Polymer Degradation and Stability 92 (2007) 1009e1015

www.elsevier.com/locate/polydegstab

Isothermal crystallization kinetics and thermal behavior of

poly(3-caprolactone)/multi-walled carbon nanotube composites
Erh-Chiang Chen, Tzong-Ming Wu*
Department of Materials Science and Engineering, National Chung Hsing University, 250 Kuo Kuang Road, Taichung 402, Taiwan
Received 23 January 2007; received in revised form 25 February 2007; accepted 28 February 2007
Available online 12 March 2007

Abstract

This study describes the preparation of poly(3-caprolactone) (PCL)/multi-walled carbon nanotube (MWCNT) composites by ultrasonically
mixing the PCL and as-fabricated MWCNT in a tetrahydrofuran solution. The TEM images show that the MWCNT is well separated and uniformly
distributed in the PCL matrix. Differential scanning calorimetry (DSC), thermogravimetric analyzer (TGA), X-ray diffraction (XRD) and polarized
optical microscopy (POM) were used to investigate the isothermal crystallization kinetics, crystalline structure and thermal behavior of PCL and
PCL/MWCNT nanocomposites. DSC isothermal results revealed that the activation energy of PCL extensively decreases with increasing MWCNT
contents, suggesting that the loading of MWCNT into PCL matrix probably induced heterogeneous nucleation during crystallization processes.
From TGA data, the addition of small amount of MWCNT into PCL matrix can improve the thermal stability of PCL matrix. TGA isothermal deg-
radation data illustrate that the activation energy Ed of the composites is smaller than that of PCL. This phenomenon can be attributed to the
incorporation of more MWCNT loading into PCL caused a decrease in the degradation rate and an increase in the residual weight for PCL/MWCNT
nanocomposites.
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Poly(3-caprolactone); Multi-walled carbon nanotube; Isothermal crystallization kinetics; Isothermal degradation

1. Introduction degraded by hydrolysis of its ester linkages alone in the polymer

Multi-walled carbon nanotubes (MWCNT) first reported by
Iijima in 1991 [1] possess exclusive structural, mechanical and
electrical properties [2]. Due to the exceptional properties of
MWCNTs, many investigations have focused on exploiting
these surprising characteristics for engineering applications,
including polymer nanocomposites, field emitters, nanoelec-
tronic devices, chemical sensors and biomaterials [3e5]. For
polymer/MWCNT nanocomposites, the addition of MWCNT
could extensively change the mechanical, thermal, and elec-
tronic properties of polymer matrix, which would remarkably
extend their application regions.
Biodegradable poly(3-caprolactone) (PCL) is an aliphatic
polyester with semi-crystalline characteristic. PCL can be
* Corresponding author. Tel.: þ886 4 22840500x806; fax: þ886 4 22857017.
E-mail address: tmwu@dragon.nchu.edu.tw (T.-M. Wu).
chains in physiological conditions and has therefore received
a wide range of possible applications, such as biodegradable
packaging materials, implantable biomaterials and microparti-
cles for drug delivery. In particular, it is specially interesting
for the purpose of long term implantable devices because of
their lower degradation rates as compared to that of polylactide
[6e8]. Recently, some investigations have used the function-
alized MWCNTs as reinforced fillers to prepare the PCL/
MWCNT nanocomposites [9e12]. They have demonstrated
that the functionalized MWCNTs were well separated and ran-
domly distributed in the PCL matrix. Because the unmodified
MWCNT containing a lot of Van der Waals forces among them-
selves is very difficult to be dispersed in a polymer matrix, there
is little information about the preparation and characterization
of polymer/MWCNT composites using the original MWCNT.
In this study, the PCL/MWCNT nanocomposites have been
fabricated by mixing the PCL and as-prepared MWCNT in tet-
rahydrofuran (THF) solution. In order to overcome the Van der

0141-3910/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymdegradstab.2007.02.019
1010 E.-C. Chen, T.-M. Wu / Polymer Degradation and Stability 92 (2007) 1009e1015

Waals forces among CNTs, the MWCNTs were first added in
tetrahydrofuran (THF) solution under ultrasonic treatment for
24 h to obtain the well distributed MWCNT in solution. The
PCL/MWCNT composites were then prepared by mechanically
mixing the MWCNT and PCL in THF solution with magnetic
stirring for 24 h. Effect of MWCNT on the isothermal crystalli-
zation kinetics and thermal degradation of PCL and PCL/
MWCNT nanocomposites were performed by the differen-
tial scanning calorimetry (DSC), thermogravimetric analyzer
(TGA) and polarized optical microscopy (POM) analysis.
The parameters of the isothermal crystallization kinetics
and thermal degradation of PCL and PCL/MWCNT nano-
composites were also studied.
2. Experimental
2.1. Specimens
Poly(3-caprolactone) (PCL) with Mn ¼ 42,500 were pur-
chased from Aldrich. The as-prepared multi-walled carbon
nanotubes (MWCNTs) were synthesized by ethylene chemical
vapor deposition process and their diameters are in the range
of 20e40 nm. In order to distribute the as-prepared MWCNT
in polymer matrix well, the as-prepared MWCNT was added
in tetrahydrofuran (THF) solution under ultrasonic treatment
for 24 h. The PCL/MWCNT nanocomposites were then fabri-
cated by mechanically mixing the MWCNT and PCL in THF
solution with magnetic stirring for 24 h.
Samples of pure PCL and 0.25, 0.5 and 1 wt% PCL/MWCNT
nanocomposites were sandwiched between two cover glasses
and heated on a Mettler Toledo FP82HT hot stage at 90  C.
The specimens were pressed into thin films with thickness in
the range of 0.03 mm, kept for 20 min to eliminate any thermal
history and then rapidly quenched into the proposed crystalliza-
tion temperatures (Tcs) for 2 h.
For thermal stability, the specimens were heated from 30 to
600  C at a rate of 10  C/min. For isothermal degradation,
each sample was heated from 30 to 100  C for 3 min to remove
the residual water and then was used a heating rate of 10  C/min
to reach the predetermined degradation temperatures (Tds) at
320, 330, 340, 350, 360, and 380  C for 40 min.
2.3. Structural analysis
X-ray patterns were recorded using a 3 kW Rigaku
D/MAX 2000 diffractometer equipped with the Cu Ka radia-
tion in the reflection mode. For XRD analysis, specimens of
the PCL and PCL/MWNT composites were sandwiched and
pressed between two thin Teflon sheets in the Mettler Toledo
FP-82 hot stage. Optical microscopy was performed using
a Zeiss optical microscope equipped with crossed polarizers
and a Mettler FP-82 hot stage. The crystallization process
was examined using the following temperature sequence.
The polymer was heated to melt at Tmax ¼ 90  C for
15 min on the hot stage to eliminate previous thermal history
and then rapidly transferred onto another hot stage controlled
at the proposed Tcs. The spherulitic morphology was moni-
tored between crossed polarizers and recorded at an
appropriate time interval by a digit camera mounted on the
microscope. Transmission electron microscopy (TEM) ob-
servations were made with a JEOL JEM-1200 using an
acceleration voltage of 120 keV. The specimen fixed on the
carbon-coated copper grid with a thickness of about 100 nm
was obtained by an ultra-microtome equipped with a diamond
knife. Due to the high electron density difference between
MWCNT and polymer matrix, staining of the samples was
not necessary.

2.2. Thermal analysis

The thermal analysis of the specimens was carried out using

a PerkineElmer PYRIS Diamond differential scanning calo-
rimetry (DSC) and all experiments were performed under a ni-
trogen atmosphere. The temperature scale was calibrated using
the melting point of high purity indium metal. All specimens
weighed in the range of 5e6 mg. For isothermal crystallization,
the specimens were heated to pre-melting temperatures (Tmax)
of 90  C at a rate of 100  C/min and held for 10 min to eliminate
any thermal history, then they were rapidly cooled to the pro-
posed crystallization temperatures (Tcs) in the range of 42e
50  C in steps of 2  C. Heat fusion versus time for isothermal
crystallization (DHf) was recorded. The specimens isothermally
crystallized at the crystallization temperatures were heated to
Tmax ¼ 90  C at a rate of 10  C/min.
The degradation behaviors of PCL and PCL/MWNT nano-
composites were carried out using a PerkineElmer thermo-
gravimetric/differential thermal analyzer (TG/DTA) and all
experiments were performed under a nitrogen atmosphere at
a purge rate of 100 ml/min. All specimens weighed about 8 mg. Fig. 1. TEM micrograph of 1 wt% PCL/MWCNT nanocomposites.
 E.-C. Chen, T.-M. Wu / Polymer Degradation and Stability 92 (2007) 1009e1015 1011

where the Xt is the relative crystallinity at time t, k is the crys-

tallization rate constant, and n is the Avrami exponent constant
depending on the nucleation and growth mechanism. The time
at which half-crystallization occurs can be defined as t1/2. To
convert conveniently with the operation, Eq. (1) can be trans-
formed into

Fig. 2. Avrami plots of ln[ln(1Xt)] versus ln t for 1 wt% PCL/MWCNT
nanocomposites.
3. Results and discussion
Fig. 1 shows the TEM image of 1 wt% PCL/MWCNT nano-
composites; the tube like morphology represents the MWCNT
in the PCL matrix (bright). From this image, it can be clearly
seen that the as-prepared MWCNT is well separated and ran-
domly distributed in PCL matrix. This result indicates that
MWCNT was successfully separated in PCL/MWCNT nano-
composite via solution mixing method under ultrasonic treat-
ment. As expected from the high aspect ratio of MWCNT,
experimental introduction of MWCNT into PCL matrix might
enhance the crystallization and thermal properties of PCL.
Therefore the crystallization kinetics of PCL and PCL/
MWCNT nanocomposites can be analyzed using the well
known Avrami equation [13,14], described as follows:
1  Xt ¼ exp ½  kðtÞn  ð1Þ
ln ½  ln ð1  Xt Þ ¼ nln t þ ln k ð2Þ
Fig. 2 shows the plots of ln[ln(1Xt)] versus ln t for 1 wt%
PCL/MWCNT nanocomposites at various Tcs. The n and k
values were directly obtained from the slope and intercept of
the best fitting line, respectively. Several crystallization param-
eters, such as t1/2, n, and k, for PCL are listed in Table 1. The plots
of ln[ln(1Xt)] versus ln t for PCL, 0.25 wt% PCL/MWCNT
and 0.5 wt% PCL/MWCNT nanocomposites show similar ten-
dency, and their crystallization parameters are also recorded in
Table 1. In general, t1/2 can be used to characterize the crystalli-
zation rate. The longer the t1/2, the slower the crystallization rate.
The values of t1/2 of the PCL/MWCNT nanocomposites are
much smaller than that of PCL. These results suggest that overall
isothermal crystallization rates of the PCL/MWCNT nanocom-
posites at the same Tc were much faster than that of PCL, indi-
cating that the addition of small amount MWCNT can behave
as effective nucleating agents and significantly speed up the
crystallization rate of PCL. In all conditions, the k values de-
creased as Tc increases. In addition, it is necessary to compare
the Avrami exponent n of PCL/MWCNT nanocomposites
with that of the pure PCL. The n values for the pure PCL samples
corresponding to the primary crystallization stage are in the
range of 2.6e2.9, which reveal that the PCL polymer chains
are apt to take the three-dimensional crystallization growth
with heterogeneous athermal nucleation. Nevertheless, the
mechanism of nucleation and the growth of PCL crystallite
are slightly influenced by the addition of MWCNT into PCL,
leading to a slight decrease of n values in the PCL/MWCNT
nanocomposites. The n values of PCL/MWCNT nanocom-
posites are in the range of 2.4e2.6, which are explained by

Table 1
Values of t1/2, n, and k at various Tcs for PCL and PCL/MWCNT nanocomposites
Tc ( C)
42 44 46 48 50
PCL t1/2 (min) 6.5 10.1 18.0 46.7 125.2
n 2.62 2.72 2.76 2.86 2.58
k 6.2E  03 1.3E  03 2.3E  04 1.2E  05 4.6E  06

0.25 wt% PCL/MWCNT t1/2 (min) 0.8 2.5 4.7 10.1 22.7
n 2.57 2.52 2.63 2.55 2.51
k 1.2E þ 00 7.0E  02 1.2E  02 2.0E  03 3.1E  04

0.5 wt% PCL/MWCNT t1/2 (min) 0.8 1.5 2.7 6.1 13.3
n 2.51 2.50 2.54 2.52 2.44
k 1.5E þ 00 2.9E  01 5.9E  02 7.3E  03 2.1E  03

1.0 wt% PCL/MWCNT t1/2 (min) 0.7 1.4 2.5 5.3 10.2
n 2.45 2.44 2.49 2.42 2.39
k 1.8E þ 00 3.3E  01 7.3E  02 1.3E  02 4.7E  03
1012 E.-C. Chen, T.-M. Wu / Polymer Degradation and Stability 92 (2007) 1009e1015
Fig. 3. Arrhenius plots of 1/n (ln k) versus 1/Tc for PCL and PCL/MWCNT
nanocomposites.
simultaneous occurrence of two- and three-dimensional spheru-
litic growth. These results indicate that the incorporation of
MWCNT into PCL might cause heterogeneous nucleation in-
duced by a change in the crystal growth process from three-
dimensional crystal growth to mixed two-dimensional and
three-dimensional crystal growth [15,16].
The isothermal crystallization for PCL is assumed to be
thermally activated, thus the crystallization rate parameter k
Fig. 4. POM micrographs of (a) PCL and (b) 0.25 wt% PCL/MWCNT, (c) 0.5 wt% PCL/MWCNT, and (d) 1 wt% PCL/MWCNT composites isothermally
crystallized at 44  C.
can be used to determine the crystallization energy using the
Arrhenius equation approximately described as follows:
1 DEa
ðln kÞ ¼ ln k0  ð3Þ
n RTc
where k0 is a temperature-independent pre-exponential factor;
DEa is an activation energy; and R is the universal gas con-
stant. Arrhenius plots of 1/n (ln k) against 1/Tc for PCL and
PCL/MWCNT nanocomposites are shown in Fig. 3, and are al-
most linear. DEa can be determined from the slope of the plots
and is significantly dependent on the content of MWCNTs. The
values of DEa are 338.8, 312.6, 301.8 and 293.5 kJ/mol for the
PCL, 0.25, 0.5 and 1 wt% PCL/MWCNT nanocomposites, re-
spectively. The values of DEa slightly decrease with the loading
of MWCNT into PCL, and then continuously decline as
MWCNT content increases. This result indicates that the addi-
tion of MWCNT into PCL matrix probably induces the hetero-
geneous nucleation (a lower DEa). The activation energy
continuously decreases with increasing MWCNT content, sug-
gesting that the additional loading of MWCNT content into
PCL further increases the number of heterogeneous nucleation
during the crystallization processes.
Fig. 4 shows the POM micrographs of PCL and PCL/
MWCNT nanocomposites after melting at 90  C and then
quenching to Tc. For the pure PCL matrix, the POM image
exhibits a typical Maltese-cross spherulite. By adding the
MWCNT into PCL, it can be clearly seen that the amount of
spherulites of 1 wt% PCL/MWCNT nanocomposites is much
 E.-C. Chen, T.-M. Wu / Polymer Degradation and Stability 92 (2007) 1009e1015 1013

higher than that of PCL. Based on these observed POM

images, the nucleation density and nucleation rates of 1 wt%
PCL/MWCNT nanocomposites is higher than that of PCL,
which is consistent with the DSC data. These results revealed
that good dispersion of MWCNT can effectively influence the
crystallization development and change the crystallization
kinetics of PCL.
While the POM micrographs show different extent of crys-
tallization and spherulitic morphology between PCL and PCL/
MWCNT nanocomposites, it is necessary to identify their ef-
fects on the crystal structure of PCL. X-ray diffraction data
for PCL and PCL/MWCNT nanocomposites isothermally crys-
tallized at the same Tc used in the POM measurements are
shown in Fig. 5. These results indicate that PCL contains three
strong reflections at 2q ¼ 21.4 , 22.0 and 23.7 , corresponding
to the (110) (111) and (200) planes of the orthorhombic crystal
form. On the other hand, the strong reflections observed for
PCL were also present for PCL/MWCNT nanocomposites.
These results suggest that the addition of MWCNT to PCL
does not change the crystal structure of PCL.
In order to study the effect of MWCNT on the thermal
behaviors of PCL matrix, DTA/TGA analysis were used to
identify the isothermal and nonisothermal degradation of
PCL and PCL/MWCNT nanocomposites. Fig. 6 shows the
DTA and TGA curves of PCL and PCL/MWCNT nanocompo-
sites at a heating rate of 10  C/min. From the DTA data, the Tm
was slightly shifted to high temperature by adding the
MWCNT into PCL matrix. Meanwhile the peak profiles also
became broader due to the addition of MWCNT to PCL. The
TGA profiles of PCL/MWCNT nanocomposites shows similar
tendency, and the onset temperature of degradation (Tonset) can
be determined from these curves by extrapolating the curve at
the peak of degradation back to the initial weight of the poly- Fig. 6. DTA and TGA curves of (a) pure PCL, (b) 0.25 wt% PCL/MWCNT, (c)
mer. The Tonset of PCL is 142  C and extensively increases to 0.5 wt% PCL/MWCNT and (d) 1 wt% PCL/MWCNT composites (inserted
168, 174 and 183  C as the loading of 0.25, 0.5 and 1 wt% TGA data of each specimen from 80 to 290  C).
MWCNT content, respectively. From these experimental re-
sults, it can be seen that the presence of MWCNT in PCL induced better thermal stability and therefore the degradation
starting temperature clearly shifted to higher temperatures.
The isothermal degradation behaviors of PCL and PCL/
MWCNT nanocomposites were also investigated at predeter-
mined temperatures in a nitrogen flow environment. The weight
loss profiles of PCL and 1 wt% PCL/MWCNT nanocomposites
during isothermal heating in TGA at various temperatures are
shown in Fig. 7. The remaining weight decreased as the heating
times increase and the higher isothermal temperatures were be-
lieved to produce the larger weight loss. It can be clearly seen
that the residual weight of 1 wt% PCL/MWCNT nanocomposites
is larger than that of PCL. This result shows that the addition of
1 wt% MWCNT can reduce the degradation rate of PCL matrix.
The isothermal degradation kinetics of PCL and PCL/
MWCNT nanocomposites can be analyzed using the well
known Freeman and Carroll’s method [17] to determine the or-
der and activation energy of the degradation. The general form
of degradation is
Fig. 5. X-ray diffraction patterns of (a) PCL and (b) 0.25 wt% PCL/MWCNT,
dW
(c) 0.5 wt% PCL/MWCNT, and (d) 1 wt% PCL/MWCNT composites isother- ¼ kd W n ð4Þ
mally crystallized at 44  C. dt
1014 E.-C. Chen, T.-M. Wu / Polymer Degradation and Stability 92 (2007) 1009e1015
Fig. 7. Weight loss profiles under isothermal time of PCL and 1 wt% PCL/
MWCNT composites at (a) 320  C, (b) 330  C, (c) 340  C, (d) 350  C, (e)
360  C, and (f) 380  C isothermal temperatures.
where W is the remaining weight, kd is the degradation rate
constant and n is the order of the degradation. If the degradation
follows first order decomposition, then Eq. (4) may be trans-
formed into
ln W0  ln W ¼ kd $t ð5Þ
where the W0 is initial weight. Plots of ln W against t for PCL
and 1 wt% PCL/MWCNT nanocomposites are shown in
Fig. 8. It can be found that a straight line is obtained for the iso-
thermal degradation at 320  C. This result indicates that the
degradation at this temperature can be assumed as the first order
decomposition with a steady-state rate constant. As the degra-
dation temperatures continuously increase to 360  C, the plots
of ln W versus t show similar tendency. However, all the TGA
curves for the isothermal degradation at 380  C have a concave
shape with the degradation rate becoming smaller and smaller,
suggesting that either the rate constant is not a constant, or the
order of reaction is not one, or both.
Using Freeman and Carroll’s model, the kinetic parameters
of the degradation were calculated assuming a first order
Fig. 8. The plot of logarithm of weight loss profiles versus isothermal time of
PCL and 1 wt% PCL/MWCNT composites at (a) 320  C, (b) 330  C, (c)
340  C, (d) 350  C, (e) 360  C, and (f) 380  C isothermal temperatures.
decomposition. The degradation rate constant approximately
described by Arrhenius equation can be rewritten as follows:
Ed
ln kd ¼ ln A  ð6Þ
RT
where A is the pre-exponential factor, Ed is the activation
energy. Fig. 9 shows the Arrhenius plots of the ln kd versus
1/T of PCL and PCL/MWCNT nanocomposites. The values
of Ed calculated using curve fitting of the experimental data
are 222.8, 222.3, 218.4 and 213.9 kJ/mol for PCL, 0.25, 0.5
and 1 wt% PCL/MWCNT nanocomposites, respectively. The
Ed value of 0.25 wt% PCL/MWCNT composites is close to
that of PCL. This result shows that the 0.25 wt% MWCNT
loading into PCL matrix does not extensively change the deg-
radation behavior of PCL. By adding more MWCNT content
into PCL, the Ed value continuously decreased. This can be at-
tributed to the incorporation of more MWCNT loading to PCL
induced a decrease in the degradation rate and an increase in
the residual weight for PCL/MWCNT nanocomposites. Never-
theless the Ed values of PCL and PCL/MWCNT composites
did not significantly change, suggesting that the addition of
 E.-C. Chen, T.-M. Wu / Polymer Degradation and Stability 92 (2007) 1009e1015
Fig. 9. Arrhenius plots of ln kd versus 1/T for PCL and PCL/MWCNT composites.
small amount MWCNT into PCL matrix does not change the
degradation mechanism of PCL [18].
4. Conclusions
From TEM images, the well-separated and randomly-
distributed PCL/MWCNT nanocomposites have been success-
fully fabricated by mixing the PCL and as-prepared MWCNT
in THF solution under ultrasonic treatment. DSC isothermal
crystallization results have shown that the activation energy
of PCL extensively decreases with increasing loading of
MWCNT contents, suggesting that the addition of MWCNT
into PCL matrix probably induced the more heterogeneous
nucleation during crystallization processes. From TGA/DTA
results, the addition of small amount MWCNT into PCL ma-
trix can enhance the thermal stability of PCL matrix as Tonset
clearly shifted to higher temperatures. Isothermal degradation
data reveal that the activation energy Ed of nanocomposites is
smaller than that of PCL. This phenomenon can be attributed
to the fact that the incorporation of more MWCNT loading to
PCL caused a decrease in the degradation rate and an increase
in the residual weight for PCL/MWCNT nanocomposites.
Acknowledgements
The financial support provided by NSC through the project
NSC95-2218-E-005-004 is greatly appreciated and the work
1015
partially sponsored by Advanced Material of Green Energy
Project (NCHU) funded by Ministry of Education is also
gratefully acknowledged.
References
[1] Iijima S. Helical microtubules of graphitic carbon. Nature 1991;354:
56e8.
[2] Treacy MMJ, Ebbesen TW, Gibson JM. Exceptionally high Young’s
modulus observed for individual carbon nanotubes. Nature 1996;381:
678e80.
[3] Tans SJ, Verschueren ARM, Dekker C. Room-temperature transistor
based on single carbon nanotube. Nature 1998;393:49e51.
[4] Ajayan PM, Stephan O, Colliex C, Trauth D. Aligned carbon nanotube
arrays formed by cutting a polymer resinenanotube composite. Science
1994;265:1212e4.
[5] Bianco A, Prato M. Can carbon nanotubes be considered useful tools for
biological applications? Adv Mater 2003;15:1765e8.
[6] Lanza RP, Langer R, Chick WL. Principles of tissue engineering.
San Diego, CA, USA: Kluwer Academic publisher; 1998.
[7] Wang Y, Rodriguez-Perez MA, Reis RL, Mano JF. Thermal and thermo-
mechanical behaviour of polycaprolactone and starch/polycaprolactone
blends for biomedical applications. Macromol Mater Eng 2005;290:
792e801.
[8] Ikada Y, Tsuji H. Biodegradable polyesters for medical and ecological
applications. Macromol Rapid Commun 2000;21:117e32.
[9] Zeng H, Gao C, Yan D. Poly(3-caprolactone)-functionalized carbon
nanotubes and their biodegradation properties. Adv Funct Mater 2006;
16:812e8.
[10] Saeed K, Park SY, Lee H-J, Baek J-B, Huh W-S. Preparation of electro-
spun nanofibers of carbon nanotube/polycaprolactone nanocomposite.
Polymer 2006;47:8019e25.
[11] Wu T-M, Chen E-C. Isothermal and nonisothermal crystallization kinet-
ics of poly(3-caprolactone)/multi-walled carbon nanotube composites.
Polym Eng Sci 2006;46:1309e17.
[12] Wu T-M, Chen E-C. Crystallization behavior of poly(3-caprolactone)/
multi-walled carbon nanotube composites. J Polym Sci Polym Phys
2006;44:598e606.
[13] Avrami M. Kinetics of phase change. I. General theory. J Chem Phys
1939;7:1103e12.
[14] Avrami M. Kinetics of phase change. II. Transformationetime
relations for random distribution of nuclei. J Chem Phys 1940;8:
212e24.
[15] Hay JN. Application of the modified Avrami equations to polymer
crystallisation kinetics. Br Polym J 1971;3:74e82.
[16] Alamo RG, Mandelkern L. Crystallization kinetics of ethylene copolymers.
Macromolecules 1991;24:6480e93.
[17] Freeman ES, Carroll B. The application of thermoanalytical techniques
to reaction kinetics. J Phys Chem 1958;62:394e7.
[18] Sivalingam G, Karthik R, Madras G. Kinetics of thermal degradation of
poly(3-caprolactone). J Anal Appl Pyrolysis 2003;70:631e47.