You are on page 1of 8

Composites: Part A 39 (2008) 1851–1858

Contents lists available at ScienceDirect

Composites: Part A
journal homepage: www.elsevier.com/locate/compositesa

Fracture behaviour of fumed silica/epoxy nanocomposites


M. Battistella a, M. Cascione a, B. Fiedler b, M.H.G. Wichmann b, M. Quaresimin a,*, K. Schulte b
a
Department of Management and Engineering, University of Padova, Stradella San Nicola 3, 36100 Vicenza, Italy
b
Institute of Polymers and Composites, Technische Universität Hamburg-Harburg, Denickestrasse 15, D-21073 Hamburg, Germany

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

Article history: The paper shows the results of an experimental programme oriented to the manufacturing and charac-
Received 12 April 2008 terisation of nanocomposites based on epoxy resin modified with fumed silica nanoparticles (0.1, 0.3 and
Received in revised form 4 September 2008 0.5 vol.%) with different surface treatments (unmodified, AMEO and GLYMO). The resulting degree of dis-
Accepted 7 September 2008
persion and the interfacial adhesion were investigated by TEM and SEM. Previous tensile and fracture
toughness properties are discussed together with new results on the fatigue behaviour under mode I
loading. The fracture toughness KIc turned out to be significantly increased and a positive trend towards
Keywords:
the improvement in the crack propagation threshold under fatigue was observed.
A. Nano-structures
B. Fracture toughness
Ó 2008 Elsevier Ltd. All rights reserved.
B. Fatigue
Fumed silica

1. Introduction ural modulus (around 40%), strength (15%) and fracture toughness
(120%). Furthermore, the crack propagation threshold and resis-
The addition of nanoparticles has proven to exhibit a high po- tance turned out to be improved dramatically, with the crack prop-
tential for significantly improving mechanical properties of poly- agation rates for nanocomposites being orders of magnitude
mers. This behaviour justifies the great interest of the scientific slower than neat resin for the same range of SIF.
community and industry. A vast amount of experimental works Adebahr et al. [6] proposed a novel route to prepare nanocom-
has been performed and published during recent years. Neverthe- posites consisting of monodispersed SiO2 nanoparticles and reac-
less, most of the available data refer to static properties while the tive resin. The addition of 23 wt.% of particles subjected to
behaviour under cycling and dynamic loading is rarely investi- thermal anhydride curing induced a 66% increase in KIC, while
gated. A brief overview on some of the most important work is pre- UV curing led to an improvement of 82% at 50 wt.%. Lin et al. [7]
sented below. reported that tensile and impact strength of titanium dioxide and
Siegel et al. [1] obtained an increase of 15% of the strain to fail- montmorillonite filled epoxy resin reached a maximum for a filler
ure filling an epoxy resin with 10 wt.% of nanometric TiO2 particles. content of 5–8 vol.% and decreases at higher filler contents, some-
Evora et al. [2] found that adding only 1 vol.% of TiO2 nanoparticles times even below the neat resin values.
within unsaturated polyester resin increased the fracture tough- Ragosta et al. [8] improved the mechanical properties of epoxy
ness of about 57% due to the uniform and fine dispersion of the fil- resin adding 10 wt.% of silica particles with a diameter of 10–
ler within the resin at low volume contents. More significant 15 nm. The normalized elastic modulus reached the value of 1.5,
enhancements in fracture toughness (almost 100% at 4.5 vol.% of while the normalized yield strength increased up to 1.3. The addi-
Al2O3 nanoparticles in unsaturated polyester) were achieved tion of silica raised the fracture energy of the epoxy matrix by a
improving the particle-matrix adhesion through a silane surface factor of about 4, whereas the increase of KIc was twofold.
treatment [3]. Zheng et al. [9] found that the addition of 3 wt.% of silica nano-
Wetzel et al. [4] studied the effects of nano (alumina) and mi- particles within epoxy matrix leads to an increase in tensile
cro-spherical (calcium silicate) particle addition to epoxy resin strength of 115%, while the impact strength increases by 56%.
and found increases in flexural modulus (up to 35%), strength (up In the literature, the toughening effect due to the addition of
to 20%) and Charpy impact energy (up to 35%). In a following, inter- particles to polymers has been studied for a long time [10–12]. Dif-
esting work [5], neat epoxy reinforced with Al2O3 nanoparticles at ferent toughening mechanisms have been mentioned, such as the
different volume contents was investigated. The 10 vol.% epoxy/ localised inelastic matrix deformation and void nucleation, particle
Al2O3 nanocomposite exhibited significant improvements in flex- debonding, crack deflection, crack pinning, crack tip blunting, par-
ticle deformation or breaking at the crack tip. However, it is still an
* Corresponding author. Tel.: + 39 0444 998723. open question which are the effective mechanisms responsible for
E-mail address: marino.quaresimin@unipd.it (M. Quaresimin). toughening on nanocomposites [13]. Furthermore, experimental

1359-835X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compositesa.2008.09.010
1852 M. Battistella et al. / Composites: Part A 39 (2008) 1851–1858

techniques and descriptive models are based on macro-mechanical From the previous discussion, the beneficial effects of nanopar-
concepts. Thus, their application to nanocomposites is not straight- ticles in improving the mechanical properties are evident even at
forward and indeed questionable. Particle-matrix debonding and low filler contents. However, most of the work has been done on
localized deformations in the process zone ahead of the crack tip the investigation of static properties, while few papers
are probably responsible of the considerable toughening effect [5,22,23,29] can be found about the effects of nanofillers on the fa-
brought by nanomodification. Recent experimental investigations tigue behaviour of polymers. Blackman et al. [23] observed an in-
by Johnsen et al. on silica nanoparticle reinforced epoxy polymers crease of KIc in an epoxy matrix, using nanoscaled silica particles.
confirm these assumptions [14]. Because of the very high specific The KIc improvement ranged from 27% at 4.0 wt.% of nano-SiO2
surface area, even very low filler contents can significantly contrib- up to 73% at 20 wt.%. Moreover, they observed an improvement
ute to matrix reinforcement. Especially interface related effects, in fatigue behaviour, as DKth was increased significantly. A correla-
such as debonding mechanisms and void nucleation could play a tion between KIc and DKth was found.
significant role even at low volume contents. Although classical In the present work attention will be focused on the crack prop-
mechanical theories concerning particle toughening sometimes agation resistance under cyclic loading. Indeed, the few previous
even predict a decrease of toughening contribution with decreas- investigations seem to suggest that nanomodification is able to sig-
ing particle size, the increasing amount of interfacial area and nificantly enhance the fatigue performance of polymers. If this
absolute number of particles in the process zone can be reasons beneficial effect is confirmed, it will be another interesting feature
for the experimentally observed increases in KIc [15]. of polymer nanocomposites.
Xie et al. [16] reported the improvement of the mechanical
properties of PVC with the addition of CaCO3. At 5 vol.%, optimal
performances were achieved in Young’s modulus, tensile yield 2. Experimental
strength, strain to failure and Charpy impact energy. The filler en-
abled ductile fracture caused by elevated triaxial stresses at the 2.1. Materials
neck region and consequently debonding at the particle-matrix
interface. Increasing the load, the ligaments between the voids The epoxy matrix investigated was a modified bis-phenol-A-
were stretched increasing the energy consumption. based epoxy resin (MGS L135i) together with an amine curing
Lazzeri et al. [17] showed that the addition of 10 vol.% of un- agent H137i, both supplied by the Bakelite MGS Kunstharzpro-
coated CaCO3 led to an increase in Young’s modulus and yield dukte GmbH. Since this system is characterized by a low viscosity
stress and to a decrease in impact strength. On the other hand, if [gRT = 250 mPa s], it can be used in LCM manufacturing processes.
the particles were covered with stearic acid, the tensile properties The nanofiller used was AEROSILÒ 380, a hydrophilic fumed sil-
slightly dropped and the impact strength linearly increased with ica with a specific surface area of 380 m2/g and an average primary
the stearic acid surface concentration. The fracture surface analysis particle diameter of 7 nm. The AEROSIL 380 was provided by De-
showed cavities and voids due to debonding and deformation gussa AG, Germany. The surface of the fumed silica was chemically
bands in the stress whitened areas. As demonstrated in [18], the modified with two different coupling agents: 3-aminopropyltrime-
void formation allows for a plastic deformation of the interparticle thoxysilane (Dynasylan AMEO) and 3-glycidyloxypropyltrimeth-
ligaments, which is assumed to be main absorbing energy oxy-silane (Dynasylan GLYMO), both supplied by Degussa AG,
mechanism. Germany.
Yang et al. [19] investigated the fracture behaviour of polyam- The surface modification process as well as the manufacturing
ide 66 filled with TiO2 nanoparticles. With the increase of the parameters are reported elsewhere [24].
TiO2 content from 1 to 3 vol.%, the plastic zone around the crack Nine different nanocomposites were produced and analysed
tip decreased and the density of dimples near the pre-notched area since three different fumed silica volume fractions (0.1, 0.3,
increased. Thus, the energy absorbed during crack propagation 0.5 vol.%) and three different chemical modification conditions
should be higher for the nanoreinforced matrix than for the pure (non-treated fumed silica, AMEO-modified and GLYMO-modified
polyamide. fumed silica) were considered.
In the last decades the greatest part of the researches carried The fatigue performances of nanocomposites were then com-
out on nanomodification was oriented to thermoplastic matrix pared with those of the neat epoxy.
nanocomposites, however the attention of the scientific commu-
nity has recently moved to the nanomodification of thermosetting 2.2. Morphological analysis
resins, in view of their possible application as matrix for ternary,
fibre reinforced laminates. Because of their size in the nanometre The degree of dispersion of the filler within the matrix was ana-
regime, nanoparticles are small enough to penetrate the fibre rov- lysed by TEM and SEM investigations. The SEM analysis was car-
ings and to act as matrix reinforcement in FRP laminates. ried out without any gold sputtering, because the presence of the
Chisholm et al. [20] investigated the influence of a nanocom- gold-layer on the specimen surface could lead to the loss of impor-
posite matrix on a laminate composite. The presence of 1.5 wt.% tant information about the filler dispersion, since it has the same
of SiC nanoparticles within the epoxy resin increased the tensile dimensions of the particles diameter.
modulus of the nano-modified matrix of 44% and the tensile From TEM analysis only qualitative results can be gained about
strength of 15%. The tensile modulus of the corresponding lami- the dispersion and distribution of the nanofiller. In general, it can
nate increased of 23.5% and the tensile strength of 11%. Also in be said that the treatment with AMEO and GLYMO resulted in only
the flexural test the laminate containing 1.5 wt.% of SiC in the ma- slight differences regarding the state of dispersion. A generally
trix showed improvements (39% in strength and 12% in modulus). good state of dispersion was found for all three systems investi-
However, when the nano-reinforcement was increased from 1.5% gated. Both, the unmodified and the GLYMO-modified fumed silica
to 3%, a worsening of both tensile and flexural properties of the particles showed a comparable fine degree of dispersion and the
composite was observed. resulting fine network structure caused by the high attractive
Kinloch et al. [21] investigated the fracture behaviour of GFRP forces between the particles.
laminates with a nano-modified epoxy matrix and found consider- Fig. 1 shows a TEM micrograph of non-modified fumed silica at
able increase in GIc and GIIc when using silica nanoparticles alone 0.1 vol.%. It reveals the tendency of fumed silica nanoparticles to
and in combination with a CTBN toughening. create small clusters within the matrix. These aggregates tend to
M. Battistella et al. / Composites: Part A 39 (2008) 1851–1858 1853

Fig. 1. TEM analysis on the 0.3% unmodified fumed silica nanocomposite. Fig. 3. TEM analysis on the 0.5% AMEO-modified fumed silica nanocomposite.

attract each other forming greater structures called agglomerates, surrounds these clusters of aggregates and are therefore not dis-
which have a chain-like branched structure. In general, these persible anymore since they are already chemically bonded. In fact,
aggregates consist of several tens of primary particles and exhibit during the following period of curing, the matrix can react only
dimensions of several ten to one hundred nanometres. with the functional groups which cover the surface of these aggre-
The dispersion achieved with the GLYMO-modified silica parti- gates, encapsulating them within the matrix itself. Fig. 4 reports an
cles is shown in Fig. 2. It shows a very fine and uniform distribution SEM analysis on the 0.5% AMEO-modified fumed silica nanocom-
comparable to that of unmodified silica particles. The same sizes posite. Aggregates and agglomerates are clearly visible on the frac-
were found for the aggregate substructures in these nanocompos- ture surface of the tensile test specimens after testing.
ites. Figs. 1 and 2 are representative for the overall morphology of
these nanocomposites, as the structure and size of nanoparticle 2.3. Tensile and fracture toughness results
clusters observed in all the TEM specimens were very similar.
These findings were confirmed by the rheological results: in fact, In our previous work [24], the results of the static tests for the
GLYMO-modified silica particles had shown the lowest value of mechanical characterisation have already been reported, including
viscosity (compared to the neat resin) and which can be attributed DMTA. The results of tensile and fracture toughness tests are sum-
to the presence of few small aggregates and the reduced attractive marised here again to give an overall view of the mechanical per-
interparticle forces. formances of the materials.
Analysing the AMEO-modified silica particles distribution, the Tensile tests were performed on dog-bone specimens
overall dispersion was comparable to the other two systems. How- (L0 = 20 mm, thickness>=2 mm) using a Zwick Z010 universal test-
ever, certain larger clusters and agglomerates were observed, ing machine according to DIN EN ISO 527 with a crosshead speed
exhibiting diameters of several hundreds of nanometres, disturb- of 1 mm/min. The elongation of the specimen was measured using
ing the generally fine state of dispersion (Fig. 3). This phenomena a long distance encoder with a gauge length of 25 mm (Zwick Multi
may be related to the fact that during calandering AMEO-modified Sense).
silica particles, which have the NH2 groups on their surface, started Tensile properties are shown in Figs. 5–7. The addition of the fil-
cross-linking with the epoxy matrix due to the high pressure in the ler to the epoxy resin does not significantly affect the strength of
calander gap. Due to this effect, a layer of crosslinked molecules the material. Unmodified fumed silica particles slightly decrease

Fig. 4. SEM analysis on the tensile fracture surface of a 0.5% unmodified fumed
Fig. 2. TEM analysis on the 0.5% GLYMO-modified fumed silica nanocomposite. silica specimen.
1854 M. Battistella et al. / Composites: Part A 39 (2008) 1851–1858

UNMOD. FS The strain to failure of the three types of nanocomposites shows


FS-AMEO an improvement, which is highest for the AMEO-modified silica
2800 FS-GLYMO particles, with an increase of 20% compared to the neat epoxy. This
Young's Modulus [MPa]

also demonstrates the capability of the nanofiller to increase


2600
toughness at low volume fractions.
2400 The fracture toughness test was performed on CT specimens
(width 33 mm and thickness 5 mm) according to ASTM 5045/99
2200 by using an MTS 858 testing machine. The pre-crack was intro-
duced by manual tapping with a razor blade and the crosshead
2000 speed was set at 10 mm/min.
The results of the fracture toughness test are shown in Fig. 8.
1800
The main effect of the nanofiller addition is the improvement of
1600 the fracture toughness for all the nanocomposites produced. In
particular, GLYMO-modified silica particles at 0.5 vol.% show a KIC
0 of 1.34 MPa m0.5, which means an increase of the 54% compared
0.1 0.3 0.5
to the neat epoxy (KIC = 0.87 MPa m0.5). It is worth noting that mor-
Volume Content [%] phological analyses indicated the GLYMO-modified silica particles
as those with the finest and most uniform dispersion also with a
Fig. 5. Tensile test: Young’s modulus for all the nanocomposites. (Values for the
neat epoxy are given by the dashed lines) strong adhesion with the matrix.

2.4. Fatigue results


UNMOD. FS
70 FS-AMEO Fatigue tests were performed using an MTS 858 testing ma-
FS-GLYMO chine, according to procedures reported for metals in the ASTM E
65 647-00. CT specimens (width 33 mm and thickness 5 mm) were
pre-cracked by tapping using a razor blade. The length of the
60 pre-crack was about 2 mm. Then a sine wave cyclic load with
UTS [MPa]

R = 0.1 and f = 15 Hz was applied. The crack tip area was magnified
55 by a travelling microscope connected to a digital camera. Through
a specific acquisition software the image of the crack could be sent
50 from the digital camera to a computer and displayed on the mon-
itor. The crack length was measured with an accuracy of about
45 0.01 mm. The crack growth was regularly detected and the crack
length as a function of the number of cycles was reported. With
40 this acquisition method, regular and smooth crack length vs. num-
0
0.1 0.3 0.5 ber of cycles curves can be produced. The a-N plots exhibited an
Volume Content [%] exponential increase, with the crack growing very slowly in the
first thousands of cycles and then growing faster and faster. A typ-
Fig. 6. Tensile test: tensile strength for all the nanocomposites. (Values for the neat ical a-N curve is shown in Fig. 9 for a 0.3 vol.% AMEO-fumed silica
epoxy are given by the dashed lines) specimen.
The incremental polynomial method was implemented accord-
ing to ASTM E 647-00 and used to evaluate the crack growth rate
UNMOD. FS da/dN. Figs. 10–17 show the Paris curves of each nanocomposite
FS-AMEO compared to that of the reference epoxy.
7 FS-GLYMO In some cases at the DK value chosen for the beginning of the
Strain to failure ε [%]

test, the crack was already in the Paris regime, where the crack
6 growth rate linearly increases with DK. Thus, the crack propagation

UNMOD. FS
5 FS-AMEO
1.4
FS-GLYMO
4 1.2
KIc [MPa*m^0.5]

1.0
3 0.8
0.0
0.1 0.3 0.5
0.6
Volume Content [%]
0.4
Fig. 7. Tensile test: strain to failure for all the nanocomposites. (Values for the neat
epoxy are given by the dashed lines) 0.2
0.0
0.1 0.3 0.5
the strength, while modified fumed silica particles lead to re- Volume fraction [%]
stricted benefits to the tensile parameters. The strength decrease
can be explained by the tendency of the filler to clustering which Fig. 8. Fracture toughness test: KIC for all the nanocomposites. (Values for the neat
leads to weaken the matrix. epoxy are given by the dashed lines)
M. Battistella et al. / Composites: Part A 39 (2008) 1851–1858 1855

-2
22 0.3vol% AMEO-modified FS sp.1 10
AMEO mod. FS - 0.1%
20
crack length, a [mm]

-3

da/dN [mm/cycle]
10
18
16 10
-4

14
-5
10
12
10 10
-6

8
0 150000 300000 450000 600000 10
-7

No. cycles 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
0.5
Δ K [MPa*m ]
Fig. 9. Crack length vs fatigue life for a 0.3% AMEO-fumed silica specimen.
Fig. 12. Paris curves for 0.1% AMEO-modified fumed silica nanocomposite and neat
epoxy. (Yellow symbols and dotted lines are for the neat epoxy, red symbols and
-2 lines for the nanocomposite).
10
UNMOD. FS - 0.3%
-3
10
da/dN [mm/cycle]

-2
10
AMEO mod. FS - 0.3%
-4
10 10
-3

da/dN [mm/cycle]
-5 -4
10 10

-6 -5
10 10

-7 -6
10 10
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1
0.5
Δ K [MPa*m ] -7
10
Fig. 10. Paris curves for 0.3% unmodified fumed silica nanocomposite and of the 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
0.5
neat matrix. (Yellow symbols and dotted lines are for the neat epoxy, blue symbols Δ K [MPa*m ]
and lines for the nanocomposite).
Fig. 13. Paris curves for 0.3% AMEO-modified fumed silica nanocomposite and neat
epoxy. (Yellow symbols and dotted lines are for the neat epoxy, red symbols and
-2
lines for the nanocomposite).
10
UNMOD. FS - 0.5%
-3
10 -2
da/dN [mm/cycle]

10
AMEO mod. FS - 0.5%
-4 -3
10 10
da/dN [mm/cycle]

-5 -4
10 10

-6 -5
10 10

-7 -6
10 10
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 .9 1 1.1
0.5
Δ K [MPa*m ] -7
10
Fig. 11. Paris curves for 0.5% unmodified fumed silica nanocomposite and neat 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
0.5
epoxy. (Yellow symbols and dotted lines are for the neat epoxy, blue symbols and Δ K [MPa*m ]
lines for the nanocomposite).
Fig. 14. Paris curves for 0.5% AMEO-modified fumed silica nanocomposite and neat
epoxy. (Yellow symbols and dotted lines are for the neat epoxy, red symbols and
threshold could not always be clearly detected. A more accurate lines for the nanocomposite).
procedure to determine the DKth could be a DK-decreasing test.
However, the testing machine software needs a crack opening dis-
placement input in order to perform a DK-controlled test and the the COD device could damage the specimens before applying the
available COD device exerted a too high force in the notch of the testing load. Thus, a DK-decreasing approach could not be applied
small CT specimens used in this work. There was the risk that in this case.
1856 M. Battistella et al. / Composites: Part A 39 (2008) 1851–1858

10
-2 AMEO-modified fumed silica nanocomposite (Fig. 14). Moreover,
GLYMO mod. FS - 0.1% for some specimens an increase in the crack propagation threshold
-3
was observed, even if in the presence of a large scatter in the
da/dN [mm/cycle]

10 threshold values. Indeed, both in the 0.5% AMEO-modified


(Fig. 14) and in the 0.3% GLYMO-modified (Fig. 16) nanocomposites
10
-4
one specimen seemed to suggest a high crack propagation thresh-
old (around DK = 0.4 MPa m1/2) , while another specimen of the
-5 same material showed propagation also at lower DK values.
10
However, the improvement in the crack propagation threshold
is clearly related to the increase in fracture toughness observed
-6
10 for all the nanocomposites. A possible explanation for the higher
crack propagation resistance of the AMEO- modified composites
-7 could be the following: the presence of slightly larger agglomerates
10
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 with respect to the more finely dispersed GLYMO particles could
0.5
Δ K [MPa*m ] induce a more significant crack path deflection, thus resulting in
a higher resistance to crack propagation.
Fig. 15. Paris curves for 0.1% GLYMO-modified fumed silica nanocomposite and This tendency towards an improvement in the crack propaga-
neat epoxy. (Yellow symbols and dotted lines are for the neat epoxy, green symbols tion resistance is promising since it suggests that a further increase
and lines for the nanocomposite).
in filler content can lead to more significant enhancements in the
fatigue behaviour.
-2
The fracture surfaces of some nanocomposites were observed
10 through SEM and typical features like river-lines were detected
GLYMO mod. FS - 0.3%
in the region where the crack propagated under fatigue. As a rep-
-3
resentative example, the fracture surface of a specimen of the
da/dN [mm/cycle]

10
0.5% AMEO-modified fumed silica nanocomposite is displayed in
-4 Fig. 18. On the upper part of the SEM image the small smooth area
10
corresponds to the pre-cracking, while river-lines are visible on the
whole fatigue crack propagation area. The fracture surface be-
-5
10 comes again smooth where the DK approached to the KIC value
and the crack propagated statically.
-6
10

-7
10
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1
0.5
Δ K [MPa*m ]
Fig. 16. Paris curves for 0.3% GLYMO-modified fumed silica nanocomposite and
neat epoxy. (Yellow symbols and dotted lines are for the neat epoxy, green symbols
and lines for the nanocomposite).

-2
10
GLYMO mod. FS - 0.5%
-3
10
da/dN [mm/cycle]

-4
10

-5
10

-6
10

-7
10
0.2 0.4 0.6 0.8 1 1.2 1.4
0.5
ΔK [MPa*m ]
Fig. 17. Paris curves for 0.5% GLYMO-modified fumed silica nanocomposite and
neat epoxy. (Yellow symbols and dotted lines are for the neat epoxy, green symbols
and lines for the nanocomposite).

From Figs. 10–17 it can be seen that for the low filler contents
(0.1 and 0.3%) the nanomodification seems to have almost no effect
on the fatigue behaviour, but increasing the fumed silica loading, Fig. 18. SEM view of the fatigue fracture surface for a 0.5% AMEO-fumed silica
some improvements become apparent, particularly for the 0.5% specimen.
M. Battistella et al. / Composites: Part A 39 (2008) 1851–1858 1857

2.5. Discussion 2.6. Conclusions

The addition of very low volume contents of fumed silica parti- Three different surface modification conditions have been
cles led to significant improvements in the fracture toughness tested for fumed silica/epoxy nanocomposites. The morphology
behaviour of epoxy as already shown in [23]. On the other hand, analysis with TEM and SEM showed that the best dispersion was
for cyclic loading some enhancements were achieved, but not as achieved with GLYMO-modified nanoparticles, while AMEO-modi-
high as in the static case. This fact could be a hint for further fied and untreated fumed silica were slightly more agglomerated.
increasing the filler contents in order to follow the positive trend The strain to failure was generally improved through the
of the fatigue behaviour and to achieve more appreciable nanomodification with enhancements up to 20%, whereas the ulti-
improvements. mate tensile strength was not significantly affected by the fumed
Wetzel et al. [5] measured the fatigue performance of alumina- silica particles.
epoxy nanocomposites and carried out an interesting investigation The most important improvements were achieved in fracture
on the possible mechanisms leading to an improved fracture toughness, up to an increase of 54% at 0.5 vol.% of AMEO-fumed sil-
toughness as well as to a higher fatigue resistance. Nanoparticles ica, thus indicating a strong effect on crack propagation resistance
were believed to induce toughening mechanisms such as crack already at the low filler contents of the present work. In the fatigue
deflection, plastic deformation and crack pinning. test, some enhancements were visible only at 0.5 vol.%, but not so
Nevertheless, the contrasting static and dynamic behaviour significant. The different behaviour in static and dynamic loading
observed in the present work seems to suggest that different seems to suggest that different toughening mechanisms could act
crack resistance mechanisms should act in the two cases. Thus, in the two cases.
different dispersion levels could be the most appropriate to im- Further developments will include a deeper investigation on the
prove fatigue behaviour or static fracture toughness. This idea is toughening mechanisms in nanocomposites and a shift towards
not supported in the literature by experimental evidences for higher filler contents to achieve more appreciable improvements
spherical nanoparticles, but a similar concept is proposed in the also in the fatigue performance.
field of clay nanocomposites, where an intercalated structure is
suggested to be in some cases preferable to completely exfoliated
Acknowledgements
clays.
Miyagawa et al. [25] studied biobased epoxy/clay composites
Part of the activity reported was carried out in the frame of the
and found that intercalated nanocomposites led to higher fracture project A methodology for the integrated design and development
toughness than exfoliated clay. Single nanoplatelets did not seem
of nanocomposite products CPDA055157 funded by University of
to be strong enough to prevent crack from propagating, while Padova, Marino Quaresimin warmly acknowledge the financial
the aggregates of intercalated clays deflected crack on the micro-
support.
metre scale and led to a rougher fracture surface and to a higher
critical strain energy release rate. Crack deflection at agglomerates
of intercalated clays in the micron range is proposed as main References
toughening mechanism also in [26], where improvements in
[1] Ng CB, Schadler LS, Siegel RW. Synthesis and mechanical properties of TiO2 -
fracture toughness up to almost 60% were detected in MMT-epoxy epoxy nanocomposites. NanoStuct Mater 1999;12:507–10.
nanocomposites. An exfoliated structure seems to favour high [2] Evora VMF, Shukla A. Fabrication, characterization and dynamic behavior of
modulus, whereas an intercalated structure leads to high polyester/TiO2 nanocomposites. Mater Sci Eng 2003;A361:358–66.
[3] Zhang M, Singh RP. Mechanical reinforcement of unsatured polyester by Al2 O3
toughness. nanoparticles. Mater Lett 2004;58:408–12.
Similar results were found in [27] with strongly intercalated [4] Wetzel B, Haupert F, Zhang MQ. Epoxy nanocomposites with high mechanical
organoclay aggregates inducing fracture toughness improvements and tribological performance. Compos Sci Technol 2003;63:2055–67.
[5] Wetzel B, Rosso P, Haupert F, Friedrich K. Epoxy nanocomposites – fracture and
of 55% with respect to neat epoxy. toughening mechanisms. Eng Fract Mech 2006;73:2375–98.
Blends of intercalated and exfoliated organoclay layers are re- [6] Adebahr T, Roscher C, Adam J. Reinforcing nanoparticles in reactive resin. Eur
ported [28] to lead to improvements both in tensile modulus (up Coatings J 2001;4:144–9.
[7] Lin JC, Chang LC, Nien MH, Ho HL. Mechanical behavior of various nanoparticle
to 20% of increase with 10% of organoclay loading), in fracture
filled composites at low- velocity impact. Compos Struct 2006;74(1):30–6.
toughness (up to 100%) and in another case [29] also in the fatigue [8] Ragosta G, Abbate M, Musto P, Scarinzi G, Mascia L. Epoxy–silica particulate
strength of organoclay/epoxy nanocomposites. nanocomposite: Chemical interaction, reinforcement and fracture toughness.
Polymer 2005;46:10506–16.
The fatigue behaviour of clay/polypropylene nanocomposites
[9] Zheng Y, Zheng Y, Ning R. Effects of nanoparticle SiO2 on the performance of
was investigated in [22] and an improvement of 13% in the fatigue nanocomposites. Mater Lett 2003;57:2940–4.
strength coefficient with respect to the neat PP was found. Failure [10] Moloney AC, Kausch HH, Kaiser T, Beer HR. Review. Parameters determining
was initiated at agglomerates of several nanoplatelets, which in- the strength and toughness of particulate filled epoxide resins. J Mater Sci
1987;22:381–93.
duced crazing on the polypropylene matrix and caused a large [11] Bandyopadhyay S. Review of the microscopic and macroscopic aspects of
number of dimples on the fracture surface. fracture of unmodified and modified epoxy resins. Mater Sci Eng
A hypothesis could be that also in the nanocomposites of the 1990;A125:157–84.
[12] Norman DA, Robertson RE. Rigid-particle toughening of polymers. Polymer
present study at some extent a slightly agglomerated structure 2003;44:2351–62.
could be preferable to improve some mechanical properties, partic- [13] Fiedler B, Gojny FH, Wichmann MHG, Nolte MCM, Schulte K. Fundamental
ularly the fatigue behaviour. In the fracture process under quasi aspects of nano-reinforced composites. Compos Sci Technol 2006;66:3115–25.
[14] Johnsen BB, Kinloch AJ, Mohammed RD, Taylor AC, Sprenger S. Toughening
static loading once the crack nucleates, it suddenly propagates, mechanisms of nanoparticle-modified epoxy polymers. Polymer
thus the resistance to crack nucleation is important and a fine dis- 2007;48:530–41.
persion on the nanoscale could be useful or even necessary to pre- [15] Wichmann MHG, Schulte K, Wagner HD. On nanocomposite toughness.
Compos Sci Technol 2008;68:329–31.
vent crack from initiating. On the other hand, in cyclic loading a
[16] Xie XL, Liu QX, Li RKY, Zhou XP, Zhang QX, Yu ZZ, et al. Rheological and
large part of the fatigue life can be spent to propagate the crack mechanical properties of PVC/ CaCO3 nanocomposite prepared by in situ
up to a critical length and therefore a slightly coarser dispersion polymerisation. Polymer 2004;45:6665–73.
[17] Lazzeri A, Zebarjad SM, Pracella M, Cavalier K, Rosa R. Filler toughening of plastic.
can provide a higher resistance to crack propagation. However, this
Part 1- The effect of surface interaction on physico-mechanical properties and
hypothesis should be confirmed by further experimental rheological behaviour of ultrafine CaCO3/HDPE nanocomposites. Polymer
investigations. 2005;46:827–44.
1858 M. Battistella et al. / Composites: Part A 39 (2008) 1851–1858

[18] Bartczak Z, Argon AS, Cohen RE, Weinberg M. Toughness mechanism in semi- [24] Wichmann MHG, Cascione M, Fiedler B, Quaresimin M, Schulte K. Influence of
crystalline polymer blends: II. High-density polyethylene toughened with surface treatment on mechanical behaviour of fumed silica/epoxy resin
calcium carbonate filler particles. Polymer 1999;40:2347–65. nanocomposites. Compos Interf 2006;13:699–715.
[19] Yang JL, Zhang Z, Zhang H. The essential work of fracture of polyamide 66 filled [25] Miyagawa H, Jurek RJ, Mohanty AK, Misra M, Drzal LT. Biobased
with TiO2 nanoparticles. Compos Sci Technol 2005;65(15–16):2374–9. epoxy/clay nanocomposites as a new matrix for CFRP. Compos A
[20] Chisholm N, Mahfuz H, Rangari VK, Ashfaq A, Jeelani S. Fabrication and 2006;37:54–62.
mechanical characterization of carbon/SiC–epoxy nanocomposites. Compos [26] Qi B, Zhang QX, Bannister M, Mai YW. Investigation of the mechanical
Struct 2005;67(1):115–24. properties of DGEBA-based epoxy resin with nanoclay additives. Compos
[21] Kinloch AJ, Masania K, Taylor AC, Sprenger S, Egan D. The fracture of glass– Struct 2006;75:514–9.
fibre-reinforced epoxy composites using nanoparticle-modified matrices. J [27] Le Pluart L, Duchet J. Sautereau H Epoxy/montmorillonite nanocomposites:
Mater Sci 2008;43:1151–4. influence of organophilic treatment on reactivity, morphology and fracture
[22] Zhou Y, Rangari V, Mahfuza H, Jeelani S, Mallick PK. Experimental study on properties. Polymer 2005;46:12267–78.
thermal and mechanical behavior of polypropylene, talc/polypropylene and [28] Becker O, Varley R, Simon G. Morphology, thermal relaxations and mechanical
polypropylene/clay nanocomposites. Mater Sci Eng A 2005;402:109–17. properties of layered silicate nanocomposites based upon high-functionality
[23] Blackman BRK, Kinloch AJ, Sohn Lee J, Taylor AC, Agarwal R, Schueneman G, epoxy resins. Polymer 2002;43:4365–73.
et al. The fracture and fatigue behaviour of nano-modified epoxy polymers. J [29] Juwono A, Edward G. Fatigue performance of clay/epoxy nanocomposites. Int J
Mater Sci 2007;42:7049–51. Nanosci 2005;4(4):501–7.

You might also like