Physica Scripta

Phys. Scr. 93 (2018) 105801 (11pp) https://doi.org/10.1088/1402-4896/aad749

Structural and electrical properties of

ferroelectric poly(vinylidene fluoride) and
mechanically activated ZnO nanoparticle
composite films
A Peleš1,8, O Aleksić2, V P Pavlović3, V Djoković4, R Dojčilović4, Z Nikolić5,
F Marinković5, M Mitrić4, V Blagojević1 , B Vlahović6,7 and V B Pavlović1
1
Institute of Technical Sciences of SASA, Knez Mihailova 35/IV 11000 Belgrade, Serbia
2
Institute for Multidisciplinary Research, University of Belgrade, 11000 Belgrade, Serbia
3
Faculty of Mechanical Engineering, University of Belgrade, 11000 Belgrade, Serbia
4
Vinča Institute of Nuclear Sciences, University of Belgrade, PO Box 522, 11001 Belgrade, Serbia
5
Faculty of Physics, University of Belgrade, Studentski trg 12-16, 11000 Belgrade, Serbia
6
North Carolina Central University, Durham, NC 27707, United States of America
7
NASA University Research Center for Aerospace Device Research and Education and NSF Center of
Research Excellence in Science and Technology Computational Center for Fundamental and Applied
Science and Education, NC, United States of America

E-mail: [email protected]

Received 6 February 2018, revised 27 July 2018

Accepted for publication 1 August 2018
Published 24 August 2018

Abstract
The influence of the mechanical activation of ZnO nanoparticle fillers on the structural and
electrical properties of the matrix of poly(vinylidenefluoride)–ZnO (PVDF–ZnO) films was
investigated. Transmission electron microscopy and scanning electron microscopy analyses
showed that mechanical activation in a high energy planetary ball mill reduces the size of ZnO
particles. X-ray diffraction and Raman spectroscopy revealed that PVDF crystallized
predominantly as the γ-phase. Non-activated ZnO filler reduces the degree of the crystallinity of
the matrix and promotes crystallization of α-phase of PVDF in the film, while the fillers
activated for 5 and 10 min induce crystallization of β-phase, indicating that mechanical
activation of the filler can be used as a general method for fabrication of PVDF composites with
increased content of piezoelectric β-phase crystals. Dielectric spectroscopy measurements show
that polymer composite with the high content of β-phase (with ZnO filler activated for 5 min)
exhibits the highest value of dielectric permittivity in 150–400 K range of temperatures. Kinetic
analysis shows combined effects of increased surface area and increased concentration of surface
defects on the interactions between polymer chains and activated nanoparticles.

Keywords: polymer composites, mechanical activation, crystallization, dielectric properties,

PVDF, ZnO
(Some figures may appear in colour only in the online journal)

Introduction

Poly(vinylidene fluoride) (PVDF) is a semi-crystalline poly-

mer with excellent pyro-and piezo-electric properties [1, 2].
8
Author to whom any correspondence should be addressed. Compared to piezoelectric ceramics, PVDF has lower

0031-8949/18/105801+11$33.00 1 © 2018 IOP Publishing Ltd Printed in the UK

Phys. Scr. 93 (2018) 105801 A Peleš et al

dielectric permittivity but offers certain advantages for sen- nanocomposite films [32, 33]. However, in spite of improved
sing applications due to low density, flexibility and tough- properties of the films after the introduction of the filler, ZnO
ness. Also, being an engineering plastic, it can easily be nanoparticles have not induced changes in phase crystal-
processed into technologically useful forms [3, 4]. For this lization (i.e. post-crystallization phase composition) of PVDF
reason, polyvinylidene fluoride (PVDF) has been widely used matrix. Another approach suggested was to use PVDF co-
in MEMS technology as well as for ultrasound transducer and polymers such as poly(vinylidene fluoride-trifluoroethylene)
medical ultrasound applications. PVDF may crystallize in five (PVDF-TrFE) as a matrix for ZnO nanoparticles [34, 35]. The
distinct crystalline phases, usually denoted as α, β, γ, δ and ε presence of the trifluoroethylene co-monomers in the mac-
[5, 6]. The most common α-phase, typically formed during romolecular chains favours β-phase crystallization of PVDF
the crystallization from melt, is characterized by alternating monomers and the ZnO will additionally improve the piezo-
s-trans and s-gauche bonds (TGTG), hence it is non-polar. electric properties of the films [36].
Both β and γ phases are electroactive, but β-phase exhibits In our previous studies [37, 38], it was shown that
the strongest ferroelectric response due to a particular crystal mechanical activation of BaTiO3 filler particles prior to
structure, with all C–C bonds in the trans conformation mixing with PVDF may favour β-phase crystallization in the
(TTTT) [6, 7]. Therefore, considerable research effort has matrix. Mechanical activation was used to reduce the size of
been devoted to optimization of the conditions that will BaTiO3 particles, without affecting their ferroelectric tetra-
induce β-phase crystallization in PVDF films [8–13]. gonal crystal structure [39]. In this way, the filler could
One of the methods suggested was mixing of PVDF with increase the dielectric permittivity due to its higher dielectric
nanostructured fillers (montmorillonite, SiO2 etc), which permittivity, and through pronounced crystallization of polar
would act as nucleation centres for crystallization of the β- β-phase crystals, achieving the same effect on the dielectric
phase [12, 13]. Addition of carbon nanofibers was found to permittivity using lower concentrations of the filler. Unlike in
have a considerable impact on phase composition and crys- the case of nanostructured organosilicate fillers [12, 13],
tallization of the β-phase [14]. It was also found that the reduction of the size of the ceramic particles prior to their
combination of two fillers (multi-walled carbon nanotubes mixing with PVDF does not necessarily lead to a significant
and polypyrrole) may exhibit a synergistic effect on the increase in dielectric permittivity of the composite, suggesting
hybrid system’s electrical properties, due to specific changes that the small size of ceramic particles of the filler is not the
in morphology [15–17]. In addition, the application of a sole necessary condition for the β-phase crystallization of the
stretching process resulted in crystallization of almost pure β- PVDF [23]. In addition to the small diameter of BaTiO3
phase (>96%) in both PVDF and PVDF-carbon nanofiber particles, their active surfaces play a crucial role in the
composites [18]. Given that the combination of dielectric nucleation processes that result in formation of β-phase
fillers and PVDF matrix has been shown to be a good route crystals [37, 38]. Here we conducted a study with another
towards increasing the dielectric permittivity of the material, type of particles, zinc oxide (ZnO), to investigate whether
the introduction of high-dielectric permittivity fillers such as mechanical activation can be used as a general method for the
barium titanate, BaTiO3, represented a next logical step surface modification of the filler that may induce crystal-
[19–23]. However, it was shown that high BaTiO3 contents lization of β-phase upon its mixing with PVDF. In addition,
(20%–40%) are necessary to achieve a significant increase in the effect of mechanical activation of the filler on dielectric
dielectric permittivity [24], while the particles tend to properties of the composites will be investigated.
agglomerate in the low surface energy polymer matrix, which
can reduce the dielectric strength by triggering percolation
effects. In order to improve the dispersion of the BaTiO3 Experimental
particles in the PVDF matrix, and consequently improve the
dielectric strength, their surface can be modified with penta- Poly(vinylidene fluoride) (Mw∼530 000) and ZnO powder
fluorobenzylphosphonic acid [20] or fluoroalkylsilane [25]. (99.9% p.a.) were purchased from Sigma Aldrich and used
The observed increase in the values of dielectric permittivity without further purification. ZnO powder was mechanically
of the composites was mainly due to the contribution of the activated for 5 and 10 min in a Fritch Pulverissete 5 high
filler, however, these studies do not report on the effect of energy planetary ball mill with wolfram carbide ball. Ball to
ceramic nanoparticles on the crystallization of different pha- sample mass ratio was 40:1, rotation speed 390 rpm. PVDF
ses in the PVDF. solution (2 wt%) was prepared by dissolving 3 g of polymer
ZnO is semiconductor with wide direct band gap in 150 ml dimethyleformamide (DMF). Non-activated and
(3.37 eV) and large excitation binding energy at room temp- mechanically activated ZnO powders (0.1 g) were dispersed
erature and finds applications as a material for gas sensors, in 50 ml DMF and sonicated in a BransonW-450 D Digital
solar cells, display screens, photocell electrodes and UV-light Sonifier for 20 min at 20% amplitude (80 W). After that,
emitting diodes [26–30]. Nanostructured ZnO [31] has been 18.75 ml of ZnO dispersions were mixed with 30 ml of the
considered as a good filler material for the fabrication of 2 wt% PVDF solutions. PVDF and nanocomposite films (50
polymer nanocomposites, due to high dielectric permittivity μm) were obtained by casting the solutions into Petri dishes.
and good physical properties. Recent studies report that the Morphology of ZnO powders and PVDF–ZnO nano-
mixing of ZnO nanoparticles with PVDF may result in composite films was investigated by a transmission electron
interesting optical and electrical properties of the obtained microscope (JEOL 100CX II instrument) at 100 kV and

2
Phys. Scr. 93 (2018) 105801 A Peleš et al

Table 1. Notation of the samples, preparation conditions and the average crystal sizes and microstrain of the filler powders. Complete
Rietveld refinement is shown in figure 3.
Sample wt% ZnO powder Mechanical activation time Average crystallite size of ZnO (nm) Microstrain (%)
PVDF 0 / / /
P5Z0 5 0 61±4 0.53±0.02
P5Z5 5 5 49±3 0.74±0.02
P5Z10 5 10 44±3 0.93±0.02

Figure 1. TEM micrographs of (a) non-activated ZnO powder, (b) ZnO powder mechanically activated for 10 min (both images).

electron scanning electron microscope (JEOL JSM-6390) at (S=πD2/4), ε′ is the real part of dielectric permittivity and
10 kV. For TEM measurements ZnO powders were dispersed ε0 is the vacuum permittivity.
in alcohol and deposited on the carbon coated copper grid Fourier transform infrared (FTIR) spectroscopy was
mash. For SEM analyses, PVDF–ZnO nanocomposite films conducted using KBr pellets in 500–4000 cm−1 region with
were fractured after immersion in liquid nitrogen and covered Nicolet 6700 FTIR spectrometer (Thermo Scientific, USA).
with gold. X-ray diffraction patterns of the composites were
obtained on a Philips PW-1050 diffractometer with Cu–Kα
radiation and a step/time scan mode of 0.05° s−1, in Bragg–
Brentano geometry. Notation of the samples and preparation Results and discussion
conditions are given in table 1.
Raman scattering measurements were performed on a Mechanical activation of ZnO powder is expected to cause a
fully automated Raman microscope (Horiba JobinYvon Lab decrease in particle size [40] and an increase in mesopore
volume and non-monotonic trends of specific surface area
Ram ARAMIS) using He–Ne laser at 633 nm. The samples
with prolonged milling, caused by powder particle adhesion
were measured under a microscope using a 100× objective.
and breakage due to mechanical activation [41]. TEM
The data were collected over the Raman shift range of
micrographs of non-activated and mechanically activated
200–3200 cm−1, using a count time of 5 s with 5 averaging
(30 min) ZnO powders are shown in figure 1. Non-activated
cycles and 1800 lines mm−1 grating.
ZnO powder consists mostly of particles of regular rectan-
Dielectric spectroscopy measurements were performed in
gular or hexagonal shape, which is consistent with hexagonal
a shielded cell, with circular electrodes (D=13 mm). Sam-
ZnO phase (figure 1(a)). Mechanically activated sample
ples in the form of discs (thickness d=0.03 mm) were cut shows irregularly shaped particles where smaller particles
from the middle of the sheets. Conductance (G) and suscep- exhibit, most likely due to highly activated surfaces, a ten-
tance (B) were measured as the real part, G, and the imaginary dency to agglomerate around larger particles (figure 1(b)).
part, B, of AC conductivity, using the Cp mode of the Rietveld analysis of XRD spectra (presented below) showed a
instrument (Y=G+iB). The measurements were performed decrease in ZnO crystallite size from ∼61 to 44 nm and an
on a Hameg 8118 instrument in the frequency range from 60 increase in microstrain up to 0.93%, due to creation of defects
Hz to 60 kHz and in the temperature range from 150 to 400 K during prolonged mechanical activation (table 1). Figure 2
using Temperature Controller Lake Shore 340. The heating shows SEM micrographs of fracture surfaces of PVDF and
rate and the acquisition step were 2 K min−1 and 10 K, PVDF–ZnO composites with non-activated and mechanically
respectively, while the applied voltage was 1.5 V. The char- activated (10 min) ZnO fillers. The fracture surfaces of the
acteristic dielectric parameters were calculated using the fol- composites show presence of ZnO agglomerates on the sur-
lowing relations: tan δ=G/B and B=2πfC, where f is face of the polymer and inside the polymer matrix, which
frequency and C is the capacity. In addition, C=ε′ε0S/d, appear as micrometer-sized particles in SEM images. Sample
where S/d describes the geometry of the samples with non-activated filler exhibits more surface agglomerates

3
Phys. Scr. 93 (2018) 105801 A Peleš et al

Figure 2. SEM micrographs of the fractured surfaces of (a) PVDF, (b) P5Z0 and (c) P5Z10.

(figure 2(b)), while sample with activated filler appears to ∼1431 cm−1 in our spectrum comes from γ-phase. The fact
exhibit smoother surface (resulting in lower contrast), sug- that one of the strongest lines in the Raman spectrum of the
gesting much better dispersion of ZnO particles inside the pure α-phase at ∼800 cm−1 is present only as a shoulder in
polymer matrix for filler activated for 10 min (figure 2(c)). our spectrum indicates that the peaks at ∼838 and
XRD patterns of the PVDF and PVDF–ZnO composites with ∼1431 cm−1 should be assigned primarily to the β and γ
5 wt% of non-activated and mechanically activated (5 and phases [42–45]. Furthermore, the peak at ∼1431 cm−1 is
10 min) filler are shown in figure 3. In the diffraction pattern more intensive than the broad peak at ∼(874–882) cm−1—in
of PVDF, all three phases, α, β and γ (corresponding JCPDS accordance with the assumption of the dominance of γ phase.
(Joint Committee on Powder Diffraction Standards) database The analyses of the Raman spectra of the composites
card numbers are JCPDS 42–1650, 38–1638 and 42–1649), showed that non-activated ZnO filler reduces crystallinity of
are present, with the pronounced contribution of γ-phase. the polymer. The intensity of crystalline polymer lines gen-
There is also a broad peak from the amorphous phase at low erally decreases in P5Z0 sample and strong peaks at 1431 and
2θ, suggesting relatively significant volume of the amorphous 2980 cm−1, which represent the combination of all three
phase in the samples. In addition, the peaks of PVDF phases crystalline phases, almost disappear (figure 4). The decrease
appear to be relatively broad, making phase composition in intensity of the peaks at ∼811 cm−1 (the γ phase) and
analysis difficult. XRD patterns of PVDF–ZnO composites ∼838 cm−1 (primarily β and γ phase) is larger than that of the
show additional peaks of hexagonal ZnO phase. ZnO dif- α peak at ∼800 cm−1, indicating a reduction in the contrib-
fraction lines, assigned to (100), (002), (101) reflections, ution of γ and β phases. Raman spectra of the PVDF–ZnO
become broader as the activation time increases due to composites with fillers activated for short times (5 and
reduced size of the ZnO particles and more pronounced lattice 10 min) show that mechanical activation results in significant
distortions [39]. Taking into account very small differences changes in crystalline morphology of the PVDF matrix. The
between the positions of γ (110) and β (110) reflections results suggest that the presence of activated ZnO particles
(∼20.2° versus ∼20.6°), and no visible peak at β (220) induces a general rise in the intensity of crystalline polymer
position, XRD analysis could not provide information on the lines and an enhancement of β-phase crystallization in PVDF.
changes in phase composition of PVDF, induced by non- In the spectra of P5Z5 and P5Z10 samples (figure 4) the
activated and mechanically activated fillers. For this reason, intensity of the line at ∼811 cm−1 (the γ phase) decreases
Raman spectroscopy, which is much more sensitive to relative to the intensity of the peak at ∼838 cm−1. Having in
changes in the crystal structures within polymer, was mind that the strongest α line at 800 cm−1 is still relatively
performed. weak, the peak at ∼838 cm−1 can be attributed primarily to
Figure 4 shows Raman spectra of PVDF and the com- the contribution of β and γ phases. Furthermore, taking into
posites (with 5 wt% of filler). The most pronounced changes account that an increase in intensity of the peak at ∼838 cm−1
in the Raman spectra were observed in the range from 750 to occurs if the excess of β phase is present in the sample [43],
900 cm−1. In order to better illustrate these effects, the spectra the observed change in peak-height ratio of 811 cm−1 peak to
are shown in the range from 700 to 1200 cm−1 (figure 4(b)). 838 cm−1 peak is most likely due to enhanced formation of β
The analysis of the Raman spectrum of PVDF suggested that crystals. This indicates that the particles mechanically acti-
γ-phase is dominant: peak at ∼811 cm−1 is pronounced, vated for 5 and 10 min act as nucleation agents for β phase
which is typical of high content of γ-phase [42, 43]. Peaks at crystallization in PVDF.
∼838 cm−1, ∼(874–882) cm−1 and ∼1431 cm−1 may origi- Fourier-transform infrared (FTIR) spectra of pure PVDF
nate from all three phases (α, β and γ), however, in the and PVDF–ZnO composites show that the pure PVDF is
spectrum of the pure γ-phase, the line at∼1431 cm−1 is very synthesized as predominantly γ-phase with prominent peaks
strong, unlike the corresponding lines in the spectra of α- and at 833 and 1234 cm−1 (figure 5). Introduction of ZnO filler
β-phase, suggesting that the main contribution to the line at promotes crystallization of α- and β-phases, with reduction in

4
Phys. Scr. 93 (2018) 105801 A Peleš et al

Figure 3. XRD patterns of PVDF and PVDF–ZnO composites with 5 wt% of non-activated and mechanically activated fillers. Rietveld
analysis of non-activated and activated ZnO filler powders is shown right (results are shown in table 1).

content of γ-phase and overall reduction in crystallinity of the generally exhibit relatively poor crystallinity of the PVDF
samples. The presence of α-phase is established from peaks at phase. However, FTIR indicates that the introduction of ZnO
766 and 976 cm−1, while β-phase is identified using peaks at filler suppresses the crystallization of γ-phase and promotes
840 and 1279 cm−1. Enhancement of region from 800 to crystallization of α- and β-phases. The apparent differences
900 cm−1 in figure 5 shows that there is a shift in the position between the Raman and FTIR results can be attributed to the
of maximum of the peak at 833 cm−1 towards 840 cm−1 in fact that Raman spectroscopy only probed the surface of the
samples with ZnO fillers, suggesting formation of β-phase PVDF/ZnO composite films, which is where the effects of
[58], which is supported by a shift in the position of CH2 strain on PVDF fibres is the most pronounced. This is also
modes in this region towards lower energies, similar to what indicated by the fact that PVDF sample with non-activated
was observed in PVDF doped with gold nanoparticles [59]. filler exhibits extremely poor Raman spectrum, while its FTIR
Lower intensities of the peaks and lower shift of CH2 modes spectrum is not much different than that of other samples.
suggests that the content of crystalline β-phase is most likely This suggests that the distribution of phases in the PVDF/
lower than α- and γ- phases, and overall FTIR spectrum ZnO films would be inhomogeneous, with β-phase being
suggests that both PVDF and PVDF–ZnO composite more prevalent at the surface of the film, due to increased

5
Phys. Scr. 93 (2018) 105801 A Peleš et al

Figure 4. (a) Raman spectra of PVDF and PVDF–ZnO composites with 5 wt% of the fillers mechanically activated for various activation
times. (b) The same spectra in the region from 700 to 1200 cm−1.

Figure 5. FTIR spectra of pure PVDF and samples with non-activated (P5Z0) and activated ZnO fillers (P5Z5 and P5Z10). Smaller sections
show enhanced 800–900 cm−1 and 2900–3100 cm−1 regions.

6
Phys. Scr. 93 (2018) 105801 A Peleš et al

Figure 6. Dielectric permittivity (ε′) versus temperature curves for the pure PVDF and P5Z0, P5Z5, and P5Z10 composites measured at
several frequencies: (a) 60 Hz, (b) 200 Hz, (c) 6 kHz and (d) 60 kHz.

strain favoring its formation, with α-phase occurring mostly entire temperature range. In contrast, the composite with the
on the inside of the film. commercial ZnO powder as a filler exhibits lower value of the
The observed changes in crystal structure of PVDF after dielectric permittivity than the pure PVDF (figure 6). This
introduction of non-activated and mechanically activated ZnO behaviour is a direct consequence of different crystallization
filler are also reflected in changes in dielectric properties. behaviour of the PVDF matrix in the presence of these two
Dielectric spectroscopy measurements were carried out in type of fillers. Since inorganic content in the composites is the
150–400 K temperature range at four different frequencies same (5 wt%), an increase in dielectric permittivity can be
(60 Hz, 200 Hz, 6 kHz and 60 kHz). Figure 6 shows the related to a higher content of β-phase crystals in the com-
dielectric permittivity versus temperature curves of the pure posites with activated fillers (figure 4) [48, 49]. The fact that
PVDF and PVDF–ZnO composites with non-activated and the dielectric permittivity of the composite with non-activated
activated (5 and 10 min) fillers. All four materials exhibit ZnO powder is lower than that of the pure PVDF suggests
typical dielectric behaviour of PVDF e.g. at fixed frequency that the introduction of the particles with higher dielectric
of the external field, the dielectric permittivity increases with permittivity into PVDF matrix is not the sole condition for
increasing in temperature [46, 57]. Also, the broad dielectric improvement of dielectric permittivity of the polymer. Rather,
permittivity peak shifts to a higher temperature as the fre- the dielectric properties will depend on the change in the
quency increases, which is typical for a relaxation process in PVDF matrix crystal polymorphism during the crystallization
ferroelectrics [47]. On the other hand, the effects of the non- in the presence of the particles as well as on the strength of
activated and activated ZnO particles on the dielectric the interaction of the fillers and matrix chains. Finally, the
permittivity of the polymer are different. The introduction of results in figure 6 show that the composite with the ZnO filler
mechanically activated ZnO powders into the matrix leads to mechanically activated for 5 min exhibits the highest di-
an increase in the dielectric permittivity of the polymer in the electric permittivity, suggesting that this is the optimal

7
Phys. Scr. 93 (2018) 105801 A Peleš et al

Figure 7. Loss tangent versus temperature curves for the pure PVDF and P5Z0, P5Z5 and P5Z10 composites measured at several frequencies:
(a) 60 Hz, (b) 200 Hz, (c) 6 kHz and (d) 60 kHz.

activation time of the filler prior to mixing with the polymer increase at temperatures above 325 K, which is more pro-
matrix. It seems that prolonged mechanical activation has a nounced at lower frequencies and in the composites with
detrimental effect on the surface activity of the particles, activated fillers (figures 6 and 7). It most likely contains
which results in their reduced activity as nucleating agents contributions from multiple processes, like electrode polar-
within the polymer matrix. Loss tangent curves that corre- ization processes [51] and αc-transition [52]. Electrode
spond to the dielectric permittivity (ε′) curves in figure 6 are polarization is a consequence of the accumulation of the
shown in figure 7. charges at the surface of the electrode while under the force of
All loss tangent curves show a typical low temperature the electric field. The electrode processes usually mask
relaxation observed at 200 K<T<300 K [49], corresp- Maxwell–Wagner–Sillars polarization processes, which are
onding to glass transition processes of PVDF, involving typical for heterogeneous systems and occur through polar-
crystalline and amorphous-crystalline regions. It is influenced izations at the interfaces of the constituents. The shift of the
by the composition and preparation of the composite [50]. onset of the increase of dielectric permittivity to higher tem-
The position of this relaxation peak shifts towards higher peratures with increase in frequency is typical of αc-transition
temperature (at fixed frequency of the external field) after [53, 54], which is indicated by the decrease in maximum
introduction of mechanically activated particles. This indi- value of dielectric loss at the end of the temperature interval at
cates strong interaction of the activated particles and PVDF 400 K with increase in frequency (table 2). In addition, the
matrix chains, which reduces segmental mobility of PVDF onset of αc-transition appears to shift towards lower tem-
chains and affects the cooperative motions in the polymer peratures in samples containing mechanically activated fillers,
matrix, necessary for the glass transition to take place. Both compared to pure PVDF and the sample with non-activated
dielectric permittivity and loss tangent curves exhibit a sharp filler. This indicates that the mechanical activation of the filler

8
Phys. Scr. 93 (2018) 105801 A Peleš et al

Table 2. Values of dielectric permittivity and loss tagnent for

different frequencies at 30 and 400 K.
Sample T=300 K T=400 K
αc-
Dielectric Loss Dielectric Loss onset
60 Hz permittivity tangent permittivity tangent (K)
PVDF 2.400 0.005 4.060 0.251 330
P5Z0 1.690 0.005 2.630 0.218 330
P5Z5 3.630 0.050 6.540 0.388 310
P5Z10 3.110 0.056 5.250 0.431 310

200Hz

PVDF 2.395 0.058 3.681 0.179 330

P5Z0 1.682 0.005 2.412 0.151 330
P5Z5 3.486 0.060 5.318 0.270 320
P5Z10 2.944 0.070 4.527 0.303 320

6kHz
Figure 8. Arrhenius plots for the pure PVDF and P5Z0, P5Z5 and
PVDF 2.355 0.010 3.017 0.073 350 P5Z10 composites.
P5Z0 1.663 0.009 2.088 0.049 350
P5Z5 2.807 0.060 3.822 0.163 340 Table 3. The overall average activation energy values of the low
P5Z10 2.562 0.040 3.209 0.132 330 temperature relaxation processes, calculated from the linear
regression of the plots in figure 7.
60kHz
Sample Ea (kJ mol−1)
PVDF 2.355 0.043 3.017 0.037 360
P5Z0 1.643 0.039 2.051 0.025 360 PVDF 206±8
P5Z5 2.647 0.054 3.184 0.120 350 P5Z0 214±8
P5Z10 2.443 0.048 2.835 0.083 350 P5Z5 223±9
P5Z10 221±9

Conclusions

reduces the perfection of the lamellar PVDF crystals, making Surface modification of ZnO nanoparticle fillers by means of
αc-transition easier [55, 56], leading to lower crystallinity of mechanical activation has pronounced effect on the structural
PVDF in the samples containing activated fillers. properties of the PVDF polymer matrix. TEM and SEM
In order to determine whether strong particle–chain analyses showed that mechanical activation reduces particle
interaction was also reflected in the change in activation size, and Raman and FTIR measurements showed that ZnO
energy of this transition, the values of overall activation particles mechanically activated for short time (5 and 10 min)
energy of the relaxation peak were calculated, since its influence the crystallization of PVDF in different way than
position shifts to higher temperatures with increase in fre- non-activated ZnO particles. Non-activated ZnO filler
quency, allowing the application of Arrhenius equation, decreases the crystallinity of the PVDF matrix and slightly
increases the content of the α phase. On the other hand,
Ea
f = f0 e- RT , (1 ) PVDF–ZnO composites with 5 and 10 min activated fillers
contain more β (ferroelectric) phase. These results, together
where f is the frequency used for measurement of dielectric with our previously published work [37, 38], show that
permittivity, R is the universal gas constant, T is the position mechanical activation of ceramic and semiconductor fillers
of the peak maximum and Ea is the overall average activation prior to mixing with PVDF matrix can be used as general
energy of the transition. According to equation (1), Ea is method for the fabrication of β-phase rich composite films.
calculated from the slope of the plot log f=f (1/T). Figure 8 Dielectric spectroscopy measurements showed that the
shows the obtained Arrhenius plots for the pure PVDF and introduction of mechanically activated ZnO fillers resulted in
P5Z0, P5Z5 and P5Z10 composites. The activated energies an increase in dielectric permittivity, most likely due to an
(Ea) of the relaxation peak determined from the linear increase in β-phase content, while the introduction of non-
regression of the plots in figure 8 are presented in table 3. activated ZnO powder resulted in a decrease in dielectric
These suggest that the introduction of activated fillers in the permittivity, likely due to the disruption of crystallization of
PVDF matrix probably increases the value of the average PVDF matrix caused by the filler. The position of the broad
activation energy, which would be consistent with the strong relaxation peak in tan δ spectra of the composites with acti-
particle–chain interaction in these composites. vated fillers shifted towards higher temperatures with respect

9
Phys. Scr. 93 (2018) 105801 A Peleš et al

to its position in the pure polymer. This suggests strong [21] Kim P, Doss N M, Tillotson J P, Hotchkiss P J, Pan M-J,
interaction of activated particles and polymer chains, due to a Marder S R, Li J, Calame J P and Perry J W 2009 ACS Nano
combination of increased surface area and increased con- 3 2581–92
[22] Li J, Claude J, Norena-Franco L E, Seok S I and Wang Q 2008
centration of surface defects in the activated fillers. This Chem. Mater. 20 6304–6
would be consistent with the obtained values of the activation [23] Fan B-H, Zha J-W, Wang D-R, Zhao J, Zhang Z-F and
energies of the relaxation peak via Arrhenius equation. Dang Z-M 2013 Compos. Sci. Technol. 80 66–72
[24] Niu Y, Yu K, Bai Y and Wang H 2015 IEEE Trans. Ultrason.
Ferroelectr. Freq. Control 62 108–15
[25] Kamezawa N, Nagao D, Ishii H and Konno M 2015 Eur.
Acknowledgments Polym. J. 66 528–32
[26] Wang Z L 2004 J. Phys.: Condens. Matter 16 R829–58
This work was supported in part by the Ministry of Educa- [27] Dong L, Cui Z L and Zhang Z K 1997 Nanostruct. Mater. 8
815–23
tion, Science and Technological Development of the Republic [28] Ginley D and Bright C 2000 MRS Bull. 25 15–8
of Serbia (Project Nos. 172056, 172057, and 45020) and [29] Haraa K, Horiguchib T, Kinoshitab T, Sayamaa K,
projects NSF CREST (HRD-0833184) and NASA Sugiharaa H and Arakawaa H 2000 Sol. Energy Mater. Sol.
(NNX09AV07A). Cells 64 115–34
[30] Keis K, Vayssieres L, Lindquist S and Hagfeldt A 1999
Nanostruct. Mater. 12 487–90
[31] Alibe I M, Matori K A, Saion E, Ali A M and Zaid M H M
ORCID iDs 2017 The influence of calcination temperature on structural
and optical properties of ZnO nanoparticles via simple
polymer synthesis route Sci. Sinter. 49 263–75
V Blagojević https://orcid.org/0000-0001-8102-989X [32] Indolia A P and Gaur M S 2013 J. Polym. Res. 20 43
[33] Indolia A P and Gaur M S 2013 J. Therm. Anal. Calorim. 113
821–30
[34] Nguyen V S, Rouxel D, Vincent B, Badie L, Dos Santos F D,
References Lamouroux E and Fort Y 2013 Appl. Surf. Sci. 279 204–11
[35] Dodds J S, Meyers F N and Loh K J 2013 Smart. Struct. Syst.
[1] Seiler D A 1997 PVDF in the chemical process industry 12 055–71
Modern Fluoropolymers ed J Scheirs (Chichester: Wiley) [36] Lu Y, Claude J, Neese B, Zhang Q and Wang Q 2006 J. Am.
pp 487–505 Chem. Soc. 128 8120–1
[2] Taylor G W, Gagnepain J J, Meeker T R, Nakamura T and [37] Mofokeng T G, Luyt A S, Pavlović V P, Pavlović V B,
Shuvalov L A 1985 Piezoelectricity (New York: Gordon and Dudić D, Vlahović B and Djoković V 2014 J. Appl. Phys.
Breach Science Publishers) 115 084109
[3] Xunlin Q 2010 J. Appl. Phys. 108 011101 [38] Pavlović V P, Pavlović V B, Vlahović B, Božanić D K,
[4] Park Y J, Kang Y S and Park C 2005 Eur. Polym. J. 41 Pajović J D, Dojčilović R and Djoković V 2013 Phys. Scr.
1002–12 T157 014006
[5] Lovinger A J 1983 Science 220 1115–21 [39] Pavlović V P, Krstić J, Šcepanović M J, Dojčilović J,
[6] Humphrey J S and Amin-Sanayei R (ed) 2002 Vinylidene Minić D M, Blanuša J, Stevanović S, Mitić V and
fluoride polymers Encyclopedia of Polymer Science and Pavlović V B 2011 Ceram. Int. 37 2513–8
Technology (New York: Wiley) [40] Vojisavljević K, Šćepanović M, Srećković T,
[7] Furukawa T 1989 Phase Transit. 18 143–211 Grujić-Brojčin M, Branković Z and Branković G 2008
[8] Matsushige K, Nagata K, Imada S and Takemura T 1980 J. Phys.: Condens. Matter 20 475202
Polym. J. 21 1391–7 [41] Peleš A et al 2015 J. Alloy. Compd. 648 971–9
[9] Davis G T, McKinney J E, Broadhurst M G and Roth S C 1978 [42] Satapathy S, Pawar S, Gupta P K and Varma K B R 2011 Bull.
J. Appl. Phys. 49 4998–5002 Mater. Sci. 34 727–33
[10] Matsushige K and Takemura T 1978 J. Polym. Sci. Polym. [43] Boccaccio T, Bottino A, Capannelli G and Piaggio P 2002
Phys. Ed. 16 921–34 J. Membr. Sci. 210 315–29
[11] Leonard C, Halary J L and Monnerie L 1988 Macromolecules [44] Nallasamy P and Mohan S 2005 Indian J. Pure. Appl. Phys. 43
21 2988–94 821–7 (http://nopr.niscair.res.in/handle/
[12] Shah D, Maiti P, Gunn E, Schmidt D F, Jiang D D, 123456789/8879)
Batt C A and Giannelis E P 2004 Adv. Mater. 16 1173–7 [45] Kobayashi M, Tashiro K and Tadokoro H 1975
[13] Ramasundaram S, Yoon S, Kim K J, Lee J S and Park C 2009 Macromolecules 8 158–71
Macromol. Chem. Phys. 210 951–60 [46] Fu J, Hou Y, Zheng M, Wei Q, Zhu M and Yan H 2015 ACS
[14] Tang C W, Li B, Sun L, Lively B and Zhong W H 2012 Eur. Appl. Mater. Interfaces 7 24480–91
Polym. J. 48 1062–72 [47] Ishai P B, Talary M S, Caduff A, Levy E and Feldman Y 2013
[15] da Silva A B, Marini J, Gelves G, Sundararaj U, Meas. Sci. Technol. 24 102001
Gregório R Jr and Bretas R E S 2013 Eur. Polym. J. 49 [48] Sencadas V, Barbosa R, Mano J F and Lanceros-Méndez S
3318–27 2003 Ferroelectrics 294 61–71
[16] Shehzad K et al 2013 J. Mater. Sci. 48 3737–44 [49] Grimau M, Laredo E, Bello A and Suarez N 1997 J. Polym.
[17] Bahrami A, Talib Z A, Yunus W M M, Behzad K and Sci. B 35 2483–93
Soltani N 2012 Adv. Mater. Res. 364 50–4 [50] El Mohajir B E and Heymans N 2001 Polymer 24 5661–7
[18] Sun L L, Li B, Zhang Z G and Zhong W H 2010 Eur. Polym. J. [51] Gregorio R Jr and Ueno E M 1999 J. Mater. Sci. 34 4489–500
46 2112–9 [52] Yang L, Ji H, Qiu J, Zhu K and Shao B 2014 J. Intell. Mater.
[19] Ducharme S 2009 ACS Nano 3 2447–50 Syst. Struct. 25 858–64
[20] Kim P, Jones S C, Hotchkiss P J, Haddock J N, Kippelen B, [53] Omote K and Ohigashi H 1997 J. Appl. Phys. 81 2760
Marder S R and Perry J W 2007 Adv. Mater. 19 1001–5

10
Phys. Scr. 93 (2018) 105801 A Peleš et al

[54] Ozkazanc E, Guney H Y, Oskay T and Tarcan E 2008 J. Appl. [57] Liu S, Xue S, Zhang W, Zhai J and Chen G 2014 J. Mater.
Polym. Sci. 109 3878–86 Chem. A 2 18040–6
[55] Yu L and Cebe P 2009 J. Polym. Sci. B 47 2520–32 [58] Martins P, Lopes A C and Lanceros-Mendez S 2014 Prog.
[56] Mijovic J, Sy J W and Kwei T K 1997 Macromolecules 30 Polym. Sci. 39 683–706
3042–50 [59] Mandal D, Henkel K and Schmeißer D 2012 Mater. Lett. 73
123–5

11