Polymer Testing 59 (2017) 75e83

Contents lists available at ScienceDirect

Polymer Testing
journal homepage: www.elsevier.com/locate/polytest

Material Properties

Property correlations for composites based on ethylene propylene

diene rubber reinforced with flax fibers
Maria Daniela Stelescu a, Anton Airinei b, *, Elena Manaila c, Gabriela Craciun c,
Nicusor Fifere b, Cristian Varganici b
a
National Research and Development Institute for Textile and Leather, Leather and Footwear Institute, 93 Ion Minulescu Street, 031215 Bucharest, Romania
b
”Petru Poni” Institute of Macromolecular Chemistry, Aleea Grigore Ghica Voda 41A, 700487 Iasi, Romania
c
National Institute for Lasers, Plasma and Radiation Physics, 409 Atomistilor Street, 077125 Magurele, Romania

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

Article history: EPDM composites filled with short flax fibers were prepared by melt blending procedure. The effects of
Received 3 November 2016 fiber loading on the mechanical, thermal and water uptake characteristics were studied. The physico-
Received in revised form mechanical, morphological thermal properties and water absorption behavior were discussed using
18 January 2017
tensile testing, differential scanning calorimetry, thermogravimetrical analysis and scanning electron
Accepted 18 January 2017
Available online 19 January 2017
microscopy. Scanning electron microscopy revealed that the flax fibers were well dispersed in the
polymer matrix. The tensile strength and hardness of the composites were found to be improved at
higher fiber loading. The water absorption pattern of EPDM/fiber composites at room temperature fol-
Keywords:
EPDM composites
lows a Fickian behavior for composites with 10, 15 and 20 phr flax fiber.
Flax fibers © 2017 Elsevier Ltd. All rights reserved.
Mechanical and thermal properties
Crosslink density
Water uptake

1. Introduction world wide availability [5e7]. Natural fibers-reinforced green

Green composites are a specific class of materials, in which at
least one of components proceeds from natural resources [1,2].
Among them the natural fiber reinforced polymer composites
represent an emerging area in the polymer science because these
materials are both environmentally friendly and sustainable [3].
After years of high-tech developments of the synthetic fibers
(aramid, glass, carbon, etc), the natural fibers (wood fibers, flax,
hemp, jute, sisal, kenaf, ramie and others) have now attracted a
renewed interest [4,5]. These natural cellulosic fibers have shown a
great potential as substitutes for synthetic fibers, in particular glass
fibers, in composites that are extensively used in the automotive
and construction industries. Natural fillers as raw materials for
polymer reinforcement exhibit many advantages relating to min-
eral fillers, glass fibers or carbon fibers such as low cost, low den-
sity, high specific strength and modulus, ease of fiber surface
modification, non abrasion, renewability and biodegradability,
good thermal and acoustic insulating properties, recyclability and
* Corresponding author.
E-mail address: airineia@icmpp.ro (A. Airinei).
composite materials are utilized in different applications, namely
door components, furniture, deck surfaces, window or automotive
components [1,8].
However, before to utilize these fiber-reinforced composites
into real life applications, especially for massive production, the
characteristics of the materials have to be deeply studied to assure
that repeatable and reliable results can be obtained. In spite of
these obvious advantages, there are several impediments to over-
come for using the natural fibers as reinforcements in the com-
posite materials, including lower compatibility between the
hydrophobic matrix and hydrophilic fibers, relative high moisture
of the fibers, dispersion properties of the resultant composites
(uniform dispersion and extreme agglomeration), manufacturing
process due to their low thermal stability which limits the appli-
cations in engineering thermoplastics. The change in microstruc-
ture of these fibers subjected to loading can also significantly affect
the final properties of the composites [9e11].
The natural flax fibers are widely used for textiles (linen) and for
technical applications, such as specialty papers, composites or
insulating materials due to its renewable nature, low cost, easy
availability, environmental benefit (i.e. biodegradability), high
specific tensile stiffness [8,12,13]. Natural fibers have less impact on

http://dx.doi.org/10.1016/j.polymertesting.2017.01.017
0142-9418/© 2017 Elsevier Ltd. All rights reserved.
76 M.D. Stelescu et al. / Polymer Testing 59 (2017) 75e83

the health of composite manufacturers because they do not bring Table 1

about skin irritations, lung cancer [7,14]. Flax fiber is a composite, in Composite formulations.

which the major constituents are cellulose (around 71%), hemi- Ingredients (phr)a/Code P0 PIn5 PIn10 PlnI5 PIn20
cellulose (around 2.2%), lignin (around 18.6e20.6%) and pectin EPDM 100 100 100 100 100
(around 2.3%) [8,15]. The unidirectional cellulose microfibers Flax 0 5 10 15 20
constitute reinforcing elements in polymer matrix/cellulosic fiber PEG 400 3 3 3 3 3
composite. Irganox1010 1 1 1 1 1
Perkadox 14e40B 8 8 8 8 8
It is known that ethylene propylene diene terpolymer (EPDM) is
Total 112 117 112 127 132
one of the most widely used and fastest growing synthetic rubber
a
Parts to 100 parts rubber.
[16]. EPDM exhibits remarkable characteristics such as high heat
resistance, ozone resistance, low temperature flexibility, cold and
moisture resistance to permanent deformation and impact, excel-
2.3. Measurements
lent electrical properties, color stability. EPDM rubber has been
widely used in automotive industry as insulation materials for
Tensile strength tests were performed with a Schopper strength
wires and cables, as floor coverings in metro train carriages, joints
test machine on dumbbell shaped specimens according to ISO 37/
materials in nuclear plants [17e19]. Even though many interesting
2012. The hardness was determined using a hardness tester ac-
papers exist in literature concerning the composites based on
cording to ISO 7619-1/2011 on samples with thick of 6 mm. The unit
polymers and flax fibers [7,8,11,20], a limited research has been
of hardness was expresses in ShoreA. The elasticity was estimated
conducted on structural characteristics of rubber/flax fiber com-
with a Schob test instrument on samples of 6 mm thick, according
posites [21e24].
to ISO 4662/2009.
The aim of this paper was to obtain and to investigate the
Thermogravimetric analysis (TGA) of the composites was per-
physico-mechanical properties of some EPDM based composites
formed on a STA449 F1 Jupiter thermal analyzer (Netzsch, Ger-
reinforced with flax fibers. Attention has been given to the effects of
many) under nitrogen atmosphere at a heating rate of 10  C/min.
fiber loading on the final properties of composites. The mechanical
The samples were heated from room temperature to 700  C at a
and thermal properties, water uptake, rubber-fiber interactions and
nitrogen flow rate of 50 ml/min. The onset of degradation tem-
morphology of the composites were compared.
perature, the temperature for which the weight loss is maximum
(Tm) and the residual weight were estimated. Differential scanning
calorimetry (DSC) measurements were carried out using a DSC200
2. Experimental
F3 Maia apparatus (Netzsch, Germany) under nitrogen atmosphere
and a heating rate of 10  C/min from room temperature to 300  C.
2.1. Materials
The gel fraction of the crosslinked EPDM/flax fiber composites
(with and without flax) was determined by the content of insoluble
EPDM monomer (Nordel 4760) was supplied by Dow Chemical
fraction from crosslinked composite after solvent extraction. The
Company (Mooney viscosity: 70 ML1þ4 at 120  C, 70% ethylene
samples were swollen in toluene and extracted after 72 h. The
content, 5-ethylidenenorbornene 4.9 wt%, density 0.88 g/cm3,
extracted samples were dried in air for 6 days and then in a labo-
crystallinity degree 10%). Polyethylene glycol, PEG 4000, was ob-
ratory over at 80  C for 3 h to constant weight and finally were
tained from Advance Petrochemicals LTD (density 1.128 g/cm3,
reweighed. The gel fraction is given by the relation
melting point range 4e8  C). Irganox 1010 (pentaerythritol tetra-
kis(3-(3,5 di-tert-butyl-4-hydroxyphenyl)propionate) was pro- ms
duced by BASF Schweiz AG (active ingredient 98%, melting point of
Gel fraction ð%Þ ¼ $100 (1)
mi
40  C). Dibenzoyl peroxide (Perkadox 14e40B) as vulcanizing agent
was supplied by Akzo Nobel Chemicals (density 1.60 g/cm3, 3.8% where mi and ms are the initial weight and the weight of the
active oxygen content, 40% peroxide content, pH 7). Ground flax insoluble portion of the composite in gel. The crosslink density (n)
fiber wastes thread length of max 3 mm were used as reinforcing of EPDM composites was estimated by equilibrium solvent swelling
agent. measurements using the modified Flory-Rehner equation [25].
Pieces of 2 mm thickness (initial weight mi) were prepared and
immersed in toluene for 72 h in order to achieve the equilibrium
2.2. Composite preparation swelling conditions. Then the swollen sample was taken out from
solvent and dried to remove the solvent excess and reweighed (mg).
EPDM rubber/flax fiber composites were obtained by melt The traces of solvent were removed by drying in air for six days and
blending using a laboratory electrically heated roller mill equipped in an oven at 80  C for 3 h. Then the sample was again weighed (ms).
with a cooling system at a friction ratio 1:1.1 and temperature of The volume fraction of polymer in the swollen network (n2m) was
60e80  C. EPDM (100 parts) was firstly melted 1e2 min, then calculated from swelling ratio G by the relation:
antioxidant (Irganox 1010) and PEG 400 incorporated and melt-
blended (2 min). The mixing was continued until a uniform v2m ¼ ð1 þ GÞ1 (2)
mixture was obtained. When a uniform mixture was realized,
different amounts of ground flax fibers 0, 5, 10, 15 and 20 phr (parts With
to 100 parts rubber), respectively were introduced (4 min) and then

8 phr of dibenzoyl peroxide as vulcanizing agent was added mg  mg rr
G¼ $ (3)
(1 min). The mixing was continued for another 5 min. The samples ms rs
were then removed from the roll in form of sheets about 2 mm
thick. Test specimens were prepared by compression molding at and rr, rs are the densities of EPDM sample and solvent (0.942 g/
160  C and a pressure of 150 MPa by using an electrical press and cm3 (EPDM) and 0.865 g/cm3 (toluene)). The densities of EPDM
then cooled under pressure at room temperature. The compound- samples were measured by hydrostatic weighing method, accord-
ing recipe is given in Table 1. ing to ISO 2781/2010. The crosslink density was calculated
 M.D. Stelescu et al. / Polymer Testing 59 (2017) 75e83 77

employing the Flory-Rehner equation [25]: increases significantly with increasing fiber loading suggesting that
  the introduction of ground flax fibers in EPDM matrix led to the
lnð1  y2m Þ þ y2m þ cy22m reinforcement of the rubber compositions. The variation of hard-
y¼ h  i (4)
1=3
V1 y2m  0:5y2m ness in composites was more important for 5 and 10 phr fiber
loading with 7 and 15 oShA, respectively. At the same time, an in-
crease of the tensile strength was noticed as the fiber level in-
where c denotes the interaction parameter between EPDM
creases in composites. The modulus at 100% elongation is near or
network and solvent (c ¼ 0.49 for EPDM-toluene system [26,27]),
higher than of the P0 sample and the enhancement is more pro-
V1 is the molar volume of toluene (106.52 cm3/mol), n2m represents
nounced for the composites containing 10 and 15 phr flax fiber
the volume fraction of polymer in the swollen specimen. The in-
(Table 2). Elasticity decreases from 64% (sample without fiber) to
teractions between EPDM matrix and flax fiber were analyzed
42% in the case of sample containing 20 phr flax fiber. The addition
applying the Kraus equation [28e31].
of flax fiber in composites does not improve the elongation at break
Vro f leading to the decrease of elongation at break because higher
¼1m (5) amounts of fibers determine a reduced mobility of the polymer
Vrf 1f
chains and induces more stiffness to the composite [11,22]. The
lowest values of elongation at break were obtained for composites
where Vro and Vrf are the volume fractions of EPDM rubber in
containing 15 and 20 phr flax fiber.
vulcanized sample and in fiber swollen recipe, f denotes the volume
The gel fraction and crosslink density of EPDM/fiber composites
fraction of fiber, m is the polymer-fiber interaction parameter.
were obtained from swelling tests and the results are in Table 3. As
The volume fraction of EPDM rubber in the swollen samples Vrf,
it can be seen in Table 3, the gel fraction presents high level of more
can be estimated by the following expression [29]:
than 97% having a slight increase for PIn5 and PIn15.
 The crosslink density of the EPDM/fiber composites increases as
ðD  FTÞ ðD  FTÞ Ao
Vrf ¼ þ (6) the flax fiber content increases. The results show that at lower fiber
rs rr rs
levels the highest increase of crosslink density was observed.
where rr and rs are the densities of EPDM rubber and solvent, D Further increases of the fiber loading led to a lower increase of the
represent the deswollen weight of the sample, F is the volume crosslink density (Table 3). Thus, the increase in the network
fraction of fiber, T is the initial weight of the sample and Ao denotes density could be due to an increase in stiffness of the composite due
the weight of the solvent absorbed by specimen. to the flax fibers or some chemical interactions of the fibers with
Water uptake tests were performed in accordance with ISO EPDM matrix evidenced by SEM observations and DSC data. A
20344/2004. Rectangular specimens from each sample were dried similar behavior was found out in the hardness dependence on the
in an oven at 80  C for 2 h, cooled in a desiccator and then weighed. fiber content. In consideration of the data relative to the mechanical
Samples were immersed in distilled water in test bottles at room properties and crosslink density we can conclude that flax fibers act
temperature (23  C), then samples were removed from the bottles as active fillers in EPDM composites leading to the reinforcement of
at periodic intervals and were wiped properly with tissue paper. material and to improve its characteristics. Our results are in good
After the water excess on the sample surface was removed the agreement with previous works which proved that the tensile
specimens were reweighed until no increase in water uptake was strength is closely related to the crosslink density [30,32e35]. The
observed. The water uptake for composite samples was determined tensile strength increases as the crosslink density increases and the
by a weight difference procedure. The water absorption (W) was majority of energy dissipation occurs in the EPDM matrix. But, at
calculated in accordance to the relation (7): higher crosslink densities, the macromolecular chains have lower
mobility and the material becomes stiffer and the elasticity will
ðms  m1 Þ decrease. The decrease in elongation at break can be explained by
Wð%Þ ¼ $100 (7) the increase of crosslinking degree and thus, the elongation at
m1
break will decrease when the formation of a network structure
where ms is the weight of the sample after immersion in water and
m1 represents the weight of the sample before immersion. The
morphology of the cryofractured surfaces obtained by breaking Table 3
EPDM/fiber composites frozen in liquid nitrogen was examined Gel values and crosslink density of EPDM/fiber composites.
with a Quanta 200 scanning electron microscope (FEI) operating at Sample Gel fraction Ga n2mb n. 104c
20 kV in low vacuum mode using a secondary electron detector (mol/ml)
LFD.
P0 98.07 ± 0.25 1.26 ± 0.04 0.443 ± 0.006 6.54 ± 0.32
PIn5 97.41 ± 0.08 0.82 ± 0.02 0.550 ± 0.004 14.23 ± 0.44
3. Results and discussion PIn10 97.06 ± 0.05 0.67 ± 0.03 0.598 ± 0.009 19.15 ± 1.18
PIn15 97.24 ± 0.06 0.66 ± 0.03 0.604 ± 0.011 19.25 ± 1.39
PIn20 97.18 ± 0.10 0.59 ± 0.04 0.629 ± 0.013 22.53 ± 1.98
The mechanical characteristics of the flax-fiber-reinforced
a
composites are given in Table 2. It is observed that the hardness G e swelling ratio.
b
y2me volume fraction of polymer in the swollen network.
c
n e crosslink density.
Table 2
Mechanical properties of the EPDM/fiber composites.

P0 PIn5 PIn10 PIn15 PIn20

Hardness, oShA 62 ± 0.71 69 ± 0.84 77 ± 0.56 81 ± 0.56 83 ± 0.45

Tensile strength, N/mm2 1.9 ± 0.20 1.8 ± 0.23 1.8 ± 0.06 2.0 ± 0.07 2.5 ± 0.23
Modulus at 100% elongation, N/mm2 1.1 ± 0.06 1.2 ± 0.00 1.8 ± 0.06 2.0 ± 0.06 e
Elasticity, % 64 ± 0.45 58 ± 0.71 46 ± 0.44 46 ± 0.89 42 ± 0.55
Elongation at break, % 287 ± 12 140 ± 10 100 ± 7 100 ± 7 80 ± 3
78 M.D. Stelescu et al. / Polymer Testing 59 (2017) 75e83

takes place [33e35].

The fiber-polymer matrix interaction can be estimated using
Kraus equation (5) from swelling data [28,31]. The ratio, Vro/Vrf,
indicates the restriction degree of swelling of EPDM matrix due to
the presence of fibers [30]. The values of Vro and Vrf calculated with
Eq (6) are listed in Table 4. It is observed from Table 4 that the ratio
Vro/Vrf decreases as the fiber loading increases as a result of the
increase in Vrf. According to Kraus theory [28], the reinforcing fillers
must have a negative slope of the plot Vro/Vrf against f/(1-f). The
decreased Vro/Vrf values at higher content of fiber reveal the rein-
forcement effect given by the fiber presence. A better adhesion
between filler and polymer matrix determines a restricted entry of
the solvent and the composites containing fibers presented lower
Vro/Vrf values [36,37], suggesting an improved interfacial interac-
tion between flax fiber and EPDM matrix. The higher interaction
between EPDM and flax fiber can provide enhanced mechanical
properties as seen previously.
Fig. 1 presents the plots of the evolution of the percentage water
uptake of EPDM fiber composites immersed in water as a function
of time. It is observed that the sample without flax fiber (P0)
attained the lowest water absorption (around 2%), whereas uptake
of EPDM/fiber composites increases as the fiber level increases in
samples reaching a percentage of water absorption around 9.2%
after 1200 h (PIn20) (Fig. 1). The increases in water absorption
follow a similar behavior like the dependence of the hardness and
crosslink density on fiber loading. Firstly, the increase was more
significant, then for higher fiber loadings of 15 and 20 phr,
respectively, the increase becomes smaller (Fig. 1). The EPDM flax
fiber composites exhibit a weight gain due to the absorbed water of
<10% even at 20 phr flax fiber suggesting an excellent behavior of
these composites in water which can extend the utilization range of
these new materials as consumer rubber goods.
The polymer composites based on natural fibers having an
elastomer matrix can contain a vulcanizing agent, namely diben-
zoyl peroxide. At higher temperature dibenzoyl peroxide de-
composes and reacts with EPDM matrix leading to the elastomer
crosslinking. At the same time the vulcanizing agent can react with
the flax fibers leading to the diminution of the hydroxyl group
content which determines the strong hydrophilic character of the
natural fibers improving the compatibility and the properties of
composites [21,38,39]. It is observed that for all samples water
uptake was faster during the first 200 h and then it develops slowly.
This is due to the nature of water absorption, because the driving
force is the water gradient of outside and in composite, and over
time this gradient is reduced, thus resulting in lower uptake rates.
In order to find out the mechanism of water diffusion into
polymer/fiber composites the power law relation can be used
[40,41]:
Fig. 1. Water uptake of EPDM/flax fiber composites.
representing the experimental data in log-log plots, according to eq
(9):
Mt
log ¼ log k þ n log t (9)
M∞
where the slope gives the diffusion mode.
A Fickian diffusion mechanism takes place if the value n is close
to 0.5. In this case the diffusion rate of the water molecules is much
smaller than the relaxation rate of the polymer chains. If the value
of n is 1, a non-Fickian diffusion mode is termed. Here the diffusion
process is much faster than the polymer chain relaxation. The
diffusion mode is named anomalous, if the value of n is in the range
0.5e1.0, where the diffusion and relaxation rates are comparable
[42e44]. Typical plots of fitting of the experimental data to eq. (9)
are depicted in Fig. 2. The values of the parameters n and k have
been evaluated by linear regression analysis and they are listed in
Table 5. It can be seen that the values of n for P0 and PIn5 EPDM/
fiber composites are over 0.5 indicating that the diffusion mecha-
nism belongs to an anomalous mode. For composites with 10, 15
and 20 phr flax fiber the diffusional mode approaches the Fickian
diffusion. The value of k for unfilled EPDM (P0) is much lower than
the values for EPDM/fiber composites due to the weak interaction
and uptake of non-polar matrix in a polar solvent (water) [45,46]. A
systematic increasing trend of k value was noticed in filled com-
posites as the flax loading increases due to the increase of the hy-
droxyl groups in composites with higher fiber level, which lead to
enhanced hydrogen bonding between flax fiber and water mole-
cules [46,47].
The morphology of the composite fracture surface was analyzed
by scanning electron microscopy in order to evaluate the particle
dispersion and potential interactions between natural fibers and

Mt
¼ kt n (8)
M∞

where Mt and M∞ denotes the water amount diffused in sample at

time t and at equilibrium state, k represents a constant depending
on the composite and exponent n indicates the mechanism of
diffusion. The constant k and exponent n can be determined by

Table 4
Vrf and Vro/Vrf values of EPDM/flax fiber composites.

Sample Vrf Vro/Vrf

PIn5 0.523 0.805

PIn10 0.564 0.747
PIn15 0.573 0.735
PIn20 0.582 0.724
Fig. 2. Diffusion fitting plots of EPDM/fiber composites immersed in water.
 M.D. Stelescu et al. / Polymer Testing 59 (2017) 75e83 79

Table 5 It can be seen that the fibers are uniformly distributed and inte-
Values of n and k for EPDM/fiber composites. grated in the polymer matrix regardless of the fiber level. This
Sample n k distribution plays an important role in improving the mechanical
P0 0.859 0.004
characteristics of composites as can be seen from the data sum-
PIn5 0.726 0.011 marized in Table 2. The surface of all fractured samples exhibits a
PIn10 0.543 0.033 moderate roughness due to the unfilled EPDM matrix morphology
PIn15 0.518 0.041 and to the presence of flax fibers. During the fracture process the
PIn20 0.418 0.075
fibers undergo a rupture without pulled out fibers suggesting an
acceptable adhesion between fibers and polymer matrix. Also, the
absence of some large gaps between EPDM matrix and fibers in-
polymer matrix. Micrographs of P0 and EPDM/fiber composites dicates the interfacial adhesion between two components.
reinforced with 5, 10, 15 and 20 phr flax fiber are illustrated in Fig. 3. Polyethylene glycol (PEG) can be used to enhance the

Fig. 3. SEM images of EPDM-based flax fiber composites with different fiber loadings: a,b) e P0; c) PIn5; d) PIn10; e) PIn15; f) PIn20.
80 M.D. Stelescu et al. / Polymer Testing 59 (2017) 75e83

interactions at the polymer matrix-fiber interface. PEG can be not Table 6

only a plasticizer for rubber, but also it can act as a compatibilizing DSC data of flax fiber reinforced composites.

agent between polymer matrix and hydrophilic flax fibers. At the Sample Tg Tm DHm Tc DHc
same time PEG can prevent the fiber aggregation, so that the flax ( C) ( C) (J/g) ( C) (J/g)
fiber disperses homogeneously in the polymer matrix to form a P0 36.49 36.34; 47.80 18.30 23.20; 15.82 53.85
network structure. Also, PEG can improve the intermolecular PIn5 36.07 34.43; 48.69 7.10 26.41; 16.17 45.42
interaction due to the presence of hydrogen bonds among the PIn10 34.61 34.06; 46.49 6.49 30.31; 14.31 43.65
PIn15 36.63 32.72; 45.61 4.96 28.71; 15.41 41.72
polymer matrix, PEG and flax fibers [48e50].
PIn20 34.61 36.10; 48.39 5.32 34.21; 15.12 41.33
Representative DSC patterns of the EPDM/fiber composites with
different flax loadings are shown in Fig. 4. The second heating runs
are given in Fig. 4a and the cooling curves in Fig. 4b. A glass tran-
and Tm as well as the maximum crystallization temperature (Tc),
sition temperature (Tg) can be observed in all samples
melt enthalpy (DHm) and heat of crystallization (DHc) are sum-
around 36  C followed by a large melting endotherm
marized for all composites. The melting temperatures of EPDM
between 30  C and 60  C with maximum located in the range of
composites decreased by 2e4  C compared with the corresponding
30e37  C (Table 6). EPDM/fiber composites present similar thermal
sample containing fiber, while the crystallization temperature
profiles with double melting endotherms (Fig. 4a). The double
shifted to a higher temperature in composites. These effects can be
peaked melting profile corresponds to the high content of ethylene
determined by the interaction between polymer chains and flax
component (70%). EPDM witnesses crystalline perfection and/or
fibers. The values of DHm and DHc in composites present a signifi-
thermal fractionation upon cooling, resulting in selective crystal-
cant decrease relating to the values of P0. Brostow et al. reported
lization of different molecular weight components, hence the
similar observations for EPDM/polypropylene composites [53].
appearance of two melting endotherms. Also, the crystallization of
The thermogravimetric analysis (TGA) shows that the pure flax
the large content of ethylene component is reflected in the high
fibers are stable up to 300  C, decomposing up to 400  C in a single
values of DHc, slowly decreasing from 53.85 J/g to 41.33 J/g with
mass loss stage of 84% and a mass residue of 15% (Fig. 5a). The
increasing flax loadings (Fig. 4b and Table 6). This aspect, correlated
EPDM composite without flax fiber (P0) has an excellent thermal
with the small decrease in DHm values (Fig. 4a and Table 6), depicts
stability up to about 400  C. As can be seen from Fig. 5b, the
an increase in the amorphous phase upon components interactions
maximum decomposition temperature (Tmax) obtained for EPDM
during the obtaining of the composite materials [51].
composite P0 appeared at 472  C and a char residue of 4.72% at
In general, the values of Tg and of melting temperature (Tm) of
700  C was obtained. Compared to the degradation pattern of P0,
the composites without (P0) and with flax fibers are quite close.
the thermal degradation of EPDM/fiber composites occurred in two
Similar behavior was reported in literature for natural fiber-based
steps, as shown in Fig. 5. The first derivative curves (DTG) (Fig. 5b)
composites [52]. In Table 6 the experimental data relating to Tg
confirmed the number of thermal decomposition stages. The initial

Fig. 4. DSC second heating (a) and cooling (b) thermograms of EPDM/fiber composites. Fig. 5. TG (a) and DTG (b) curves of flax fiber and EPDM/fiber composites.
 M.D. Stelescu et al. / Polymer Testing 59 (2017) 75e83 81

minor weight loss started at about 320e330  C and it was


completed at around 387  C and it was due to the presence of da A E
¼ exp f ðaÞ (11)
volatile matters from fibers. By incorporating the flax fibers in the dT b RT
polymer matrix, the thermal stability of the flax fibers significantly
After the integration of eq (11) between the limits To and Tp, the
increased, hence the second step of degradation occurred in the
integral function of conversion g(a), is given by eq (12)
temperature range 490e494  C, as compared to 300  C for the pure
fibers. In this stage the degradation of polymer chains of EPDM
occurred with a Tmax around 473  C and the degradation was Zap ZTp

da A E
completed at about 590  C. It can be seen that the weight loss of gðaÞ ¼ ¼ exp  dT (12)
f ðaÞ b RT
composites decreases gradually with the increase of fiber loading. 0 T0
As the flax fiber level increases in composite the char at 700  C in
the second weight loss process is higher. where To represents the initial temperature corresponding to a ¼ 0
The isoconversional Flynn-Wall-Ozawa (FWO) method was used and Tp is the temperature corresponding to the peak from DTG
to evaluate the kinetic parameters describing the thermal decom- curve, when a ¼ ap. The integral conversion function depicts the
position process [54e56]. Isoconversional methods rely on the mechanism of thermal degradation [57].
assumption that the applied heating program is independent on The kinetic parameters of thermal degradation behavior of
the kinetic reaction model and uses the shifts in thermogravimetric EPDM/fiber composites were evaluated by the isoconversional
(TG) curves to higher temperature ranges when the heating rate Flynn-Wall-Ozawa method [54e56] according to eq (13)
increases. For this purpose Fig. 6 illustrates the TG curves of some
EPDM composites without fiber (P0) and with 10 phr flax fiber AE R
log b ¼ log  log gðaÞ  5:3305  1:052 (13)
(PIn10) at four different heating rates. R RT
The expression of the rate of a thermal decomposition reaction From applying eq. (13), a plot of log b versus 1/T at different
is given by eq. (10), conversions was constructed for P0 and PIn10 (Fig. 7). The slopes of
these straight lines were utilized to obtain the activation energies
for EPDM composite decomposition. The values of kinetic param-
eters determined by FWO method are gathered in Table 7. As it can

 be observed from Table 7, the values of the kinetic parameters
da E
¼ A exp  ð1  aÞn (10) decreased as the conversion degree of EPDM composites increases,
dt RT

where a is the conversion degree, A is the pre-exponential factor, t

denotes time, E is the activation energy of thermal degradation, R
represents the gas constant, T is the absolute temperature, f(a) ¼
(1-a)n is the conversion function and n is the reaction order.
Under non-isothermal conditions the degradation rate is esti-
mated if the heating rate, b ¼ dt/dl, is taken into account:

Fig. 6. Thermograms of EPDM composites at four different heating rates: a) P0; b) Fig. 7. FWO model application in the determination of kinetic parameters for EPDM
PIn10. composites: a) P0; b) PIn10.
82 M.D. Stelescu et al. / Polymer Testing 59 (2017) 75e83

Table 7 for the reinforced samples, which occur in the range 330e420  C,
Non-isothermal kinetic parameters of EPDM composites determined by FWO the reaction rate is controlled by the diffusion of a volatilized
method.
initially solid product into the structure of the other component,
a P0 PIn10 with respect to changes in boundary geometry of the participating
E log A E log A reactions interfaces. The gaseous molecules diffuse outwards be-
(kJ mol1) (s1) (kJ mol 1
) (s1) tween the unmodified and stabile solid structures. The very high E
0.1 324 20.91 418 37.27 values corresponding to lower values of a, correlated with initial
0.2 302 19.17 351 22.97 decomposition temperatures exceeding 330  C, suggest the occur-
0.3 287 18.08 308 19.67 rence of a boiling process regarded to volatilization. One must keep
0.4 276 17.32 287 18.11
in mind that the determined kinetic parameters are apparent and
0.5 266 16.62 271 17.01
0.6 256 15.87 258 16.04 should not exclude the possibility of autocatalysis phenomena
0.7 244 15.05 244 15.0 5 occurrence simultaneously with the diffusion processes. The
0.8 229 14.05 228 13.93 decreasing in E values with a indicates that the reaction interface(s)
0.9 212 12.87 206 12.42 advance more facile into the yet unreacted material with temper-
ature increase.
The single thermal decomposition stage of the non-reinforced
suggesting a complex decomposition pattern [54]. FWO procedure polymer scaffold (P0) was best described by an n-th order reac-
imposes a first order reaction model for g(a) (g(a) ¼ 1 - a). The tion model. This stage of thermal decomposition occurs at tem-
straight lines in the FWO plots do not obey the same parallelism peratures above 420  C, exhibits the highest mass loss and is
throughout the whole decomposition process (Fig. 7) due to the described by an n-th order kinetic model in the case of all three
changes occurring in the thermal degradation mechanism and thus presented samples. This aspect is an indication that the thermal
the first order reaction model is not the best option to describe the decomposition process starts with the volatilization and diffusion
process kinetics. of the reinforced fiber material into the cross-linked EPDM matrix
In order to determine the real form of the reaction model for before the degradation of the final product takes place.
four heating rates the multivariate nonlinear regression method
was applied [58]. A model of thermal decomposition in two stages 4. Conclusions
(Eq. (14)) was proposed for the reinforced structures:
In this paper the mechanical, thermal and morphological
1 2 properties of some EPDM/flax fiber composites were investigated.
A/B/C (14) An enhancement in the tensile strength, hardness and 100%
modulus was observed with the fiber content increasing due to the
where A is the initial structure, C is the thermostable residue and B interaction between EPDM matrix and flax fibers. The gel content
represents a solid intermediate. A similar model (A/B) was sug- value was over 97% for all composites. The crosslink density in-
gested for the non-reinforced structure which thermally de- creases slightly as the fiber loading in composites increased. Two
composes in a single step. After the testing of 16 kinetic models steps of thermal degradation were remarked at 320e390 and
thoroughly discussed in the literature [59], the reaction types 390e495  C, respectively. The results suggest a good thermal sta-
which best characterize the thermal decomposition process of each bility of polymer/fiber composites within the classical processing
stage of sample PIn10 are presented in Table 8. It can observed that temperature range. The non-isothermal kinetic parameters of
the first stage of thermal decomposition shows a complex pattern thermal decomposition were determined by applying Flynn-Wall-
described very well by three simultaneous diffusion processes (D4, Ozawa method. The activation energy decreased with the conver-
D2, D3), having Fexp < Fcrit [60,61], while the second stage is better sion degree indicated a complex mechanism of thermal decom-
described by a n-th order reaction model (Fn ¼ (1 - a)n). position. The water absorption tests indicated that the water
The kinetic models coded D4, D3, D2, D1F and D3F correspond to uptake increased with fiber level and the saturation appeared after
the GinstlingeBrounshtein model, the Jander type model, the two 700 h.
dimensional diffusion controlled reaction and the one- and three-
dimensional forms of Fick's law, respectively, [62e65]. The main
References
idea is that in the case of the first stages of thermal decomposition
[1] V.K. Thakur, M.K. Thakur, R.K. Gupta, R. Prasanth, M.R. Kessler, Green com-
posites: an introduction, in: V.K. Thakur (Ed.), Green Composites from Natural
Table 8 Resources, CRC Press, 2013.
Tested kinetic models for PIn10 composite. [2] A.N. Netravali, S. Chabba, Composites get greener, Mater. Today 6 (2003)
22e29.
Kinetic model code Fcrit (0.95) Fexp [3] S. Ouajai, R.A. Shanks, Biocomposites of cellulose acetate butyrate with
Thermal decomposition modified hemp cellulose fibres, Macromol. Mater. Eng. 294 (2009) 213e221.
stage [4] J. Summerscales, N.P.J. Dissanayake, A.S. Virk, W. Hall, A review of bast fibres
and their composites. Fibres as reinforcements, Compos. Part A 41 (2010)
I II 1329e1335.
[5] V.K. Thakur, A.S. Singh, Nanotechnology in Polymers, Studium Press LLC
D4 Fn 1.11 1.00 Houston, TX, 2012.
D2 Fn 1.11 1.00 [6] M.D. Stelescu, E. Manaila, G. Craciun, M. Dumitrascu, New green polymeric
D3 Fn 1.11 1.00 composites based on hemp and natural rubber processed by electron beam
D3F Fn 1.11 1.00 irradiation, Sci. World J. 2014 (2014), 684047.
D1F Fn 1.11 1.02 [7] M. Soleimani, L. Tabil, S. Panigrahi, A. Opoku, The effect of fiber pretreatment
CnB Fn 1.11 1.19 and compatibilizer on mechanical and physical properties of flax fiber-
F2 Fn 1.11 1.20 polypropylene composites, J. Polym. Environ. 16 (2008) 74e82.
An Fn 1.11 1.23 [8] H.L. Bos, The Potential of Flax Fibres as Reinforcement for Composite Mate-
R3 Fn 1.11 1.32 rials, Technische Universiteit Eindhoven, Eindhoven, 2004.
R2 Fn 1.11 1.40 [9] A.K. Bledzki, H.P. Fink, M. Sain, Biocomposites reinforced with natural fibers:
2000-2010, Progr. Polym. Sci. 37 (2012) 1552e1596.
Fn Fn 1.11 1.44
[10] D. Aydemir, A. Kiziltas, E.E. Kiziltas, D.J. Gardner, G. Gunduz, Heat treated
 M.D. Stelescu et al. / Polymer Testing 59 (2017) 75e83
woodenylon 6 composites, Compos. Part B 68 (2015) 414e423.
[11] F. Puch, C. Hopmann, Experimental investigation of the influence of the
compounding process and the composite composition on the mechanical
properties of a short-flax fiber reinforced polypropylene composite, Polym.
Compos 36 (2015) 2282e2290.
[12] A.G. Valladares Ju arez, G. Rost, U. Heitmann, E. Heger, R. Muller, Development
of a biotechnological process for the production of high quality linen fibers,
Bioprocess Biosyst. Eng. 34 (2011) 913e921.
[13] J. Merrotte, A. LeDuigou, A. Bourmaud, K. Behlouli, C. Baley, Mechanical and
acoustic behaviour of porosity controlled randomly dispersed flax/PP bio-
composite, Polym. Test. 51 (2016) 174e180.
[14] G. Ausias, A. Bourmaud, G. Coroller, C. Baley, Study of the fibre morphology
stability in polypropylene-flax composites, Polym. Degrad. Stab. 98 (2013)
1216e1224.
[15] C. Albano, J. Gonzales, M. Ichazo, D. Kaiser, Thermal stability of blends of
polyolefins and sisal fiber, Polym. Degrad. Stab. 66 (1999) 179e190.
[16] M. Homocianu, A. Airinei, D.M. Stelescu, D. Timpu, A. Ioanid, Morphological
structure and surface properties of maleated ethylene propylene diene
monomer/organoclay nanocomposites, Polym. Compos 33 (2012) 379e387.
[17] M.D. Stelescu, E. Manaila, G. Craciun, Vulcanization of ethylene-propylene-
terpolymer-based rubber mixtures by radiation processing, J. Appl. Polym.
Sci. 128 (2013) 2325e2336.
[18] A. Perejon, P.E. Sanchez-Jimenez, E. Gil-Gonzales, L.A. Perez-Maqueda,
J.M. Criado, Pyrolysis kinetics of ethylene-propylene (EPM) and ethylene-
propylene-diene (EPDM), Polym. Degrad. Stab. 98 (2013) 1571e1577.
[19] A.B. Nair, U.G. Kalappura, P. Kurian, R. Joseph, Dielectric behavior of
aluminium hydroxide-filled oil-extended ethylene rubber composites in mi-
crowave fields, Polym. Eng. Sci. 53 (2012) 699e706.
[20] S.K. Garkhail, R.W.H. Heijenrath, T. Peijs, Mechanical properties of natural-
fibre-reinforced thermoplastics based on flax fibres and polypropylene,
Appl. Compos. Mater 7 (2000) 351e372.
[21] E. Ma na
ila, M.D. Stelescu, L. Surdu, G. Cra ciun, C.L. Dinca , D. Gurau,
A.M. Vasilescu, Polymeric composites based on flax wastes and natural rub-
ber, Ind. Text. 65 (1) (2014) 53e60.
[22] M. Cristea, A. Airinei, D. Ionita, M.D. Stelescu, Relaxation behavior of flax-
reinforced ethyleneepropyleneediene rubber, High. Perform. Polym. 27
(2015) 676e682.
[23] L. Mathew, J. Ulahannan, R. Joseph, Effect of curing temperature, fibre loading
and bonding agent on the equilibrium swelling of isora-natural rubber com-
posites, Compos. Interf. 13 (2006) 391e401.
[24] M. Natali, M. Rallini, D. Puglia, J. Kenny, L. Torre, EPDM based heat shielding
materials for solid rocket motors: a comparative study of different fibrous
reinforcement, Polym. Degrad. Stab. 98 (2013) 2131e2139.
[25] P.J. Flory, J. Rehner, Statistical mechanics of crosslinked polymer networks II.
Swelling, J. Chem. Phys. 11 (1943) 521e526.
[26] P. Anelli, S. Baccaro, M. Carenza, G. Palma, Radiation grafting of hydrophilic
monomers onto ethylenepropylene rubber, Rad. Phys. Chem. 46 (1995)
1031e1035.
[27] K. Nashar, U. Gosh, G. Heinrich, Influence of molecular structure of blend
components on the performance of thermoplastic vulcanisates prepared in
electron induced reactive processing, Polymer 91 (2016) 204e210.
[28] G. Kraus, Swelling of filler-reinforced vulcanizates, J. Appl. Polym. Sci. 7 (1963)
861e871.
[29] M. Jacob, S. Thomas, K.T. Varughese, Mechanical properties of sisal/oil palm
hybrid fiber reinforced natural rubber composites, Compos. Sci. Technol. 64
(2004) 955e965.
[30] Z.X. Dong, M.X. Liu, D.M. Jia, Y.H. Zhou, Synthesis of natural rubber-g-maleic
anhydride and its use as a compatibilizer in natural rubber/short nylon fiber
composites, Chin. J. Polym. Sci. 31 (2013) 1127e1138.
[31] U. Wisittanawat, S. Thanawan, T. Amornsakchai, Mechanical properties of
highly short pineapple leaf fiber reinforced e nitrile rubber composite: effect
of fiber and bonding agent, Polym. Test. 35 (2014) 20e27.
[32] H.N. Dhakal, Z.Y. Zhang, M.O.W. Richardson, Effect of water absorption on the
mechanical properties of hemp fibre reinforced unsaturated polyester com-
posites, Compos. Sci. Technol. 67 (2007) 1674e1683.
[33] Q. Wang, F. Wang, K. Cheng, Effect of crosslink density on some properties of
electron beam-irradiated styreneebutadiene rubber, Rad. Phys. Chem. 78
(2009) 1001e1005.
[34] P.S. Majumder, A.K. Bhowmick, Structure-property relationship of electron-
beam-modified EPDM rubber, J. Appl. Polym. Sci. 77 (2000) 323e337.
[35] V. Vijayabaskar, V.K. Tikku, K.A. Bhowmick, Electron beam modification and
crosslinking: influence of nitrile and carboxyl contents and level of unsatu-
ration on structure and properties of nitrile rubber, Rad. Phys. Chem. 75
(2006) 779e792.
[36] J. Wang, W. Wu, W. Wang, J. Zhang, Effect of a coupling agent on the prop-
erties of hemp-hurd-powder-filled styrene-butadiene rubber, J. Appl. Polym.
Sci. 121 (2011) 681e689.
[37] D. De, D. De, B. Adhikari, Curing characteristics and mechanical properties of
83
alkali-treated grass-fiber-filled natural rubber composites and effects of
bonding agent, J. Appl. Polym. Sci. 101 (2006) 3151e3160.
[38] X. Li, L.G. Tabil, S. Panigrahi, Chemical treatments of natural fibers for use in
natural fiber-reinforced composites: a review, J. Polym. Environ. 15 (2007)
25e33.
[39] M.S. Sreekala, M.G. Kumaran, S. Joseph, M. Iacob, S. Thomas, Oil palm fibre
reinforced phenol formaldehyde composites: influence of fibre surface mod-
ifications on the mechanical performance, Appl. Compos. Mater 7 (2000)
295e329.
[40] N.A. Peppas, N.M. Franson, The swelling interface number as a criterion for
prediction of diffusional solute release mechanisms in swellable polymers,
J. Polym. Sci. Polym. Phys. 21 (1983) 983e997.
[41] F. Ganji, S. Vasheghami-Farahami, E. Vasheghami-Farahami, Theoretical
description of hydrogel swelling: a review, Iran, Polym. J. 19 (2010) 375e398.
[42] A.R. Khare, N.A. Peppas, Swelling/deswelling of anionic copolymer gels, Bio-
materials 16 (1995) 559e567.
[43] W.-C. Lin, D.G. Yu, M.-C. Yang, pH sensitive polyelectrolyte complex gel mi-
crospheres composed of chitosan/sodium tripolyphosphate/dextran sulfate:
swelling kinetics and drug delivery properties, Colloids Surf. B 44 (2005)
143e151.
[44] S. George, P. Kumari, U.G. Panicher, Influence of static and dynamic cross-
linking techniques on the transport properties of ethylene propylene diene
monomer rubber/poly(ethylene-co-vinyl acetate) blends, J. Polym. Res. 17
(2010) 161e169.
[45] K.P. Dasan, A.P. Haseena, G. Unnikrishnan, R. Alex, E. Purushothaman,
Transport characteristics of filled poly(ethylene-co-vinyl acetate) membranes,
Polym. Polym. Compos 12 (2004) 599e617.
[46] F. Chigondo, P. Shoko, B.C. Nyamunda, U. Guyo, M. Moyo, Maize stalk as
reinforcement in natural rubber composites, Int. J. Sci. Technol. Res. 2 (2013)
263e271.
[47] D. Bachtiar, S.M. Salit, E. Zainudin, K. Abdan, K.Z.H.M. Dahlan, Effects of
alkaline treatment and a compatibilising agent on tensile properties of sugar
palm fibre-reinforced high impact polystyrene composites, BioResources 6
(2011) 4815e4823.
[48] P. Qu, Y. Gao, G.F. Wu, L.P. Zhang, Nanocomposites of poly (lactic acid) rein-
forced with cellulose nanofibrils, BioResources 5 (2010) 1811e1823.
[49] S. Luo, J. Cao, X. Wang, Investigation of the interfacial compatibility of PEG and
thermally modified wood flour/polypropylene composites using the stress
relaxation approach, BioResources 8 (2013) 2064e2073.
[50] S.P. Luo, J.Z. Cao, X. Wang, Properties of PEG/thermally modified wood flour/
polypropylene (PP) composites, For. Stud. China 14 (2012) 307e314.
[51] https://shop.perkinelmer.com/content/applicationnotes/app_
thermalepdmelastomers usingdsc.pdf.
[52] C.S.R. Freire, A.J.D. Silvestre, C.P. Veto, A. Gandini, L. Martin, I. Mondragon,
Composites based on acylated cellulose fibers and low-density polyethylene.
Effect of the fiber content, degree of substitution and fatty acid chain length
on final properties, Compos. Sci. Technol. 68 (2008) 3358e3364.
[53] W. Brostow, T. Datashvili, J. Geodakyan, J. Lou, Thermal and mechanical
properties of EPDM/PP thermal shock-resistant ceramic composites, J. Mater.
Sci. 46 (2011) 2445e2455.
[54] J.H. Flynn, L.A. Wall, A quick, direct method for the determination of activation
energy from thermogravimetric data, J. Polym. Sci. Part B Polym. Lett. 4 (1966)
323e328.
[55] T. Ozawa, A new method of analyzing thermogravimetric data, Bull. Chem.
Soc. Jpn 38 (1965) 1881e1886.
[56] J. Opfermann, E. Kaisersberger, An advantageous variant of the Ozawa-Flynn-
Wall analysis, Thermochim. Acta 203 (1992) 167e175.
[57] M. Villanueva, J.L. Martin Iglesias, J.A. Rodriguez-Anon, J. Proupin-Castineiras,
Thermal study of an epoxy system DGEBA (n¼o)/MXDA modified with POSS,
J. Therm. Anal. Calorim. 96 (2009) 575e582.
[58] J. Opfermann, Kinetic analysis using multivariate non-linear regression. I.
Basic concepts, J. Therm. Anal. Calorim. 60 (2000) 641e658.
[59] A.K. Galwey, M.E. Brown, Kinetic background to thermal analysis and calo-
rimetry, in: M.E. Brown (Ed.), Handbook of Thermal Analysis and Calorimetry,
vol. 1, Elsevier, Amsterdam, 1998, pp. 150e179.
[60] D. Rosu, L. Rosu, M. Brebu, Thermal stability of silver sulfathiazole-epoxy resin
network, J. Anal. Appl. Pyrolysis 92 (2011) 10e18.
[61] C.D. Varganici, A. Durdureanu-Angheluta, D. Rosu, M. Pinteala,
B.C. Simionescu, Thermal degradation of magnetite nanoparticles with hy-
drophilic shell, J. Anal. Appl. Pyrolysis 96 (2012) 63e68.
[62] W. Jander, Reaktionen in festen Zustande bei hoheren Temperaturen,
Z. Anorg. Allg. Chem. 163 (1927) 1e30.
[63] W. Jander, Neuere Forschungen uber die Reaktionen in festen Zustand,
Angew. Chem. 41 (1928) 73e79.
[64] A.M. Ginstling, B.I. Brounshtein, Zh. Prikl. Khim 23 (1950) 1327.
[65] J.B. Holt, J.B. Cutler, M.E. Wadsworth, Rate of thermal dehydration of kaolinite
in vacuum, J. Amer. Ceram. Soc. 45 (1962) 133e136.