Materials Science and Engineering A 491 (2008) 492–500

Contents lists available at ScienceDirect

Materials Science and Engineering A

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

Isothermal crystallization and mechanical behavior of ionomer treated

sisal/HDPE composites
Arup Choudhury ∗
Department of Polymer Engineering, Birla Institute of Technology, Mesra, Ranchi 835215, India

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

Article history: Isothermal crystallization behavior and mechanical properties of ionomer-treated sisal fibres reinforced
Received 3 January 2008 high-density polyethylene (HDPE) composites have been investigated. Avrami analysis of the crystalliza-
Received in revised form 18 February 2008 tion kinetics showed high crystallization rate (Kn ) and short crystallization half time (t0.5 ) for composites
Accepted 7 March 2008
compared to neat HDPE, indicating strong nucleating ability of sisal fibres. The crystallization activation
energy (Ea ) of the composites decreased with increasing fibre content. The polarizing optical microscopy
Keywords:
showed the occurrence of transcrystallization at the fibre–matrix interface. The HDPE/sisal composites
Natural fibre–HDPE composites
displayed superior tensile and flexural properties compared to neat HDPE. For these composites, predic-
Isothermal crystallization
Surface morphology
tions of Young’s modulus using the Modified Rule of Mixtures and Cox model correlated well with the
Mechanical properties experimental data.
Analytical modelling © 2008 Elsevier B.V. All rights reserved.

1. Introduction compatible with both the matrix and fibre, and thus could act as
coupling agents [17,18].
In recent years, natural fibres have found use as potential For composites based on semicrystalline polymers, the mor-
resources for making high-performance composite material. The phological features like degree of crystallinity, spherulite size,
attention in natural fibre reinforced polymer composites is growing lamella thickness and crystallite orientation play an important
rapidly due to their high mechanical performance, wide ver- role in influencing the ultimate properties of the polymer matrix,
satility, good processing advantages, low cost, and low density and thus the composite. It was found that both morphology and
[1,2]. Lignocellulosic fibres such as jute, sisal, hemp, coir, and crystallinity of thermoplastic polymer matrices deeply affected by
banana have been successfully used as reinforcing materials in fibre reinforcement [19–21]. In the case of thermoplastic com-
many polymeric resin matrices [3–6]. Among these fibres, sisal posites, the growth of highly oriented transcrystalline layers at
(Agave sisalana) is of particular interest in that its composites fibre–matrix interphase is thought to have a critical influence
have high impact strength besides having moderate tensile and on improvement of the mechanical properties, due to better
flexural properties compared to other lignocellulosic fibres [7]. fibre–matrix adhesion [22]. The occurrence of transcrystallization
However, some studies have been reported in the literature on the depends on the type of fibre used and the crystallization tem-
use of sisal fibre as a reinforcing agent in thermoplastic matrices perature [23,24]. A transcrystalline layer can be formed at the
[8–13]. The major drawback of these natural fibre/thermoplastic fibre–matrix interface regions if the fibres with high nucleating
composites is the inherent incompatibility of the hydrophilic lig- ability were employed. However, there were not many studies
nocellulosic fibres with the hydrophobic thermoplastic matrices, devoted to investigate the crystallization behavior and microstruc-
which calls for improving the fibre–matrix interfacial adhesion ture of the natural fibre reinforced thermoplastic composites
by using suitable compatibilisers and coupling agents [14–16]. [23–28].
Among many coupling agents currently used for improvement The mechanical properties of fibre-reinforced composites
of the fibre–matrix interfacial bond strength, ionomers based on strongly depend on several factors such as fibre size, fibre load-
copolymers of acids and olefin monomer units are worthy of explo- ing, fibre dispersion, fibre orientation and fibre–matrix interfacial
ration. The amphiphilic nature may allow these ionomers to be bond strength [29,30]. In many occasions, the effects of fibre rein-
forcement on the mechanical properties have been explained by
comparing the experimental results with existing micromechan-
ical models such as Parallel and Series, Hiesch, Cox, Halpin–Tasi,
∗ Tel.: +91 9430 732461; fax: +91 651 2276184. Modified Halpin–Tasi, Bowyer–Bader models, etc. [31–35]. It was
E-mail address: [email protected].

0921-5093/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.msea.2008.03.011
 A. Choudhury / Materials Science and Engineering A 491 (2008) 492–500 493

Table 1 Table 2
Physical, chemical and mechanical properties of sisal fibres Formulation of the different HDPE/sisal composites

Fibre Sisal Composites Weight percentage (wt%) Volume fraction

HDPE matrix Ionomer-treated sisal fibre of fibre, Vf (%)
Average diameter (␮m) 100
Average length (mm) 4.0 (95:5) HDPE/fibre 95 5 3.22
Density (g/cm3 ) 1.3–1.5 (90:10) HDPE/fibre 90 10 6.44
Cellulose (%) 56.52 (85:15) HDPE/fibre 85 15 9.66
Hemi-cellulose (%) 16.49 (80:20) HDPE/fibre 80 20 12.88
Lignin (%) 10.62
Tensile strength (MPa) 580
Tensile modulus (MPa) 773
Elongation (%) 4.3
2.4. Composites fabrication

Composites were prepared by the melt mixing of the ionomer-

concluded that the mechanical properties of fibre-reinforced com- treated short sisal fibres with HDPE in a HAAKE rotor mixture
posites strongly depend upon the fibre size distribution and fibre (Model Rheomix 600, Dreieich, Germany) at 120 ◦ C with roller
orientation distribution. However, a limited number of research blades and a mixing chamber with a 60 cm3 volumetric capacity.
works was dedicated to assess the viability of experimental data The process was carried out for 10 min at an optimum speed of
by using theoretical models for randomly oriented short fibre rein- 50–60 rpm, which depended on the quantity of the fibres by weight.
forced composites. Each batch contained various plastic to fibre weight ratio (95:5,
The aim of this work is to investigate the effect of sisal fibre 90:10, 85:15 and 80:20). The composite formulations are shown in
reinforcement on the crystallization kinetics and morphology of Table 2.
high-density polyethylene (HDPE). Surlyn ionomer was used as Each composite mixture was then homogenized in a two-
a coupling agent to improve the composites performance. The roll mill (150E-400 Collins, Germany) at 130 ◦ C and compression
variations of tensile, flexural and impact properties as a func- molded with a Delta Malikson 100TY pressman (Mumbai, India)
tion of fibre contents were studied. Moreover, a comparison to produce composite sheets (3 ± 0.2-mm thick). Test specimens
between experimental and estimated Young’s modulus using var- were prepared from these sheets according to ASTM (American
ious micromechanical models was presented for the HDPE/sisal Society for Testing and Materials) standard D with a counter-cut
composite. copy-milling machine (6490, Ceast, Italy) with calibrated tem-
plates.

2. Experimental
2.5. Thermal analysis

2.1. Materials
The thermal behavior of the neat HDPE and formulated
HDPE/sisal composites was measured with a differential scanning
The pure high-density polyethylene (M6805U: melt flow
calorimetric (DSC) analyzer (DSC-10Q, TA Instruments, New Cas-
index = 0.5 g/10 min at 190 ◦ C and density = 0.968 g/cm3 at 23 ◦ C)
tle, DE, USA) under nitrogen atmosphere. The samples were heated
was obtained from Haldia Petrochemical Ltd. India. The coupling
to 170 ◦ C with a heating rate of 10 ◦ C/min and kept at this temper-
agent used in this investigation was Surlyn® 1650 (den-
ature for 10 min to eliminate previous thermal history and then
sity = 0.94 g/cm3 ; melt flow index = 1.5 g/10 min at 190 ◦ C; melting
rapidly cooled (at 100 ◦ C/min) to crystallization temperature (Tc ),
temperature = 97 ◦ C with a heating rate of 10 ◦ C/min; methacrylic
and maintained at Tc till the time necessary for complete crys-
acid co-monomer content = 3.7–4.0 mol% and 57 mol% acid was
tallization of the polymer matrix. The heat evolved during the
neutralized by sodium salt), procured from Du Pont USA. Sisal fibres
isothermal crystallization (Hc ) was recorded as a function of time
(A. sisalana) were collected form local sources. Table 1 presents the
(t), at different crystallization temperatures. After crystallization,
physical, chemical and mechanical properties of sisal fibres. Sodium
the samples were heated to melting point at a rate of 10 ◦ C/min. The
hydroxide (NaOH) and acetic acid used for chemical treatment of
crystalline melting temperatures (Tm ) and degree of crystallinity
the fibres were obtained from E. Merck, India.
(Xcr ) of the samples were determined from the endotherm. The
enthalpy of crystallization of 100% crystalline polyethylene was
2.2. Washing and alkali treatment taken as 290 J/g [36].

The short sisal fibres (length ∼4.0 mm) were thoroughly washed 2.6. Mechanical properties
by aqueous detergent solution to remove dirt followed by cleaning
with distilled water and dried in vacuum oven at 70 ◦ C for 1 day. Rectangular bar specimens for tensile and flexural tests were
The washed fibres were immersed in 5% aqueous NaOH solution prepared according to ASTM D638 and ASTM D790, respectively.
for a period of 1 h at 40 ◦ C, and this was followed by washing with The measurements of the samples were conducted with an Instron
0.1N acetic acid and distilled water. The alkali-treated fibres were 3366 tensile test machine (Grove City, PA, USA) at a crosshead speed
than dried in vacuum oven at 70 ◦ C for 2 days to obtain mercerized of 2 mm/min. The Izod impact strength of notched samples (2 mm
fibres. depth) was measured with a Davenport Izod impact tester as per
ASTM D256 method.

2.3. Coupling agent treatment 2.7. Morphological analysis

The mercerized sisal fibres were immersed in hot 5% Surlyn The surface morphology of the composites was analyzed
ionomer solution (in toluene) at 70 ◦ C for 30 min to obtain ionomer- using a polarized optical microscopy (Leika Metalographic Aris-
treated fibres. tomet model), which attached with automatic hot-stage thermal
494 A. Choudhury / Materials Science and Engineering A 491 (2008) 492–500

Table 3
Melting and crystallization kinetic parameters of neat HDPE and HDPE/sisal composites obtained from DSC measurements

Material HDPE/fibre Thermal properties

Tc a (◦ C) nb Kn c t0.5 d (s) Tm e (◦ C) Hm f (J/g) Eg (kJ/mole) % Xcr h

100/0 116 2.06 0.05643 374 126.6 82.94 28.6

100/0 119 2.19 0.00255 868 127.1 86.42 29.8
85.8
100/0 121 2.43 0.00021 1288 129.8 89.03 30.7
100/0 124 2.62 0.00003 1405 131.4 90.48 31.2

95/5 116 2.32 0.14148 254 124.7 86.42 29.8

95/5 119 2.51 0.01014 628 125.5 90.48 31.2
108.3
95/5 121 2.74 0.00091 726 127.6 91.35 30.5
95/5 124 2.93 0.00008 945 130.4 96.86 33.4

90/10 116 2.28 0.46613 232 123.7 92.22 31.8

90/10 119 2.79 0.01588 511 124.4 95.41 33.9
127.6
90/10 121 2.83 0.00179 610 126.4 97.15 34.5
90/10 124 2.88 0.00017 855 128.7 100.63 34.7

85/15 116 2.68 0.52967 208 123.3 98.6 33.0

85/15 119 2.73 0.02548 458 123.8 102.66 33.4
134.8
85/15 121 2.96 0.00231 485 125.5 107.01 36.9
85/15 124 3.01 0.00016 642 127.9 109.04 37.6

80/20 116 2.85 0.72664 153 122.8 108.87 36.5

80/20 119 2.93 0.02734 376 124.5 110.35 37.0
151.5
80/20 121 3.25 0.00355 422 125.3 117.79 39.5
80/20 124 3.51 0.00033 541 127.6 116.49 39.0
a
Crystallization temperature.
b
Avrami exponent.
c
Crystallization kinetic constant.
d
Crystallization half time.
e
Crystalline melting temperature.
f
Heat of fusion.
g
Crystallization activation energy.
h
Crystallinity.

unit (Mettler FP-90). The composite samples were sandwiched 3. Results and discussion
between microscope glass slides and inserted into the hot stage
unit. The samples were heated to 170 ◦ C at 10 ◦ C/min, held 3.1. Melting and crystallization behavior of HDPE/sisal composites
at this temperature for 10 min, then rapidly cooled to 116 ◦ C
at 100 ◦ C/min and allowed to crystallize isothermally. Images The quantitative results of melting and crystallization behavior
were obtained using a Nikon digital camera connected to the as determined from DSC measurements are summarized in Table 3.
microscopy. The crystalline melting temperature of HDPE matrix in the compos-
ites was recorded lower compared to neat HDPE. The Tm values in
the composites decreased with increasing fibre content (Table 3).
2.8. Fourier transform infrared (FT-IR) spectroscopy The occurrence of lower melting peak in the composites could
be attributed to the fact that strong nucleation on fibre surfaces
Infrared spectroscopic analysis of the ionomer-treated and shortened the time required for HDPE crystallization, therefore,
untreated sisal fibres was carried out using an FT-IR Thermo Nicolet limiting the isothermal thickening of HDPE crystals and decreasing
(NEXUS 870) spectrophotometer at 25 ◦ C. their melting temperature. For all examined samples, an increase

Fig. 1. Heat flow curves of (a) neat HDPE and (b) 80:20 HDPE/sisal composite during isothermal crystallization.
 A. Choudhury / Materials Science and Engineering A 491 (2008) 492–500 495

Fig. 2. Plots of relative crystallinity (Xt ) versus time (t) at various Tc obtained for (a) neat HDPE, (b) 90:10 HDPE/sisal composites and (c) 80:20 HDPE/sisal composites.

of the melting temperature (Tm ) with increasing Tc was observed. nucleating effects of sisal fibres in the composites. Fig. 3 illustrates
The overall crystallization rate (1/t0.5 ) at higher Tc was found to the dependence of overall crystallization rate (Rc ), expressed as the
slow (Table 3), which might be caused the formation of thick reciprocal of crystallization half time (1/t0.5 ), on the crystallization
HDPE crystals having higher melting temperature. The percent temperature for all examined samples. The crystallization rate (Rc )
crystallinity of HDPE (%Xcr ) was significantly increased by fibre of composites was recorded higher compared to neat polymer at
reinforcement, indicating strong heterogeneous nucleating ability the same Tc , and it can be observed to increase with increasing fibre
of the sisal fibres. In addition, the nucleating effects of the fibre content (Fig. 3). One possible reason was that strong nucleation on
surfaces promoted the growth and formation of interfacial tran-
scrystalline layer when the composite sample was cooled from the
melt, as observed in the optical microscopy images [37,38].
The effect of sisal fibre reinforcement on the crystallization
behavior and thermodynamics of HDPE matrix was evaluated by
analyzing the isothermal crystallization kinetics of neat HDPE and
its composites in the temperature range of 116–124 ◦ C. Fig. 1 shows
the heat flow curves of 100% HDPE and 20% fibre filled HDPE/sisal
composite during isothermal crystallization at various Tc . It was
seen from Fig. 1 that at any crystallization temperature, the pri-
mary crystallization of HDPE matrix in the composites was almost
completed before the crystallization of neat HDPE had reached to
50–70% conversion. The fast crystallization conversion in the com-
posites could be attributed to the transcrystallization as a result of
nucleating effects of fibre surfaces. The short induction time and
high crystallization rate are supporting this hypothesis. The plots
of relative crystallinity (Xt ) as function of time t at various crystal-
lization temperatures (Tc = 116, 119, 121 and 124 ◦ C) are shown in
Fig. 2. From the comparison among the crystallization isotherms
at the same crystallization temperature it can be clearly pointed
out that the crystallization rate of HDPE matrix in the compos-
Fig. 3. Reciprocal of crystallization half time (t0.5 ) versus Tc for neat HDPE and
ites was faster than that of neat HDPE, indicating the occurrence of different HDPE/sisal composites.
496 A. Choudhury / Materials Science and Engineering A 491 (2008) 492–500

Fig. 4. Avrami plots of the (a) neat HDPE, (b) 90:10 HDPE/sisal composites and (c) 80:20 HDPE/sisal composites at various Tc .

fibre surfaces shortened the time required for HDPE crystallization. Avrami exponent in the composites was varied around 3.0 imply-
For all examined samples, the crystallization process was hindered ing a heterogeneous nucleation with three-dimensional growth of
with increasing Tc , as anticipated on the basis of the nucleation HDPE crystals. In contrast, the value of n was varied about 2.0 for the
theory [39]. neat HDPE indicating two-dimensional diffusion controlled crystal
growth. At any crystallization temperature, the crystallization rate
3.2. Isothermal crystallization kinetics constant of HDPE exhibited a significant increment as the fibre con-
tent increase (Table 3). The higher crystallization rate constant of
The crystallization kinetics of the neat polymer and its sisal fibre the composites could be ascribed to the nucleating effect of the
composites are analyzed by using the Avrami model [40]: fibres.
The crystallization rate constant Kn can be presented by Arrhe-
Xt = 1 − exp(−Kn t n ) (1) nius equation:
 −E 
The linearisation of Eq. (1) leads to Eq. (2): 1/n
Kn = K0 exp (3)
RTc
ln[− ln(1 − Xt )] = n ln t + ln Kn (2)
where K0 , R and E are the pre-exponential factor, universal gas
where Kn and n are the crystallization rate constant and Avrami constant and activation energy, respectively. The value of activa-
exponent, respectively. The Avrami approach is the most applied tion energy for the crystallization process was calculated from the
tool to describe the phenomenology of the crystal growth. The slopes of the linear plots of (1/n) ln Kn versus 1/Tc . The values
Avrami exponent depends on growth geometry and nucleation of E of neat HDPE and composites are summarized in Table 3.
type of the crystals. According to Eq. (2), the values of Kn and n The activation energy of HDPE crystallization was found to be sig-
were obtained from the intercept and slope of the linear plots of nificantly enhanced by the addition of short sisal fibre to HDPE
ln[−ln(1 − Xt )] versus ln(t). Fig. 4 demonstrates the Avrami plots matrix. The crystallization activation energy is the sum of the
of the neat HDPE, (90:10) HDPE/sisal and (80:20) HDPE/sisal com- transport activation energy and the nucleation activation energy.
posites at various crystallization temperatures. For all examined As mentioned above, the nucleation effect of sisal fibres leads to
samples, the values of Avrami parameters (n and Kn ) in the tem- enhance the crystallization of HDPE in the composites. Hence, the
perature range 116–124 ◦ C are complied in Table 3. As observed in higher crystallization activation energy of the composite might be
Table 3 that n increased with increasing Tc , and it recorded slightly originated from their greater transport activation energy because
higher for the HDPE/sisal composites compared to neat HDPE. The the mobility of HDPE chain segments was strongly restricted
 A. Choudhury / Materials Science and Engineering A 491 (2008) 492–500 497

Fig. 5. Polarized optical micrographs of the isothermally crystallized 80:20 HDPE/sisal composites (Tc = 116 ◦ C).

Table 4
Mechanical properties of HDPE/sisal composites at different fibre content

Sample Mechanical properties

Tensile strength (MPa) Tensile modulus (MPa) Flexural strength (MPa) Flexural modulus (MPa) Impact strength (J/m)

Pure HDPE 26.0 ± 0.5 860 ± 0.3 37.0 ± 0.8 928 ± 0.3 37.0 ± 0.16
(95:5) HDPE/fibre 28.5 ± 1.9 952 ± 2.3 40.7 ± 1.5 1131 ± 2.7 20.3 ± 0.19
(90:10) HDPE/fibre 30.7 ± 1.8 1086 ± 3.7 43.2 ± 1.2 1194 ± 3.7 17.7 ± 0.24
(85:15) HDPE/fibre 32.0 ± 2.3 1107 ± 3.3 45.0 ± 1.9 1354 ± 2.9 15.2 ± 0.22
(80:20) HDPE/fibre 32.8 ± 2.8 1219 ± 3.4 46.7 ± 1.9 1468 ± 3.5 14.3 ± 0.26

by the fibres and consequently hindered the crystallization pro- culty in proper dispersion of short sisal fibres and random fibre
cess. alignment in the matrix. However, all the composites exhibited
improved mechanical properties compared to neat HDPE. When
3.3. Optical microscopy study the fibre content raised up to 20%, the tensile strength, tensile
modulus, flexural strength and flexural modulus of the HDPE/sisal
As observed in the polarized optical microscopic analysis, the composite increased about 27, 41, 26, and 58%, respectively. The
growth and morphology of the HDPE crystals were significantly strong interfacial adhesion between HDPE and ionomer-treated
influenced by the addition of sisal fibres. Fig. 5 shows the opti- fibres lead to improve the stress transfer from the matrix to the
cal photomicrographs of the isothermally crystallized (80:20) fibres, which caused better mechanical performance. At lower fibre
HDPE/sisal composites (Tc = 116 ◦ C). As shown in the photomi- content (<10%) the dispersion of fibres was very poor that resulted
crographs, the nucleation of HDPE spherulites as well as regular poor stress transfer, whereas a strong tendency of the fibre–fibre
transcrystalline growth occurred along the fibre–matrix interfaces. interaction at higher fibre content (>15%) significantly enhanced
However, the formation and growth of transcrystallinity in a com- the stress transfer. A hypothetical model of the interface between
posite depend upon the performances of fibre–matrix interfaces ionomer-treated sisal fibre and HDPE is shown in Scheme 1. The PE
[41]. As seen in Fig. 5, the crystalline morphology at the fibre–matrix chain of ionomer diffuses into the HDPE matrix through interchain
interface does not look like the morphology of the crystals in the entanglements. On the other hand, the acrylic groups of ionomer
bulk of the matrix. In this situation, the transcrystalline layer was molecules form ester bonds with the hydroxyl groups of the fibres.
grown isothermally at the fibre–matrix interface while the bulk FT-IR spectra of ionomer-treated and untreated sisal fibres are
of the matrix still remained in the molten form, thus the growth reported in Fig. 6. The spectra of ionomer-treated sisal fibres show
transcrystallinity was practically undisturbed. The thickness of the a significant absorption peak in the carbonyl region (1715 cm−1 ),
transcrystalline layer was roughly constant of about 30 ␮m. The which is associated with the stretching vibration of methacrylic
fibre–matrix interface region was made of needle-like compact ester bond.
crystals propagated vertically on the surface of fibre. In earlier
investigations, the needle-like crystals have also been identified on
polyethylene with a certain molecular weight range under isother- 3.5. Modelling of mechanical property
mal conditions [42,43].
The elastic properties of the HDPE/sisal composites can be
3.4. Mechanical properties theoretically determined or derived by using a variety of microme-
chanical composite models. These models are derived based on
The properties of matrix and fibre are very important aspects the properties of the individual components of the composite and
in achieving good mechanical properties of the composites. The their arrangement [45]. Thus, a micromechanical model might
melt strength and processability are the most sensitive parame- be used to find the best combination of constituent materials to
ters to matrix polymer, whereas the modulus is dependent on fibre satisfy material design considerations. In this study, four differ-
properties [44]. The mechanical properties (tensile strength, tensile ent micromechanical composite models have been compared with
modulus, flexural strength, flexural modulus, and impact strength) experimental results in order to assess the influences of fibre dis-
of neat HDPE and its formulated sisal fibre composites are sum- persion on the Young’s modulus of HDPE/sisal composites. The
marized in Table 4. The composites have shown scattered response micromechanical models of the Young’s modulus for randomly dis-
to mechanical properties, as presented in Table 4, because of diffi- tributed short fibre-based polymer composites are as follows.
498 A. Choudhury / Materials Science and Engineering A 491 (2008) 492–500

Scheme 1. Hypothetical model of fibre modification with Surlyn ionomer and reaction with HDPE matrix.

3.5.1. Voigt–Reuss model 3.5.2. Halpin–Tsai model

The Voigt–Reuss model [46,47] has been used to estimate the The semi-empirical equation developed by Halpin and Tsai [48]
Young’s modulus for a lamina of short random fibre composites. has been widely used for predicting the elastic properties of a lam-
This model is independent on the geometry of reinforcement. It ina of short random fibre composites. The following form of the
can be expressed as: Halpin and Tsai equation was used to determine the Young’s mod-

3 5
 εf εm
 ulus of the composites:
εVR = [V ε + (1 − Vf )εm ] + (4)  3 1 + 2(l /d ) (1 − V ) 
8 f f 8 εf (1 − Vf ) + εm Vf 5 1 + 2T Vf
f f L f
εHT = + εm (6)
where εf and εm are the Young’s moduli of fibre and matrix, respec- 8 1 − L Vf 8 1 − T Vf
tively. Vf is the volume fraction of fibre in the composites, calculated
where lf and df are the average length and diameter of the fibres. In
using the following equation:
Eq. (6) the parameters L and T are given as:
Wf m
Vf = (5)
Wf f + (1 − Wf )f (εf /εm ) − 1
L = (7)
(εf /εm ) + 2(lf /df )
where Wf , f and m denote the weight fraction of fibres in the
composite, the density of fibre and the density of matrix, respec- (εf /εm ) − 1
tively. T = (8)
(εf /εm ) + 2

3.5.3. Modified Rule of Mixtures

The Rule of Mixtures can be used to determine the Young’s mod-
ulus for three-dimensional random short fibre composites [49,50].
This model works extremely well for aligned continuous fibre
composites. The Young’s modulus was determined by using this
equation:

1
εMRM = V ε + (1 − Vf )εm (9)
5 f f

3.5.4. Cox model

This model introduces the influence of the reinforcement aspect
ratio into the expression of Rule of Mixtures [51]. The Young’s
modulus of three-dimensional random short fibre composites was
evaluated using this equation:

1
εCox = L εf Vf + (1 − Vf )εm (10)
Fig. 6. FT-IR spectra of ionomer-treated and untreated sisal fibres. 5
 A. Choudhury / Materials Science and Engineering A 491 (2008) 492–500 499

Table 5
Experimental and predicted Young’s modulus for neat HDPE, sisal fibre and HDPE/sisal composites

Sample Young’s modulus, ε (GPa)

Experimental Voigt–Reuss model Halpin–Tsai model Modified mixture law Cox model

Pure HDPE 1.195 ± 0.1 – – – –

Sisal fibre 12.40 ± 2.7 – – – –
(95:5) HDPE/fibre 2.002 ± 0.2 3.333 3.003 2.313 2.207
(90:10) HDPE/fibre 2.425 ± 0.3 4.688 4.115 3.033 2.544
(85:15) HDPE/fibre 2.975 ± 0.2 5.439 4.652 3.553 3.037
(80:20) HDPE/fibre 3.124 ± 0.3 5.947 5.277 3.771 3.281

indicating that the assumptions made to derive these theoretical

equations were closely agreeing with experimental conditions. The
relative deviation between experimental and theoretical results
increased with increasing fibre content, but it was well within
agreeable limits for the Modified Rule of Mixtures and the Cox
model. This suggested that the Modified Rule of Mixtures and Cox
model can be applied to predict Young’s modulus of HDPE/sisal
composites. However, the best theoretical–experimental correla-
tion was obtained with the Cox relationship because it takes into
account the fibre aspect ratio. Apart from that, the modulus of com-
posite depends on the homogeneity of fibre dispersion at a given
volume fraction of sisal fibre. Obviously, the higher modulus was
obtained for better fibre distribution. The composite Young’s mod-
ulus was nearly the matrix modulus for poor fibre dispersion.

4. Conclusions

The present investigation showed the usefulness of sisal fibre as

a good reinforcing agent and Surlyn ionomer as an effective cou-
Fig. 7. Comparison of experimental Young’s modulus values of HDPE/sisal compos-
pling agent for composite fabrication with polyethylene matrix.
ites with different micromechanical models.
DSC analysis demonstrated that the presence and the concentra-
tion of short sisal fibres affect markedly the crystallization behavior
The independent parameters of Eq. (10) are calculated using the of HDPE matrix. A noticeable decrease in the half time of HDPE
following equations: crystallization was observed by the incorporation of short sisal
  fibre to HDPE. This could be attributed to the fact that the surfaces
tan(ˇs)
L = 1 − (11) of sisal fibres act as nucleating sites for the crystallization of the
ˇs
matrix polymer, promoting the growth and formation of transcrys-
2lf talline layer in the composites. The concentration of nucleating
s= (12) sites in the HDPE/sisal composites increased as the fibre content
0.5df
increase, which was reflected in the increase of the crystallinity
2εm of the composites. The Avrami exponent in the composite sam-
ˇ2 = (13)
εf (1 + m ) ln(1/Vf ) ples was varied around 3.0 implying three-dimensional growth
of HDPE crystals, whereas the n-values of about 2.0 for the neat
where m is the Poisson’s ratio of HDPE matrix.
HDPE suggest two-dimensional diffusion controlled crystal growth.
Generally, Young’s modulus of the composite material reflects
A hot-stage polarized optical microscopy was used to follow the
the capability of both fibre and matrix to transfer the elastic
development of transcrystalline layer along the fibre–matrix inter-
deformation under small strain without interface fracture. Table 5
face. The mechanical properties of HDPE/sisal composites were
presents the predicted Young’s modulus values obtained from
found to increase with increasing fibre content. This was ascribed
micromechanical models and the experimental data for all for-
to the reinforcing effect imparted by the ionomer-treated fibres,
mulated samples. It can be seen that the Young’s modulus of
which allowed a uniform stress distribution from the polymer
the composites increased as the fibre volume fraction increase. A
matrix to the dispersed fibre phase. A comparison between the
comparison between the experimental data and predicted Young’s
experimental results and the theoretically predicted Young’s mod-
modulus values is demonstrated in Fig. 7. The predicted Young’s
ulus has been presented. Modified Rule of Mixtures and Cox model
modulus values were recorded higher than experimental values
give more close agreement to experimental data for HDPE/sisal
(Fig. 7). Fig. 7 shows a greater degree of divergence between the
composite.
experimental Young’s modulus values and the predicted values
using Voigt–Reuss and Halpin–Tsai models, because these models
were developed for lamina composite (two-dimensional structure), References
not for three-dimensional composite. The differences of these two
[1] K.G. Satyanarayana, B.C. Pai, K. Sukumaran, S.G.K. Pillai, Handbook of Ceramics
micromechanical models with the experimental Young’s modu- and Composites, Marcel Decker, New York, 1990, pp. 339–360.
lus were recorded ∼65% for the Voigt–Reuss model and ∼50% for [2] K.G. Satyanarayana, K. Sukumaran, P.S. Mukherjee, C. Pavithran, S.G.K. Pillai,
the Halpin–Tsai model. Although, the comparison between experi- Natural Fibre-Polymer Composites, Haworth Press, England, 1990, pp. 117–136.
[3] L. Dányádi, K. Renner, Z. Szabó, G. Nagy, J. Móczó, B. Pukánszky, Polym. Adv.
mental Young’s modulus and those obtained from Modified Rule of Technol. 17 (2006) 967–974.
Mixtures or Cox model demonstrated a close agreement within 10%, [4] S. Mohanty, S.K. Verma, S.K. Nayak, Compos. Sci. Technol. 66 (2006) 538–547.
500 A. Choudhury / Materials Science and Engineering A 491 (2008) 492–500

[5] I. Baroulaki, J.A. Mergos, G. Pappa, P.A. Tarantili, D. Economides, K. Magoulas, [28] M. Pracella, D. Chionna, I. Anguillesi, Z. Kulinski, E. Piorkowska, Compos. Sci.
C.T. Dervos, Polym. Adv. Technol. 17 (2006) 954–966. Technol. 66 (2006) 2218–2230.
[6] T. Li, N. Yan, Compos. A 38 (2007) 1–12. [29] A.K. Rana, A. Mandal, S. Bandyopadhyay, Compos. Sci. Technol. 63 (2003)
[7] C. Pavithran, P.S. Mukharjee, M. Brahmakumar, A.D. Damodaran, J. Mater. Sci. 801–806.
Lett. 6 (1987) 882–884. [30] A.P. Kumar, R.P. Singh, B.D. Sarwade, Mater. Chem. Phys. 92 (2005) 458–469.
[8] K. Jayaraman, Compos. Sci. Technol. 63 (2003) 367–374. [31] A.G. Facca, M.T. Kortschot, N. Yan, Compos. Sci. Technol. 67 (2007) 2454–
[9] X. Colom, F. Carrasco, P. Pages, J. Canavate, Compos. Sci. Technol. 63 (2003) 2466.
161–169. [32] R.G. de Villoria, A. Miravete, Acta Mater. 55 (2007) 3025–3031.
[10] P.A. Sreekumar, K. Joseph, G. Unnikrishnan, S. Thomas, Compos. Sci. Technol. 67 [33] T.T.L. Doan, S.L. Gao, E. Mäder, Compos. Sci. Technol. 66 (2006) 952–963.
(2007) 453–461. [34] G. Kalaprasad, K. Joseph, S. Thomas, J. Mater. Sci. 32 (1997) 4261–4267.
[11] S. Mohanty, S.K. Nayak, Mater. Sci. Engg. A443 (2007) 202–208. [35] G. Kalaprasad, K. Joseph, S. Thomas, J. Compos. Mater. 31 (1997) 509–527.
[12] N. Chand, D. Jain, Compos. A 36 (2005) 594–602. [36] J. Brandup, E.H. Immergut, Polymer Handbook, Wiley, New York, 1988 (chapter
[13] M. Idicula, A. Boudenne, L. Umadevi, L. Ibos, Y. Candau, S. Thomas, Compos. Sci. 8).
Technol. 66 (2006) 2719–2725. [37] Y. Huang, J. Petermann, Polym. Bull. 36 (1996) 517–524.
[14] M. Brahmakumar, C. Pavithran, R.M. Pillai, Compos. Sci. Technol. 65 (2005) [38] A. Amash, P. Zugenmaier, Polymer 41 (2000) 1589–1596.
563–569. [39] J.D. Hoffman, G.T. Davis, J.I. Lauritzen, The Rate of Crystallization of Linear Poly-
[15] J.M. Park, S.T. Quang, B.S. Hwang, D.V.K. Lawrence, Compos. Sci. Technol. 66 mers with Chain Folding, Crystalline and Noncrystalline Solids, Plenum Press,
(2006) 2686–2699. New York, 1976, p. 497.
[16] L.M. Arzondo, C.J. Perez, J.M. Carella, Polym. Engg. Sci. 45 (2005) 613–621. [40] M. Avrami, J. Chem. Phys. 7 (1939) 1103–1112.
[17] H. Akimoto, T. Kanazawa, M. Yamada, S. Matsuda, G.O. Shonaike, A. Murakami, [41] S. Incardona, C. Migliaresi, D. Wagner, A. Gilbert, G. Marom, Compos. Sci. Tech-
J. Appl. Polym. Sci. 81 (2001) 1712–1720. nol. 47 (1993) 43–50.
[18] E.R. Sadiku, R.D. Sanderson, Macromol. Mater. Engg. 286 (2001) 472–479. [42] L. Mandelkern, M. Glotin, R.A. Benson, Macromolecules 14 (1981) 22–34.
[19] M. Arroyo, R. Zitzumbo, F. Avalos, Polymer 41 (2000) 6351–6359. [43] J. Maxfield, L. Mandelkern, Macromolecules 10 (1977) 1141–1153.
[20] E. Piorkowska, Macromol. Symp. 169 (2001) 143–148. [44] D. Hull, An Introduction to Composite Materials, 2nd ed., Cambridge University
[21] H. Quan, Z.-M. Li, M.-B. Yang, R. Huang, Compos. Sci. Technol. 65 (2005) Press, Cambridge, 1996, pp. 326–345.
999–1021. [45] R.M. Jones, Mechanics of Composite Material, Hemisphere Publishing Corpora-
[22] H. Schonhorn, F.W. Ryan, J. Polym. Sci. 6 (1968) 231–240. tion, New York, 1975, pp. 86–104.
[23] C. Wang, L.M. Hwang, J. Polym. Sci. B 34 (1996) 1435–1442. [46] W. Voigt, Ann. Phys. 38 (1889) 573–587.
[24] C. Wang, C.R. Liu, Polymer 38 (1997) 4715–4718. [47] A. Reuss, Z. Angrew. Math. Mech. 9 (1929) 49–58.
[25] P.V. Joseph, K. Joseph, S. Thomas, C.K.S. Pillai, V.S. Prasad, G. Groeninckx, M. [48] J.C. Halpin, J. Compos. Mater. 3 (1969) 732–734.
Sarkissova, Compos. A 34 (2003) 253–266. [49] H. Fukuda, T.W. Chou, J. Mater. Sci. 16 (1981) 1088–1096.
[26] P.V. Joseph, K. Joseph, S. Thomas, Compos. Sci. Technol. 59 (1999) 1625–1640. [50] H. Fukuda, T.W. Chou, J. Mater. Sci. 17 (1982) 1003–1011.
[27] J.L. Thomason, A.A. Van Rooyen, J. Mater. Sci. 27 (1992) 889–896. [51] H.L. Cox, Brit. J. Appl. Phys. 3 (1952) 72–79.