Nanocomposites

ISSN: 2055-0324 (Print) 2055-0332 (Online) Journal homepage: http://www.tandfonline.com/loi/ynan20

Long-term performance and durability of

polycarbonate/carbon nanotube nanocomposites

Leonid V. Pastukhov, Frans P. M. Mercx, Ton Peijs & Leon E. Govaert

To cite this article: Leonid V. Pastukhov, Frans P. M. Mercx, Ton Peijs & Leon E. Govaert
(2019): Long-term performance and durability of polycarbonate/carbon nanotube nanocomposites,
Nanocomposites, DOI: 10.1080/20550324.2018.1558799

To link to this article: https://doi.org/10.1080/20550324.2018.1558799

© 2019 The Author(s). Published by Informa

UK Limited, trading as Taylor & Francis
Group.

Published online: 19 Jan 2019.

Submit your article to this journal

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=ynan20
NANOCOMPOSITES
https://doi.org/10.1080/20550324.2018.1558799

RESEARCH ARTICLE

Long-term performance and durability of polycarbonate/carbon nanotube

nanocomposites
Leonid V. Pastukhova,b, Frans P. M. Mercxc, Ton Peijsd and Leon E. Govaerta
a
Eindhoven University of Technology, Department of Mechanical Engineering, Materials Technology Institute, Eindhoven, MB,
Netherlands; bDutch Polymer Institute (DPI), Eindhoven, AB, Netherlands; cSABIC, Bergen op Zoom, PX, Netherlands; dSchool of
Engineering and Materials Science, and Materials Research Institute, Queen Mary University of London, London, United Kingdom

ABSTRACT ARTICLE HISTORY

Due to their good mechanical properties, low density, and ease of processing polymer nano- Received 23 November 2018
composites are of interest for a multitude of applications in the automotive, electronics, and Revised 9 December 2018
leisure industry. Besides having an impact on short-term mechanical performance of poly- Accepted 11 December 2018
mers, the addition of nanoreinforcements can have also a significant effect on long-term
KEYWORDS
properties such as the resistance to static (creep) and cyclic (fatigue) loadings. However, des- Carbon nanotube; creep;
pite its significance there is a shortage of long-term mechanical performance data for durability; fatigue; failure
thermoplastic-based polymer nanocomposites. Reason being that existing characterization prediction; nanocomposite;
methods for long-term performance and durability are time consuming and limited in their thermoplastic
applicability. Here, an engineering approach to predict long-term time-to-failure of polycar-
bonate/carbon nanotube (PC/CNT) nanocomposites is presented based on short-term experi-
mentation with an application to both creep and fatigue. Results showed that the addition
of CNTs had an opposite effect on two important long-term failure mechanisms. Addition of
CNTs lead to improvements in durability in the plasticity-controlled failure regime, whereas
it had an adverse effect in the slow crack growth-controlled regime, meaning that in the lat-
ter regime nanocomposite performance was significantly less than that of the neat poly-
mer matrix.

1. Introduction
Polymer-based materials are being used in a variety
of load bearing applications where structural integ-
rity and long-term durability is of great importance.
Examples of such systems can be found in a wide
range of industrial applications in the transport,
electrical, leisure, and medical industry. In many
applications, severe loading conditions (e.g. creep
and fatigue) and environmental conditions (e.g.
temperature and humidity) have put an increased
demand on material performance [1]. In all of these
areas, endeavours to achieve performance increases
together with weight reductions has intensified
attempts to improve physical properties of poly-
meric systems, which among others have led to the

CONTACT Ton Peijs t.peijs@qmul.ac.uk School of Engineering and Materials Science, and Materials Research Institute, Queen Mary University
of London, London E1 4NS, United Kingdom
ß 2019 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
2 L. V. PASTUKHOV ET AL.
development of engineering composites. Traditional
polymer composites are multiphase materials made
by combining polymer matrices with various rein-
forcements. Aim of the reinforcement is to improve
mechanical properties of the polymer matrix by
utilising key properties of the filler such as high
stiffness and high strength. Traditional reinforce-
ments have dimensions at the micron scale (10
6
m). Examples of such fillers are glass, carbon, and
aramid fibres [2]. Typically, a large amount of fibers
is needed to substantially improve the overall
properties of a composite material (tens of weight
percent) often leading to loss of processability [3].
In recent years, advances in nanotechnology have
promoted the development of reinforcements with
dimensions at the nano-scale (10
9 m). This dimen-
sional shift has opened new possibilities for altering
physical properties of polymer matrices (mechanical,
electrical, thermal, etc.) by incorporating relatively
small amounts of nanofillers (several weight percent
or less) [3, 4]. Nanocomposites are being considered
as a potential replacement of traditional composites
or when combined with traditional reinforcements
as a means to add other functionalities (electrical,
thermal, sensing, etc.) or further improve the mech-
anical performance of traditional composites by cre-
ating hybrid or hierarchical composites [5–10].
Among potential materials that can be used as a
matrix for composite materials, thermoplastic poly-
mers offer advantages such as low density, high spe-
cific strength and toughness, together with ease of
processing. In order to achieve good dispersion of
nanofillers in the thermoplastic matrix, several
methods are reported in literature including solution
mixing, melt blending, and in-situ polymerization
[3, 11]. Some of the most common thermoplastic
materials used are acrylonitrile butadiene styrene
(ABS), polyethylene (PE), polypropylene (PP), poly-
amides (PA), polystyrene (PS), poly(methyl meth-
acrylate) (PMMA), polycarbonate (PC), etc. The
most common nanofillers are carbon nanotubes
(CNTs), nanoclays, and graphene based nanoplate-
lets (GNPs). CNTs can be considered as one atom
thick carbon layers rolled in small cylinders. Most
common types of CNT are: (i) single wall carbon
nanotubes (SWCNT) and (ii) multi-wall carbon
nanotubes (MWCNT). CNTs have often been used
to improve the stiffness of polymer matrices. The
Young’s modulus of a thermoplastic matrix varies
from 2000 MPa to 4000 MPa for amorphous poly-
mers and from 300 to 6000 MPa for semi-crystalline
polymers. Addition of CNTs which have superior
mechanical properties (Young’s moduli up to 1TPa
and tensile strengths up to 100 GPa [12]) can not
only greatly improve the mechanical properties of
polymer matrices but will also affect other physical
properties (electrical conductivity, thermal conduct-
ivity, fire retardancy, air and liquid permeability,
fatigue resistance etc.). These multi-functionalities
have opened new possibilities for potential industrial
applications of polymer nanocomposites. It has been
well-documented that the addition of CNTs signifi-
cantly affects short-term properties of thermoplastic
polymers and an increase in elastic modulus and
yield stress has been reported for a wide range of
thermoplastic matrices including PE [13–17], PS
[18–20], PMMA [21–24], PP [25–28], PA [29–31],
PC [32–35], and many more.
Being a relatively novel material, polymer nano-
composites have attracted a lot of attention; however,
due to the lack of long-term mechanical performance
data for polymer nanocomposites, the use of these
materials in industrial applications has been limited.
Therefore, development of material characterisation
methods and reliable long-term performance estima-
tion tools is of utmost importance for this class of
materials. Since the long-term mechanical perform-
ance of polymer composites reinforced with nanofil-
lers is matrix dominated [36], it is very likely that,
similar to unfilled polymer systems, under some
applied stress different failure mechanisms can be
encountered. On this account, it seems reasonable to
investigate the time- and stress-dependence of poly-
mer matrices and the influence of nanoreinforcements
on time-dependent failure of such polymer matrices.
There are well-established methodologies for the
characterisation of time-dependent failure of various
polymer systems. These methods are often based on
characterising creep rupture (plasticity-controlled
failure) [37, 38] and fatigue rupture (slow crack
growth-controlled failure) [39–41]. From these
approaches for a large selection of polymers
(unfilled and with traditional reinforcements), it has
been shown [42] that there is a clear distinction in
response under static (creep) and cyclic (fatigue)
loading for the same maximum applied stress as
shown in Figure 1. At higher applied stresses and
shorter times, the dominant failure mechanism is
plasticity-controlled failure. Here, the applied stress
leads to plastic flow within the material and these
plastic deformations accumulate over time until they
reach a critical deformation level, leading to material
failure. At lower applied stresses (and longer times)
another process is active: slow crack growth.
Imperfections within a polymeric material act as
precursors for cracks and under an applied load
these cracks will grow until they reach a critical size
resulting in failure [43–46]. By plotting in a double
logarithmic graph, the maximum applied stress ver-
sus time-to-failure for both static and cyclic loadings
(Figure 1b), an important observation can be made:
cyclic loading leads to an increase in resistance to

Figure 1. (a) Static vs cyclic loading. (b) Schematic illustration of the effects of cyclic loading on plasticity-controlled and crack
growth-controlled failure mechanisms in a thermoplastic polymer matrix.
plastically-controlled failure but a decrease in resist-
ance to crack growth. Hence, cyclic loading has an
opposite effect on both failure mechanism [42];
therefore, providing a powerful characterisation tool
for failure mechanism identification.
In this study, we will investigate how the addition
of CNTs affects plasticity-controlled failure and slow
crack growth-controlled failure of a polycarbonate
(PC) matrix under both static and cyclic loadings
along with the basic principles of phenomenological
models that can be used to estimate the lifetime of
polymer nanocomposites.
2. Background
2.1. Plasticity-controlled failure
Under an applied load, a polymer flows due to
increased polymer chain mobility [47]. During con-
stant strain rate experiments (tensile tests) plastic flow
is increasing with time until at some point (the yield
point) plastic flow within the material exactly matches
the experimentally applied strain rate (condition of
steady state plastic flow). As has been demonstrated
by Bauwens-Crowet et al. [48], steady state plastic
flow at the yield point in a constant strain rate tensile
test is identical to steady state flow during secondary
creep when a polymer is subjected to a constant load.
Therefore, stress and temperature dependence of plas-
tic flow under creep loading can be measured by
means of tensile tests at different applied strain rates.
This dependence can be described using Eyring’s acti-
vated flow theory [49]. For instance, in a simple case
where a single process governs the deformation (as is
the case for PC at room temperature and above), the
magnitude of plastic flow is defined by:
   
DU rV
e_ pl ðr; TÞ ¼ e_ 0 exp
sinh
|{z} RT kT
I |fflfflfflfflfflfflfflfflffl{zfflfflfflfflfflfflfflfflffl} |fflfflfflffl{zfflfflfflffl}
II III
where part I, e_ 0 is a rate factor, which is a par-
ameter that depends on the thermodynamic state of
NANOCOMPOSITES 3
the material (age of the amorphous phase). In part
II of the equation, the Arrhenius type exponent, DU
is the activation energy, R is the universal gas con-
stant, and T is the absolute temperature. These fac-
tors cover the temperature dependency of the plastic
flow rate. In part III, the hyperbolic sine, r is the
applied stress, V  is the activation volume, and k is
the Boltzmann’s constant. This part captures the
stress dependency of the flow rate. However, poly-
mers cannot flow indefinitely, and eventually plastic
deformations will localise and failure will occur.
Therefore, time-to-failure in the plasticity-controlled
region can be estimated by calculating the total
accumulated strain. Once this reaches its critical
value ecr , failure will occur:
ðt0
epl ðtÞ ¼ e_ pl ðr; T; t 0 Þdt with failure if epl ¼ ecr (2)
0
where epl is the plastic strain at a certain time, e_pl
is the plastic flow rate according to Equation 1, and
ecr is the plastic strain at failure. Note, that although
ecr is usually smaller than the actual strain-at-break,
this phenomenological measure has been shown to
be a reliable measure for predicting time-to-failure
of polymers [42, 48, 50]. Under a constant load
assuming constant plastic flow rate during a creep
test, Equation 2 can be simplified to:
ecr
tf ¼ (3)
e_ pl
Thus, knowing the deformation kinetics which
can be determined by performing tensile tests at
different strain rates, consequent time-to-failure
predictions can be made using the critical strain
concept based on Equation 3.
(1)
2.2. Crack growth-controlled failure
Crack growth-controlled failure manifests itself as a
propagation and growth of cracks under an applied
stress inside the material. Cracks are initiated within
4 L. V. PASTUKHOV ET AL.
the material due to the presence of flaws and imper-
fections that result in local stress concentrations.
The stress state around a crack tip can be described
using linear elastic fracture mechanics (LEFM) [51].
Considering Mode I loading (crack opening), the
stress state at the crack tip is defined by a stress
intensity factor as KI :
pffiffiffiffiffiffi
KI ¼ Yr pa (4)
where r is the applied remote stress, a is the
crack length, and Y is a geometry factor. The evolu-
tion of crack length with time or, in other words,
crack propagation rate, a_ is related to the stress
intensity factor by a power law; this relation is
known as the Paris’ law [52]:
a_ ¼ AKI m (5)
It has been shown by Gray et al. [53], that, by
knowing the crack propagation kinetics (parameters
A and m in the Paris’ law), time-to-failure can be
predicted, assuming some initial crack length and
neglecting the time of crack initiation by means of
the following equation:
 
m
r crosshead speed, resulting in constant engineering
tf ¼ (6)
cf
where cf is a scaling factor, which can be under-
stood as the level of stress leading to a failure time of
1 s. Therefore, according to Equation 5, a log-log plot of
time-to-failure versus applied stress results in a straight
line. Dependence of the time-to-failure versus applied
stress is also a straight line in a double logarithmic plot
with a slope
1=m according to Equation 6.
3. Experimental
3.1. Materials
The following materials were selected for this study:
polycarbonate (PC) and polycarbonate reinforced
with carbon nanotubes (PC þ CNT). The grade of
polycarbonate for both unfilled and nanocomposite
samples was - LEXANTM resin 141 R, a bisphenol-A
based thermoplastic from SABIC (Bergen op Zoom,
Netherlands), while multiwalled carbon nanotubes
(MWCNT) were obtained from Hyperion. The PC/
CNT nanocomposites were in-house compounded at
SABIC by diluting a 10 wt.% PC/CNT masterbatch
using twin-screw extrusion (Coperion ZSK25,
L/D ¼ 32, 300 rpm, 20 kg/hr throughput, melt
temperature 300  C) to compounds with 1 wt.% and
3 wt.% CNT loading. Subsequently these nanocom-
posites were injection molded (Engel 45, melt
temperature 300  C, mold temperature 80  C) into
tensile bars with a geometry according to ISO 527
Type 1 A. In order to avoid effects of moisture,
samples were stored in a sealed bag with silicone gel
moisture absorbent.
3.2. Heat treatment
Annealing was performed in an air circulated oven at
a temperature of 120  C (i.e. below the glass transi-
tion temperature, Tg). Initial equilibration of the
sample temperature was measured to take approxi-
mately 10 min. After a predefined ageing time, the
samples were removed from the oven and allowed to
cool to room temperature in air before being tested.
3.3. Mechanical testing
All uniaxial tensile and creep measurements at
room temperature were carried out using an Instron
5566 universal tensile machine equipped with a
10 kN load cell. The temperature in the lab was held
constant at 24  C throughout all tensile and creep
experiments. Engineering stresses were calculated
using the average of the cross-sectional surface area
as measured at three locations in the gauge section.
Tensile experiments were carried out at a constant
strain rates ranging from 10
5 s
1 to 0.05 s
1. Creep
tests were performed at constant load and, therefore,
constant engineering stress. To avoid overshoot
effects, the load was slowly applied within 10 s.
Clamps were carefully aligned to ensure that tor-
sional stresses did not affect the results. The creep
rupture tests were performed in uniaxial extension
over a wide range of stresses. These ranges were
carefully chosen for each material and temperature
in order to obtain failure within a time-scale rang-
ing from 102 s to 105 s (longest experimental time-
scale is approximately 2 days). Time-to-failure was
determined by the time of fracture. For unfilled PC
and PC þ 3 wt.% CNT two sets of experiments were
carried out at different temperatures (40  C and
60  C). These tests were performed on an Instron
8801 universal tensile machine equipped with an
environmental chamber. The samples were placed in
the chamber and once the temperature had reached
a set value, the tests were started after 10 min (once
the temperature inside the chamber equilibrated).
Cyclic tests were carried out on an Instron 8872 ser-
vohydraulic fatigue testing system. For each set of
samples, the stiffness was determined first in order
to ensure the correct tuning of the measured signal.
Load ratio and frequencies were changed according
to the test requirements. For all tests, a sinusoidal
cyclic load was applied, while for each test the load
amplitude, maximum load, and frequency were kept
constant. Temperature in the lab was at a constant
20  C throughout all cyclic fatigue experiments.

Figure 2. (a) Strain rate dependence of the yield stress of PC at various temperatures. The solid lines represent the best-fit
using Equation 1. (b) Stress dependence of the time-to-failure of PC in uniaxial loading. The solid lines represent predictions
using Equation 3.
Table 1. Eyring-parameters and the average critical strain
for unfilled PC.
Material V  [nm3] DU [kJ/mol] e_ 0 [s
1]
Unfilled PC (as-moulded) 3.4 315 1.53 1031
Unfilled PC (annealed) 3.4 315 1.09 1028
3.4. Scanning electron microscopy
Dispersion of CNTs in the PC matrix was evaluated
using a FEI Inspect F scanning electron microscope
(SEM). For this, injection molded tensile bars
(untested) were cold fractured in liquid nitrogen.
Samples were coated with a thin layer of gold
before imaging.
3.5. Dynamic mechanical thermal analysis
Dynamic mechanical analysis (DMTA) was per-
formed on a TA Instruments Q800. Rectangular
samples of PC and PC þ CNT nanocomposites were
prepared with dimensions of 1  4.5  30 mm3. A
frequency of 1 Hz was used in a strain-controlled
experiment. The storage and loss modulus were
measured in a temperature window ranging from

80  C to 200  C.
4. Results and discussion
4.1. Plasticity-controlled failure
4.1.1. Unfilled PC
Applied stress leads to increased polymer mobility
and stimulates molecular rearrangements [54],
resulting in a constant rate of plastic flow [49]. The
magnitude of this plastic flow rate depends on strain
rate (time) and temperature. Consequently, if the
applied strain rate and temperature changes, the
yield stress of the polymeric material will change
accordingly in order to find an equilibrium between
applied strain rate and intrinsic plastic flow rate.
The accumulated critical strain [50], connecting
NANOCOMPOSITES 5
deformational kinetics of plastic flow rate to time-
to-failure predictions, can be used to make life-time
ecr [
] predictions based on constant strain-rate test data.
0.014 Figure 2 shows results of this type of modelling for
0.014
unfilled PC at different temperatures, proving that
accurate predictions for different temperatures can
be made using the concept of critical strain.
Corresponding Eyring parameters for the stress and
temperature dependency for unfilled PC are listed
in Table 1.
The yield stress of an amorphous polymer is
strongly influenced by physical ageing [55]. Physical
ageing leads to an increased resistance against plas-
tic deformation, resulting in a higher yield stress at
the same strain rate in comparison to unaged
material [55, 56].
Figure 2 also shows that the rate and temperature
dependence of an amorphous polymeric material
can be described well by Equation 1 by only chang-
ing the preexponential factor e_ 0 . Figure 3a shows
the results of as-molded samples in comparison to
annealed samples (3 hrs at 120  C). The change in
deformation kinetics is accurately captured by
changing e_ 0 as shown in Figure 3a. Using the same
value of ecr , time-to-failure of annealed samples can
be successfully predicted as demonstrated in
Figure 3b.
4.1.2. PC/CNT nanocomposites
In order to understand the effect of CNTs on short-
and long-term performance of the PC matrix, first
tensile tests at a constant strain rate of 10
3 s
1 and
creep tests at different levels of applied stress at
room temperature have been performed. Results of
these initial tensile tests are shown in Figure 4a and
reveal that the addition of nanotubes leads to an
increase in yield stress (10% at 3 wt.% CNT) and
Young’s modulus (40% at 3 wt.% CNT) compared
to unfilled PC.
6 L. V. PASTUKHOV ET AL.

The evolution of plastic strain rate during creep
experiments can be visualised in a so-called Sherby-
Dorn plot [57] (Figure 4b). A Sherby-Dorn plot for
PC þ 3 wt.% CNT at four different creep stresses
shows that three distinct regions of creep can be
recognised: (i) a region of decreasing strain rate
(primary creep), (ii) a region where plastic flow rate
remains approximately constant (secondary creep),
followed by (iii) a region with increasing strain rate
(tertiary creep) accompanied by strain localisation
in the form of necking. These observations indicate
that the influence of the addition of CNTs in the
polymer matrix can be captured using the same
concept as used for describing deformation kinetics
and time-to-failure prediction of unfilled PC.
Strain rate dependence of the yield stress and
time-to-failure dependence of PC/CNT nanocompo-
sites has been compared with results of unfilled PC.
This comparison is shown in Figure 5. A notable
observation from this comparison is that the add-
ition of CNTs does not affect the activation volume
V  , i.e. the slopes of the lines remain the same with
only a parallel shift. To further elaborate on this
idea PC/CNT nanocomposites were tested at differ-
ent temperatures. By performing tests at different
temperatures similar to the tests on unfilled polycar-
bonate and subsequently using Eyring modelling,
deformation kinetics of PC/CNT nanocomposites
can be accurately described as can be seen in Figure
6a and c for PC þ 1 wt.% CNT and PC þ 3 wt.%
CNT, respectively. Eyring parameters for PC/CNT
nanocomposite materials are presented in Table 2.
Time-to-failure predictions using the critical
strain concept are shown in Figure 6b and d for PC
þ 1 wt.% CNT and PC þ 3 wt.% CNT, respectively.
In creep predictions for nanocomposites, values of
the critical strain were found to be slightly less than
those for unfilled PC as shown in Figure 7. At the
same time, it was observed that the strain-at-break
in tensile tests was lower for the nanocomposite
samples. Neat PC matrix typically has an elongation
at break of around 100%, while PC þ 3 wt.% CNT
showed a significantly reduced elongation at break
of approximately 15% as a result of an embrittle-
ment effect by the nanofiller. Some deviation of
experimental data from predictions at higher tem-
peratures and longer times can be explained by
stress and temperature induced physical ageing dur-
ing the long-term creep measurements (for instance
as in the case of the longest measurement at 60  C

Figure 3. (a) Strain rate dependence of the yield stress of PC for as-moulded and annealed samples. The solid lines represent
a best-fit using Equation 1. (b) Stress dependence of the time-to-failure of PC in uniaxial loading. The solid lines represent
predictions using Equation 3.

Figure 4. (a) Stress versus strain for unfilled PC and PC/CNT nanocomposite with 1 wt.% and 3 wt.% CNT. Markers indicate
the yield stress. (b) Sherby-Dorn plot for several applied stresses.
 NANOCOMPOSITES 7

in Figure 6b). Interestingly, the values of activation
volume and activation energy are similar for both
unfilled PC and PC/CNT nanocomposites. The only
parameter that has to be changed to capture the
deformation kinetics of the nanocomposites is the
preexponential factor_e 0 . This is similar to the case of
annealed PC and therefore, interestingly, the addition
of CNTs has the same effect on yield stress and
deformation kinetics as changing the thermodynamic
state of the unfilled matrix. To further elaborate on
this observation, two sets of unfilled PCs were
annealed at 120  C for two different periods of time
(20 min and 2 hrs, respectively). A temperature of
120  C was chosen in order to observe effects of
annealing at shorter times based on the time-tem-
perature superposition principle (annealing at higher
temperature allows one to achieve the same increase
in yield stress at a shorter time). Annealing times
were chosen in such a way that the resulting yield
stresses were close to those of the nanocomposite
materials. Ageing kinetics of PC at different tempera-
tures was extensively studied by Klompen et al. and

Figure 5. (a) Strain rate dependence of the yield stress of unfilled PC and PC/CNT nanocomposites with 1 wt.% and 3 wt.%
CNT. The solid lines are guides to the eye. (b) Stress dependence of the time-to-failure of unfilled PC and PC/CNT nanocompo-
sites in uniaxial loading. The solid lines are guides to the eye.

Figure 6. Strain rate dependence of the yield stress of PC þ 1 wt.% CNT (a) and (c) PC þ 3 wt.% CNT. The solid lines
represent the best-fit using Equation 1. Stress dependence of the time-to-failure of PC þ 1 wt.% CNT (b) and PC þ 3 wt.%
CNT (d) in uniaxial loading. The solid lines represent predictions using Equation 3.
8 L. V. PASTUKHOV ET AL.
Engels et al. [58–60], providing all the necessary
information regarding the time required to obtain a
certain level of yield stress at different temperatures.
Having annealed unfilled PC samples for a prede-
fined time, yield stresses were measured at different
strain rates and results are presented in Figure 8 for
both 1 wt.% and 3 wt.% nanocomposites.
Experimental observations in plasticity-controlled
failure have shown that the addition of CNTs in the
PC matrix leads to an increase of the yield stress by
10%, which results in a 100-fold lifetime improve-
ment compared to unfilled PC. Clearly, in the plasti-
city-controlled failure regime the addition of CNTs
is highly beneficial with respect to long-term mech-
anical behaviour. Moreover, the deformation kinet-
ics of the PC/CNT nanocomposite systems can be
described by the same set of values of activation vol-
ume and activation energy in the Eyring equation,
with the only parameter that changes being the
value of e_ 0 , similarly to that of aged (annealed) sam-
ples. Therefore, the increase in yield stress and elas-
tic modulus and subsequent improvement in
lifetime for PC/CNT nanocomposites under a con-
stant load are similar to those of annealed unfilled
PC samples. An increased yield stress (either due to
annealing or the addition of nanofiller) results in an
increased resistance to plastic deformations.
Consequently, at the same applied stress it will take
longer time to reach the critical amount of plastic
strain, thus leading to a marked increase in creep
lifetime at the same applied stress.
4.2. Crack growth-controlled failure
In order to fully characterise the long-term perform-
ance of polymer systems, one needs to be aware
Table 2. Eyring parameters and the average critical strain
for PC/CNT nanocomposites.
Material V  [nm3] DU [kJ/mol] e_ 0 [s21]
PC þ 1 wt.% CNT 3.4 315 1.5 1030
PC þ 3 wt.% CNT 3.4 315 9.25 1028
Figure 7. (a) Plastic flow rate during secondary creep versus time-to-failure for unfilled PC, PC þ 1 wt.% CNT, and PC þ
3 wt.% CNT. (b) Critical strain dependence on CNT loading. Dotted line is a guide for the eye.
that there are two stress dependent failure mecha-
nisms. As previously discussed, there is plasticity-
controlled failure, which is typically observed during
basic mechanical characterisation or tensile testing.
To observe crack growth-controlled failure under a
static load, long-term measurements are required
with the more durable materials requiring longer
times to observe and identify slow crack growth-
controlled failure. One of the ways to accelerate
crack propagation, and consequently, reduce the
experimental time-scale required to observe crack
growth, is to apply a high amplitude tension-tension
cyclic (fatigue) load thus increasing the crack propa-
gation rate [61]. In unreinforced polymers, typically
the change in failure kinetics can be observed
macroscopically by looking at the way the material
has failed: in case of plasticity-controlled failure
clear strain localisation (necking) is visible, whereas
in crack growth-controlled failure, failure occurs in
a brittle manner with some cracks being visible (as
in a transparent material such as PC) [42].
In the current study, first a cyclic load with two
 
different R – values R ¼ rrmaxmin
of 0.1 and 0.25 was
applied to unfilled PC to demonstrate the effect of
cyclic load on both plasticity-controlled and crack
growth-controlled failure. Results are shown in
Figure 9a. Under high amplitude cyclic loading both
failure mechanisms can be identified with distinct
differences in macroscopic failure as shown in
Figure 9b. Plasticity-controlled failure was accompa-
nied with large plastic deformations and strain local-
isation (necking), whereas under lower applied
stress less plastic deformation has been accumulated,
and clear brittle fracture can be observed. Moreover,
a clear transition from one mechanism to the other
is visible when plotting the data in a log-log plot.
Similar to unfilled PC, static and cyclic tests were
performed for PC/CNT nanocomposite systems.
ecr [
] Results are shown in Figure 10 for both PC þ
0.012 1 wt.% CNT and PC þ 3 wt.% CNT (Figure 10a and
0.07
b, respectively). In Figure 11, nanocomposite
 NANOCOMPOSITES 9

samples after failure are illustrated having similar
macroscopic failure features as unfilled PC. With
regards to plasticity-controlled failure, as expected,
nanocomposite samples show superior behaviour
compared to unfilled PC (for PC þ 3 wt.% CNT
improvements in time-to-failure are in the order of
two decades). However, a very noteworthy observa-
tion is that in the crack growth-controlled failure
region the addition of CNTs has reduced the life-
time under cyclic loading (for PC þ 3 wt.% CNT the
shift in time-to-failure is in the order of one
decade). Having seen that in the plasticity-controlled
failure region behaviour of annealed unfilled PC
and PC nanocomposites were similar, the same two
sets of annealed unfilled PC samples have been
prepared and were subsequently tested under
a cyclic load at R ¼ 0.1.
Remarkably, the nanocomposite samples did not
only show similar behaviour in the plasticity-con-
trolled failure region but showed similar behaviour
in the crack growth region as well. Results of such
comparison are shown in Figure 12 for both PC þ
1 wt.% CNT and PC þ 3 wt.% CNT (Figure 12a and
b, respectively).
The reason for these similarities might be due to
the fact that the incorporation of CNTs in a poly-
mer matrix can lead to the local reduction of mobil-
ity of polymer chains due to a decrease in free
volume, which in turn can explain the apparent
increase in yield stress and elastic modulus.
As can be seen from the SEM images in Figure 13,
the CNTs are well and homogeneously dispersed
in the PC matrix. Many nanotubes are visible owing
to the high contrast between them and the polymer
matrix, with the PC/CNT nanocomposites demon-
strating a relatively uniform filler distribution across
the cold fractured surface. Immobilization of poly-
mer chains near the CNT surface has been observed
to lead to an increase in Tg, and, hence, a decrease
in molecular mobility [62].
In order to verify this influence of CNT content
on polymer mobility, the Tg of unfilled PC and PC

Figure 8. (a) Strain rate dependence of the yield stress of unfilled PC, PC þ 1 wt.% CNT and annealed PC (20 min at 120  C).
(b) Strain rate dependence of the yield stress of unfilled PC, PC þ 3 wt.% CNT and annealed PC (2 hrs at 120  C).

Figure 9. (a) Time-to-failure versus maximum applied load under static (creep) and cyclic (fatigue) loading conditions using
several R-values for unfilled PC. Test frequency was 1 Hz and temperature 20  C. (b) Macroscopically observed differences
between plasticity-controlled (necking) and crack growth controlled (brittle fracture) failure of unfilled PC.
10 L. V. PASTUKHOV ET AL.

þ 3 wt.% CNT were measured using DMTA (Figure
14). Based on the observed increase in yield stress
an increase in Tg would be anticipated for these
nanocomposite systems. However, similar to a previ-
ous study [35], only a very small increase in Tg was
observed for the PC/CNT nanocomposites (159  C
for PC þ 3 wt.% CNT compared to 158  C for neat
PC, determined by the maximum of the tan delta
peak). This increase is by far not sufficient to
explain the strong increase in yield stress observed
in Figure 5. It has been suggested that the addition
of nanofillers can lead to an anti-plasticization
effect, which promotes a higher resistance to plastic
deformation [63]. However, it should be noted that
anti-plasticization is generally accompanied by a
marked decrease in Tg, which is not observed in the
PC/CNT nanocomposites. The fact that the response
of the PC/CNT nanocomposites resembles that of
an annealed PC closely, rather seems to suggest that
the presence of CNTs influences the evolution
of the physical ageing process. Once cooled below
Tg, the polymer glass will no longer be in a thermo-
dynamic equilibrium state and will display a con-
tinuous strive to regain it. Even during fast cooling,
this leads to a considerable increase of the yield
stress compared to unaged samples [59]. In the

Figure 10. (a) Time-to-failure versus maximum applied load under static (creep) and cyclic (fatigue) loading for unfilled PC
and PC þ 1 wt.% CNT. (b) Time-to-failure versus maximum applied load under static and cyclic loading for unfilled PC and PC
þ 3 wt.% CNT. Both sets of experiments are performed at 3 Hz and 20  C.

Figure 11. Samples after failure of PC þ 3 wt.% CNT, corresponding to both failure regions; (bottom) plasticity-controlled fail-
ure (necking), and (top) crack growth-controlled failure (brittle fracture).

Figure 12. (a) Time-to-failure versus maximum applied load under static (creep) and cyclic (fatigue) loading for unfilled PC, PC
þ 1 wt.% CNT, and annealed PC (20 min at 120  C). (b) Time-to-failure versus maximum applied load under static and cyclic
loading for unfilled PC, PC þ 3 wt.% CNT, and annealed PC (2 hrs at 120  C). Both sets of experiments are performed at 3 Hz,
R ¼ 0.1, and 20  C.
 NANOCOMPOSITES 11

unfilled matrix, the physical ageing process involves
localised, subtle changes in chain conformation that
allows chain segments to rearrange with respect to
each other and strengthen their interaction.
Apparently, interactions between (segments of)
polymer chains and nanofillers intensify markedly
during physical aging, leading to a stronger increase
of yield stress with time. Close to Tg, chain mobility
is much higher due to thermal activation, and con-
formational changes occur over a much larger scale.
This enhanced mobility will make that the stronger
interactions between chain segments and nanopar-
ticles, as created during ageing, will have disap-
peared close to Tg.
With regard to the observed differences in the
crack growth region of failure: at the initial stage of
deformation crack tip blunting occurs during a
period of crack initiation. The stress state in such a
case can be described by yielding of material in the
vicinity of a crack tip, and the subsequent formation
of a plastic zone. The size of this plastic zone is dir-
ectly related to the value of the yield stress; a low
yield stress will produce a large plastic deformation
zone at the crack tip that will effectively lead to
crack blunting which increases fracture toughness
[64] and is hypothesized to decrease fatigue crack
propagation. Vice versa, an increase in yield stress
due to the incorporation of CNTs will reduce the
size of the plastic zone and, therefore, lowers the
effectiveness of this crack tip blunting mechanism
which, in turn, leads to an increased crack propaga-
tion rate and a lowered failure resistance of the
polymer nanocomposites. An increase in yield stress
due to annealing has been shown to have a similar
detrimental effect on PC’s ability to stabilise crack
growth [65]. As mentioned earlier, nanocomposites
showed also a more brittle failure behaviour in
tensile tests, with PC þ 3 wt.% CNT having an
elongation at break of approximately 15% as com-
pared to 100% for neat PC matrix.
Relatively large amounts of nanofiller reinforce-
ment can lead to the formation of agglomerates and
very significant embrittlement effects as a result of
strain concentrations due to the stiffness mismatch
between polymer and nanofiller [66]. The strength
of the interphase can also be a factor affecting

Figure 13. (a) SEM micrographs of PC þ 1 wt.% CNT nanocomposites at a magnification of 30,000. (b) SEM micrographs
of PC þ 3 wt.% CNT nanocomposites at a magnification of 30,000.

Figure 14. DMTA result for (a) unfilled PC and (b) PC þ 3 wt.% CNT.
12 L. V. PASTUKHOV ET AL.

overall fatigue resistance [67]. Moreover, it has been
demonstrated that samples generally display reduced
fracture toughness in the case of both annealing
[68] and the addition of CNTs (for instance, in
PMMA the reduction in Mode I fracture toughness
with the addition of 1 vol.% MWCNTs is 47%) [69].
Lower fracture toughness will lead to a reduced
critical crack length at the same applied stress level
and, therefore a shorter time-to-failure.
4.3. Extrapolation to static loading in crack
growth-controlled failure
Kanters et al. [70] proposed a phenomenological,
crack-propagation based model which takes into
account the influence of both frequency and load
Table 3. Parameters to describe the time-to-failure as
function of R-value and frequency using Equation 7 for
PC þ 3 wt.% CNT.
Material a[
] ctf ;static [MPa s1/m] cN f ;cyclic [MPa]
PC þ 3 wt.% CNT 3.11 145690 2382
amplitude on the time-to-failure in cyclic tension-
tension fatigue. The model is based on multiplica-
tive decomposition of crack propagation rates by
assuming two reference limit cases: one at R ¼ 0
(cyclic) is assumed to scale with frequency and
another at R ¼ 1 (static) is frequency independent.
Time-to-failure for any combination of R-value and
frequency (for tension-tension fatigue) can be
described using Equation 7 and parameters listed in
the Table
 3:  
m
log tf ðR; f Þ ¼ Ra  log ct r
f ;static
 !
m
a 1 r
þð1
R Þ  log ð7Þ
f cNf ;cyclic
where a is the parameter to capture the non-linear
dependency on R-value (load ratio), r is the
maximum applied stress, f is the frequency of the
applied cyclic load, cNf ;cyclic cyclic and ctf ;static static are
m [
] pre-factors for two reference loading cases (R ¼ 0
2.2
and R ¼ 1, respectively).

Figure 15. PC þ 3 wt.% CNT: (a) Time-to-failure under cyclic fatigue loading versus maximum applied stress, for different load
ratios at 3 Hz. Markers represent measurements, solid lines predictions using Equation 7. (b) Time-to-failure in cyclic fatigue for
a maximum applied stress of 45 MPa versus load ratio. Lines represent the time-to-failure calculated using Equation 7.

Figure 16. PC þ 3 wt.% CNT: (a) Time-to-failure under cyclic fatigue loading versus maximum applied stress for different load
ratios at 1 Hz. Markers represent measurements, solid lines predictions using Equation 7. (b) Time-to-failure under cyclic fatigue
loading versus maximum applied stress for different stress ratios at 10 Hz. Markers represent measurements, solid lines predic-
tions using Equation 7.
 NANOCOMPOSITES 13

The model has been validated for different also signify that overall long-term performance
frequencies and R-values as presented in Figure 15 improvements of nanocomposites can be signifi-
and 16. It is shown that the model gives adequate cantly less than those observed during typical short-
predictions of time-to-failure for polymer nanocom- term material characterisation tests.
posites. Extrapolation allows one to predict also
static failure based on relatively short-term cyclic
Notes on contributors
experiments. Thus, for the first time, the long-term
mechanical performance of polymer nanocomposites Leonid Pastukhov is a PhD
was studied not only experimentally but was also student in the Polymer
described by physical models, giving new significant Technology Group at the
insights into the fundamentals of mechanical behav- Department of Mechanical
Engineering at Eindhoven
iour of polymer nanocomposites. University of Technology.
His research focuses on
experiential assessment of
5. Conclusions mechanical performance and
In this study an engineering approach for the evalu- durability of polymers and
ation of long-term performance of polymer nano- polymer composites.
composites has been introduced, with an application
to both creep and fatigue. Cyclic loading has shown
to be an effective technique for failure mechanism Frans Mercx is a chief scien-
identification due to the strong acceleration of crack tist at SABIC T&I, Specialties,
LNP in Bergen op Zoom, the
propagation. Deformation kinetics in the plasticity-
Netherlands. His expertise
controlled failure regime of polymer nanocompo- lies in polymeric compounds
sites can be characterised by performing tensile tests and the use of additive tech-
at different strain rates and can be modelled by nologies for improving/engin-
using the Eyring equation. Remarkably, the activa- eering the performance of
tion volume and activation energy of the nanocom- plastics with a recent focus on
thermally and electrically
posites were identical to those of the aged neat PC conductive composites.
matrix. Therefore, it was hypothesised that the
marked increase in yield stress and time-to-failure
in the plasticity-controlled failure region is due to Ton Peijs was a professor of
an increased resistance to plastic deformation Polymer Technology at
Queen Mary University of
mainly as a consequence of reduced chain mobility
but recently joined the
due to interactions between sections of the polymer Warwick Manufacturing
chain and nanofillers. In the crack-growth con- Group (WMG) at London
trolled region of failure this influence persisted. The the University of Warwick.
reduction in the size of the plastic zone behind a His research area is in the
field of structure-processing-
crack tip may lead to an increase in macroscopic
property relationships in pol-
crack propagation rate and/or decrease in fracture ymers and their composites.
toughness, which in turn will result in embrittle-
ment and a reduced durability of polymer nanocom-
posites. These observations confirm that a Leon Govaert is an associate
professor at the Department
comparison of long-term mechanical performance of Mechanical Engineering at
between materials or the effect of the addition of Eindhoven University of
nanoreinforcements to a polymer matrix cannot be Technology. His research
justified simply by comparing short-term properties focuses on the long-term
(yield stress, creep rate, etc.). In addition, it is mechanical performance of
solid polymers and their
important to understand the deformation and crack
composites. He is also a
propagation kinetics and to identify underlying fail- part-time professor in
ure mechanism in the polymer matrix under given Polymer Mechanics at the
conditions. This study showed that the addition of University of Twente.
CNTs has an opposite effect on two important fail-
ure mechanisms in long-term mechanical behaviour,
with CNTs having a positive effect in the plasticity- Disclosure statement
controlled regime, while having a negative effect in No potential conflict of interest was reported by
the crack growth controlled regime. These results the authors.
14 L. V. PASTUKHOV ET AL.
References
1. Sonsino C, Moosbrugger E. Fatigue design of highly
loaded short-glass-fibre reinforced polyamide
parts in engine compartments. Int J Fatigue. 2008;
30:1279–1288.
2. Thomas S, Joseph K, Malhotra SK, et al. Polymer
Composites, Macro-and Microcomposites, vol. 1.
Weinheim, Germany: Wiley-VCH Verlag & Co.; 2012.
3. Plesa I, Notingher PV, Schlogl S, et al. Properties
of polymer composites used in high-voltage applica-
tions. Polymers. 2016;8:173.
4. Tanaka T, Montanari G, Mulhaupt R. Polymer
nanocomposites as dielectrics and electrical
insulation-perspectives for processing technologies,
material characterization and future applications.
IEEE Trans Dielect Electr Insul. 2004;11:763–784.
5. Qian H, Greenhalgh ES, Shaffer MS, et al. Carbon
nanotube-based hierarchical composites: a review.
J Mater Chem. 2010;20:4751–4762.
6. Grimmer CS, Dharan C. High-cycle fatigue of
hybrid carbon nanotube/glass fiber/polymer compo-
sites. J Mater Sci. 2008;43:4487.
7. Yavari F, Rafiee M, Rafiee J, et al. Dramatic increase
in fatigue life in hierarchical graphene composites.
ACS Appl Mater Interfaces. 2010;2:2738–2743.
8. Manjunatha C, Taylor A, Kinloch A, et al. The
tensile fatigue behaviour of a silica nanoparticle-
modified glass fibre reinforced epoxy composite.
Compos Sci Technol. 2010;70:193–199.
9. Knoll J, Riecken B, Kosmann N, et al. The effect
of carbon nanoparticles on the fatigue performance
of carbon fibre reinforced epoxy. Compos Part A:
Appl Sci Manufacturing. 2014;67:233–240.
10. Jagannathan N, Anilchandra A, Manjunatha C.
A study on the fatigue performance of a glass
fiber-epoxy polymer nanocomposite under random
loads. Nanocomposites. 2015;1:138–144.
11. Haider S, Haider A, Almasry WA, et al.
Thermoplastic nanocomposites and their processing
techniques. London: INTECH Open Access
Publisher; 2012.
12. Wu Y, Huang M, Wang F, et al. Determination of
the Young’s modulus of structurally defined carbon
nanotubes. Nano Lett. 2008;8:4158–4161.
13. El Achaby M, Qaiss A. Processing and properties
of polyethylene reinforced by graphene nanosheets
and carbon nanotubes. Mater. Design. 2013;44:81–89.
14. Yang B, Pramoda KP, Xu GQ, et al. Mechanical
reinforcement of polyethylene using polyethylene-
grafted multiwalled carbon nanotubes. Adv Funct
Mater. 2007;17:2062–2069.
15. Mahfuz H, Adnan A, Rangari VK, et al.
Manufacturing and characterization of carbon
nanotube/polyethylene composites. Int J Nanosci.
2005;4:55–72.
16. Shofner ML, Khabashesku VN, Barrera EV.
Processing and mechanical properties of fluorinated
single-wall carbon nanotube polyethylene compo-
sites. Chem Mater. 2006;18:906–913.
17. Ruan S, Gao P, Yang XG, et al. Toughening high
performance ultrahigh molecular weight polyethyl-
ene using multiwalled carbon nanotubes. Polymer.
2003;44:5643–5654.
18. Blake R, Coleman JN, Byrne MT, et al. Reinforcement
of poly (vinyl chloride) and polystyrene using
chlorinated polypropylene grafted carbon nanotubes. J
Mater Chem. 2006;16:4206–4213.
19. Safadi B, Andrews R, Grulke E. Multiwalled carbon
nanotube polymer composites: Synthesis and char-
acterization of thin films. J Appl Polym Sci. 2002;
84:2660–2669.
20. Thostenson ET, Chou T-W. Aligned multi-walled
carbon nanotube-reinforced composites: Processing
and mechanical characterization. J Phys D: Appl
Phys. 2002;35:L77.
21. Gorga RE, Cohen RE. Toughness enhancements in
poly (methyl methacrylate) by addition of oriented
multiwall carbon nanotubes. J Polym Sci B Polym
Phys. 2004;42:2690–2702.
22. Zamfirova G, Gaydarov V, Faraguna F, et al.
Influence of carbon nanotubes functionalization on
the mechanical properties of polymethacrylate
nanocomposites. Colloids Surfaces A: Physicochem
Eng Aspects. 2016;510:169–175.
23. Blond D, Barron V, Ruether M, et al. Enhancement
of modulus, strength, and toughness in poly(methyl
methacrylate) based composites by the incorpor-
ation of poly (methyl methacrylate)functionalized
nanotubes. Adv Funct Mater. 2006;16:1608–1614.
24. Kim KH, Jo WH. Improvement of tensile properties
of poly (methyl methacrylate) by dispersing multi-
walled carbon nanotubes functionalized with poly
(3-hexylthiophene)-graft-poly (methyl methacryl-
ate). Compos Sci Technol. 2008;68:2120–2124.
25. Kovalchuk AA, Shevchenko VG, Shchegolikhin AN,
et al. Effect of carbon nanotube functionalization
on the structural and mechanical properties of poly-
propylene/MWCNT composites. Macromolecules.
2008;41:7536–7542.
26. Xiao Y, Zhang X, Cao W, et al. Dispersion and
mechanical properties of polypropylene/multiwall
carbon nanotubes composites obtained via dynamic
packing injection molding. J Appl Polym Sci. 2007;
104:1880–1886.
27. Deng H, Bilotti E, Zhang R, et al. Effective
reinforcement of carbon nanotubes in polypropyl-
ene matrices. J Appl Polym Sci.. 2010;118:30–41.
28. Ma J, Deng H, Peijs T. Processing of poly (propyl-
ene)/carbon nanotube composites using scco2as-
sisted mixing. Macromol Mater Eng. 2010;295:
566–574.
29. Choi EY, Roh SC, Kim C. Noncovalent functionali-
zation of multi-walled carbon nanotubes with pyr-
ene-linked nylon66 for high performance nylon66/
multi-walled carbon nanotube composites. Carbon.
2014;72:160–168.
30. Dong J, Fang Y, Gan F, et al. Enhanced mechanical
properties of polyimide composite fibers containing
amino functionalized carbon nanotubes. Compos
Sci Technol. 2016;135:137–145.
31. Zeng H, Gao C, Wang Y, et al. In situ polymeriza-
tion approach to multiwalled carbon nanotubes-
reinforced nylon 1010 composites: mechanical
properties and crystallization behavior. Polymer.
2006;47:113–122.
32. Pande S, Chaudhary A, Patel D, et al. Mechanical
and electrical properties of multiwall carbon nano-
tube/polycarbonate composites for electrostatic dis-
charge and electromagnetic interference shielding
applications. RSC Adv. 2014;4:13839–13849.
33. Poetschke P, Arnaldo MH, Radusch H-J. Percolation
behavior and mechanical properties of polycarbonate

composites filled with carbon black/carbon nanotube
systems. Polimery. 2012;57:204–211.
34. Eitan A, Fisher F, Andrews R, et al. Reinforcement
mechanisms in MWCNT-filled polycarbonate.
Compos Sci Technol. 2006;66:1162–1173.
35. Wang Z, Bilotti E, Peijs T. Effective modulus of sin-
gle and multiwalled carbon nanotubes in melt-com-
pounded polycarbonate nanocomposites. Adv
Compos Lett. 2010;19:169.
36. Xu LR, Bhamidipati V, Zhong W-H, et al. Mechanical
property characterization of a polymeric nanocompo-
site reinforced by graphitic nanofibers with reactive
linkers. J Compos Mater. 2004;38:1563–1582.
37. ISO1167, Plastics pipes for the transport of fluids -
determination of the resistance to internal pressure;
1973.
38. ISO9080, Plastic piping and ducting systems -
determination of the long-term hydrostatic strength
of thermoplastics materials in pipe form by extrapo-
lation; 2003.
39. Frank A, Pinter G, Lang R. Prediction of the
remaining lifetime of polyethylene pipes after up to
30 years in use. Polymer Testing. 2009;28:737–745.
40. Frank A, Huta P, Pinter G. Numerical assessment of
pe 80 and pe 100 pipe lifetime based on paris-erdo-
gan equation. Macromol Symp. 2012;311:112–121.
41. Lu JP, Davis P, Burn L. Lifetime prediction for abs
pipes subjected to combined pressure and deflection
loading. Polym Eng Sci. 2003;43:444–462.
42. Kanters MJ, Kurokawa T, Govaert LE. Competition
between plasticity-controlled and crack-growth con-
trolled failure in static and cyclic fatigue of thermo-
plastic polymer systems. Polymer Testing. 2016;50:
101–110.
43. Andersson U. Which factors control the lifetime of
plastic pipes and how the lifetime can be extrapo-
lated. Book Institute of Materials. 2001;759:311–320.
44. Gedde UW, Viebke J, Leijstr€ om H, et al. Longterm
properties of hotwater polyolefin pipes: a review.
Polym Eng Sci. 1994;34:1773–1787.
45. Lang R, Stern A, Doerner G. Applicability and limi-
tations of current lifetime prediction models for
thermoplastics pipes under internal pressure.
Angew. Makromol. Chemie. 1997;247:131–145.
46. Chudnovsky A, Zhou Z, Zhang H, et al. Lifetime
assessment of engineering thermoplastics. Int J Eng
Sci. 2012;59:108–139.
47. Lee H-N, Paeng K, Swallen SF, et al. Direct meas-
urement of molecular mobility in actively deformed
polymer glasses. Science. 2009;323:231–234.
48. Bauwens-Crowet C, Ots J-M, Bauwens J-C. The
strain-rate and temperature dependence of yield of
polycarbonate in tension, tensile creep and impact
tests. J Mater Sci. 1974;9:1197–1201.
49. Eyring H. Viscosity, plasticity, and diffusion as
examples of absolute reaction rates. J Chem Phys.
1936;4:283–291.
50. Visser HA, Bor TC, Wolters M, et al. Lifetime
assessment of loadbearing polymer glasses: An ana-
lytical framework for ductile failure. Macromol
Mater Eng. 2010;295:637–651.
51. Anderson TL. Fracture mechanics: Fundamentals
and applications. Boca Raton, FL: CRC press; 2017.
52. Paris P, Erdogan F. A critical analysis of crack
propagation laws. J Basic Eng. 1963;85:528–533.
NANOCOMPOSITES 15
53. Gray A, Mallinson J, Price J. Fracture behaviour of
polyethylene pipes. Plastics and Rubber Processing
Appl. 1981;1:51–53.
54. Loo LS, Cohen RE, Gleason KK. Chain mobility in the
amorphous region of nylon 6 observed under active
uniaxial deformation. Science. 2000;288:116–119.
55. Engels TA, S€ ontjens SH, Smit TH, et al. Time-
dependent failure of amorphous polylactides in
static loading conditions. J Mater Sci: Mater Med.
2010;21:89–97.
56. Janssen RP, de Kanter D, Govaert LE, et al. Fatigue
life predictions for glassy polymers: A constitutive
approach. Macromolecules. 2008;41:2520–2530.
57. Sherby OD, Dorn J. Anelastic creep of polymethyl
methacrylate. J Mech Phys Solids. 1958;6:145–162.
58. Klompen E, Engels T, Govaert L, et al. Modelling of
the post yield response of glassy polymers: influence
of thermomechanical history. Macromolecules.
2005;38:6997–7008.
59. Engels TA, van Breemen LC, Govaert LE, et al.
Predicting the long-term mechanical performance of
polycarbonate from thermal history during injection
molding. Macromol Mater Eng. 2009;294:829–838.
60. Engels TA, van Breemen LC, Govaert LE, et al.
Predicting age-induced embrittlement of glassy pol-
ymers. Polymer. 2011;52:1811–1819.
61. Zhou Y, Brown N. The mechanism of fatigue fail-
ure in a polyethylene copolymer. J Polym Sci B
Polym Phys. 1992;30:477–487.
62. Eriksson M, Goossens H, Peijs T. Influence
of drying procedure on glass transition temperature
of PMMA based nanocomposites. Nanocomposites.
2015;1:36–45.
63. Zhou J, Aguilar Ventura I, Lubineau G. Probing
the role of poly (3, 4-ethylenedioxythiophene)/poly
(styrenesulfonate) coated multiwalled carbon nano-
tubes in the thermal and mechanical properties of
polycarbonate nanocomposites. Ind Eng Chem Res.
2014;53:3539–3549.
64. Holt D, Khor P, Lai M. The relation between the
fracture toughness of plates and the thickness of the
shear lips. Eng Fracture Mechanics. 1974;6:307–313.
65. Wrublewski DT, Lesser AJ. “Annealing effects on
the yield and fracture of bisphenol-A and 4,40 -dihy-
droxydiphenyl copolycarbonates,” In ANTEC con-
ference proceedings, vol. 3, pp. 4039–4043, Society
of Plastics Engineers.
66. Bikiaris D. Microstructure and properties of poly-
propylene/carbon nanotube nanocomposites.
Materials. 2010;3:2884–2946.
67. Satapathy BK, Gan M, Weidisch R, et al. Ductile to
semi-ductile transition in PP/MWNT nanocompo-
sites. Macromol Rapid Commun. 2007;28:834–841.
68. Ho C, Vu-Khanh T. Physical aging and time-tem-
perature behavior concerning fracture performance
of polycarbonate. Theor Appl Fracture Mech. 2004;
41:103–114.
69. Varela-Rizo H, Weisenberger M, Bortz D, et al.
Fracture toughness and creep performance of PMMA
composites containing micro and nanosized carbon
filaments. Compos Sci Technol. 2010;70:1189–1195.
70. Kanters MJ, Engels TA, van Erp TB, et al.
Predicting long-term crack growth dominated static
fatigue based on short-term cyclic testing. Int J
Fatigue. 2018;112:318–327.