COMPOSITES

SCIENCE AND
TECHNOLOGY
Composites Science and Technology 65 (2005) 635645
www.elsevier.com/locate/compscitech

Silica nanoparticles lled polypropylene: eects of particle

surface treatment, matrix ductility and particle species on
mechanical performance of the composites
Chun Lei Wu a, Ming Qiu Zhang
a

b,*

, Min Zhi Rong b, Klaus Friedrich

Key Laboratory for Polymeric Composite and Functional Materials of Ministry of Education, Zhongshan University, Guangzhou 510275, PR China
b
Materials Science Institute, Zhongshan University, Guangzhou 510275, PR China
c
Institute for Composite Materials (IVW), University of Kaiserslautern, D-67663 Kaiserslautern, Germany
Received 9 November 2003; received in revised form 14 July 2004; accepted 17 September 2004
Available online 5 November 2004

Abstract
The current paper is a continuation of the authors work on mechanical performance of nano-silica/polypropylene (PP) composites. Unlike the fumed nano-silica used in the previous studies, precipitated nano-silica is employed in the present investigation. The
results indicate that graft polymerization onto the precipitated nano-silica (that has been successfully applied to the surface modication of fumed nano-silica) is still an eective method to pre-treat the particles, which leads to an overall improvement of the
composites properties. In addition to the grafting polymers covalently attached to the nanoparticles, matrix ductility and nanoparticles size are important factors that inuence the extent of performance enhancement of the composites. In the case of suitable combination of these factors, the positive eect of the nanoparticles can be maximized.

2004 Elsevier Ltd. All rights reserved.
Keywords: A. Particle-reinforced composites; B. Mechanical properties; B. Surface treatments; Nanoparticles

1. Introduction
In recent years, inorganic nanoparticles lled polymer
composites have received increasing research interests of
materials scientists because the ller/matrix interface in
these composites might constitute a much greater area
and hence inuence the composites properties to a
much greater extent at rather low ller concentration
as compared to conventional micro-particulate composites. Considering the versatility of production facilities
and raw materials, dispersive mixing of ready-made
nanoparticles and polymers is still one of the main manufacturing methods to make nanocomposites. The results of a series of thermoplastics based composites
*

Correspondence author. Tel.: +86 208 411 2283/403 6576; fax:

+86 208 403 6576.
E-mail address: ceszmq@zsu.edu.cn (M.Q. Zhang).
0266-3538/$ - see front matter
2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compscitech.2004.09.004

prepared in this way have been reported, like nanoSiO2/polypropylene (PP) [1,2], nano-SiO2/high-density
polyethylene (HDPE) [3], nano-CaCO3/PP [46], nanoSiO2/poly(ethylene terephthalate) (PET) [7], nanoTiO2/polystyrene (PS) [8], nano-SiO2/acrylic latex [9],
and nano-SiO2/polyethersulfone (PES) [10].
It is worth noting that the market available nanoparticles generally take the form of agglomerates, which are
hard to be broken apart during compounding due to the
strong interaction among the nanoparticles, the limited
shear force provided by the mixing device and the high
melt viscosity of polymer melts. Modication with coupling agents, which can only react with the exterior
nanoparticles of the agglomerates as restricted by the
larger molecules, is helpless to well disperse the nanoparticles. In some cases, as a result, the composites with
the addition of nanoparticles would exhibit properties
worse than microcomposites. To bring the eect of the

636

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

nanoparticles into play, graft polymerization onto the

particulates surface was developed by the authors as a
pre-treatment technique [13,5,6]. The low molecular
weights of the grafting monomers allow them to penetrate into the agglomerated nanoparticles and react with
the particles both inside and outside the agglomerates.
Taking the advantage of this, the following benets
can be gained: (i) the hydrophilic particles surfaces are
converted into hydrophobic in favor of improving miscibility between the components; (ii) the loosened nanoparticles agglomerates are turned into compact
nanocomposite structure consisting of the particles, the
grafting polymers and the homopolymers generated in
the course of graft polymerization; and (iii) the interfacial interaction between the ller particles and the surrounding matrix is enhanced through entanglement of
the grafting polymers attached to the nanoparticles with
the matrix molecules. Therefore, stress can be transferred to all the nanoparticles when the composites are
subjected to applied force, while stiening, reinforcing
and toughening eects are observed at very low nanoparticle content. It is believed that a double percolation
of eective stress volumes takes the responsibility for the
overall enhancement of the composites [11].
In our previous studies [13,6], fumed nano-silica
was employed as the predominant ller particles and
the feasibility of graft pre-treatment of these nanoparticles for acquiring mechanical properties improvement
has been investigated by using lab-scale and industrial
scale compounding machines, respectively. Besides, surface morphologies of the nanoparticles before and after
grafting were also characterized [12]. For conducting
systematic researches, precipitated nano-silica, which
is synthesized by a process dierent from the one for
making fumed silica, is used in the present work. On
the basis of this ller selection, some other important
factors that have not yet been reported in the literature, like the eects of matrix ductility and particulate
size on the composites mechanical performance, are
studied hereinafter. To maintain the continuity of our
work on this subject, PP acts as the matrix polymer
once more.
Precipitated silica is manufactured by a wet procedure by treating silicates with mineral acids to obtain
ne hydrated silica particles in the course of precipitation. The reaction and drying conditions determine the
porosity, surface area, surface chemistry and the degree
of impurities in the precipitated silica. In general, precipitated silicas are cheap and have a particle size higher
than 10 lm [13]. However, they can be made as tiny as
nano-scale under specic circumstances [14]. Fumed silicas are manufactured by high-temperature hydrolysis of
silicon tetrachloride in a ame. Silanol and siloxane
groups are created on the silica surface, leading to
hydrophilic nature of the particles. The use of fumed silica as llers in thermoplastics has been well documented

not only by our own works stated above but also by

other groups [7,9,10]. It is thus expected that the applicability of nano-silica in thermoplastics modication
would be further broadened if precipitated silicas prove
to be as useful as fumed ones.

2. Experimental
2.1. Materials
The precipitated nano-SiO2 with an average primary
particle size of 10 nm and a specic surface area of
640 m2/g was supplied by Zhoushan Nanomaterials
Co., China. For purposes of comparative study, fumed
nano-SiO2 with an average primary particle size of
15 nm and a specic surface area of 374 m2/g was
purchased from Shenyang Chemical Engineering Ltd.,
China. The two types of silicas are denoted by p-SiO2
and f-SiO2 for the convenience of discussion in the following text, respectively.
Isotactic polypropylene (PP) homopolymer T30S,
supplied by Qilu Petrochemical Industrial Co., China,
was used as the matrix polymer. It has a melt ow index
(MFI) of 3.2 g/10 min (2.16 kg at 230 C). To reveal the
eect of matrix ductility, however, other two types of PP
were also introduced. They are isotactic PP homopolymer PP700 (MFI = 13 g/10 min) produced by Guangzhou Petroleum Chemical Co., China, and block
copolymerized PP EPS30R (consisting of the segments
of ethylene and propylene, MFI = 1.9 g/10 min) by Qilu
Petrochemical Industrial Co., China.
For carrying out graft polymerization onto the
nano-silica particles, various commercial monomers:
styrene, methyl methacrylate, ethyl acrylate and butyl
acrylate, were used as grafting monomers without further purication, respectively. It is known that the
macromolecular chains constructed by these monomers
have rigidities ranking in the order of their appearance
as written above. The later three monomers have the
same backbones but dierent lengths of the side chains,
which might help to understand the interfacial eect
easily.
2.2. Pre-treatment of the nanoparticles through graft
polymerization and the related analysis
Modication of nano-silica proceeded according to
the following steps called gaseous graft polymerization,
which is dierent from the method used previously [1].
The nanoparticles were pre-treated at 140 C under vacuum for 5 h to eliminate possible absorbed water on the
surface of the particles. Then they were lled into a glass
vessel and absorb certain amount of monomer vapor
under vacuum. Afterwards, the sealed vessel containing
the nanoparticles was irradiated by 60Co c-ray under

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

637

atmosphere at room temperature. After exposure to a

dose of 4 Mrad, the powder was available for the subsequent compounding.
To evaluate the results of grafting and to characterize
the grafted nanoparticles, the grafting polymer and the
homopolymer, which were generated during the irradiation polymerization of the monomers, should be separated. For this purpose, a certain amount of the
irradiation products were extracted by benzene in a
Soxhlet apparatus for 36 h. In this way the homopolymer was isolated. The residual material was then dried
in vacuum at 80 C until a constant weight was reached.
By using a Shimadzu TA-50 thermogravimetre (TG),
the weight of the grafting polymer on the modied
nanoparticles was determined and the percent grafting
can be calculated accordingly. To further separate the
grafting polymer from the treated nanoparticles, nanosilica accompanied with the unextractable grafting
polymer was immersed in 20% HF solution for 72 h to
remove the inorganic particles. Table 1 lists the parameters quantifying the grafting reaction on the nanoparticles for reference.
A Micromeritics ASAP 2100 surface area analyzer
was employed to measure the micropore volumes and
specic surface areas of the nano-silica before and after
the graft treatment. Prior to the tests, the samples were
dried in vacuum under 150 C for 24 h.

Charpy impact bars (GB/T 1043-93) with a CJ150MZ

injection-molding machine at 215 C.
Room temperature tensile testing of the composites
was conducted on a Hounseld-5KN universal testing
machine. Unless otherwise specied, the crosshead speed
was set at 50 mm/min. Five samples were tested for each
case. A Hitachi S-520 scanning electron microscope
(SEM) was used to observe the fractured surfaces. Unnotched Charpy impact strengths of the composites
were measured by an XJJ-50 impact tester. Eight samples were tested for each case.
The isothermal crystallization of PP and its composites was conducted on a PerkinElmer dierential scanning calorimetry (DSC-7) instrument. The samples were
heated from room temperature to 210 C at a rate of
10 C/min in N2, and kept at 210 C for 5 min. Then,
the samples were cooled down to 130 C at a rate of
80 C/min and started to isothermally crystallize. In
the meantime, the exothermic curves were collected.
When the crystallization is completed, the samples were
heated again to 210 C at a rate of 10 C/min and the
melting behaviors were recorded.

2.3. Composites preparation and characterization

As the current paper is focused on the mechanical

performance of the PP composites lled with nanosilica, the chemistry related to the graft polymerization
on the particles surface is not discussed in detail.
Instead, the investigation results in this aspect [15] are
summarized below, providing knowledge basis for the
composites analysis. According to the infrared and
X-ray photoelectron spectroscopy studies, it is proved
that after irradiation the grafting polymers are chemically attached to silica surface through SiOC and
SiC bonds as expected. The weight average molecular

The nanoparticles were rstly compounded with PP

(1:2 by weight) using an X(S)R-160 two-roll mill at
195 C to produce composite masterbatch. Then, the
masterbatch was mixed with neat PP to dilute the ller
loading to desired values through an SHJN-25 twinscrew extruder at 210230 C. The rotation speed of
the extruder was set to 180 rpm. Finally, the resultant
pellets were molded into dog-bone-shaped tensile bars
(ASTM D638-97 Type IV specimen) and rectangular

3. Results and discussion

3.1. Eect of surface grafting onto the nanoparticles

Table 1
Irradiation induced graft polymerization of vinyl monomers onto precipitated nano-silica particlesa
Samples

Percent graftingb (%)

Grafting eciencyc (%)

Monomer conversiond (%)

p-SiO2g-PSe
p-SiO2g-PMMAf
p-SiO2g-PEAg
p-SiO2g-PBAh

11.4
12.7
15.3
13.4

57.1
63.3
76.5
66.9

100
100
100
100

a
b
c
d
e
f
g
h

Irradiation dose: 4 Mrad; monomer/silica = 20 wt%.

Percent grafting = weight of grafting polymer/weight of silica.
Grafting eciency = weight of grafting polymer/weight of grafting polymer and homopolymer.
Monomer conversion = weight of polymer/weight of monomer.
PS: polystyrene.
PMMA: polymethyl methacrylate.
PEA: polyethyl acrylate.
PBA: polybutyl acrylate.

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

280

210

1: p-SiO2 as-received
2: p-SiO2-g-PS
3: p-SiO2-g-PEA
Adsorption curve
Desorption curve

Differential pore volume [cm3/g]

Adsorbed volume[cm3/g]

350

140

70
3

0
0.0

0.2

0.4

0.6

0.8

1.0

Relative pressure
Fig. 1. Nitrogen absorption isotherms of nano-silica particles.

weights of the grafting polymers are about 105 and those

of the homopolymers are 104.
To have an image of the surface structure variation,
nitrogen absorption isotherms of the nanoparticles were
measured (Fig. 1). The curve proles are something between type II and IV isotherms as viewed from Brunauer
classication system [16]. Although hysteresis loops can
be seen on the isotherms of the nanoparticles with and
without the grafting polymers, those are not closed in
the case of p-SiO2g-PS and p-SiO2g-PEA. It means
that the capillary condensation occurring under higher
relative pressure is severer in the grafted nano-silica.
In comparison with p-SiO2 as-received, both p-SiO2gPS and p-SiO2g-PEA have signicantly reduced
absorbance. Evidently, it can be attributed to the fact
that the micropores and macropores of the nanoparticle
agglomerates are lled by the grafting polymers, which
in turn demonstrates that the grafting polymers have
been planted onto the nanoparticles. The plots showing
the micropore size distribution in Fig. 2 also provide
supporting evidence for the estimation. After grafting
treatment, the pore volumes of the nanoparticles are
about ten times smaller than that of the untreated

0.015

0.15
p-SiO2 as-received
p-SiO2-g-PS
p-SiO2-g-PEA

0.12

0.012

0.09

0.009

0.06

0.006

0.03

0.003

0.00
0.0

0.5

1.0

1.5

2.0

0.000
2.5

3
Differential pore volume [cm /g]

638

Pore diameter [nm]

Fig. 2. Micropore size distribution of nano-silica particles.

nano-silica. In addition, the pore diameters corresponding to the maximum pore volumes of the treated nanoparticles shift from 0.55 to 0.71 nm. It seems that the
graft polymerization was initiated at the smallest micropores of the nanoparticles. The porous structure of the
particles has been changed accordingly.
Figs. 3 and 4 show the typical fracture surfaces and
mechanical properties of PP composites with nano-silica, respectively. The untreated nanoparticles are severely agglomerated in PP matrix (Fig. 3(a)), while the
treated ones are well separated into tiny aggregates
(Fig. 3(b)). The SEM observations evidence the authors
expectation of graft treatment of the nanoparticles. It
means that the composites containing the untreated
nanoparticles are provided with heterogeneous microstructure and would exhibit worse reinforcing and
toughening eects as compared to those having more
homogeneous appearances due to graft pre-treament
of the particles.
As for the details of the composites mechanical performance, it is seen that both treated and untreated
nano-silicas are able to stien the matrix, as reected
by the proportional relationship between Youngs modulus and the ller loading (Fig. 4(a)). Comparatively, the
stiness of the composites with grafted nano-silica is

Fig. 3. SEM microphotos of tensile fracture surfaces of nano-silica/PP composites (PP: T30S). (a) p-SiO2 as-received/PP (nano-silica content = 2.74
vol%); (b) p-SiO2g-PS/PP (nano-silica content = 2.75 vol%).

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

1.8

p-SiO2 as-received/PP
p-SiO2-g-PS/PP
p-SiO2-g-PMMA/PP
p-SiO-g-PEA/PP
2
p-SiO2-g-PBA/PP

(c) 400
Elongation to break [%]

Young's modulus [GPa]

(a) 2.0

1.6

1.4

1.2
0.0

0.5

1.0

1.5

2.0

300

200
p-SiO2 as-received/PP
p-SiO2-g-PS/PP
p-SiO2-g-PMMA/PP
p-SiO-g-PEA/PP
2
p-SiO2-g-PBA/PP

100

0
0.0

2.5

0.5

39

38

37

36

0.5

1.0

1.5

2.0

(d)
Area under stressstrain curve [MPa]

Tensile strength [MPa]

p-SiO2 as-received/PP
p-SiO2-g-PS/PP
p-SiO2-g-PMMA/PP
p-SiO-g-PEA/PP
2
p-SiO2-g-PBA/PP

40

35
0.0

2.0

2.5

90

60
p-SiO2 as-received/PP
p-SiO2-g-PS/PP
p-SiO2-g-PMMA/PP
p-SiO-g-PEA/PP
2
p-SiO2-g-PBA/PP

30

0
0.0

2.5

0.5

1.0

1.5

2.0

2.5

SiO2 [vol%]
p-SiO2 as-received/PP
p-SiO2-g-PS/PP
p-SiO2-g-PMMA/PP
p-SiO-g-PEA/PP
2
p-SiO2-g-PBA/PP

125

Impact strength [kJ/m ]

1.5

120

SiO2 [vol%]
(e)

1.0

SiO2 [vol%]

SiO2 [vol%]

(b)

639

100

75

50

25

0
0.0

0.5

1.0

1.5

2.0

2.5

SiO2 [vol%]
Fig. 4. Mechanical properties of nano-silica/PP composites as a function of silica content (PP: T30S): (a) Youngs modulus; (b) tensile strength; (c)
elongation at break; (d) area under tensile stressstrain curve; and (e) impact strength.

lower than that of the untreated nanoparticles composites. As interfacial stress transfer eciency depends on
the stiness of the interphase, higher interfacial stiness
favors improvement of the composites modulus [17].
The grafting polymer adhering to the nano-silica as well
as the surrounding homopolymer establish a compliant
interlayer between the particles and the matrix, and
hence decrease the stiening eect of the particles. On
the other hand, it is interesting to note that the moduli
of the composites with grafted nano-silica at identical
particle content are arranged in the following order:
p-SiO 2 g-PS/PP > p-SiO 2 g-PMMA/PP > p-SiO 2 g-

PEA/PP > p-SiO2g-PBA/PP. It exactly coincides with

the exibilities of the grafting polymer chains. That is,
PBA possesses the highest exibility and consequently
masks the high stiness of the particles in the composites
to the greatest extent. The phenomenon again proves the
above analysis concerning the dependence of composites
moduli on interfacial stiness.
With respect to the composites under higher nano-silica content, it is found that the moduli increment of the
composites with treated nano-silica can no longer keep
the linearity. For p-SiO2g-PEA/PP and p-SiO2
g-PBA/PP composites, their moduli even decline. This

640

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

should be the result of the increasing amount of the soft

interphase comprised of the grafting polymers and
homoploymers with a rise in the content of the grafted
nanoparticles.
Besides stiening eect, the precipitated nano-silica
can also provide PP with reinforcing eect at rather
low ller concentration (Fig. 4(b)), which is similar to
the situation of fumed nano-silica [1,2]. In the case of
p-SiO2g-PS/PP and p-SiO2g-PBA/PP, for example,
tensile strengths of the composites are higher than that
of neat PP. This is dierent from what is observed in
conventional micrometer particles/polymer composites,
i.e., tensile strength of the composites remarkably decreases with the addition of the particulate llers due
to the poor bonding at the interface [18,19]. Jancar
et al. [20] suggested that a strong ller/matrix adhesion
would lead to enhanced strength of particulate composites. According to their consideration, it is known that
the improvement of tensile strength of the composites
exhibited in Fig. 4(b) should also be interpreted as the
improvement of the interfacial interaction. Especially
when PS- and PBA-grafted nano-silica is incorporated,
the chain entanglement between the grafting polymers
and the matrix polymer guarantees eective interfacial
bonding over the whole ller content range of interests.
However, grafting treatment does not always take eect
as revealed by the composition dependence of tensile
strength of p-SiO2g-PMMA/PP and p-SiO2g-PEA/
PP. This contradicts the results of PP composites with
fumed nano-silica grafted by the same species of polymers [1]. It means that in addition to the nature of the
grafting polymers on the nanoparticles, species of nanoparticles themselves and other unknown factors also
greatly inuence the reinforcing eect of the grafted
nanoparticles. More detailed study in this direction is required to have a reasonable conclusion.
From engineering point of view, elongation-to-break
is an important parameter describing the rupture behavior of composite materials. The addition of mineral particulates into polymers used to lower it, even though the
matrix has high impact toughness [21]. Fig. 4(c) clearly
indicates that this is not the case when nano-silica is
used. Either untreated or treated nano-silica is able to
increase elongation-to-break of PP. Compared to the
untreated and PS-grafted ones, acrylic polymers-grafted
nano-silicas are more eective, in particular when
p-SiO2g-PMMA is concerned. It implies that in principle compliant grafting polymers can induce more matrix
polymer to be involved in plastic deformation. However,
chain exibility of the grafting polymers on the
nanoparticles is not the only prerequisite, otherwise pSiO2g-PBA/PP should have the highest value of elongation-to-break. In contrast to tensile strength of the
composites that needs strong ller/matrix bonding, the
micro-deformation mechanism involved in elongationto-break of the composites depends upon extensionality

of the nanoparticles agglomerates in the composites [22].

Balanced viscoelasticity of the grafting polymers on the
nanoparticles might thus be necessary for the improvement of elongation-to-break of the composites.
Fig. 4(d) shows the areas under tensile stressstrain
curves of the composites, another parameter characterizing the static toughness. The magnitude order of the
values at given ller content is similar to that in Fig.
4(c), suggesting that the deformation features of the
composites still govern their areas under tensile stress
strain curves. Grafted nano-silica again performs well
than the untreated nanoparticles. The maximum value
of p-SiO2g-PMMA/PP is about 3.8 times higher than
that of neat PP. The results demonstrate the role of
grafting polymers, i.e., interconnecting the nanoparticles
through chemical bonding and correlating the grafted
nanoparticles with the matrix by chain entanglement.
Under the applied force, plastic deformation of large
amount of the matrix polymer beside the grafted nanoparticles is induced, leading to signicantly high elongation-to-break and areas under tensile stressstrain curve
of the composites. In the case of untreated nanoparticles
lled PP, voiding and disintegration of the nanoparticle
agglomerates are the main ways of energy dissipation
due to the lack of extensionality.
Unnotched impact strengths of the composites are given in Fig. 4(e) as a function of silica fraction. It is clear
that after grafting treatment the toughening ability of
the nanoparticles is greatly increased. At the silica content of 0.45 vol%, the impact ductility of p-SiO2g-PBA/
PP is over 3 times higher than that of unlled PP, and all
the impact strengths of the composites at this ller loading are arranged in the order contrary to what observed
in Fig. 4(a) concerning Youngs modulus: p-SiO2/
PP < p-SiO2g-PS/PP < p-SiO2g-PMMA/PP < p-SiO2>
g-PEA/PP < p-SiO2g-PBA/PP. Since unnotched impact
strength reects the energy consumed by the plastic deformation prior to crack initiation, the above results manifest
that the exible macromolecular chains grafted onto the
nanoparticles surfaces must have made contribution to this
part of energy. In comparison to the data shown in
Fig. 4(c) and (d), it is known that the dynamic toughness
of the composites (i.e., impact strength) is more sensitive
to the dispersion status of the grafted nanoparticles than
the static toughness (i.e., elongation-to-break and area under tensile stressstrain curve). With a rise in the ller content, homogeneity of the particles distribution might be
worse as a result of the increased viscosity of the composite
systems. Under this circumstance, the grafting polymers
could not response in step with the impact load and hence
decline of impact strength is observed when the content of
nano-silica exceeds 0.45 vol%.
Since PP is a semi-crystalline polymer and its
mechanical properties would change with the crystalline
structure and crystallinity, the inuence of the addition
of the nanoparticles should be known. As listed in

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

641

Table 2
Kinetic parameters of isothermal crystallization of PP and its composites at 130 C
Samples

t1/2a (min)

tfb (min)

nc

kd ( 10 3min n)

DHce (J/g)

PP
p-SiO2 as-received/PP (nano-silica content = 0.86 vol%)
p-SiO2g-PS/PP (nano-silica content = 1.066 vol%)
p-SiO2g-PEA/PP (nano-silica content = 0.82 vol%)

8.68
4.69
6.17
9.79

17.8
8.89
11.39
17.7

2.59
3.04
3.04
2.92

2.57
6.31
2.75
1.08

97.1
97.0
95.0
97.0

a
b
c
d
e

t1/2: half-crystallization time.

tf: the time at which the crystallization is completed.
n: Avrami index.
k: rate constant of crystallization.
DH: enthalpy of crystallization.

Table 2, the untreated nano-silica remarkably accelerates the crystallization of PP matrix as revealed by the
values of t1/2, tf and k. This nucleating activity can be explained by the thermodynamic model proposed by
Ebengou [23]. That is, when PP chains were absorbed
on the silica surface, the congurational entropy of the
entire chain decreased, forming a nucleus of a certain
volume within the adsorbed chains costs less energy.
In the case of grafted nanoparticles, the nucleation effects are less profound because the grafting polymers
shielded the nanoparticles from the direct contacts with
PP, which coincides with the results of PP/elastomer/
particles composites characterized by core-shell microstructures [24]. Fig. 5 further shows the melting behaviors of PP and its composites. In comparison with neat
PP, the endothermic peak proles and temperatures of
the composites are almost the same, suggesting the crystal microstructure of PP has not been changed. Relatively, the crystallinity of PP in the composites is
slightly reduced. On the whole, however, it cannot be
concluded that the aforesaid mechanical performance
variations of PP composites with the incorporation of
silica nanoparticles results from the variations in PP
crystalline structure and crystallinity. Although un-

Endo >

Samples
1
2
3
4

Tm ( C)
166.5
166.4
166.5
166.1

Xc (%)
46.6
45.7
45.4
44.5

4
3
2

140

150

160

170

180

190

200

Temperature [oC]
Fig. 5. DSC heating traces of PP and its composites having been
isothermally crystallized at 130 C (PP: T30S). (1) Neat PP; (2) p-SiO2
as-received/PP (nano-silica content = 0.86 vol%); (3) p-SiO2g-PS/PP
(nano-silica content = 1.06 vol%); (4) p-SiO2g-PEA/PP (nano-silica
content = 0.82 vol%). Tm: peak melting temperature; Xc: crystallinity.

treated nano-silica has signicant nucleating ability, it

fails to improve PP properties equivalently. Similarly,
the grafted nanoparticles only lead to marginal decrease
of PP crystallinity, which is also out of proportion to the
performance enhancement.
3.2. Eects of matrix ductility and nanoparticle species
Mechanical performance of composite materials is a
function of ller and matrix characteristics. When studying nano-CaCO3/PP composites, for example, Ren
et al. [25] showed the importance of the matrix toughness. A much more remarkable increase in impact resistance was observed in the composites with a PP
copolymer possessing higher ductility as matrix, while
the same nanoparticles did not result in a similar
improvement in a PP homopolymer with lower ductility.
As a result, it was suggested that the polymer to be
toughened by nanoparticles should possess at least a certain toughness, which is dierent from the case when
elastomer acts as toughener.
To nd out whether nano-silica/PP composites follow
the same law, three types of PP: EPS30R, T30S and
PP700, are used in the current work. Relatively,
EPS30R has the highest toughness and PP700 the lowest. For making comparative tests, the crosshead speed
of the tensile tests was raised to 200 mm/min to match
the large failure strain of EPR30S. As illustrated by
Fig. 6(a) and (b), the addition of nano-silica into either
EPR30S or PP700 results in reduced values of elongation-to-break and area under tensile stressstrain curve,
no matter the particles have been treated or not. It
resembles the behavior of micro-particles lled polymer
composites. However, nano-silica lled T30S has acquired signicantly high static toughness. The untreated
nanoparticles have already certain ability to increase
elongation-to-break and area under tensile stressstrain
curve of T30S PP. It means that under tensile loading,
considerable matrix yielding of nano-silica lled PP
occurs only in the case of matrix polymer possessing
moderate toughness. The synergetic eect brought about
by the grafted nanoparticles depends on the matching
of matrix toughness and exibility of the grafting

642

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

Fig. 6. Mechanical properties of nano-silica/PP composites with

dierent PP as matrices: (a) elongation at break; (b) area under tensile
stressstrain curve; and (c) impact strength. The percentage numerals
quantify the relative variation of the mechanical property with
reference to the value of the corresponding neat PP. Nano-silica
content = 0.5 vol%, crosshead speed of the tensile tests = 200 mm/min.

polymers. For the composites based on EPR30S and

PP700, the nanoparticles are ineective to induce localized matrix drawing.
Fig. 6(c) shows the impact strength of the composites.
Although EPS30R and its composites are too tough to
be broken by the impact load under the current testing
conditions, some useful hints can still be yielded by comparing the performance of T30S and PP700 based composites. It is interesting to see that the nanoparticles
exert toughening eect on both types of PP. However,
the relative increments of the impact strengths of T30S
based composites are much higher than those of

PP700 based composites. Relating the results of

Fig. 6(a)(c) to each other, it is clear that the matrix ductility or the capability of the matrix to plastically deform
is a key factor inuencing the toughening eect of the
nanoparticles. In the case of suitable matrix ductility like
T30S, large scale of matrix polymer is successfully involved in plastic deformation as induced by the nanoparticles, leading to high impact toughness of the
composites.
On the other hand, it has been known that the interface between the ller particles and the matrix in a polymer nanocomposite constitutes a much greater area
within the bulk material as compared with conventional
composites containing micrometer-sized particles, and
hence inuences the composites properties to a much
greater extent. However, there is no information about
the eect of nanoparticles having dierent sizes. In the
current work, fumed nano-silica (15 nm) acts as a reference for the precipitated nano-silica (10 nm). To bring
the positive eect of the nanoparticles into full play,
both types of the particles were grafted with PBA (the
percent grafting of f-SiO2g-PBA is 8.26%, close to that
of p-SiO2g-PBA: 13.4%) and compounded with PP
(T30S) under the same conditions.
First Youngs moduli of the composites are compared
in Fig. 7(a). The ller content dependences of modulus
of the two types of composites are almost the same,
but the precipitated nano-silica leads to higher composites stiness. It must result from the increased interfacial
area in the composites with ner nano-silica, which promotes the stress transfer eciency within small strain
range. The reduction of Youngs modulus at higher
silica content manifests the softening eect of the compliant PBA interlayer, which becomes more and more
evident with a rise in content of the grafted
nanoparticles.
With respect to tensile strength of the composites, on
the contrary, the larger particles (fumed silica) seem to
have more remarkable reinforcing ability (Fig. 7(b)).
Generally, composites strength depends on the ller/matrix bonding under large strain condition. Since the present nanoparticles are grafted with the same polymer
(i.e., PBA) and the amounts of grafting polymer are similar in the two types of nano-silica, the interfacial adhesion in the composites should be almost identical at
given ller concentration [26]. The signicant dierence
in the tensile strengths of the composites might be attributed to the load bearing ability of the nano-silica itself.
That is, the fumed nano-silica particles might be stronger than the precipitated ones. To the authors knowledge, the strength of nano-silica particles has not yet
been reported because of the experimental diculties,
but one might nd some traces on the basis of their synthesis processes. Precipitated silica particles are obtained
in low-temperature wet process and might form soft
agglomerates (more physical bonding) during drying.

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

1.7

(b) 40
p-SiO2-g-PBA/PP
f-SiO2-g-PBA/PP

Tensile strength [MPa]

Young's modulus [GPa]

(a) 1.8

1.6

1.5

1.4

1.3
0.0

0.5

1.0

1.5

2.0

39

38

37

36
0.0

2.5

p-SiO2-g-PBA/PP
f-SiO2-g-PBA/PP

0.5

(c) 300

1.5

2.0

2.5

(d) 90

Area under stressstrain curve [MPa]

Elongation at break [%]

1.0

SiO2 [vol%]

SiO2 [vol%]

250

200

150
p-SiO2-g-PBA/PP
f-SiO2-g-PBA/PP

100

50
0.0

643

0.5

1.0

75

60

45
p-SiO2-g-PBA/PP
f-SiO2-g-PBA/PP

30

1.5

2.0

15
0.0

2.5

0.5

SiO2 [vol%]

1.0

1.5

2.0

2.5

SiO2 [vol%]

Impact strength [kJ/m ]

(e) 140

105

70

35

0
0.0

p-SiO2-g-PBA/PP
f-SiO2-g-PBA/PP

0.5

1.0

1.5

2.0

2.5

SiO2 [vol%]
Fig. 7. Mechanical properties of PP composites lled with dierent nano-silicas as a function of silica content (PP: T30S): (a) Youngs modulus; (b)
tensile strength; (c) elongation at break; (d) area under tensile stressstrain curve; and (e) impact strength.

As for fumed silica particles, they are produced in hightemperature gaseous process, which would made them
form hard agglomerates (more chemical bonding). During grafting pre-treatment, the former particle agglomerates are easier to be broken apart so that the grafted
precipated nano-silicas can be well dispersed than the
fumed ones. When the composites are subjected to applied stress, the intrinsic high strength of the fumed silica
aggregates oers higher strength for the composites.
In contrast to Youngs modulus and tensile strength,
elongation-to-break and area under tensile stressstrain
curve of the composites shown in Fig. 7(c) and (d) indi-

cate that the species of the nanoparticles nearly have

nothing to do with the two parameters when the ller
content 60.84 vol%. The characteristics of the grafting
polymer play the leading role in this case. This is reasonable because under tensile loading the extent of matrix
stretching is mainly related to the chain entanglement
between the grafting polymers and the matrix molecules.
For the composites with nano-silica content higher than
0.84 vol%, the decrease in elongation-to-break and area
under tensile stressstrain curve of f-SiO2g-PB/PP
should be due to the poor dispersion of the nanoparticles. Compared with precipitated nano-silica, fumed

644

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

nano-silica has much lower packing density. As a result,

homogeneous distribution of fumed nano-silica during
compounding is more dicult to be achieved than precipitated nano-silica especially at higher ller content.
It explains that the two parameters reecting static
toughness maintain almost unchanged over the whole
ller content range of interests in the case of p-SiO2
g-PB/PP (Fig. 7(c) and (d)).
When the composites specimens are subjected to impact load, however, dierent nanoparticles give dierent
responses (Fig. 7(e)). Similar to the results in Fig. 7(a),
precipitated nano-silica lled PP has higher impact
strength than fumed nano-silica/PP in most cases. This
implies that besides the interphase formed by the grafting polymers and matrix, the interfacial area between
the nanoparticles and the surrounding polymers is also
important to the toughening eect. The latter factor
might facilitate the generation of crazes during the impact test, absorbing certain amount of the input energy
additionally. Therefore, the composites lled with nanosilica with smaller particle size exhibit greater resistance
to impact loading.

4. Conclusions
Precipitated nano-silica is able to provide PP with
stiening, reinforcing and toughening eects at rather
low ller concentration as fumed nano-silica. Having
been grafted with dierent polymers onto the surfaces
in terms of gaseous graft polymerization, the nanoparticles become more ecient to improve the strength and
toughness of the composites. The nature of the grafting
polymer chains plays an important role in the properties
enhancement.
Ductility of the matrix PP determines the toughening eect of the nanoparticles. Only in the case of
moderate matrix ductility, the composites can receive
the highest extent of toughness increase. Besides, the
size and surface area of the nanoparticles are also
important inuencing factors. The smaller nanoparticles lead to higher Youngs modulus and impact
strength of the composites, and reduce the sensitivity
of the static toughness to the status of ller
distribution.

Acknowledgements
The authors are grateful to the support of the Deutsche Forschungsgemeinschaft (DFG FR675/40-4) for
the cooperation between the German and Chinese institutes on the topic of nanocomposites. Further thanks
are due to the National Natural Science Foundation
of China (Grant No. 50133020), the Team Project of
the Natural Science Foundation of Guangdong, China

(Grant No. 20003038), and the Key Program of the Science and Technology Department of Guangdong, China
(Grant No. A10172).

References
[1] Rong MZ, Zhang MQ, Zheng YX, Zeng HM, Walter R,
Friedrich K. Structureproperty relationships of irradiation
grafted nano-inorganic particle lled polypropylene composites.
Polymer 2001;42:16783.
[2] Wu CL, Zhang MQ, Rong MZ, Friedrich K. Tensile performance
improvement of low nanoparticles lled-polypropylene composites. Compos Sci Technol 2002;62:132740.
[3] Zhang MQ, Rong MZ, Zhang HB, Friedrich K. Mechanical
properties of low nano-silica lled high-density polyethylene
composites. Polym Eng Sci 2003;43:490500.
[4] Wang G, Chen XY, Huang R, Zhang L. Nano-CaCO3/polypropylene composites made with ultra-high-speed mixer. J Mater Sci
Lett 2002;21:9856.
[5] Zhang MQ, Rong MZ, Pan SL, Friedrich K. Tensile properties of
polypropylene lled with nanoscale calcium carbonate particles.
Adv Compos Lett 2002;11:2938.
[6] Rong MZ, Zhang MQ, Zheng YX, Zeng HM, Walter R,
Friedrich K. Irradiation graft polymerization on nano-inorganic
particles: an eective means to design polymer based nanocomposites. J Mater Sci Lett 2000;19:115961.
[7] Chung SC, Hahm WG, Im SS, Oh SG. Poly(ethylene terephthalate) (PET) nanocomposites lled with fumed silicas by melt
compounding. Macromol Res 2002;10:2219.
[8] Zhang J, Wang X, Lu L, Li D, Yang X. Preparation and
performance of high-impact polystyrene (HIPS)/nano-TiO2 nanocomposites. J Appl Polym Sci 2003;87:3815.
[9] Xiong MN, Wu LM, Zhou SX, You B. Preparation and
characterization of acrylic latex/nano-SiO2 composites. Polym
Int 2002;51:6938.
[10] Jana SC, Jain S. Dispersion of nanollers in high-performance
polymers using reactive solvents as processing aids. Polymer
2001;42:6897905.
[11] Rong MZ, Zhang MQ, Zheng YX, Zeng HM, Friedrich K.
Improvement of tensile properties of nano-SiO2/PP composites in relation to percolation mechanism. Polymer 2001;
42:33014.
[12] Zhang MQ, Rong MZ, Zeng HM, Schmitt S, Wetzel B, Friedrich
K. An atomic force microscopy study on structure and properties
of irradiation grafted silica particles in polypropylene based
nanocomposites. J Appl Polym Sci 2001;80:221827.
[13] Torro-Palau AM, Fernandez-Garca JC, Orgiles-Barcelo AC,
Martn-Martnez JM. Characterization of polyurethanes containing dierent silicas. Int J Adhes Adhes 2001;21:19.
[14] Wang G, Zhang L, He G. Chinese Patent (CN1046921); 1996.
[15] Wu CL. PhD thesis, Zhongshan University; 2002.
[16] Brunauer S, Deming L, Deming W, Teller E. On a theory of the
van der Waals adsorption of gases. J Am Chem Soc
1940;62:172332.
[17] Kerner EH. Electrical conductivity of composites media. Proc
Phys Soc B 1956;69:8027.
[18] Nielsen LE. Simple theory of stressstrain properties of lled
polymer. J Appl Polym Sci 1966;10:97103.
[19] Nicolais L, Narkis M. Stressstrain behavior of styreneacrylonitrile/glass bead composites in the glassy region. Polym Eng Sci
1971;11:1949.
[20] Jancar J, Dianselmo A, Dibenedetto AT. The yield strength of
particulate reinforced thermoplastic composites. Polym Eng Sci
1992;32:13949.

C.L. Wu et al. / Composites Science and Technology 65 (2005) 635645

[21] Manson JA, Sperling LH. Polymer blends and composites Plenum Press. New York; 1976.
[22] Wu CL, Zhang MQ, Rong MZ, Lehmann B, Friedrich K.
Deformation characteristics of nano-SiO2 lled polypropylene
composites. Polym Polym Compos 2003;11:55962.
[23] Ebengou RH. Adsorption as a mechanism for nucleating activity:
a thermodynamic explanation. J Polym Sci Polym Phys
1997;35:13338.

645

[24] Long Y, Shanks RA. PP/elastomer/ller hybrids. II. Morphologies and fracture. J Appl Polym Sci 1996;62:63946.
[25] Ren X, Bai L, Wang G. Reinforcement and toughening of
polypropylene composites by nanoparticle CaCO3. Chem World
2000;41(2):837. [in Chinese].
[26] Rong MZ, Zhang MQ, Pan SL, Lehmann B, Friedrich K.
Analysis of the interfacial interaction in polypropylene/silica
nanocomposites. Polym Int 2004;53:17683.