrymak AN, Kamal MR. Microinjection molding of
polypropylene/multi-walled carbon nanotube nanocomposites: the influence
of process parameters. Polym Eng Sci. doi: https://doi.org/10.1002/pen.24682.
[42] Abbasi S, Derdouri A, Carreau PJ. Properties of microinjection molding of
polymer multiwalled carbon nanotube conducting composites. Polym Eng Sci
2011;51(5):9921003.
95
[43] Zhou S, Hrymak AN, Kamal MR. Electrical and morphological properties of
microinjection molded polypropylene/carbon nanocomposites. J Appl Polym
Sci 2017;134(43):45462.
[44] Zhou S et al. Enhanced thermal conductivity of polyamide 6/polypropylene
(PA6/PP) immiscible blends with high loadings of graphite. J Compos Mater
2016;50(3):32737.
[45] Kamal MR, Derdouri A, El-Otmani R. Presentation at PPS Americas Conference.
Niagara Falls; May 2012.
[46] Logakis E et al. Structure-property relationships in polyamide 6/multi-walled
carbon nanotubes nanocomposites. J Polym Sci Part B: Polym Phys
2009;47:76474.
[47] Zhou S et al. Thermally conductive composites obtained by flake graphite
filling immiscible polyamide 6/polycarbonate blends. Thermochim Acta
2013;566:8491.
[48] Zhou S et al. High thermally conducting composites obtained via in situ
exfoliation process of expandable graphite filled polyamide 6. Polym Compos
2013;34(11):181623.
[49] Zhou S et al. Preparation of highly thermally conducting polyamide 6/graphite
composites via low-temperature in situ expansion. J Appl Polym Sci 2014;131
(1):39596.
[50] Motaghi A, Hrymak A, Motlagh GH. Electrical conductivity and percolation
threshold of hybrid carbon/polymer composites. J Appl Polym Sci 2015;132
(13):41744.
[51] Ptschke P et al. Melt mixing of polycarbonate/multi-wall carbon nanotube
composites. Compos Interface 2003;10(45):389404.
[52] Li Y, Shimizu H. High-shear processing induced homogenous dispersion of
pristine multiwalled carbon nanotubes in a thermoplastic elastomer. Polymer
2007;48(8):22037.
[53] Chen GX, Li Y, Shimizu H. Ultrahigh-shear processing for the preparation of
polymer/carbon nanotube composites. Carbon 2007;45(12):233440.
[54] Krause B et al. Characterization of dispersability of industrial nanotube
materials and their length distribution before and after melt processing. In:
Tasis D, editor. Carbon nanotube-polymer composites. Cambridge: RSC
Publishing; 2013. p. 21233.
[55] Wang L, Dang ZM. Carbon nanotube composites with high dielectric constant
at low percolation threshold. Appl Phys Lett 2005;87(4):042903.
[56] Psarras GC. Charge transport properties in carbon black/polymer composites. J
Polym Sci Part B: Polym Phys 2007;45:253545.
[57] Min C et al. The electrical properties and conducting mechanisms of carbon
nanotube/polymer nanocomposites: a review. Polym-Plast Technol
2010;49:117281.
[58] Alig I et al. Establishment, morphology and properties of carbon nanotube
networks in polymer melts. Polymer 2012;53:428.
[59] Krause B, Ptschke P, Huler L. Influence of small scale melt mixing
conditions on electrical resistivity of carbon nanotube-polyamide
composites. Compos Sci Technol 2009;69(10):150515.
[60] Phang IY et al. Crystallization and melting behavior of multi-walled carbon
nanotube-reinforced nylon-6 composites. Polym Int 2006;55:719.
[61] Ferreira T et al. Microinjection molding of polyamide 6. Polym Adv Technol
2014;25(8):8915.
[62] Kamal MR et al. Morphology of microinjection moulded polyoxymethylene.
Plast Rubber Compos 2010;39(7):33241.
[63] Zhang N, Choi SY, Gilchrist MD. Flow induced crystallization of poly(ether-
block-amide) from the microinjection molding process and its effect on
mechanical properties. Macromol Mater Eng 2014;299(11):136283.