JOURNAL OF

POLYMER SCIENCE

WWW.POLYMERPHYSICS.ORG

FULL PAPER

Nanoindentation, Nanoscratch, and Nanotensile Testing of

Poly(vinylidene fluoride)-Polyhedral Oligomeric Silsesquioxane
Nanocomposites
Fanlin Zeng,1 Yizhi Liu,1 Yi Sun,1 Enlai Hu,1 Yu Zhou2
1

Department of Astronautic Science and Mechanics, Harbin Institute of Technology, Harbin, Peoples Republic of China

Material Science and Engineering, Harbin Institute of Technology, Harbin, Peoples Republic of China

Correspondence to: F. Zeng (E-mail: zengfanlin@hit.edu.cn)

Received 19 June 2012; revised 9 August 2012; accepted 13 August 2012; published online 14 September 2012
DOI: 10.1002/polb.23159

ABSTRACT: Nanocomposites composed of a poly(vinylidene fluoride) (PVDF) matrix and 0, 3, 5, and 8 wt % fluoropropyl polyhedral oligomeric silsesquioxane (FP-POSS) were prepared by
using the solvent evaporation method. The morphology and the
crystalline phase of the nanocomposites were investigated by
digital microscopy, scanning probe microscopy, X-ray diffractometer, and Fourier transform infrared spectroscopy. FP-POSS
acted as nucleating agent in PVDF matrix. A small content of FPPOSS resulted in an incomplete nucleation of PVDF and generated bigger spherical particles, whereas higher contents led to a
complete nucleation and formed more separate and less-crosslinked particles. Nanoindentation, nanoscratch, and nanotensile
tests were carried out to study the influence of different contents
of FP-POSS on the key static and dynamic mechanical properties of different systems. The nanocomposite with 3 wt % FPPOSS was found to possess enhanced elastic properties and
hardness. However, with the increase of the FP-POSS content,

the elastic modulus and hardness were found to decrease, and

the improvement on stiffness was negative at contents of 5 and
8 wt %. Compared with neat PVDF, the scratch resistance of the
PVDF/FP-POSS nanocomposites was decreased due to a
rougher surface derived from the bigger spherulites. Nanotensile testing results showed both the stiffness and toughness of
PVDF-FP3% were enhanced and further additions of FP-POSS
brought dramatic enhancements in toughness while associated
with a decline in stiffness. Dynamical mechanical properties
indicated the viscosity of the nanocomposites increased with
C 2012 Wiley Periodicals, Inc.
the increasing FP-POSS contents. V
J Polym Sci Part B: Polym Phys 50: 15971611, 2012

INTRODUCTION As a semicrystalline polymer, poly(vinyli-

thermal stability, and dielectric properties.1012 Specifically,

polyaniline (PANI),12 PZT,13 polyamide,14 LiClO4, and TiO2,15
thermoplastic polyurethane (TPU),16 poly(methylmethacrylate) (PMMA),17 and so on have been used to improve the
mechanical properties of PVDF. It appears from current experimental investigations that improvements in mechanical
properties of PVDF composites depend on a set of factors in
terms of filler size, shape, mass ratio, degree of dispersion,
and so on. However, it is difficult to find plausible correlations between specific factors, for example, the filler mass ratio, and mechanical properties from the reported results. It
seems that the improvement effects depend not only on the
filler mass ratio but also on the filler type.

dene difluoride) (PVDF) is widely used in smart structure

sensors, actuators, and transducers because it exhibits excellent piezoelectric and pyroelectric properties which were
derived from the two polar phases b and c of PVDF.13
Because PVDF is not as hard as some other piezoelectric
materials, such as lead zirconate titanate (PZT), PbTiO3, and
so on, it can easily be prepared to the soft and thin transducers which are dictated by the shape of the structures
they coat. This property is very useful in fields of energy
harvesting4,5 and adaptive inflatable structures in aerospace
engineering.6,7 But unfortunately, neat PVDF cannot completely meet the mechanical, thermal, and oxidation resistance property requirements of some harsh environments.79
Many efforts have been taken to improve the properties of
PVDF. For example, incorporation of organic polymer or inorganic fillers into the PVDF matrix to produce composites has
been extensively studied with the objective of further
improving its properties, such as mechanical performance,

KEYWORDS: mechanical properties; morphology; nano-compo-

sites; polyhedral oligomeric silsesquioxanes; poly(vinylidene

difluoride)

Polyhedral oligomeric silsesquioxanes (POSS) are a unique

class of organicinorganic hybrid materials that can be
depicted by the formula (RSiO1.5)n (where n is an even number and R H, Cl, or a variety of organic groups). Because
of the robust inorganic cage-like core structure with the
SiAO atoms, POSS exhibits many superior thermomechanical

C 2012 Wiley Periodicals, Inc.

V

WWW.MATERIALSVIEWS.COM

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

1597

FULL PAPER

WWW.POLYMERPHYSICS.ORG

properties in terms of wearability, thermal stability, oxidation

resistance, and high strength1820 among others. Fluorinated
POSS are a new kind of POSS compounds with high molecular weight and density.21 Theoretically, PVDF/POSS nanocomposites with highly improved mechanical properties are
expected to be produced if fluorinated POSS is incorporated
into PVDF, assuming they are miscible. For a polymer matrix,
it is widely accepted that inorganic fillers tend to improve in
stiffness while organic ones in toughness. But as a kind of
organicinorganic hybrid compounds, how will different contents of fluorinated POSS influence the key mechanical properties of PVDF is hard to evaluate and needs to be
researched by experiments.
As a standard mechanical test method, nanoindentation has
been developed and widely used to characterize some important mechanical properties (hardness, elastic modulus,
and so on) of nanostructural materials and thin films or
coatings.2225 Nanoindentation method measures the hardness and elastic modulus (Youngs modulus) of a material
from indentation load-displacement data obtained during
one cycle of loading and unloading.23 Besides the standard
method for hardness and elastic modulus, other nanoindentation methods have also been devised for evaluating some
other properties. For example, although not well developed,
the yield stress and strain-hardening characteristic of metals;26,27 parameters characteristic of damping and internal
friction in polymers, such as the storage and loss modulus;28,29 and the activation energy and stress exponent for
creep30,31 have been determined by these methods. A nanoscratch test with the nanoindentation instrument is widely
used to characterize the films and coatings for scratch resistance.25,32 The scratch resistance is measured from the depth
at a given load or the load at which material fails (the critical load for a thin film adhering on a substrate). There are
three scan steps in a typical scratch procedure: (1) prescan,
the indenter tip only approaches and scans from the scratch
area to achieve the original topography of the surface; (2)
scratch scan, the tip scratches on the sample with ramping
loads; and (3) postscan, the tip unloads from the scratch end
to obtain the residual deformation. From the original topography of the surface, morphology compensation can be fulfilled at the latter two steps and the effective penetration
depth can be obtained. A tensile test has been widely used
to determine some important mechanical properties such as
Youngs modulus, fracture toughness, and yield stress of
materials. However, a traditional tensile test is not capable of
a sample with nanoscale size, where a very small tensile
load and very precise measure equipments are needed. In
recent years, the nanotensile test has been developed and
used to characterize the tensile behaviors of very small samples such as the nanotubes, silk, fibers, and so on.3335 The
nanotensile test is also used to determine some dynamic mechanical properties by adding a very small harmonic force
with specific frequency on the sample.36
Specifically, the crystallization behaviors, morphology, and
physical properties of PVDF or its nanocomposites have been
investigated by many researchers. Chang et al.,37 for exam-

1598

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

JOURNAL OF
POLYMER SCIENCE

ple, conducted an experimental study to address the thermomechanical and optical characteristics of PVDF for flexible
electronic applications. They found that thermomechanical
characteristics, thermal elongation, and expansion were
greatly influenced at stretching ratios of over four in the
stretching direction, and the optical properties were greatly
influenced in stretched films. Chae et al.38 mixed different
mass ratios of silver nanoparticles into PVDF and tested
their physical properties, the crystallization behavior under
shear, and the consequential crystalline morphology. They
claimed that the overall crystallization process of PVDF
under shear was accelerated by reducing both induction
time and crystallization time, when Ag loading level was
increased. Guney,39 who researched the influence of temperature on the elastic properties of PVDF, found that the relaxation behavior of PVDF was affected from the form of mechanical disturbance. Ma et al.40 mixed different small
amount of nanoparticles, such as montmorillonite (MMT),
SiO2, CaCO3, or polytetrafluoroethylene (PTFE), into PVDF
and researched their crystallization and melting behaviors.
They found that addition of small amount of these four types
of nanoparticles would not affect the original crystalline
phase obtained in the neat PVDF sample, but accelerated the
crystallization rate because of the nucleation effect. In these
four blend systems, MMT or PTFE nanoparticles could be
applied well for PVDF nanocomposite preparation because of
stronger interactions between particle surface and PVDF
molecules. The nucleation enhancement and the growth rate
of the spherulites were decreased in the order SiO2 >
CaCO3 > PTFE > MMT. Fang et al.41 conducted an uniaxial
tension experiment to study the deformation and fracture
behavior of poly(vinylidene fluoride-trifluorethylene) ferroelectric copolymer films. They found that the polymer film
samples prepared by stretching the solution-cast films and
then annealing fracture at a much larger maximum strain
and a higher tensile strength than those prepared by solution casting and then annealing. Salimi et al.42 researched
the conformational changes and phase transformation mechanisms in PVDF solution-cast films. They found that the lowtemperature crystallization of PVDF in dimethylacetamide solution, mainly resulted in the formation of trans states (b
and c phases), whereas at higher temperatures, gauche states
become more populated (a phase). Moreover, the uniaxial
stretching greatly enhanced piezoelectric properties of the
films, due to formation of oriented b phase crystals, which
were of more uniform distribution of dipole moments. Very
recently, Martins et al.43 conducted a study about the morphological, viscoelastic, and thermal properties of PVDF/
POSS nanocomposites. They found that the crystallinity of
PVDF was little influenced at low POSS contents, and the
methacryl POSS were acting as lubricant in the nanocomposite system. It seems from current investigations that the
crystallization behaviors, morphological, and physical properties of PVDF and its blends will be influenced by many different factors. In addition, these properties of PVDF and
POSS nanocomposites, however, were seldom researched.
In previous work, we have researched the miscibility in mixtures of PVDF and several kinds of POSS compounds

JOURNAL OF
POLYMER SCIENCE

WWW.POLYMERPHYSICS.ORG

FULL PAPER

FIGURE 1 The nanotensile testing system UTM T150 (a) and the sample image measured by optical microscopy (b).

including four kinds of fluorinated POSS compounds using

molecular simulations.44 The simulation results show four
kinds of fluorinated POSS compounds and PVDF are fully
miscible at any temperature and the miscibility is derived
from the polar CAF bonds and the electrostatic interactions
in the POSS and PVDF molecules. We have also investigated
the elastic properties of mixtures of PVDF and the (3,3,3-trifluoropropyl)8Si8O12 (FP-POSS)21 at different temperatures
using molecular dynamics (MD) simulations.45 We found that
the glass transition temperature of PVDF was significantly
improved with FP-POSS, and the moduli of PVDF were
improved and the improvement effect, in general, nearly
decreased with the increase of the mass ratio of FP-POSS.
However, experiments are indispensable to this work. On the
one hand, simulation results need to be verified from experiments, and on the other hand, some other key mechanical
properties, such as the scratch resistance, fracture toughness,
and some dynamic mechanical properties can also be
obtained. In this work, the PVDF/FP-POSS nanocomposites
were first prepared by the solvent evaporation method. And
then, the microscopy characterization, nanoindentation,
nanoscratch, and nanotensile testing were carried out. The
focus of this work is on the effects of the FP-POSS contents
on the morphology, the key mechanical properties of PVDF/
FP-POSS nanocomposites, such as the hardness, elastic modulus, scratch resistance, tensile strength, fracture toughness,
loss tangent, and so on, and the aim is to determine how
these properties would be influenced by varying the amount
of FP-POSS. Furthermore, the deformation mechanism of the
nanocomposites was analyzed, and it would be helpful to
understand why different materials exhibited different mechanical properties from microscale.
EXPERIMENTAL

Materials and Sample Preparation

Commercial PVDF (Mw 534,000; 99.99% purity; Aldrich),
(3,3,3-trifluoropropyl)8Si8O12 (FP-POSS; Hybrid Plastics cod.
FL0578), and N,N-dimethylformamide (DMF; 99%; Aldrich)
were used to prepare the PVDF/FP-POSS mixtures. PVDF

WWW.MATERIALSVIEWS.COM

and FP-POSS powder were first dissolved in DMF in a

beaker; next, the solution was kept at room temperature
under mechanical stirring until PVDF and FP-POSS were dissolved, and air was removed fully. The solution containing
PVDF and FP-POSS was homogeneously transparent. Then,
the total evaporation was carried out at 70  C to remove the
solvent, and composite films with thicknesses of around 70
lm were obtained. The whole process of the preparation
was carried out in a vacuum. Samples for nanoindentation,
nanoscratch, and nanotensile testing were cut from these
films. The size of the samples is about 5  5 mm2 for nanoindentation and nanoscratch, and about 100 lm  10
mm for nanotensile testing (shown as Fig. 1, but not the
same size in each test). In order to be comparable to the
previous simulation work,45 the mass ratios of FP-POSS
added to the PVDF matrix were 0, 3, 5, and 8%, and nanocomposites with different FP-POSS contents are denoted as
PVDF-FPi% (i 0, 3, 5, and 8).
Morphology and Structure Characterization
Optical microscopy was performed on the nanocomposite
films using a VHX-600E digital microscope (Keyence Company) with a real-time depth composition, two/three-dimensional functions, and 20 to 5000 zoom. Scanning probe
microscopy (SPM) was performed using a SPM9500J2 microscope (Shimadzu Company) equipped with software (version
2.30) for images processing and profile analysis. During the
scanning, the contact scanning mode was used and both the
deflection and height traces were obtained.
X-Ray diffractometer (XRD) analysis was applied using a
Rigaku D/max-rB rotating anode XRD with CuKa radiation, k
0.15418 nm, at a generator voltage of 60 kV and a current
of 200 mA. The data were collected from 10 to 80 intervals. Fourier transform infrared (FTIR) analysis was carried
out using a FTIR spectrometer (Avatar360, Nicolet), with a
scan range 4000500 cm1 and a resolution of 4 cm1.
Nanoindentation and Nanoscratch Testing
The hardness and elastic modulus of the nanocomposite
films were determined by using the OliverPharr

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

1599

FULL PAPER

WWW.POLYMERPHYSICS.ORG

method23,4648 in the nanoindentation testing with a G200

Nano Indenter (Agilent Technologies). A diamond triangular
pyramid Berkovich indenter (TB20114 ISO) with a tip radius
of about 20 nm, face angle (the centerline-to-face angle) of
65.3 , Youngs modulus of 1141 GPa, and Poissons ratio of
0.07 was used. From the simulation results,45 the Poissons
ratio of the nanocomposite films was estimated as 0.3. It
may seem counterintuitive that one must know the very
accurate Poissons ratio of the material to compute its modulus, even a rough estimate. However, in fact, a very precise
value of this parameter is not necessary, because when it is
given as 0.25 6 0.1, produces only about a 5% uncertainty
in the calculated value of the elastic modulus for most materials.46 Two kinds of specific test methods: (1) G-Series Basic Hardness, Modulus at a Depth; and (2) G-Series Hardness and Modulus via Cycles Load Control were used to
determine the hardness and elastic modulus at different
maximum depths and loads. In the former method, tests
with different maximum depths of 500, 1000, 1500, and
2000 nm were carried out on each sample. During the test,
the loading speed was controlled as the same strain rate of
0.05 s1, which means the indenter was driven into the sample at a same speed; thus, the loads would increase very
slowly at the initial stage and very fast at the end. In the latter method, the maximum load of 10 mN and eight loading
cycles were carried out at each test. Thus, eight relevant
tests were carried out at the same position, and the maximum loads at each test increased as an exponential function
(0.075, 0.15, 0.31, 0.62, 1.25, 2.5, 5.0, and 10.0 mN for each
test, respectively). In this method, the size effect can be
clearly displayed; from which one can judge what maximum
loads are appropriate to obtain the reliable results. During
the test in this method, the loading speed was kept as a constant force increasing per second (loading time of 20 s
adopted here), thus, the indenter was driven into the sample
at different speeds, fast at the initial stage and slowly at the
end. Both in the two methods, each kind of test was
repeated at six different positions with intervals of 50 lm
(defined as a 3  2 array). The peak-hold time, that is, the
time the indenter held at the maximum load (load control)
or depth (depth control) was set as 10 s. The thermal drift
was considered after the tip detected the surface with a custom stiffness criteria of 125N/m at each test. The allowable
drift rate was set as 0.5 nm/s, which means the test will not
begin until the drift of the indenter influenced by the thermal fluctuation is less than 0.5 nm/s. At the end stage of the
unloading (90%), the indenter would be held for about 1
minute again to compute and correct the final drift. All tests
were carried out in a clean-air environment with a relative
humidity of 30%, while the temperature was kept constant
at 20 6 0.5  C. Hardness and elastic modulus were measured from indentation load-displacement data obtained during one cycle of loading and unloading. From the maximum
load and the corresponding projected contact area hardness
can be determined and by measuring the elastic contact stiffness (measured from the upper portion of the unloading
data) and the projected contact area the elastic modulus can
thus be derived. In the real testing work, the projected con-

1600

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

JOURNAL OF
POLYMER SCIENCE

tact area A has been calibrated as A 24.0808h2 382h (h

is the contact depth). In addition, the two main factors, the
adhesion between the tip and the specimen surface and the
pile-up (or sink-in) of the specimen surface under the tip,
will bring spurious results in the nanoindentation testing.
For the adhesion, its effect on the nanoindentation results is
only obvious at low loads or low indentation depths. Thus,
the overestimation in the measured indentation moduli/
hardness can be corrected by performing indentations with
loads high enough so that the modulus/hardness is independent of the applied load. When pile-up (or sink-in)
occurs, the contact area is greater (or less) than that predicted by the OliverPharr method, and both the hardness
and the modulus are overestimated.23 Fortunately, Pharr
et al.23 found that when hf/hmax < 0.7 (hf, the final depth,
the permanent depth of penetration after the indenter is
unloaded fully; hmax, the maximum displacement), very little
pile-up (or sink-in) is observed no matter what the workhardening behavior of the material. Thus, the contact areas
are independent of the work-hardening characteristics in this
case. Therefore, in an indentation experiment, care must be
taken when hf/hmax > 0.7, as using the OliverPharr method
can lead to large errors in the contact area.
Scratch testing was carried out using a Nano Indenter G200
instrument with the method of G-Series Ramp Load Scratch
with Topography Compensation. The test procedure was
similar to that presented elsewhere,25,49 but the topography
compensation was considered here because the samples surface was not very smooth. Before the real scratching, the
original topography and surface roughness of the scratching
area were detected by prescanning the surface with the indenter under a profiling load 20 lN. In the real scratching,
the indenter tip scratched on the sample with linear ramping
loads from 0 to maximum value of 100 mN. The scratch
length and speed were set as 250 lm and 10 lm/s, respectively. Depths of scratches with increasing scratch load were
measured by profiling the surface during and after the
scratch test, resulting in a total length of about 350 lm for
each test (including 50-lm prescanning and postscanning at
the two ends of the real scratch path), whereas the real
effective scratch length was 250 lm as applied to all specimens. The effective penetration depth during (scratch depth)
and after (residual depth) the scratch test at different
scratch distance then could be obtained. Three independent
scratch tests were carried out for each sample, and the distance interval of different scratches was 100 lm. The test
environment and the allowable thermal drift rate were similar to those in the nanoindentation test.
Nanotensile Testing
A commercial nanotensile testing system (Nano UTMTM Universal Testing System T150, Agilent Technologies) with the
method of UTM-Bionix Standard Toecomp CDA was used to
conduct the tensile test. The instrument consists of a loading
frame, moving crosshead, two grips, and a nanomechanical
actuating transducer (NMAT) as shown in Figure 1(a). The
nominal maximum load of this system is 500 mN (can reach
to 750 mN in real test), and the maximum crosshead

JOURNAL OF
POLYMER SCIENCE

WWW.POLYMERPHYSICS.ORG

FULL PAPER

FIGURE 2 Optical microscopy surface images of PVDF-FPi% obtained by solution, i (a) 0, (b) 3, (c) 5, and (d) 8.

extension is 150 mm (nominal, stages with larger extensions

can be obtained on a custom basis). There is a sensitive spring
installed in NMAT that can provide a very small harmonic
force with specific frequency on the sample; thus, the dynamic
mechanical properties, meanwhile, can be obtained during the
tensile procedure. The samples were prepared as slender belts
cut from PVDF-FPi% films and fixed on a perforated cardboard
as shown in Figure 1(b). The length and height of belt-like
samples were measured by a caliper and screw micrometer,
and the width was measured by the optical microscopy mentioned in Morphology and Structure Characterization section.
Three individual samples have been prepared and tested in
each category (12 in all). The tensile strain rate was set as 1.0
 103 s1, and the harmonic force and the frequency were
typically 4.5 mN and 20 Hz, respectively.
RESULTS AND DISCUSSION

Microscopic Morphology
The morphologies of the free surface of PVDF-FPi% films,
that is, the surface opposite to the one in contact with the
glass substrate, were analyzed. Figure 2 shows the microstructural conformation of PVDF-FPi% measured by optical
microscopy. Samples of neat PVDF [Fig. 2(a)] show a microporous structure composed by micron-sized spherulites. This
structure is very similar to that in the work reported by
Nunes et al.50 and Serrado et al.,51 where it was identified as
the porous b-PVDF. Compared with the image of neat PVDF,
spherulites in samples of PVDF-FPi% (i 3, 5, and 8) are
visibly bigger. Maybe it would be considered to the adhesion
of FP-POSS on PVDF spherulite, but the content of FP-POSS
obviously determined that there is no so much FP-POSS can

WWW.MATERIALSVIEWS.COM

be provided. Further microscopy details would be needed to

explain this satisfactorily. Furthermore, the nonuniformity of
spherulites in Figure 2(bd), that is, some spherulites look
much smaller than others, may be noticed. POSS aggregates
have taken place in these samples. However, we want to
explain that it is not this case since these images were
obtained at a very high magnification (2000 zoom) and a
very shallow depth of field. The smaller spherulites are
just those with only a small top part located on the focal
plane. These results also show that the samples of PVDFFPi% (i 3, 5, and 8) have bigger spherulites and rougher
surfaces compared with neat PVDF.
The surface SPM height trace images taken within one spherulite for PVDF-FPi% at scan ranges of 2.0  2.0 lm2 are
shown in Figure 3(ad), respectively. Height data correspond
to the change in piezo height needed to keep the cantilever
deflection constant, which is always used to reflect the topography of the surface. It can be observed clearly that all
spherulites consist of small particles with different sizes for
different samples. Compared with the image of neat PVDF
[Fig. 3(a)], particles in PVDF-FPi% (i 3, 5, and 8) are similar but possess larger size and more uniform dispersion. No
POSS aggregate could be observed in all PVDF-FPi% (i 3, 5,
and 8). It seems that the effect of FP-POSS is just to form
bigger PVDF particles and very similar to the nucleating
agents in some PVDF mixtures.40,52,53 This may be understood from the simulation results:44 strong electrostatic
forces arise owing to the same polar CAF bonds on PVDF,
and the fluorinated POSS molecules tend to make more
PVDF molecules be nucleated around FP-POSS. Further addition of nucleating agents probably has little effect on forming

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

1601

FULL PAPER

WWW.POLYMERPHYSICS.ORG

JOURNAL OF
POLYMER SCIENCE

FIGURE 3 SPM height trace images taken within a spherulite for PVDF-FPi%, i (a) 0, (b) 3, (c) 5, and (d) 8, at scan ranges
of 2.0  2.0 lm2.

even bigger particles since they possess similar size for

PVDF-FPi% (i 3, 5, and 8). However, some changes on the
morphology can still be observed, particles are exhibited
more clearly with the increasing of the FP-POSS content.
This may be derived from a more complete nucleation due
to a relative high content of FP-POSS forming stronger nucleating agents. The surface SPM deflection trace images taken
within one spherulite for PVDF-FPi% at scan ranges of 5.0 
5.0 lm2 are shown in Figure 4(ad), respectively. Deflection
data come from the differential signal off of the top and bottom photodiode segments. They are collected with low gain
feedbacks so the piezo remains at a constant position relative to the sample. In this case, the tip and cantilever are
deflected by the features on the sample surface. The output
fluctuations in the cantilever deflection voltage from the top
and bottom photodiode segments are recorded as a measure
of the variation in the sample surface. Thus, the surface
details can be shown more clearly from the deflection trace.
It can be found clearly from these figures that the particles
are more separate and the domain size of the particles
becomes larger with the increasing of the FP-POSS content.
As mentioned earlier, it should be derived from the more

1602

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

complete nucleation of PVDF due to the higher content of

FP-POSS. These morphology changes would greatly influence
some mechanical properties of PVDF/FP-POSS nanocomposites, which will be discussed later.
Both optical and SPM images for all samples show that all
spherulites and particles are uniformly dispersed and no
POSS aggregates could be observed. All these proved that
FP-POSS compounds and PVDF are fully miscible, which is in
good agreement with the simulation results.44
Phase Analysis
XRD patterns for PVDF-FPi% (i 0, 3, 5, and 8) and neat FPPOSS are shown in Figure 5. It can be found that the b and a
phases are obviously exhibited in both neat PVDF and PVDFFPi% (i 3, 5, and 8) via reflections at (110) and (020) crystal planes, occurring at 2y 20.6 and 18.5 , respectively.
No obvious difference can be found in the diffraction patterns between neat PVDF and PVDF-FPi% (i 3, 5, and 8),
which means there may be similar crystallization process in
four different systems, and the incorporation of FP-POSS has
little influence on the crystalline phase of PVDF. Although a
very intense main diffraction peak at 2y 19.4 reveals that

JOURNAL OF
POLYMER SCIENCE

WWW.POLYMERPHYSICS.ORG

FULL PAPER

FIGURE 4 SPM deflection trace images taken within a spherulite for PVDF-FPi%, i (a) 0, (b) 3, (c) 5, and (d) 8, at scan ranges of
5.0  5.0 lm2.

the neat FP-POSS is highly crystalline, and the XRD spectra

of PVDF-FPi% (i 3, 5, and 8) show their crystalline features
that are only referable to neat PVDF matrix. This result provides evidence that FP-POSS with mass ratio of 3 and 5%,

FIGURE 5 XRD patterns for PVDF-FPi% (i 0, 3, 5, and 8) and

neat FP-POSS.

WWW.MATERIALSVIEWS.COM

even up to 8%, are almost dispersible in PVDF matrix, just

as already observed from AFM images.
The FTIR spectra for PVDF-FPi% (i 0, 3, 5, and 8) and neat
FP-POSS are shown in Figure 6. It can be found that all FTIR
curves for PVDF-FPi% (i 3, 5, and 8) are very similar to

FIGURE 6 FTIR spectra for PVDF-FPi% (i 0, 3, 5, and 8) and

neat FP-POSS. The peaks for a phase, b phase, and SiAOASi in
FP-POSS are located.

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

1603

FULL PAPER

WWW.POLYMERPHYSICS.ORG

JOURNAL OF
POLYMER SCIENCE

FIGURE 7 Influence of different FP-POSS contents on (a) hardness and (b) elastic modulus of PVDF-FPi% (i 0, 3, 5, and 8) as
determined from the G-Series Hardness and Modulus via Cycles Load Control method.

that of neat PVDF, indicating that FP-POSS has not changed

the crystalline phase of PVDF. The band at 1120 cm1 is
attributed to the stretching vibration of SiAOASi in POSS.54
However, it is not obvious in PVDF-FPi%, revealing that no
POSS agglomerates in the nanocomposites. The characteristic
transmission peaks of a phase are 611, 766, and 1404 cm1,
and the b phase can be proven through the presence of 870,
1064, and 1182 cm1 transmission peaks.43 Compared with
other nanocomposites, PVDF-FP5% presents more intense
peaks for the same a and b phases, indicating FP-POSS may
be leading to the formation of more a and b phase crystals
with 5 wt % content, which is very similar to the result in
literature.43
Hardness and Elastic Modulus
The local values of hardness and elastic modulus of PVDFFPi% as a function of the maximum load and indentation
depth are shown in Figures 7 and 8, respectively. The size

effect revealed in Figure 7 is so evident that it is hard to

judge what values are reliable enough as determined from
the G-Series Hardness and Modulus via Cycles Load Control
method. It may be related to the rough surface of PVDF-FPi%
films and constant loading rate derived from this method. A
rough surface determines that the real projected contact
area is always different with that computed from the penetration depth. And a constant loading rate means a changing
strain rate during the test, that is, the indenter is driven into
the sample with different speed as explained previously, will
bring strain rate effect to the test results,55 especially to
some strain rate sensitive polymers and viscoelastic materials.56 Even so, approximate orders of PVDF-FP3% > PVDFFP0% > PVDF-FP5%/PVDF-FP8% for the values of hardness
and elastic modulus are shown definitely in Figure 7. The
order for PVDF-FP5% and PVDF-FP8% is hard to give since
these two pairs of curves are so intertwined. It appears that
the hardness and elastic modulus obtained by the G-Series

FIGURE 8 Influence of different FP-POSS contents on (a) hardness and (b) elastic modulus of PVDF-FPi% (i 0, 3, 5, and 8) as
determined from the G-Series Basic Hardness, Modulus at a Depth method.

1604

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

JOURNAL OF
POLYMER SCIENCE

WWW.POLYMERPHYSICS.ORG

TABLE 1 The Hardness (H), Elastic Modulus (E), Plasticity Index

E/H, and H/E Ratio of PVDF-FPi% (i 5 0, 3, 5, and 8)
FP-POSS (wt %)

H (GPa)

E (GPa)

E/H

H/E

0.174 (0.012)

2.61 (0.12)

14.99

0.0667

0.177 (0.018)

2.77 (0.16)

15.63

0.0640

0.158 (0.018)

2.33 (0.26)

14.68

0.0681

0.130 (0.036)

2.21 (0.26)

16.94

0.0590

Values in parentheses indicate (6) standard deviations.

Basic Hardness, Modulus at a Depth method (Fig. 8) are

much more uniform than the values in Figure 7. As the same
very low strain rate of 0.05 s1 was carried out at each test,
the strain rate effect could be ignored. In addition, the size
effect is likely to be eliminated by indentation depths with at
least 500 nm. Compared with Figure 7, we think the results
in Figure 8 are more reliable, which were used to compute
the intrinsic average values of hardness and elastic modulus
of PVDF-FPi%. Similar orders for the values of hardness and
elastic modulus are shown in Figure 8 and proved this result
is not accidental. This needs to be identified by specific
values.
Table 1 lists the values of hardness (H), elastic modulus (E),
plasticity index (E/H),25,57 and H/E ratio of PVDF-FPi% averaged from the results as determined from the G-Series Basic
Hardness, Modulus at a Depth method (Fig. 8). It is clear
that the hardness and elastic modulus values follow the
orders PVDF-FP3% > PVDF-FP0% > PVDF-FP5% > PVDFFP8%. PVDF-FP3% possesses enhanced hardness and elastic
properties, which is in good agreement with the simulation
results.45 This may be understood from the role of FP-POSS.
A small content of FP-POSS probably made the particles bigger and more condensed, and thus, the mechanical properties were improved in this case. While a large content of FPPOSS addition perhaps resulted in fully PVDF nucleation and
formed more separate spherical particles, which would
largely decrease the crosslinkage among PVDF particles.
Thus, although stronger particles were probably brought in
this case, the whole mechanical properties dropped. In addition, all the experimental values are substantially less than
the simulated ones, probably caused by the amorphous models used in the simulation work, in which the nucleation was
not considered thus the models are different from the structures in this work. Plasticity index values do not show
obvious differences or changing rules, and indicate that the
plasticity of PVDF was not observably influenced by FPPOSS.
Typical loadunload cycles for PVDF-FPi% as obtained from
two different methods are shown in Figures 9 and 10,
respectively. First, hf/hmax in all the curves is distinctly less
than 0.7, indicating little pile-up occurred; thus, its influence
on the computing of contact areas could be ignored. In the
method G-Series Hardness and Modulus via Cycles Load
Control, as the same maximum loading was carried out in
each test and the loadunload cycles are similar thus the
hysteresis (area between load and unload curves) can be

WWW.MATERIALSVIEWS.COM

FULL PAPER

used to measure the plastic deformation produced during

the loading part of the cycle. First, the area of PVDF-FP3% is
smallest, indicating a small plastic deformation and better
hardness for this film. Then the areas of PVDF-FPi% follow
the order PVDF-FP3% > PVDF-FP0% > PVDF-FP8% > PVDFFP5%, very similar to the variation trend of the hardness and
elastic modulus as listed in Table 1. In the method G-Series
Basic Hardness, Modulus at a Depth, as the same maximum
indentation depth was kept in each test and the projected
contact areas are similar, then the loading at this point is a
measure of hardness, and the slope of the upper portion of
the unload curve corresponds to the elastic modulus.48 It
can be found from Figure 10 that the curves of PVDF-FP3%
always exhibit the maximum loading among others, indicating a better hardness for it. It is hard to compare the slope
of the upper portion of the unload curve directly; however,
PVDF-FP3% and PVDF-FP0% always reveal the bigger slope,
indicating the elastic moduli for these two films are greater
than those of the other two. This is in accordance with the
results in Table 1.
Nanoscratch Profiles
The influence of different FP-POSS contents on the effective
scratch profiles of PVDF-FPi% is shown Figure 11(ad),
where the original morphology profile in prescan is not
shown since it has been compensated for computing the
effective scratch and residual depth in the scratch-scan and
postscan, respectively. The scratch-scan and postscan profiles
can be divided into two regions in the horizontal displacement: (1) the superficial region (050 lm), in which the tip
does not penetrate the sample and the case is very similar
to the prescan; (2) the scratch (for scratch-scan) or residual
(for postscan) depth region, in which the profile is essentially determined by the scratch force, hardness, and plasticity, showing the scratch resistance, recovery behavior, and
the elastoplastic deformation of a sample.58 In case of a

FIGURE 9 Loadunload cycles for PVDF-FPi% (i 0, 3, 5, and 8)

as obtained from the G-Series Hardness and Modulus via
Cycles Load Control method.

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

1605

FULL PAPER

WWW.POLYMERPHYSICS.ORG

JOURNAL OF
POLYMER SCIENCE

FIGURE 10 Loadunload cycles for PVDF-FPi% (i 0, 3, 5, and 8) at maximum indentation depth of (a) 500 nm, (b) 1000 nm, (c)
1500 nm, and (d) 2000 nm as obtained from the G-Series Basic Hardness, Modulus at a Depth method.

coating or thin film on a substrate, there should be another

more region in which the fluctuant profile is related to the
adhesive strength between the film and the substrate.59 This
is always used to compute the critical load (Lc), which is
very valuable to characterize some failure behaviors caused
by scratch testing, such as the adhesion strength, coating
cracking, delamination and brittle fracture, and so on.60 In
this work, only the former two regions are involved and no
any critical load values were obtained.
The effective scratch depth reveals the deformation resistance and hardness of the samples at a specific load, the
harder the sample is and the shallower the depth should be.
However, nearly all the samples possess similar effective
scratch depth [Fig. 11(ad)] (within the range of 50300
lm), which indicates FP-POSS is probably not helpful to
enhance the scratch resistance of PVDF. The effective residual depth show that the additions of FP-POSS even reduce
the elasticity recovery of PVDF because larger plastic deformations (residual depth) can be found in Figure 11(c, d). It
seems that the content of FP-POSS has brought little differ-

1606

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

ence to this influence because the areas between the residual

depth curve and the upper horizontal axis in these three figures almost are the same size. Specific values of residual
scratch depth and pile up height at cross profile (located at
the scratch load of 0.5 mN) obtained after the test (listed in
Table 2) have provided further evidence for this.
The additions of FP-POSS will not increase but probably
weaken the scratch resistance of PVDF to some extent,
seems to be against the nanoindentation results in hardness
because PVDF-FP3% exhibits the best hardness and elastic
modulus properties among all samples. This should be
derived from the surface roughness, which has been proved
to affect the abrasion or scratching profile of a film.25,61,62
The rougher the surface is, the weaker the scratch resistance
would be exhibited.63 All the surface roughness values for
PVDF-FPi% are listed in Table 2. Not surprisingly, PVDF-FP3%
possesses the roughest surface, which has been confirmed
by several more tests at different positions. Although the
best hardness is shown for PVDF-FP3%, the roughest surface
determines a weaker scratch resistance for it. To PVDF-FP5%

JOURNAL OF
POLYMER SCIENCE

FULL PAPER

WWW.POLYMERPHYSICS.ORG

FIGURE 11 Influence of different FP-POSS contents on the scratch profiles of PVDF-FPi%, i (a) 0, (b) 3, (c) 5, and (d) 8.

and PVDF-FP8%, their relative softer hardness makes it be

easily understood that they possess weaker scratch resistance. Moreover, neat PVDF (PVDF-FP0%) possesses the best
scratch resistance may be caused by two factors: a smoother
surface and a higher crosslinkage among particles, which is
very likely to exist in PVDF.
Engineering StressStrain Response
The engineering stressstrain responses for PVDF-FPi% as
obtained from the nanotensile testing are shown in Figure
12. Some key points on these curves, such as the offset yield
strain (0.2% offset), peak stress, and break point, have been
labeled as letters Y, P, and B, respectively. The subscript
numbers point out which sample these values belong to. The
full stressstrain curves can be divided into three regions:
(1) the elastic region (from the start point to Y), in which
recoverable (pure elastic) deformation happens and the
Hookes law is satisfied between the stress and strain, and
thus, the Youngs modulus can be computed. The stress at
the offset yield strain point (ry) is called the tensile yield
strength,63 at which a material begins to deform plastically;
(2) the strain hardening region (Y!P), in which the stress is
no longer proportional to strain and permanent nonrecoverable (plastic) deformation occurs. The stress at the maximum
on the engineering stressstrain curve (point P) is defined as
the tensile strength or ultimate strength.64 The stressstrain
behavior in this region is always described by the formula r
Ken,63 where K (the strength coefficient) and n (the strain

WWW.MATERIALSVIEWS.COM

hardening exponent) are constants. The value of the strain

hardening exponent lies between 0 and 1. A value of 0
means that a material is a perfectly plastic solid, while a
value of 1 represents a 100% elastic solid; (3) the unstable
failure region (P!B), in which a small constriction or neck
begins to form at some point, and fracture ultimately occurs
at the neck. The fracture strength63 corresponds to the stress
at fracture and the area under the stressstrain curve up to
the point of fracture is always used to ascertain the fracture
toughness [for the quasi static (low strain rate) situation],52,63 which is a measure of the ability of a material to
absorb energy up to fracture and always used to describe
the brittleness or ductility of a material. All these key mechanical properties, for PVDF-FPi% in the three regions, as
well as the brittleness index (H/Kc)65 are listed in Table 3.
TABLE 2 The Surface Roughness (Ra), Residual Scratch Depth
(Dr), Pile-Up Height (Hp), and Scratch Width (Wr) at cross
profile of PVDF-FPi% (i 5 0, 3, 5, and 8)
FP-POSS
(wt %)

Ra (nm)

Dr (nm)

Hp (nm)

Wr (um)

655 (138)

1,130 (350)

469 (322)

41.7 (9.7)

1,031 (163)

1,552 (1,191)

744 (806)

32.6 (17.2)

587 (143)

1,065 (1,032)

789 (682)

53.9 (21.1)

689 (103)

1,366 (182)

635 (232)

45.7 (10.1)

Values in parentheses indicate (6) standard deviations.

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

1607

FULL PAPER

JOURNAL OF
POLYMER SCIENCE

WWW.POLYMERPHYSICS.ORG

FIGURE 12 Engineering stressstrain responses for PVDF-FPi%, i (a) 0, (b) 3, (c) 5, and (d) 8 at strain rate of 103 s1. The letters
Y, P, and B locate the positions of offset yield strain, peak stress, and break point, respectively.

The Youngs modulus of PVDF-FPi% shows the same variation trend with the indentation results, which follows the
orders PVDF-FP3% > PVDF-FP0% > PVDF-FP5% > PVDFFP8%. However, the specific values are all much lower here.
This is probably resulted from two factors: (1) the uncertainty in measuring the sizes of samples, which always
brings positive errors (greater than the real value) in computing the equivalent diameters of cross sections because
the deformations and fracture always occur at the position
with a minimum size of the cross section but hard to be
found when the sizes of the samples were measured; (2)

some micro defects in the samples, which will influence the

results in a tensile test while perhaps have no effect on a
nanoindentation one because only some small local areas are
involved in such case. All these two factors are likely to
make the test values of Youngs modulus are lower than real
ones. Anyway, suppose these influences are similar to all
samples, it is obvious that the elastic modulus of PVDF was
reinforced by a small content of FP-POSS (3 wt %) but weakened by further additions of FP-POSS (5 and 8 wt %). The
values of yield strength (YS), tensile strength (TS), and fracture strength (FS) almost show the same trend and indicate

TABLE 3 The Youngs Modulus (E), Offset Yield Strain (ey), Yield Strength (YS), Tensile Strength (TS), Fracture Strength (FS),
Fracture Strain (ef), Fracture Toughness (Kc), and Brittleness Index (H/Kc) of PVDF-FPi% (i 5 0, 3, 5, and 8) as obtained from
stressstrain responses and hardness values
FP-POSS (wt %)

E (GPa)

ey (%)

YS (MPa)

TS (MPa)

FS (MPa)

1.71 (0.13)

1.77 (0.15)

26.67 (1.13)

36.51 (2.28)

35.76 (2.07)

2.00 (0.11)

1.77 (0.21)

30.44 (1.36)

40.75 (2.28)

39.61 (2.37)

1.59 (0.12)

1.97 (0.31)

27.94 (4.11)

36.96 (4.86)

1.17 (0.26)

1.90 (0.26)

19.96 (2.74)

30.64 (2.22)

Values in parentheses indicate (6) standard deviations.

1608

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

ef (%)

Kc (MPa)

H/Kc

5.90 (1.00)

1.70 (0.40)

102.35

7.23 (1.03)

2.42 (0.51)

71.43

32.98 (5.32)

13.87 (3.02)

4.52 (1.38)

34.96

26.63 (2.70)

24.20 (1.61)

6.50 (0.93)

20

JOURNAL OF
POLYMER SCIENCE

the FP-POSS contents have the same effect based on PVDF.

The values of offset yield strain (ey) are almost the same for
PVDF-FP3% and PVDF-FP0% and a little higher for PVDFFP5% and PVDF-FP8%, shows FP-POSS will not decrease but
slightly increase the elastic deformation range of PVDF.
The elongation at break (fracture strain) and fracture toughness values reveal an undoubtedly influence of FP-POSS on
PVDF. The fracture strain and fracture toughness increase
with the increasing content of FP-POSS. The value of ef
increases from 5.9 to 24.2%, giving PVDF-FP8% the toughness 300% higher than that of neat PVDF. It can also be
seen that the brittleness index of PVDF-FP3%, PVDF-FP5%,
and PVDF-FP8% are about 43, 193, and 412% lower than
that of neat PVDF, respectively. Dramatic enhancements in
fracture toughness have been observed when the FP-POSS
content reaches to some extent. This result is very valuable
in some applications of PVDF, such as energy harvesting,
adaptive inflatable structures in aerospace engineering, large
deformational piezoelectric actuators, and so on, in which a
better toughness is vital and needs to be strictly satisfied.
The mechanism to the toughness enhancements can probably be understood from the morphology change in different
PVDF-FPi% images. In contrast to the neat PVDF, more separate and less crosslinked particles in PVDF nanocomposites
(especially in PVDF-FP5% and PVDF-FP8%) lead to a structure
much more conductive to plastic flow under applied stress
and probably result in a more efficient energy dissipation,
thereby reducing crack formation. This is similar to the work
by Shah et al.52 and may be useful to explain the research by
Martins et al.,43 where they found that POSS were acting as
lubricant in PVDF matrix. More interpretations can be
quoted from MD studies.66 PVDF nanocomposites might possess remarkable toughness derived from the FP-POSS nanoparticles to dissipate energy due to their mobility under
applied stress. Comparable time scales for motion of FPPOSS molecules and PVDF chains results in the enhanced
toughness. Our results can also provide more evidence of a
good nanoscale dispersion of FP-POSS in PVDF matrix
because it has been proved that good nanoscale dispersion
is critical for improved toughness.52
It should be noticed that the properties of PVDF-FP3% are
unique, as it shows increases in both stiffness and toughness.
Most rigid fillers produce increase in stiffness versus that of
the unfilled polymer. However, it is generally associated with
a significant decline in toughness.67 On the contrary, an
improvement in toughness is always accompanied by a
decline in stiffness for soft organic fillers.68 It seems that FPPOSS has acted as both rigid and soft fillers in PVDF matrix
simultaneously. However, this has been influenced greatly by
the content of POSS. When the content is low, the role of FPPOSS is thus the rigid fillers to stiffen the material. While
when the content is high, the toughness enhancements are
dominated, and the stiffness declines; thus, the role of FPPOSS is the soft fillers. Similar results were obtained by
Kopesky et al.69 when cyclohexyl-POSS was incorporated
into the poly(methyl methacrylate) matrix; however, the
mechanism was not analyzed. It seems that this effect is, to

WWW.MATERIALSVIEWS.COM

FULL PAPER

WWW.POLYMERPHYSICS.ORG

TABLE 4 The Storage Modulus (G0 ), Loss Modulus (G00 ), and

Loss Tangent (tan d) of PVDF-FPi% (i 5 0, 3, 5, and 8) as
obtained from the nanotensile testing at a harmonic force of
4.5 mN and a frequency of 20 Hz
FP-POSS (wt %)

G0 (GPa)

G00 (MPa)

tan d

1.80 (0.11)

159.6 (31.5)

0.089 (0.022)

2.28 (0.18)

221.3 (48.0)

0.097 (0.021)

2.13 (0.20)

232.3 (56.7)

0.109 (0.025)

1.86 (0.26)

207.9 (28.9)

0.113 (0.015)

Values in parentheses indicate (6) standard deviations.

some extend, a common pattern of POSS when it is used as

fillers to a polymer matrix.
Dynamic Mechanical Properties
Dynamic mechanical properties of PVDF-FPi%, such as the
storage modulus, loss modulus, and loss tangent can also be
obtained in the nanotensile testing. Dynamic mechanical
properties are helpful in understanding and confirming the
energy dissipation mechanism for PVDF-FPi% as analyzed
previously. Table 4 lists the values of storage modulus (G0 ),
loss modulus (G00 ), and loss tangent (tan d) for PVDF-FPi% (i
0, 3, 5, and 8) obtained from the nanotensile testing at
the harmonic force of 4.5 mN and the frequency of 20 Hz. It
can be observed that all the values of storage modulus are
higher than the corresponding values of Youngs modulus as
obtained from the elastic portion of stressstrain responses
(listed in Table 3). This is not surprising as the storage modulus is averaged within the whole range of stressstrain
responses, in which modulus with higher values is kept after
the yield strain due to a full elastic deformation. Loss modulus values for PVDF-FPi% (i 3, 5, and 8) show that the
addition of FP-POSS obviously increases the viscosity of
PVDF because they are much higher than that of the neat
PVDF. Compared with the storage and loss modulus, loss tangent is more reliable to measure viscoelastic properties
because it is the ratio of loss to storage modulus thus less
influenced by the uncertain factors in measuring or some
unpredictable defects in materials. The values of loss tangent
in Table 4 show an evidently viscosity increasing with the
increasing content of FP-POSS. The variation of loss tangent
is in good agreement with the previous results in toughness.
A higher value of the loss tangent, reflecting greater viscous
flow, is corresponded with a more efficient energy dissipation. This is expected, since particles with lower crosslinkage
level resulted from further addition of FP-POSS lead to a
structure much more favorable for plastic flow under applied
stress.
CONCLUSIONS

In this article, the morphology and some key mechanical

properties of PVDF and FP-POSS nanocomposites were
explored using experiments. PVDF/FP-POSS nanocomposites
can be prepared by the solvent evaporation method. Optical
and SPM microscopy results showed FP-POSS and PVDF
were fully miscible and FP-POSS acted as the nucleating

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

1609

FULL PAPER

WWW.POLYMERPHYSICS.ORG

agent in PVDF matrix. A small content of FP-POSS could

make PVDF chains be nucleated to form bigger spherical particles and a further addition of FP-POSS formed more separate and less crosslinked particles due to a more complete
nucleation. XRD patterns and FTIR spectra showed that there
are two different crystalline phases in the nanocomposites
and FP-POSS has not changed the crystalline phase of PVDF.
As in XRD patterns, up to 5-wt % FP-POSS content, there
was an increase in both a and b phase. Hardness and elastic
modulus of PVDF/FP-POSS nanocomposites can be measured
by nanoindentation test with a very low strain rate and the
adequate indentation depths. PVDF-FP3% possessed the best
elastic properties and hardness and further additions of FPPOSS made the stiffness of PVDF/FP-POSS nanocomposites
decrease. Nanoscratch testing can be used to measure the
scratch resistance of PVDF/FP-POSS nanocomposites. The
scratch-testing profiles were clearly influenced by different
FP-POSS contents in PVDF-FPi%. Neat PVDF exhibited the best
scratch resistance while the additions of FP-POSS weakened
the scratch resistance due to a rough surface or a less crosslinked structure. Nanotensile testing results showed both the
stiffness and toughness of PVDF-FP3% were enhanced and further additions of FP-POSS brought surprisingly enhancements
in toughness of the nanocomposites while associated with a
decline in stiffness. Dynamical mechanical properties indicated the viscosity of the nanocomposites increased with the
increasing FP-POSS contents in PVDF-FPi%. Because of the
unique organicinorganic structure, FP-POSS probably acted
as rigid inorganic fillers in PVDF matrix at a low content while
soft organic fillers at relative high contents. The great
improvement effects on toughness and the increasing viscosity resulted in further additions of FP-POSS should be derived
from the more separate and less crosslinked particles in the
nanocomposites, and the nanoscale size of POSS compounds,
which made the materials, are easier to form plastic flow and
more efficient on energy dissipation. The results in this work
are helpful in realizing the improvement effects of different
FP-POSS contents on PVDF and understanding the mechanism
for these different effects. Another important advantage of FPPOSS on PVDF, we found recently, is that the piezoelectric
properties of PVDF are only very slightly influenced probably
due to the similar CAF bonds in their molecules. This is very
important to the application of PVDF and will be discussed in
our future work.

1610

JOURNAL OF
POLYMER SCIENCE

3 Kerur, S. B.; Ghoshy, A. Int. J. Struct. Stab. Dyn. 2011, 11,

237255.
4 Sun, C. L.; Shi, J.; Bayerl, D. J.; Wang, X. D. Energy Environ.
Sci. 2011, 4, 45084512.
5 Vatansever, D.; Hadimani, R. L.; Shah, T.; Siores, E. Smart
Mater. Struct. 2011, 20, 055019.
6 Celina, M.; Dargaville, T. R.; Assink, R. A.; Martin, J. W. High
Perform. Polym. 2005, 17, 575592.
7 Dargaville, T. R.; Celina, M.; Chaplya, P. M. J. Polym. Sci.
Part B: Polym. Phys. 2005, 43, 13101320.
8 Dargaville, T. R.; Celina, M.; Martin, J. W.; Banks, B. A.
J. Polym. Sci. Part B: Polym. Phys. 2005, 43, 25032513.
9 Dargaville, T. R.; Elliott, J. M.; Celina, M. J. Polym. Sci. Part
B: Polym. Phys. 2006, 44, 32533264.
10 Xu, H. P.; Dang, Z. M.; Jiang, M. J.; Yao, S. H.; Bai, J. J.
Mater. Chem. 2008, 18, 229234.
11 Li, Y. J.; Shimizu, H. Macromolecules 2008, 41, 53395344.
12 Malmonge, L. F.; Langiano, S. D. C.; Cordeiro, J. M. M.; Mattoso, L. H. C.; Malmonge, J. A. Mater. Res.-Ibero-Am. J. Mater.
2011, 13, 465470.
13 Thongsanitgarn, P.; Watcharapasorn, A.; Jiansirisomboon,
S. Surf. Rev. Lett. 2010, 17, 17.
14 Ratajska, A. M.; Kulak, W. P.; Poeppel, A.; Seyler, A.; Roslaniec, Z. Pol. J. Chem. Technol. 2009, 11, 2734.
15 Wang, Y. J.; Kim, D. Electrochim. Acta 2007, 52, 31813189.
16 Ma, H. Y.; Yang, Y. M. Polym. Test. 2008, 27, 441446.
17 Nicotera, I.; Coppola, L.; Oliviero, C.; Castriota, M.; Cazzanelli, E. Solid State Ionics 2006, 177, 581588.
18 Fu, B. X.; Gelfer, M. Y.; Hsiao, B. S.; Phillips, S.; Viers, B.;
Blanski, R.; Ruth, P. Polymer 2003, 44, 14991506.
19 Deng, J.; Polidan, J. T.; Hottle, J. R.; Farmer-Creely, C. E.;
Viers, B. D.; Esker, A. J. Am. Chem. Soc. 2002, 124, 15194.
20 Fu, B. X.; Hsiao, B. S.; White, H.; Rafailovich, M.; Mather, P.
T.; Jeon, H. G.; Phillips, S.; Lichtenhan, J.; Schwab, J. Polym.
Int. 2000, 49, 437440.
21 Mabry, J. M.; Vij, A.; Iacono, S. T.; Viers, B. D. Angew.
Chem. Int. Ed. 2008, 47, 41374140.
22 Poon, B.; Rittel, D.; Ravichandran, G. Int. J. Solids Struct.
2008, 45, 60186033.
23 Oliver, W. C.; Pharr, G. M. J. Mater. Res. 2004, 19, 320.
24 Li, X.; Diao, D.; Bhushan, B. Acta Mater. 1997, 45,
44534461.
25 Zhang, X. W.; Hu, L. J.; Sun, D. Z. Acta Mater. 2006, 54,
54695475.
26 Field, J. S.; Swain, M. V. J. Mater. Res. 1995, 10, 101112.
27 Swain, M. V. Mater. Sci. Eng. A 1998, 253, 160166.

ACKNOWLEDGMENTS

28 Wright Wendelin, J.; Nix, W. D. J. Mater. Res. 2009, 24,

863871.

The authors thank the National Natural Science Foundation of

China (11102053), the Science and Technology Innovation Talents Special Fund of Harbin (Grant No. 2012RFQXG001), and
the Fundamental Research Funds for the Central Universities
(Grant No. HIT. NSRIF. 2010070) for the financial support of
this research.

29 Loubet, J. L.; Oliver, W. C.; Lucas, B. N. J. Mater. Res. 2000,

15, 11951198.
30 Oyen, M. L. Acta Mater. 2007, 55, 36333639.
31 Goodall, R.; Clyne, T. W. Acta Mater. 2006, 54, 54895499.
32 Koch, T.; Evaristo, M.; Pauschitz, A. Thin Solid Films 2009,
518, 185193.

REFERENCES AND NOTES

33 Tan, E. P. S.; Ng, S. Y.; Lim, C. T. Biomaterials 2005, 26,

14531456.

1 Park, G.; Ruggiero, E.; Inman, D. J. Smart Mater. Struct.

2002, 11, 147155.

34 Kiuchi, M.; Matsui, S.; Isono, Y. J. Microelectromech. Syst.

2007, 16, 191201.

2 Seminara, L.; Capurro, M.; Cirillo, P.; Cannata G.; Valle, M.

Sens. Actuators A Phys. 2011, 169, 4958.

35 Seydel, T.; Knoll, W.; Greving, I.; Dicko, C.; Koza, M. M.;
Krasnov, I.; Muller, M. Phys. Rev. E 2011, 83, 016104.

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

JOURNAL OF
POLYMER SCIENCE

WWW.POLYMERPHYSICS.ORG

FULL PAPER

36 Wu, C. L.; Lin, H. C.; Hsu, J. S.; Yip, M. C.; Fang, W. L. Thin
Solid Films 2009, 517, 48954901.

54 Zhou, Z. Y.; Cui, L. M.; Zhang, Y.; Zhang, Y. X.; Yin, N. W.

Eur. Polym. J. 2008, 44, 30573066.

37 Chang, W. Y.; Fang, T. H.; Lin, Y. C. J. Polym. Sci. Part B:

Polym. Phys. 2008, 46, 949958.

55 Burgess, T.; Laws, K. J.; Ferry, M. Acta Mater. 2008, 56,

48294835.

38 Chae, D. W.; Hong, S. M. J. Polym. Sci. Part B: Polym. Phys.

2010, 48, 23792385.

56 Fujisawa, N.; Swain, M. V. J. Mater. Res. 2008, 23, 637641.

39 Guney, H. Y. J. Polym. Sci. Part B: Polym. Phys. 2005, 43,

28622873.

58 Tayebi, N.; Polycarpou, A. A.; Conry, T. F. J. Mater. Res.

2004, 19, 17911802.

40 Ma, W. Z.; Wang, X. L.; Zhang, J. J. Polym. Sci. Part B:

Polym. Phys. 2010, 48, 21542164.

59 Sanchez, J. M.; El-Mansy, S.; Sun, B.; Scherban, T.; Fang,

N.; Pantuso, D.; Ford, W.; Elizalde, M. R.; Martnez-Esnaola, J.
M.; Martn-Meizoso, A.; Gil-Sevillano, J.; Fuentes, M.; Maiz, J.
Acta Mater. 1999, 47, 44054413.

41 Fang, F.; Zhang, M. Z.; Huang, J. F. J. Polym. Sci. Part B:

Polym. Phys. 2005, 43, 32553260.
42 Salimi, A.; Yousefi, A. A. J. Polym. Sci. Part B: Polym. Phys.
2004, 42, 34873495.
43 Martins, J. N.; Bassani, T. S.; Oliveira, R. V. B. Mater. Sci.
Eng. C 2012, 32, 146151.
44 Zeng, F. L.; Sun, Y.; Zhou, Y.; Li, Q. K. Modell. Simul. Mater.
Sci. Eng. 2009, 17, 075002.
45 Zeng, F. L.; Sun, Y.; Zhou, Y.; Li, Q. K. Modell. Simul. Mater.
Sci. Eng. 2011, 19, 025005.
46 Hay, J. L.; Pharr, G. M. Instrumented Indentation Testing;
ASM International: Ohio, 2000; pp 232243.
47 Oliver, W. C.; Pharr, G. M. MRS Bull. 2010, 35, 897907.
48 Pharr, G. M.; Oliver, W. C. MRS Bull. 1992, 17, 2833.
49 Huang, L.; Bonifacio, C.; Song, D.; Benthema, K. V.; Mukherjeea, A. K.; Schoenung, J. M. Acta Mater. 2011, 59, 51815193.
50 Nunes, J. S.; Wu, A.; Gomes, J.; Sencadas, V.; Vilarinho, P.
M.; Lanceros-Mendez, S. Appl. Phys. A 2009, 95, 875880.
51 Sencadas, V.; Gregorio Filho, R.; Lanceros-Mendez, S. J.
Non-Cryst. Solids 2006, 352, 22262229.
52 Shah, D.; Maiti, P.; Gunn, E. Adv. Mater. 2004, 16, 11731177.
53 Li, X. F.; Lu, X. L. J. Appl. Polym. Sci. 2006, 101, 29442952.

WWW.MATERIALSVIEWS.COM

57 Munz, M. J. Phys. D: Appl. Phys. 2006, 39, 40444058.

60 Kim, B. R.; Ko, M. J. Thin Solid Films 2009, 517, 32163221.

61 Chang, Y. M.; Wen, H.-C.; Yang, C. S.; Lian, D.; Tsai, C. H.;
Wang, J. S.; Wu, W. F.; Chou, C. P. Microelectron. Reliab. 2010,
50, 11111115.
62 Kim, H. H.; Cho, S. H.; Kang, C. G. Mater. Sci. Eng. A 2008,
485, 272281.
63 Callister, W. D. Jr. Materials Science and Engineering: An
Introduction, 7th ed.; Wiley: New York, 2006.
64 Dowling, N. E. Mechanical Behavior of Materials, 2nd ed.;
Prentice Hall PTR: Paramus, New Jersey, 1998.
65 Taylor, L. J.; Papadopoulos, D. G.; Dunn, P. J.; Bentham, A.
C.; Dawson, N. J.; Mitchell, J. C.; Snowden, M. J. Org. Process
Res. Dev. 2004, 8, 674679.
66 Gersappe, D. Phys. Rev. Lett. 2002, 89, 058301.
67 Nielsen, L. E. J. Appl. Polym. Sci. 1966, 10, 97103.
68 Spirkova, M.; Duchek, P.; Strachota, A.; Poreba, R.; Kotek,
J.; Baldrian, J.; Slouf, M. J. Coat. Technol. Res. 2011, 8,
311328.
69 Kopesky, E. T.; McKinley, G. H.; Cohen, R. E. Polymer 2006,
47, 299309.

JOURNAL OF POLYMER SCIENCE: PART B: POLYMER PHYSICS 2012, 50, 15971611

1611