Polymer Degradation and Stability 87 (2005) 411e418

www.elsevier.com/locate/polydegstab

Investigation of interfacial modication for ame

retardant ethylene vinyl acetate copolymer/alumina
trihydrate nanocomposites
Xingui Zhang, Fen Guo*, Jianfeng Chen, Guoquan Wang, Hui Liu
Department of Chemical Engineering, Key Lab for Nanomaterials, Ministry of Education,
Beijing University of Chemical Technology, Beijing 100029, PR China
Received 18 July 2004; received in revised form 30 August 2004; accepted 31 August 2004
Available online 11 November 2004

Abstract

Ethylene vinyl acetate copolymer (EVA)/alumina trihydrate (ATH) nanocomposites were prepared by melt-blending. Two kinds
of interfacial modiers were used in this research, a titanate coupling agent and maleated ethylene vinyl acetate copolymer (MEVA),
which acted as a compatibiliser. The eects of modiers on the properties of EVA nanocomposites were studied by
thermogravimetric analysis, tensile and combustion tests. The dispersion and adhesion patterns of the ATH nano-particles in
EVA matrix were characterized through Molau solution test, TEM and SEM. The experimental results demonstrated that the use of
untreated ATH could clearly decrease the tensile properties of EVA composites. It was found that the combined addition of the two
kinds of interfacial modiers led to a dramatic increase in tensile and ame retardant properties of the nanocomposites. Also the
thermal stability of this system, in terms of the thermal degradation test, was enhanced. These eects would be diminished if only
one modier was used. The improvement in the properties of interfacial modied EVA/ATH nanocomposites is mostly attributed to
a better dispersion of surface modied ller and a strong adhesion between the ller and matrix.
2004 Elsevier Ltd. All rights reserved.

Keywords: Interfacial modication; ATH; EVA nanocomposites; Mechanical properties; Flame retardancy

1. Introduction alumina trihydrate (ATH) composites have been widely

In recent years, a great deal of attention has been
paid to the use of ame retardant (FR) polymeric
composites with halogen-free (HF) ame retardants, for
example conventional alumina trihydrate or magnesium
hydroxide, because these materials do not cause
environmental pollution, such as toxicity, heavy
smoke and corrosion [1e3]. In the wire and cable
industry, ethylene vinyl acetate copolymer (EVA)/
* Corresponding author. Tel.: C8601064451035; fax: C860106443
4784.
E-mail address: guof@mail.buct.edu.cn (F. Guo).
researched and used [4,5].
To achieve the required ame retardancy, high
loadings of about 60 wt% of ATH are always adopted,
which could lead to a poor mechanical performance of
the resulting materials [4]. This problem can be solved
through improving the dispersion and compatibility of
the inorganic ller in polymer matrix. While a variety of
surface modiers can be eective dispersants for llers,
only coupling agents and polymer compatibilisers are
found to be able to form a chemical connection
between the two phases. The coupling agents are
designed to be a molecular bridge, which has limited
adhesion eect [6]. The compatibilisers consist of two
parts, the hydrophobic polyolen copolymer or grafting

0141-3910/$ - see front matter 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.polymdegradstab.2004.08.013
412 X. Zhang et al. / Polymer Degradation and Stability 87 (2005) 411e418

with polar hydrophilic monomers, such as acrylic acid
(AAC), methyl methacrylate (MMA) and maleic anhy-
dride (MAH). These reactive groups interact readily
with functional groups on the inorganic ller, and the
polyolen phase is miscible or compatible with the
polymer matrix. A durable connection between the ller
and the matrix is thereby established. So compatibilisers
are also termed polymer coupling agents [6,7]. Many
studies about polyolen/ATH composites were focused
on the eects of coupling agents or compatibilisers.
Liauw et al. systematically studied the eects of various
coupling agents on the mechanical properties of PP/
ATH composites [8]. A recent work reported by Hippi
et al. showed that the presence of compatibilisers could
improve both stiness and toughness of PE/ATH
composites while FR properties of the composites
remained unchanged [3].
The present investigation was concerned with the
tensile properties and ame retardancy of EVA FR
composites with nano-scale ATH, which is a material
likely to agglomerate in a polymer matrix. Appropriate
coupling agent and compatibiliser were selected in order
to enhance the dispersion and adhesion between the
ller and matrix, so as to improve the properties of
EVA/ATH nanocomposites.
After thorough drying at 100  C for 10 h, ATH
particles were charged into a bench top mixer and the
modier solution was added slowly over a period of
10 min to ensure uniform distribution of the coupling
agent. After completion of the titanate addition,
the ller was continuously mixed for 25 min, and the
temperature was in the range of 100e110  C. Then the
surface treated ATH-T particles were dried at 100  C
for at least 3 h.
2.2.2. Polymer compounding
The EVA cable compositions are shown in Table 1.
These raw materials were rst mixed using blending
equipment. Then they were formed into pellets by
compounding the polymer via a Prism 16 mm twin
screw extruder at operating temperatures of 110  C at
zone 1, 140  C at zone 2, 150  C at zone 3, 160  C at
zone 4 and 60 rpm screw speed. Subsequently the pellets
were fed into a single screw extruder to prepare the test
specimens, at temperatures of 125  C at zone 1, 155  C
at zone 2, 170  C at zone 3, 165  C at zone 4 and 50 rpm
screw speed.
2.3. Analysis of samples

2.3.1. Combustion test

2. Experimental
Two standard test methods were used to evaluate the
FR properties of all samples prepared.
2.1. Materials
The limiting oxygen index (LOI) test was adopted to
giving a measure of the relative ammability of these
An ethylene vinyl acetate copolymer (EVA14-2)
samples in which case higher LOI values represented
resin, produced by Beijing Organic Chemical Plant with
better ame retardancy. Measurements were undertaken
melt ow rate of 2 g/10 min containing 14 wt% vinyl
within a specimen of a dimension of 120 ! 7 ! 5 mm
acetate, was used. The alumina trihydrate (ATH), from
using a HC-22 type instrument (made in China) in
Zhongzhou Branch, Aluminium Corporation of China
accordance with ISO 4589-1964.
Limited, has an average particle size of 100e125 nm.
The UL-94 test was performed to evaluate the ame
The coupling agent T, isopropyl tris (dioctyl pyrophos-
spread of the test specimens in air atmosphere, using
phate) titanate (KR-38S, MW Z 1310.4, chemical struc-
a CZF-1 type instrument (made in China) on sheets
ture shown in Fig. 5), was supplied by Yangzhou Lida
according to ISO 1210-1992.
Resin Company, China. The compatibiliser MEVA,
a type of EVA grafted with 2 wt% maleic anhydride,
was manufactured by Ningbo Nengzhiguang New Table 1
Compositions of all samples
Materials Technology Cooperation, China.
Samples EVA/wt% ATH/wt% ATH-T/wt% MEVA/wt%
1 100 0 0 0
2.2. Sample preparation 2 80 20 0 0
3 60 40 0 0
2.2.1. Treating the nano-ATH surface 4 40 60 0 0
with the coupling agent T 5 80 0 20 0
6 60 0 40 0
The dosage of the coupling agent was 0.5 wt% (based
7 40 0 60 0
on ller mass), and the modier was diluted in atoleine 8 90 0 0 10
to make up a 5 wt% solution. Note that the dosage of 9 70 0 20 10
the modier was selected based on the earlier experi- 10 50 0 40 10
mental results on ame retardancy and mechanical 11 30 0 60 10
12 30 60 0 10
behaviours of EVA composites [9].
 X. Zhang et al. / Polymer Degradation and Stability 87 (2005) 411e418
2.3.2. Thermogravimetric analysis test
The thermogravimetric analysis (TGA) data were
obtained using a STA-449C thermogravimetric analyser
(made in Germany). In each case, a 10e20 mg sample
was tested in air or nitrogen atmosphere at a heating
rate of 10  C/min.
2.3.3. Mechanical properties test
Before testing the mechanical properties, the extrud-
ed samples were conditioned at 23  C and a relative
humidity of 50% for 48 h. Tensile properties were
characterized with an Instron 4706 universal testing
machine (made in USA) with a test speed of 2 mm/min
and with specimen-type 1BA in line with the standard
ISO 527-1993(E).
2.3.4. Assessment of the dispersion and
adhesion qualities of EVA nanocomposites
Two methods were introduced to evaluate the
dispersion and adhesion qualities of the composites.
First, Molau solution test was carried out. Samples
(3 g) were dissolved in 50 ml of cyclohexane solution
with heating. The mixtures were then left at room
temperature for 80 h. Because cyclohexane is a favour-
able solvent for EVA resin and a non-solvent for ATH,
it is expected that if the compatibility between the two
components is poor, the separation and deposition of
the non-soluble component will be observed. Otherwise
the solution should be in a homogeneous and stable
state.
Fig. 1. TEM micrographs of dierent EVA/ATH nanocomposites.
413
In another way, transmission electron microscopy
(TEM Hitachi H-800, Japan) and scanning electron
microscope (SEM Hitachi LTDx-650, Japan) were used
for the visualization of ller dispersion and interaction
patterns in polymer matrix.
3. Results and discussion
3.1. TEM observation
Dispersion patterns of the unmodied, the coupling
agent treated and the compatibilised composites of EVA
with nano-ATH at high loadings (samples 4, 7 and 11)
are shown in Fig. 1. It is evident that the untreated ATH
was poorly dispersed in EVA matrix (Fig. 1(a)). The
eect of interfacial modication on ller dispersion is
apparent from Fig. 1(b) and (c). The coupling agent T
acted as an eective dispersant, and MEVA made the
dispersed phase more homogeneous in size.
3.2. Flame retardancy
Table 2 lists the LOI values, UL-94 test results,
Youngs moduli, tensile strengths, and elongations at
break of all samples, whose compositions are listed in
Table 1. It is seen that pure EVA resin (sample 1) is
ammable with the LOI value of 17.0, and ATH exhibits
ame retardancy at high loadings. For instance, with
addition of 60 wt% the untreated ATH, the LOI value
414 X. Zhang et al. / Polymer Degradation and Stability 87 (2005) 411e418

Table 2
Combustion and tensile properties of all samples
Samples Combustion tests Tensile test
LOI/%vol UL-94/mm Youngs modulus/MPa Tensile strength/MPa Elongation at break/%
1 17.0 Fail 0.80 G 0.02 15.7 G 0.1 1635.6 G 20.6
2 18.2 Fail 0.98 G 0.04 12.3 G 0.5 1425.3 G 30.7
3 22.4 Fail 47.10 G 2.91 7.5 G 0.2 128.5 G 12.3
4 30.6 V-2 103.33 G 5.76 4.2 G 0.1 23.5 G 2.5
5 19.8 Fail 0.89 G 0.05 15.1 G 0.6 1610.7 G 21.8
6 24.8 HB (70 G 5.6) 28.27 G 1.83 9.9 G 0.3 600.1 G 20.1
7 37.2 V-1 77.46 G 3.52 8.2 G 0.2 78.1 G 4.7
8 17.2 Fail 0.80 G 0.05 15.6 G 0.2 1651.1 G 32.3
9 20.0 Fail 0.75 G 0.01 16.1 G 0.5 1620.3 G 24.7
10 27.3 V-2 19.71 G 3.14 12.2 G 0.4 709.1 G 19.8
11 39.1 V-0 45.82 G 1.03 11.0 G 0.6 182.0 G 12.5
12 37.9 V-1 56.89 G 3.97 9.3 G 1.1 123.7 G 21.6

of sample 4 was 30.6 and the UL-94 test only passed the
V-2 rating. In comparison to the previous case, the ame
retardancy of the coupling agent treated system (sample
7) was improved, i.e., this specimen passed the UL-94
V-1 rating with the LOI value of 37.2. After 10 wt%
MEVA was added (sample 11), the LOI value increased
slightly to 39.1, and UL-94 test rating improved from
V-1 to V-0.
It is inferred that the raising of LOI is possibly due to
a better dispersion of the treated ATH in EVA resin
(Fig. 1(b), (c)) [6]. Consequently, a high specic surface
area of the powders is obtained. In the combustion of
these nanocomposites, ATH particles are heated homo-
geneously, and their decomposition rate is accelerated.
Furthermore, pyrophosphate, which is the composition
of the coupling agent T, could hydrolyse into phosphate
ester, which to some extent enhances ame retardancy
[10].
It also shows that the improvement in FR properties
of the compatibilised system is more evident. The TGA
results for EVA and MEVA, in air atmosphere at the
rate of 10  C/min, are shown in Fig. 2. It is obvious that
a considerable change has occurred in the thermal
behaviour of EVA upon grafting. The initial decompo-
sition temperature of MEVA was higher than that of
EVA. This suggests that the grafted MAH oers an
eect against the ammability of EVA by raising the
thermal stability of polymer matrix, and MEVA dis-
plays a FR synergistic eect with ATH [11].
3.3. Thermogravimetric analysis
Fig. 3 presents the thermogravimetric (TG) curves of
the unmodied, the coupling agent treated and the
compatibilised EVA composites in N2 atmosphere at the
heating rate of 10  C/min. The EVA/ATH blends
showed a two-step weight loss as previously reported

Fig. 2. TGA curves of EVA and MEVA at a heating rate of 10  C/min Fig. 3. TGA curves of dierent EVA composites at a heating rate of
in air atmosphere. 10  C/min in N2 atmosphere.
 X. Zhang et al. / Polymer Degradation and Stability 87 (2005) 411e418 415

Table 3
TGA results for the thermal degradation of these samples
Samples Tipa/  C Char at First step Second step
600  Cb/wt% Tbc/  C Ted/  C Weight Tb/  C Te/  C Total weight
loss/wt% loss/wt%
4 261 1.2 240 340 24.8 340 600 59.6
7 273 2.6 240 340 18.2 340 600 58.2
11 287 2.9 240 340 16.9 340 600 57.9
a
The initial pyrolysis temperature.
b
The char of ATH over 35e600  C has already been subtracted.
c
The beginning temperature.
d
The end temperature.

[5,12]. The rst step was considered to involve the
dehydration reaction of ATH and the loss of acetic acid
in the range of 240e340  C. The second step was due to
the volatilisation of the residual polymer at temper-
atures of between 340 and 600  C.
Table 3 lists the TGA analysis data of the
three samples. It was clear that the initial pyrolysis
temperatures of the interfacial modied systems were
higher than that of the unmodied one. Meanwhile,
samples 7 and 11 gave more residue than that of sample
4 during the two steps of degradation. The residue layer
insulates the underlying material and slows the escape of
the volatile products generated during decomposition
[13]. It is also worth noting that a higher residual weight

Fig. 4. Eects of ller loading and interfacial modiers on the tensile properties of EVA composites.
416 X. Zhang et al. / Polymer Degradation and Stability 87 (2005) 411e418

is also associated with higher ame retardancy [12]. In

conclusion, the compatibilised EVA nanocomposites
show the highest thermal stability.

3.4. Mechanical properties

Dierences in the eects of modiers on the tensile

properties of EVA composites are readily seen in Fig. 4.
An addition of the untreated ATH produced a clear
increase in the Youngs modulus of EVA (Fig. 4(a)).
Concurrently, however, the unmodied samples lost the
strength and toughness of pure EVA resin as shown in
the low values of tensile strength and elongation at
break in Fig. 4(b) and (c). Because a usual method to
compensate loss in toughness is to apply a coating [3],
blends of EVA with the treated ATH-T were also
prepared. As expected, the moduli of these blends were
lower than those of unmodied ones (Fig. 4(a)), and the
former showed a moderate improvement in toughness
and strength. However, at a high loading of the treated Fig. 5. Bonding mechanism of the interfacial modiers between the
ATH (sample 7), tensile strength was only 8.2 MPa, and ller and EVA matrix.
elongation at break was 78.1%, which still cannot
satisfy the performance criterion. adhesion. The mechanical properties of the coupling
When the coupling agent treatment was combined agent treated system performed deciently at high ller
with an addition of the compatibiliser MEVA in loadings due to limited short-range intermolecular
blending step, pronounced improvement was achieved entanglements between the ller and matrix, just as
in the tensile properties of EVA composites. The can be seen from step 2. When the loading of the ller
compatibilised EVA/ATH system, whose values of was low, the treated ATH also showed good dispersion
Youngs modulus were the lowest in the three kinds of and compatibility in polymer matrix, so the usefulness
EVA composites as depicted in Fig. 4(a), remains the of MEVA did not appear.
soft and tough characteristics of the original EVA resin.
According to the experimental data of sample 11, tensile
3.5. Study of llerematrix compatibility
strength improved 34%, and elongation at break
increased approximately 232% compared with sample
3.5.1. Molau solution test
7. However, there is nearly no dierence between pure
Fig. 6 illustrates the results of Molau solution tests.
EVA resin (sample 1) and EVA/MEVA blends (sample
Since cyclohexane is a solvent for EVA, the pure resin
8), which claries that the compatibiliser itself is not
was dissolved rapidly when heated, then formed a trans-
a reinforcing agent.
parent colourless colloid (Fig. 6(a)). There is a distinct
It should be pointed out once again that mechanical
phase separation in Fig. 6(b). In the bottom part of the
behaviours depend strongly on interfacial dispersion
tube (the black area) there are deposits of inorganic
and adhesion qualities [7,14]. The interfacial properties
ATH, in the middle part the solution dissolving partial
are greatly inuenced by the additions of various
EVA composites is observed, and in the upper the
interfacial modiers. Fig. 5 represents the reaction
cyclohexane solvent exists. It indicates that there is
mechanism of the coupling agent and compatibiliser
scarcely any chemical bonding between EVA matrix and
on EVA/ATH composites. At rst the pyrophosphate
the untreated ATH (sample 4).
groups of the coupling agent react with the free protons
in the inorganic ATH interphase (step 1), giving rise to
the formation of organic monomolecular layers on the
surface of the substrate. When MEVA is added (step 2),
these layers are compatible and have esterication
reactions with the anhydride on the compatibiliser (step
3). Moreover, there is a possibility of miscibility between
MEVA and EVA matrix [15]. So there exists a rm
bonding between the polymer resin and ller particles as
(a) sample 1 (b) sample 4 (c) sample 7 (d) sample 11
depicted in Fig. 5. The tensile test results obtained from
the compatibilised system reect enhanced llerematrix Fig. 6. The results of Molau solution test.
 X. Zhang et al. / Polymer Degradation and Stability 87 (2005) 411e418
The coupling agent treated composite (sample 7) was
rst dissolved into homogeneous colloid. After 50 h,
a small quantity of ATH began to deposit, as shown in
Fig. 6(c), indicating that traditional coupling agent was
inecient to form durable llerematrix adhesion.
Fig. 6(d) reveals that the compatibilised composite
(sample 11) maintained a stable colloid state in
cyclohexane solvent. It is attributed to the network
structure caused by MEVA in EVA/ATH nanocompo-
site, as depicted in Fig. 5. So this structure retards ATH
deposition, and enhances the compatibility between the
ller and matrix.
3.5.2. SEM observation
The adhesion quality of the ller with EVA matrix is
readily visible from Fig. 7. Fig. 7(a) shows the fracture
surface of the unmodied EVA composite (sample 4), in
which case ATH particles were distinctly separated from
EVA matrix and no adhesion existed between them. The
voids surrounding the additive particles looked like
bubbles because of the strong surface tension of the
matrix [6]. The interfacial bonding was improved for the
coupling agent treated system (sample 7) as can be seen
in Fig. 7(b), but there were still some voids around
particles.
If MEVA was directly added into the unmodied
system (sample 12), the compatibiliser was only miscible
with EVA matrix and did not enhance the interfacial
bonding, as shown in Fig. 7(c). It is obvious that the
Fig. 7. SEM photos of dierent EVA/ATH composites.
417
anhydride grafted on the compatibiliser reacts more
easily with hydroxyl of the coupling agent than that on
ATH particle surface, just as illustrated in Fig. 5. So the
properties of sample 12, which were listed in Table 2,
were inferior to those of sample 11.
Fig. 7(d) shows that the fracture-surface morphology
of the compatibilised EVA/ATH composite (sample 11)
has a great change. All the particles were encapsulated
by MEVA tightly, so a smooth transition was estab-
lished between the inorganic phase and polymer phase.
4. Conclusion
1. It is found that, if the conventional dosage of ATH
necessary for ame retardancy has been adopted, the
mechanical properties of EVA composites based on
nano-scale ATH would be poor.
2. The interfacial modied system, in combination with
the titanate coupling agent and polymer compatibil-
iser (MEVA), showed dramatic improvement in
tensile properties and ame retardancy. Also the
thermal stability of this system was increased. These
eects would be diminished if only one modier was
used.
3. The improvement in the properties of interfacially
modied EVA/ATH nanocomposites is mostly
attributed to a better dispersion of surface modied
ller and a strong adhesion between the ller and
matrix.
418 X. Zhang et al. / Polymer Degradation and Stability 87 (2005) 411e418
4. There are two causes which lead to rm connection
on interface of ATH particles and EVA matrix, the
stable miscibility of MEVA with EVA matrix and
the esterication reactions between the hydroxyl of
the coupling agent linked to ller surface and
anhydride on the compatibiliser.
Acknowledgements
This work is nancially supported by the 863 Project
Foundation of China. Grant No. 2002AA302605.
References
[1] Li ZZ, Qu BJ. Flammability characterization and synergistic
eects of expandable graphite with magnesium hydroxide in
halogen-free ame-retardant EVA blends. Polym Degrad Stab
2003;81(3):401.
[2] Camino G, Maezzoli A, Braglia M. Eect of hydroxides and
hydroxycarbonate structure on re retardant eectiveness and
mechanical properties in ethylene-vinyl acetate copolymer. Polym
Degrad Stab 2001;74(3):457.
[3] Hippi U, Mattila J, Korhonen M. Compatibilization of poly-
ethylene/aluminum hydroxide (PE/ATH) and polyethylene/mag-
nesium hydroxide (PE/MH) composites with functionalized
polyethylenes. Polymer 2003;44(4):1193e201.
[4] Cross MS, Cusack PA, Hornsby PR. Eects of tin additives
on the ammability and smoke emission characteristics of
halogen-free ethylene-vinyl acetate copolymer. Polym Degrad
Stab 2003;79(2):309.
[5] Serge B, Michel LB, Robert L. Recent advances in the use of zinc
borates in ame retardancy of EVA. Polym Degrad Stab 1999;
64(3):419.
[6] Chiang WY, Hu CH. Approaches of interfacial modication for
ame retardant polymeric materials. Composites Part A 2001;
32(3/4):517e24.
[7] Tselios Ch, Bikiaris D, Maslis V. In situ compatibilization of
polypropyleneepolyethylene blends: a hermomechanical and
spectroscopic study. Polymer 1998;39(26):6807e12.
[8] Liauw CM, Lees GC, Hurst SJ. The eect of ller surface
modication on the mechanical properties of aluminium hydrox-
ide lled polypropylene. Plast Rubber Compos Process Appl
1995;24(5):249e60.
[9] Gonzalez J, Albano C, Ichazo M. Analysis of thermogravimetric
data of blends of polyolens with calcium carbonate treated with
Lica 12. Polym Degrad Stab 2001;73(2):218.
[10] Stepek J, Daoust H. Reinforcing llers, reinforcing agents, and
coupling agents. In: Stepek J, Daoust H, editors. Additives for
plastics. New York: Springer-Verlag Press; 1982 [chapter 4].
[11] Chiang WY, Ku YA. The eect of maleic anhydride grafting on
the ame retardation of plasma pretreated polyethylene. Polym
Degrad Stab 2002;76(2):284e8.
[12] Yeh TJ, Yang MJ, Hsieh SH. Combustion of polyethylenes lled
with metallic hydroxides and ethylene vinyl acetate copolymer.
Polym Degrad Stab 1998;61(3):468e9.
[13] Wang SF, Hu Y, Song L. Preparation and thermal properties of
ABS/montmorillonite nanocomposite. Polym Degrad Stab 2002;
77(3):424.
[14] Thomas PS, Joseph K, Thomas S. Mechanical properties of
titanium dioxide-lled polystyrene microcomposites. Mater Lett
2004;58(1):281.
[15] Li B, He JM. Investigation of mechanical properties, ame
retardancy and thermal degradation of LLDPE e wood-bre
composites. Polym Degrad Stab 2004;83(2):242.