Industrial Crops and Products 72 (2015) 125–132

Contents lists available at ScienceDirect

Industrial Crops and Products

journal homepage: www.elsevier.com/locate/indcrop

Toughened polyester cellulose nanocomposites: Effects of cellulose

nanocrystals and liquid epoxidized natural rubber on morphology
and mechanical properties
Hanieh Kargarzadeh a , Rasha M. Sheltami a,c , Ishak Ahmad a,∗ ,
Ibrahim Abdullah a , Alain Dufresne b
a
Faculty of Science and Technology, School of Chemical Sciences and Food Technology, Polymer Research Center (PORCE),
Universiti Kebangsan Malaysia (UKM), 43600 Bangi, Selangor, Malaysia
b
The International School of Paper, Print Media and Biomaterials (PAGORA), Grenoble Institute of Technology, BP65,
38402 Saint Martin d’He‘res Cedex, France
c
Chemistry Department, Faculty of Science, University of Benghazi, Benghazi, Libya

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

Article history: Unsaturated polyester resin (UPR) has been modified with a liquid epoxidized natural rubber (LENR). The
Received 17 September 2014 addition of LENR improves the toughness, but it is inevitably accompanied by a significant loss in stiff-
Received in revised form ness, yield strength, and thermal resistance. Incorporation of a rigid material into the LENR–UPR blend,
14 November 2014
such as cellulose nanocrystal (CNC), provided a ternary polyester nanocomposites with desire mechani-
Accepted 22 December 2014
cal properties. In the present work, the effects of different CNC and LENR content on the nanocomposite
Available online 14 January 2015
morphology and mechanical properties were examined. Morphological studies revealed that the size of
the rubber particles is independent of the rubber content due to the chemical interaction of rubber and
Keywords:
Blend
matrix; however, the size of the LENR particles increased with increasing CNC content. Tensile tests indi-
Unsaturated polyester resin cated that the blend’s tensile strength and elastic modulus could be further improved by incorporation
Cellulose nanocrystal of the CNCs. The high crystallinity and surface area of CNCs were the main factors for improvement of
Liquid epoxidized natural rubber tensile properties of the blend. Moreover, the impact energy of the UPR improved when it was modified
Nanocomposite with the LENR, and even further improvement was achieved with the CNC addition to form nanocompos-
Toughness ites. Dynamic mechanical thermal analysis (DMTA) showed that the storage modulus of nanocomposites
with 1.5 wt% rubber content was lower than those with 4.5 wt% LENR; however, both samples showed a
higher storage modulus than the neat UPR and the unfilled blend. The glass transition temperature (Tg ) of
the UPR decreased with the LENR addition, but improved with the addition of CNCs. The nanocomposites
prepared with a higher LENR content displayed better mechanical properties than those with low LENR
content.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction equipment, containers, automobiles, and cultured marble because

Unsaturated polyester resins (UPRs) is one of the most popu-
lar thermoset materials used as a matrix for composites, electronic
Abbreviations: UPR, unsaturated polyester resin; LENR, liquid epoxidized nat-
ural rubber; CNC, cellulose nanocrystal; MEKPO, methyl ethyl ketone peroxide;
Tg , glass transition temperature; ENR, epoxidized natural rubber; DMTA, dynamic
mechanical thermal analysis; FESEM, field emission scanning electron microscope;
TEM, transmission electron microscopy; CTBN, carboxyl-terminated butadiene-co-
acrylonitrile.
∗ Corresponding author. Tel.: +603 8921 5431/5424; fax: +603 8921 5410.
E-mail addresses: hanieh.kargar@gmail.com (H. Kargarzadeh),
gading@ukm.edu.my (I. Ahmad).
of its clarity and excellent chemical and corrosion resistance. The
major drawback of UPRs is that they are brittle with poor impact
resistance in the cured state. Therefore, toughening UPRs has
been the subject of intense investigation throughout the world
(Kargarzadeh et al., 2014; Thomas et al., 2013).
The most common technique for toughening brittle UPR is incor-
porating a dispersed elastomeric phase in the resin. In this process, a
rubber is initially miscible with the resin and a curing agent. When
the reaction begins, the rubber generally forms an initial copoly-
mer with the resin before phase separation, resulting in the cured
thermosetting polymer possessing a dispersed rubbery phase. The
addition of a rubber modifier improves the fracture toughness, but
it is inevitably accompanied by a significant loss in stiffness, yield

http://dx.doi.org/10.1016/j.indcrop.2014.12.052
0926-6690/© 2014 Elsevier B.V. All rights reserved.
126 H. Kargarzadeh et al. / Industrial Crops and Products 72 (2015) 125–132
strength, and thermal resistance. This is not unexpected, as the elas-
tic modulus of the toughness modifier is significantly lower than
that of the thermosetting matrix (Kargarzadeh et al., 2014). There-
fore, rigid inorganic nanoparticles, such as silica, clay, and carbon
have been used to toughen polyester without adversely affecting
the elastic modulus. Although the polyester nanocomposites show
better balance properties, the toughening effect of the inorganic
nanoparticles was not so prominent compare with the liquid rubber
(Afina et al., 2013; Vijayan et al., 2012; Xu et al., 2013).
Cellulose nanocrystals (CNCs) are a very important biopolymer
that have attracted a great deal of interest in the nanocompos-
ite field, due to their high modulus, strength, low density, high
crystallinity, high surface area, unique optical properties, and their
renewability and environmental friendly (Dufresne, 2012; Habibi
et al., 2010). CNCs have been widely used as a reinforcement for a
variety of thermoplastic and thermoset polymeric matrices. How-
ever, reports on the preparation of CNC-reinforced thermosetting
matrices are relatively rare. There have been only a few researchers
who have reported utilizing CNCs in an epoxy resin (Tang and
Weder, 2010; Xu et al., 2013), but to the best of our knowledge,
no report has been found on using an unsaturated polyester as the
matrix to produce a CNC-reinforced, rubber-modified thermoset.
In this study, liquid epoxidized natural rubber (LENR) was used
to modify the UPR, an orthophthalic polyester resin containing 30%
styrene, and then, the resulting material was combined with CNCs.
Both, LENR and CNC have been prepared from natural source and
they are good candidates to replace synthetic liquid rubbers and
inorganic fillers to toughen UPR while keeping desirable mechan-
ical properties. The objective of this research was to investigate
the effects of the LENR and CNC content on the morphology and
mechanical properties of the nanocomposites. In particular, the
dispersion of the rubber particles and CNCs, along with extensive
discussion on the matrix interaction and toughening mechanism of
both the blends and the nanocomposites was evaluated.
2. Materials and methods
2.1. Materials and chemicals
Raw kenaf (Hibiscus cannabinus) bast fibers, from which CNCs
were prepared, were supplied by KFI Sdn. Bhd. (Malaysia). The UPR
was supplied by Revertex Sdn. Bhd. (Malaysia). All other chemi-
cals used were purchased from SYSTERM, Classic Chemicals Sdn.
Bhd. (Malaysia). LENR was prepared in our laboratory by photo-
sensitized degradation of dried epoxidized natural rubber (ENR)
(Abdullah and Zakaria, 1989; Kargarzadeh et al., 2015).
2.1.2. Cellulose nanocrystal preparation
Colloidal suspensions of cellulose nanocrystals (CNC) in water
were prepared from sulfuric acid hydrolysis of cellulose extracted
from kenaf bast fiber as reported in our previous work (Kargarzadeh
et al., 2012).
2.1.3. Preparation of LENR-modified UPR blend
To prepare the rubber-modified UPR, LENR was added to the UPR
and mixed using an electrical stirrer for approximately 3 h. After
that, methyl ethyl ketone peroxide (MEKPO) initiator at a 1.5 wt%
concentration was added and stirred for 2 min. The mixture was
poured into a 3-mm-thick aluminum mold and cured at room tem-
perature for 24 h. The rubber content was varied between 1.5 and
6% (w/w) (Kargarzadeh et al., 2015).
2.1.4. Nanocomposite processing
Freeze-dried CNC were dispersed in styrene by using an ultra-
sonicator for 30 min in an ice bath. The CNC suspension in styrene
was mixed with the UPR and sonicated for 30 min. The result-
ing mixture was stirred with a mechanical stirrer and heated at
60 ◦ C until the excess styrene evaporated. The mixture was then
cooled to room temperature, the LENR was added to the UPR–CNC
and mixed for approximately 2 h before the initiator (1.5% con-
centration relative to the resin) was added and stirred for 2 min.
The final mixture was poured into molds and cured at room tem-
perature for 24 h. Different percentages of CNC loading (2, 4 and
6 wt%) were prepared. The nanocomposites using the CNC will be
referred to as CNC–LENR–UPR. Two different nanocomposites were
prepared, with 1.5 and 4.5 wt% LENR content, chosen on the basis
of the mechanical properties (blend with 1.5 wt% LENR indicated
higher tensile strength and 4.5 wt% LENR performed higher impact
energy).
2.2. Characterization
2.2.1. Microscopy
The fractured surface morphologies of the LENR–UPR blend
and the CNC–LENR–UPR nanocomposites were investigated using
a Zeiss Supra 55VP field emission scanning electron microscope
(FESEM). All samples were sputter-coated with gold prior to obser-
vation to prevent charge build up. In the case of the blend, its rubber
phase was extracted using toluene for 12 h under ambient condi-
tions.
Transmission electron microscopy (TEM) was conducted using
a Philips CM30 microscope to investigate the morphology of
the CNC–LENR–UPR nanocomposite. A thin film (70 nm) of this
nanocomposite was prepared using cryo-ultramicrotomy (Leica
EM, FC6) under liquid nitrogen.
2.2.2. Tensile test
Tensile tests of the samples were performed using an Instron
Universal Testing Machine, model 5567 according to the ASTM D-
638-91 standard test methods for examining the tensile properties
of the unsaturated polyester at a 5 mm/min crosshead speed. The
specimens were cut from the cured sheets to a 160 × 13 × 3 mm
size, with the shown values taken from an average of 10 measure-
ments.
2.2.3. Impact test
The impact strength of unnotched specimens was measured
following the ASTM D4812 impact resistance of plastic test. The
tests were run using a RAY-RAN digital universal pendulum impact
system at room temperature, for an average of 10 samples.
2.2.4. Dynamic mechanical thermal analysis (DMTA)
Dynamic mechanical thermal analysis was performed using a
TA Instruments, DMTA instrument model Q 800 according to the
ASTM D4065. The purpose of this test was to determine the storage
modulus (E ), tan delta (ı), and glass transition temperature (Tg ) of
the samples. The peak in the tan delta vs. temperature thermogram
was taken as Tg . Measurements were taken from −100 to 150 ◦ C at a
5 ◦ C/min heating rate and a 1 Hz fixed frequency. A single cantilever
mode was used for the 35 × 13 × 3 mm samples, with values taken
from an average of 4 specimens.
3. Results and discussion
3.1. LENR-modified UPR
3.1.1. Morphological analysis
Fig. 1 shows the typical micrograph of a neat UPR and rubber-
modified polyesters. The neat UPR displayed a smooth and glassy
fractured surface with ripples, demonstrative of its poor impact
strength (Ahmad and Hassan, 2010; Kargarzadeh et al., 2015). How-
ever, the fractured surface of the LENR–UPR blend [Fig. 1(b–d)]
 H. Kargarzadeh et al. / Industrial Crops and Products 72 (2015) 125–132 127

Fig. 1. FESEM micrographs of the neat UPR and LENR–UPR blend: (a) UPR, (b) 1.5 wt%, (c) 4.5 wt%, and (d) the 4.5 wt% blend at higher magnification. White arrows in (c)
showing the rubber particles which remains in the matrix after toluene extraction. White arrows in image (d) show crack pinning effect induced by LENR.

clearly shows two distinct phases, a continues UPR matrix and the neat UPR (40 MPa) increased to 49 MPa after incorporating 1.5 wt%
dispersed rubber particles, that appear as dark spherical phase due LENR, and then it levels off with increased rubber content. The ten-
to the extraction of rubber particles after treatment with toluene sile modulus displayed a similar trend to other rubber–polyester
for 12 h. blends (Ahmad and Hassan, 2010; Ben Saleh et al., 2009; Hisham
In general, the size of the rubber phase increases with increasing et al., 2011), showing a reduction with increasing rubber content.
rubber content (Ben Saleh et al., 2009; Saadati et al., 2005; Thomas The high tensile strength of the 1.5 wt% rubber content
et al., 2010). However, for the LENR–UPR blend, the size of the rub- LENR–UPR blend was due to the enhanced compatibility of the
ber particles was independent of rubber content, resulting from LENR with the UPR matrix via chemical interaction. At this quan-
the high compatibility and miscibility of the LENR due to its chem- tity, the majority of the rubber could interact and cross-link with
ical interactions with the UPR matrix. The likely mechanism and the UPR. In contrast, due to the coagulation of rubber particles at
chemical interaction occurring between the LENR and UPR is shown higher rubber concentration, fewer rubber particles were available
in Fig. 2, illustrating the potential chemical interaction as well as for cross-linking to the UPR, as the majority remained free rubber
hydrogen bonding between the reactive terminal groups (OH or particles within the UPR matrix, thus reducing the tensile strength.
OOH) of the LENR and the carboxyl groups of the UPR, improving A lower tensile modulus of the blend compared to the neat UPR
the compatibility between the rubber phase and polyester. In addi- is attributed to the LENR’s soft segment structure as well as to the
tion, it can be seen that not all of the rubber phase was extracted by intrinsically low modulus of liquid rubbers compared to neat UPR
the toluene which is shown by white arrows in Fig. 1(c), indicating (Seng et al., 2011).
a strong degree of interaction between the LENR and UPR. The impact energy of the UPR increased significantly with the
addition of the rubber. For example, a 4.5 wt% LENR improved
3.1.2. Mechanical properties of the LENR–UPR blend the impact energy by a factor of 156% (Table 1). According to
The effects of LENR content on the tensile and impact properties Ratna (2004), both chemical interaction between the liquid rub-
of the UPR is depicted in Table 1. It is well known that the incor- ber and polyester, and a high degree of rubber particle dispersion
poration of a rubber phase into polyester matrices significantly are required to achieve effective toughening properties. It can be
reduces the tensile strength and modulus of the resultant blend seen from Fig. 1(b and c) that the LENR particles were dispersed
(Ben Saleh et al., 2009; Cherian and Thachil, 2004; Mathew et al., homogenously throughout the matrix. The LENR can also chemi-
2010). However, the LENR–UPR blend shows a different behavior cally interact with the UPR via both the epoxy group and double
for tensile strength. As seen in Table 1, the tensile strength of the band, as suggested in Fig. 2.
The increase in toughness was also due to the amount of elastic
energy stored in the rubber particles while stretching. Therefore,
Table 1
the deformation of the rubber particles in the matrix appeared
The tensile and impact properties of the neat UPR and LENR–UPR blends
(Kargarzadeh et al., 2015). to be responsible for the enhanced stress transfer and consequent
increase in impact resistance. Other toughening mechanisms, such
Sample (all Tensile strength Modulus (MPa) Impact energy
as shear yielding, crack pinning, and crack deflection are also
numbers are (MPa) (kJ/m2 )
“wt%”) responsible for the higher impact resistance of the LENR–UPR
blend. Fig. 1(d) shows crack pinning effect. According to this mech-
UPR 40 ± 3.4 821 ± 50 2.5 ± 0.4
1.5 LENR 49 ± 1.5 624 ± 28 4.0 ± 0.5
anism, the proposed role of the rubber particles is to behave
3.0 LENR 39 ± 0.5 531 ± 12 5.4 ± 0.6 as impenetrable objects that cause the crack to bow out which
4.5 LENR 35 ± 1.5 445 ± 24 6.4 ± 0.6 consumes extra energy. Indirect evidence for the crack pinning
128 H. Kargarzadeh et al. / Industrial Crops and Products 72 (2015) 125–132
Fig. 2. Likely chemical structures of the cured LENR–UPR blend.
mechanism comes from the occurrence “tails” near the particles
on the fracture surface. The blends with 1.5 wt% and 4.5 wt% LENR,
with the highest strength and highest impact energy, respectively,
were selected to prepare nanocomposites.
3.2. Morphology investigation of the nanocomposites
The cryo-ultramicrotomed CNC–LENR–UPR nanocomposite
thin film with 6 wt% CNC and 4.5 wt% LENR was examined using
TEM to study the composite’s morphology at the nanoscale. The
nanocomposite with highest amount of CNC and rubber was cho-
sen to increase the possibility of the presence of both CNC and
rubber particles in the nanocomposite thin film. Fig. 3 demon-
strates that the nanoscale rod-like CNC shape was preserved
after the nanocomposite preparation. Fig. 3 illustrates that the
Fig. 3. TEM image of a 6 wt% CNC and 4.5 wt% LENR, CNC–LENR–UPR
nanocomposite.
CNCs dispersed individually or as small aggregates within the UPR
matrix, with the majority aggregated around the rubber particles
within the interphase.
The typical fractured surfaces of a LENR–UPR sample and the
CNC–LENR–UPR nanocomposites are shown in Fig. 4. As previously
mentioned, the fractured surface of LENR modified UPR consists
of two distinct phases: spherical LENR particles (ranging from 1 to
7 ␮m) dispersed uniformly in a continuous UPR matrix as seen in
Fig. 4(a). Fig. 4(b and c) shows the fractured surface micrographs
of nanocomposites with different CNC loading and 1.5 wt% LENR,
while Fig. 4(d and e) illustrates the fractured surfaces of nanocom-
posite with 4.5 wt% of LENR with 2 wt% and 6 wt% CNC loading,
respectively. By comparing the micrographs it is easy to identify
the CNC nanoparticles. They appear dotty, corresponding to the
transverse section of the CNCs. Fig. 4(b–e) shows that the rubber
particles are clearly phase separated and formed as spherical par-
ticles, with the incorporated CNCs. The size of the rubber particles
slightly increased (2–10 ␮m) as the CNC filler loading increased.
However, this phenomenon was more obvious for the nanocom-
posites with 4.5% LENR, potentially explained by the fact that the
increased viscosity hinders phase separation of the rubber from the
UPR matrix, leading to large particle formation during curing (Tang
et al., 2012). Moreover, it can be seen that all of the nanoparticles
are embedded in the matrix, with no debonding or voids observed,
and identifying the CNCs was difficult, indicating they were well
dispersed.
3.3. Tensile properties
The effects of the CNC concentration on the LENR–UPR’s ten-
sile properties were investigated. Fig. 5(a and b) shows the tensile
strength and modulus evolution for the LENR–UPR blend and
CNC–LENR–UPR nanocomposites at different CNC and LENR con-
tent. As discussed previously, the tensile strength of the neat UPR
increased from 40 to 49 MPa with 1.5 wt% rubber, and decreased
to 35 MPa with 4.5 wt% rubber. A different behavior was observed
 H. Kargarzadeh et al. / Industrial Crops and Products 72 (2015) 125–132
Fig. 4. Fractured surface FESEM micrographs of (a) the LENR–UPR blend; (b) 2 wt% CNC–1.5 wt% LENR–UPR; (c) 6 wt% CNC–1.5 wt% LENR–UPR; (d) 2 wt% CNC–4.5 wt%
LENR–UPR; and (e) 6 wt% CNC–4.5 wt% LENR–UPR nanocomposites.
with the CNC addition into the blends with the different rubber con-
centrations. In case of the low rubber content samples, the tensile
strength reduced with the addition of 2 wt% CNC and then slightly
increased with 4 and 6 wt% CNC. The lower tensile strength of the
nanocomposites with the 1.5 wt% LENR and 2 wt% CNC content may
be due to the dilution effect of CNCs when blended with the matrix
(Mandal and Chakrabarty, 2014).
The slight increment of tensile strength with 4 and 6 wt% CNC
was due to the unique behavior of CNC, observed in the TEM image
(Fig. 3). The CNCs at the interface provide bridges between the LENR
and UPR phase, resulting in a better adhesion between the two. This
is because the CNCs can have physical interactions with both the
LENR’s functional groups (e.g., the reactive end chain group, such
as the C O, OH, or epoxy groups) (Abdullah, 1994; Abdullah and
Zakaria, 1989) as well as the UPR’s carboxyl groups (Chirayila et al.,
2014). In addition, it seems that the 3D network of CNC, started to
form at this CNC content, thus, tensile strength increased slightly.
However, the tensile strength of nanocomposites with 4 and 6 wt%
CNC, was lower than of the LENR–UPR blend.
In contrast, the tensile strength of nanocomposite with 4.5 wt%
LENR, steadily increased with increasing CNC concentration. The
increased tensile strength in the 4.5 wt% LENR samples may be
ascribed to several reasons: (1) the nanoscale CNC was dispersed
uniformly in the LENR–UPR blend; (2) strong physical interactions
occurred, potentially via hydrogen bonding between the CNC to
both the LENR and UPR, as shown in the TEM image, CNCs at the
129
interface provide bridges between the LENR and UPR phase, result-
ing in a better adhesion between the two (Fig. 3); and (3) the rigid
CNC particle effects are more noticeable with an increased elastic
rubber phase. The presence of more rubber reduces the UPR stiff-
ness, as it is an intrinsically rigid material. Thus, a large quantity of
rubber introduces a greater softening effect on the matrix, and the
addition of a small amount of rigid particles improved the tensile
strength significantly. Additionally a CNC 3D network forms, once
the CNC concentration is increased to 6 wt%; the formation of a 3D
network is an accepted theory for higher mechanical properties of
nanocomposites.
The tensile moduli of the neat UPR and its related nanocom-
posites at different LENR concentrations are depicted in Fig. 5(b).
The results revealed that the tensile moduli of the nanocompos-
ites increased with increasing CNC content for both the low and
high LENR concentrations. This increase was expected, as the CNC’s
modulus is extremely high (136–155 GPa) (Eichhorn and Davies,
2006). Moreover, the high surface area provided by the highly crys-
talline nanocellulose, increased the surface interaction between
the filler and matrix and led the nanocomposite to display higher
mechanical properties, especially tensile modulus.
As previously mentioned, the reinforcing effect of the CNCs is
more noticeable in the high rubber content samples. For com-
parison, the nanocomposites with 1.5 wt% and 4.5 wt% LENR,
reinforced with 6 wt% CNCs have modulus values around 791 and
778 MPa, corresponding to 27% and 75% improvement from the
130 H. Kargarzadeh et al. / Industrial Crops and Products 72 (2015) 125–132
Fig. 5. Mechanical properties of the neat UPR and the CNC-LENR-UPR nanocomposites relative to CNC content: (a) Tensile strength, (b) tensile modulus, and (c) impact
energy.
non-reinforced LENR–UPR samples, respectively. It can be con-
cluded that the effect of the CNC addition overshadowed that of
the LENR-based modulus reduction.
3.4. Impact properties
The impact energies of the LENR–UPR nanocomposites are plot-
ted in Fig. 5(c) relative to the neat UPR. The impact energy increases
with increasing CNC filler loading for both types of nanocompos-
ites. However, for the nanocomposites with the 4.5 wt% LENR, the
impact energy reduced at the 6 wt% CNC content, as described
below. All of the impact energies for the 4.5 wt% rubber content
nanocomposite were higher than those with 1.5 wt% LENR. For
comparison, the maximum impact energy of the 1.5 wt% LENR and
6 wt% CNC nanocomposite was 5.5 kJ/m2 , giving a 37.5% improve-
ment over the LENR–UPR sample; while the 4 wt% CNC and the
4.5 wt% LENR nanocomposite had a 7.5 kJ/m2 impact energy, giv-
ing a 17.5% improvement. These impact energy numbers become
more significant when compared to the neat UPR: ∼96.5% and 200%
improvement for the maximum impact values of nanocomposites
with 1.5 and 4.5 wt% LENR, respectively.
This significant enhancement was due to the unique properties
of the cellulose nanocrystals, as their high aspect ratio can resist
fractures, resulting in an increased energy dissipation from crack
deflection at the filler-matrix interface. In addition, the LENR par-
ticles alone can bear a load in a triaxial tension and enhance shear
localization by acting as stress concentrators. The parallel effect of
both the rubber particles and the CNCs are the primary reason for
improved impact energy in the nanocomposites.
The lower impact energy at 6 wt% CNC loading with the 4.5 wt%
LENR was likely due to the agglomeration of CNCs combined with
the formation of larger rubber particles. In fact, the large rubber
particles act as deflection sites, which can lead to catastrophic fail-
ure of the matrix (Kargarzadeh et al., 2015). However, the impact
energy at this concentration is still higher than both the neat UPR
and the LENR–UPR blends.
In the case of the CNC reinforced LENR–UPR blend, the tough-
ening mechanism resulted from both the rubber particles and
nanocellulose particles. However, the toughening mechanism from
the CNCs is quite different from those of the rubber particles. Obvi-
ously, rubber particle cavitation, shear deformation, crack pinning,
and crack deflection are the main toughening mechanisms induced
by the rubber particles. However, the role that CNCs play in the
toughening of the nanocomposites cannot be directly observed in
the FESEM images. According to Xu et al. (2013), the stress concen-
tration factor is dependent on particle size. Therefore, the stress
concentration of the CNCs is small and cannot directly induce a
localized shear deformation of the matrix. In addition, the rigid
nanoparticles cannot produce a negative hydrostatic stress in the
matrix like the rubber particles, thus, they are not able to effectively
initiate cavitation of the neighboring matrix. The nanoparticles are
also significantly smaller than the crack-opening displacement,
making them unable to induce a plane mechanism such as crack
pinning or deflection (Johnsen et al., 2007). Due to the CNCs’ high
aspect ratio, the interface area between the CNCs and the matrix is
increased, indicating that the energy consumption required for par-
ticle debonding is also increased. Therefore, it is expected that the
energy needed for CNC removal or debonding from the matrix plays
an important role for toughening the nanocomposite (Xu et al.,
2013).
Moreover, the hydrophilic nature of the CNCs and the formation
of a percolation network resulting from strong hydrogen bonding
between CNCs, especially at higher loading, is another potential
reason for the higher impact properties. The CNC network behaves
 H. Kargarzadeh et al. / Industrial Crops and Products 72 (2015) 125–132
Fig. 6. Dynamic mechanical properties of the neat UPR, LENR-UPR blend and CNC–LENR–UPR nanocomposites reinforced with 6 wt% CNC; (a) Storage modulus, (b) a close
up of a region in (a) and (c) tan ı. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
as an impact modifier, absorbing energy to prevent crack prop-
agation through the matrix, similar to the rubber particles. The
unique behavior of the CNCs, with localized agglomeration at the
LENR–UPR interphase (Fig. 3), facilitates an efficient energy trans-
fer process to toughen the nanocomposite. The network of CNCs
also act as stress concentrators during fracture, inducing a localized
plastic deformation of the surrounding matrix that result in stress-
whitened zones on the fracture surface of the nanocomposite. This
stress-whitening effect was related to a local plastic deformation
at the crack edge, which can further explain the increased fracture
toughness of the nanocomposites compared to the LENR–UPR alone
(Fig. 4).
Table 1 and Fig. 5 show that choosing a careful match between
concentrations of the two additives (LENR and CNC) can provide a
higher toughness than either the LENR–UPR alone or the neat UPR.
Therefore, using a ternary mixture of the LENR–UPR combined with
the CNC represents a good strategy for achieving balance between
the tensile modulus and toughness.
3.5. Dynamic mechanical thermal analysis (DMTA)
The dynamic mechanical properties of both the neat UPR and
the nanocomposites are shown in Fig. 6. Fig. 6(a) shows that
the neat UPR has a storage modulus of ∼4560 MPa at −100 ◦ C.
At this temperature, the polymer is in the glassy state and the
modulus remained approximately constant with temperature. At
around 50 ◦ C, the modulus began dropping sharply, corresponding
to the relaxation of the polyester. In contrast, both the LENR–UPR
blends and the nanocomposites displayed a two-step drop in the
modulus (Fig. 6(b)). The low temperature modulus drop, which
increased with increasing of rubber content, is associated with
the glass transition temperature (Tg ) of the LENR (Kargarzadeh
et al., 2015). The results revealed that the UPR’s storage modulus
increased to 4819 MPa and 4619 MPa after modification with 1.5
and 4.5 wt% LENR, due to the high density of crosslinking within the
LENR–UPR blend. However, the blend with the 4.5 wt% LENR dis-
played a lower value than that with the 1.5 wt%. A similar trend was
reported for a blend of epoxy and carboxyl-terminated butadiene-
co-acrylonitrile copolymer (CTBN) (Thomas et al., 2004).
Upon addition of the CNCs, the storage moduli increased with
increasing filler content for both the low and high rubber con-
tent nanocomposites, becoming higher than the LENR–UPR blends
alone (Fig. 6a). This phenomenon showed the CNCs’ superior
reinforcing effect in the LENR–UPR matrix at the glassy region
(−100 ◦ C). The results indicate that the well-dispersed CNCs stiffen
the LENR–UPR matrix, especially for the blend with a higher rubber
content. It is worth noting that all of the nanocomposites’ storage
131
moduli were higher than the neat UPR (4560 MPa), indicating that
the modified nanocomposites are stiffer (Fig. 6a).
The storage moduli in the rubbery region (150 ◦ C) improved
with increasing CNC content. For instance, the storage mod-
ulus of a blend with 6 wt% CNC increased by 88% and 183%
from the LENR–UPR blends alone, for the low (1.5 wt%) and
high (4.5 wt%) rubber content nanocomposites, respectively. This
immense increase can be explained by the fact that the rubber
particles become extremely soft at temperatures above their Tg
(150 ◦ C), and the CNCs’ rigid reinforcement effect became notice-
able, restricting the plastic. This phenomenon is more obvious
for the 4.5% LENR nanocomposites, especially at higher CNC con-
tent. Moreover, the formation of a CNC rigid network, with the
CNCs connected by hydrogen bonding, could be another reason for
the nanocomposites’ excellent mechanical properties (Azizi Samir
et al., 2005; Goffin et al., 2011; Lu et al., 2008; Raquez et al., 2012).
Fig. 6(c) delineates the variation of tan ı as a function of temper-
ature for the neat UPR, the LENR–UPR blends and their 6 wt% CNC
nanocomposites. The tan ı peak, associated with Tg , was obtained
at 94 ◦ C for the neat UPR. As shown in Fig. 6, the incorporation of
1.5 and 4.5 wt% LENR into the UPR matrix reduced the neat UPR’s
Tg from 94 ◦ C to 90 and 89 ◦ C, respectively, due to the low modulus
nature of elastomers (Kargarzadeh et al., 2015). Additionally, the
modified UPR matrixes had reduced tan ı values, resulting from
enhanced damping properties in the matrix induced by the LENR
particles.
Incorporation of the CNCs into the LENR–UPR blend slightly
increased the Tg values. Based on the Fig. 6, the Tg of the 1.5 wt%
LENR and 6 wt% CNC nanocomposite increased from 90 to 92 ◦ C,
while that of the 4.5 wt% LENR and 6 wt% CNC nanocomposite
shifted from 89 to 92 ◦ C. These Tg values fall between the Tg range
of the blend and neat UPR. The increased Tg may be attributed
to increased adhesion between the polymer matrix and nanocel-
lulose particles, or it could result from the CNC’s nanometer size
and homogenous dispersion through the matrix and rubber inter-
phase. In fact, the CNCs can have both a strong intramolecular
interaction and a physical interaction with the matrix and LENR
particles, which can affect the polymer segment motion, reduc-
ing polymer chain mobility, a typical effect for CNC inclusion in
a polymer system. The CNCs’ reinforcing effect is a direct result
of the interaction between the filler and matrix, as well as the
rigidity of the CNCs’ web-like structure (Lu et al., 2008). It appears
that the CNC–LENR–UPR nanocomposites primary reinforcement
occurred from this CNC web-like structure of CNCs in addition to the
CNCs’ high crystallinity. In comparison to inorganic nanoparticles,
a reduction in Tg has been reported for nanocomposites prepared
with CTBN, nanoclay, and epoxy (Chonkaew et al., 2013) as well as
CTBN, nanosilica, and epoxy (Tang et al., 2012).
132 H. Kargarzadeh et al. / Industrial Crops and Products 72 (2015) 125–132
Additionally, the magnitude of the tan ı peak increased slightly
with the incorporation of CNCs into the LENR–UPR blends. This
incremental increase in peak height could result from the same
polymer chain restriction occurring from the highly rigid web-like
CNC structure running through it, thus reducing the damping effect
provided by the rubber particles, increasing the tan ı peak intensity.
4. Conclusions
An unsaturated polyester resin was toughened by incorporating
LENR. Tensile tests showed a significant reduction in strength and
modulus with increased rubber content, with the exception of an
increase observed for the tensile strength of 1.5 wt% LENR blend.
The size of LENR particles was independent of the LENR content
due to the chemical interaction of rubber and matrix. Impact energy
increase by the factor of 156% with 4.5 wt% LENR. The CNC incor-
poration significantly improved the mechanical properties of the
LENR–UPR blend. CNC-based nanocomposites were prepared with
two different rubber concentrations (1.5 and 4.5 wt%). A signifi-
cant enhancement was observed in tensile modulus, when the CNCs
were added, due to their high crystalinity and stiffness. Impact tests
revealed that the UPR’s impact resistance increased with the LENR
addition, and it improved further with the CNC addition, due to the
strong chemical interaction between the LENR and UPR, as well as
the formation of a 3D CNC network that inhibits crack growth and
acts as a stress concentrator. The DMTA showed that the storage
modulus of the LENR–UPR blends was higher than the neat UPR due
to a high crosslinking density, and the storage modulus increased
with CNC addition. The Tg shifted to a lower temperature for the
LENR–UPR blends; however, it increased with the incorporation of
the CNCs. A better balance of material properties was observed for
the nanocomposites, i.e., a higher modulus, Tg , and impact energy.
Moreover, the effects of the CNCs were enhanced in the nanocom-
posites having higher LENR content. Finally, as the LENR and CNCs
are green, natural materials, they are good candidates to replace
synthetic liquid rubbers and inorganic fillers to toughen UPR while
keeping desirable mechanical properties.
Acknowledgments
The authors acknowledge financial support from the
ministry of Higher Education (MOHE) under FRGS grant
FRGS/2/2013/SG06/UKM/02/2 and the Ministry of Science,
Technology, and Innovation (MOSTI) under the ScienceFund grant
03-01-02-SF1106.
References
Abdullah, I., 1994. Liquid natural rubber: preparation and application. In: Ghiggino,
K. (Ed.), Progress in Pacific Polymer Science 3. Springer, Berlin, Heidelberg, pp.
351–365.
Abdullah, I., Zakaria, Z., 1989. Pendepolimeran fotokimia getah asli. Sains
Malaysiana 18, 99–109.
Afina, M.R.A., Najmi, B.N., Shakirah, S.S., Surip, S.N., 2013. Mechanical and thermal
properties of rubber toughened carbon black-filled polyester composite. Adv.
Mater. Res. 812, 163–168.
Ahmad, I., Hassan, F.M., 2010. Preparation of unsaturated polyester liquid natural
rubber reinforced by montmorillonite. J. Reinf. Plast. Compos. 29, 2834–2841.
Azizi Samir, M.A.S., Alloin, F., Dufresne, A., 2005. Review of recent research into
cellulose whiskers, their properties and their application in nanocomposite
field. Biomacromolecules 6, 612–626.
Ben Saleh, A.B., Mohd Ishak, Z.A., Hashim, A.S., Kamil, W.A., 2009. Compatibility,
mechanical, thermal, and morphological properties of epoxy resin modified
with carbonyl-terminated butadiene acrylonitrile copolymer liquid rubber. J.
Phys. Sci. 20, 1–12.
Cherian, A.B., Thachil, E.T., 2004. Modification of isophthalic unsaturated polyester
resin using elastomers bearing reactive functional groups. Prog. Rubber Plast.
Recycl. Technol. 20, 247–266.
Chirayila, C.J., Joya, J., Mathewa, L., Koetzb, J., Thomas, S., 2014. Nanofibril
reinforced unsaturated polyester nanocomposites: morphology, mechanical
and barrier properties viscoelastic behaviorand polymer chain confinement.
Ind. Crops Prod. 56, 246–254.
Chonkaew, W., Sombatsompop, N., Brostow, W., 2013. High impact strength and
low wear of epoxy modified by a combination of liquid carboxyl terminated
poly (butadiene-co-acrylonitrile) rubber and organoclay. Eur. Polym. J. 49,
1461–1470.
Dufresne, A., 2012. Nanocellulose: potential reinforcement in composites. In: John,
M.J., Thomas, S. (Eds.), Natural Polymers: Volume 2: Nanocomposites. Royal
Society of Chemistry, Cambridge, pp. 1–32.
Eichhorn, S., Davies, G., 2006. Modelling the crystalline deformation of native and
regenerated cellulose. Cellulose 13, 291–307.
Goffin, A.L., Raquez, J.M., Duquesne, E., Siqueira, G., Habibi, Y., Dufresne, A., Dubois,
P., 2011. From interfacial ring-opening polymerization to melt processing of
cellulose nanowhisker-filled polylactide-based nanocomposite.
Biomacromolecules 12, 2456–2465.
Habibi, Y., Lucia, L., Rojas, O.J., 2010. Cellulose nanocrystals: chemistry
self-assembly, and applications. Chem. Rev. 110, 3479–3500.
Hisham, S.F., Ahmad, I., Daik, R., Ramli, A., 2011. Blends of LNR with unsaturated
polyester resin from recycled PET: comparison of mechanical properties and
morphological analysis with the optimum blend by commercial resin. Sains
Malaysiana 40, 729–735.
Johnsen, B., Kinloch, A., Mohammed, R., Taylor, A., Sprenger, S., 2007. Toughening
mechanisms of nanoparticle-modified epoxy polymers. Polymer 48,
530– 541.
Kargarzadeh, H., Ahmad, I., Abdullah, I., 2014. Liquid rubbers as toughening agents.
In: Thomas, S., Sinturel, C., Thomas, R. (Eds.), Micro- and Nano-structured
Epoxy/Rubber Blends. Wiley-VCH Verlag GmbH & Co. KGaA, Germany,
pp. 31–52.
Kargarzadeh, H., Ahmad, I., Abdullah, I., Dufresne, A., Zainudin, S.Y., Sheltami, R.M.,
2012. Effects of hydrolysis conditions on the morphology, crystallinity, and
thermal stability of cellulose nanocrystals extracted from kenaf bast fibers.
Cellulose 19, 855–866.
Kargarzadeh, H., Ahmad, I., Abdullah, I., Thomas, R., Dufresne, A., Thomas, S.,
Hassan, A., 2015. Functionalized liquid natural rubber and liquid epoxidized
natural rubber: a promising green toughening agent for polyester. J. Appl.
Polym. Sci., http://dx.doi.org/10.1002/app.41292.
Lu, J., Wang, T., Drzal, L.T., 2008. Preparation and properties of microfibrillated
cellulose polyvinyl alcohol composite materials. Compos. Part A: Appl. Sci.
Manuf. 39, 738–746.
Mandal, A., Chakrabarty, D., 2014. Studies on the mechanical, thermal,
morphological and barrier properties of nanocomposites based on poly(vinyl
alcohol) and nanocellulose from sugarcane bagasse. J. Ind. Eng. Chem. 20,
462–473.
Mathew, V.S., Sinturel, C., George, S.C., Thomas, S., 2010. Epoxy resin/liquid natural
rubber system: secondary phase separation and its impact on mechanical
properties. J. Mater. Sci. 45, 1769–1781.
Raquez, J.M., Murena, Y., Goffin, A.L., Habibi, Y., Ruelle, B., Debuyl, F., Dubois, P.,
2012. Surface modification of cellulose nanowhiskers and their use as
nanoreinforcers into polylactide: a sustainably-integrated approach. Compos.
Sci. Technol. 72, 544–549.
Ratna, D., 2004. Rubber toughened epoxy. Macromol. Res. 12, 11–21.
Saadati, P., Baharvand, H., Rahimi, A., Morshedian, J., 2005. Effect of modified liquid
rubber on increasing toughness of epoxy resins. Iran. Polym. J. 14,
637–646.
Seng, Y.L., Ahmad, S.H., Rasid, R., Noum, S.Y., Hock, Y.C., Tarawneh, M.A., 2011.
Effect of liquid natural rubber (LNR) on the mechanical properties of LNR
toughened epoxy composite. Sains Malaysiana 40, 679–683.
Tang, L., Weder, C., 2010. Cellulose whisker/epoxy resin nanocomposites. ACS Appl.
Mater. Interfaces 2, 1073–1080.
Tang, L.C., Zhang, H., Sprenger, S., Ye, L., Zhang, Z., 2012. Fracture mechanisms of
epoxy-based ternary composites filled with rigid–soft particles. Compos. Sci.
Technol. 72, 558–565.
Thomas, R., Abraham, J., Thomas, P.S., Thomas, S., 2004. Influence of
carboxyl-terminated (butadiene-co-acrylonitrile) loading on the mechanical
and thermal properties of cured epoxy blends. J. Poly. Sci. 42,
2531–2544.
Thomas, R., Ahmad, I., Ahmad, S.H., Koshy, S., 2013. Blend and IPNs of natural
rubber with thermosetting polymers. In: Thomas, S., Chan, C.H., Pothen, L.A.,
Rajisha, K.R., Maria, H.J. (Eds.), Natural rubber materials: Volume 1: Blend and
IPNs. RCS Publishing, Cambridge, UK, pp. 336–348.
Thomas, R., Boudenne, A., Ibos, L., Candau, Y., Thomas, S., 2010. Thermophysical
properties of CTBN and HTPB liquid rubber modified epoxy blends. J. Appl.
Poly. Sci. 116, 3232–3241.
Vijayan, P.P., Puglia, D., Jyotishkumar, P., Kenny, J.M., Thomas, S., 2012. Effect of
nanoclay and carboxyl-terminated (butadiene-co-acrylonitrile) (CTBN) rubber
on the reaction induced phase separation and cure kinetics of an epoxy/cyclic
anhydride system. J. Mater. Sci. 47, 5241–5253.
Xu, S.A., Wang, G.T., Mai, Y.W., 2013. Effect of hybridization of liquid rubber and
nanosilica particles on the morphology, mechanical properties, and fracture
toughness of epoxy composites. J. Mater. Sci. 48, 3546–3556.