Silicone-barium titanate composites with increased electromechanical sensitivity. the effects of the filler morphology

Article abstract of DOI:10.1039/c4ra09903f

A high molecular weight polydimethylsiloxane-α,ω-diol, synthesized in the lab, was used as a matrix for nanocomposites. Barium titanate nanoparticles, either cubic or nanorods, were obtained by a hydrothermal procedure and used as a filler. In order to ensure a good compatibility with the matrix, the filler was surface treated with a commercial surfactant. A highly reactive trifunctional silane was used as crosslinking agent in the presence of organometallic catalyst. Two other samples, the first consisting of pure crosslinked polydimethylsiloxane and the second being the polymer matrix filled with commercial barium titanate, were prepared and used as references to evaluate the influence of the presence of barium titanate nanoparticles and their shape on some characteristics of the resulting crosslinked composites: morphology, thermal behavior, moisture sorption, mechanical and dielectric characteristics. The electromechanical sensitivity and energy output were calculated on the basis of appropriate experimental data in order to estimate the potential of the composites for future electromechanical applications.

Full text of DOI:10.1039/c4ra09903f

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Silicone–barium titanate composites with
increased electromechanical sensitivity. The effects
of the filler morphology†
Cite this: RSC Adv., 2014, 4, 58522
*
A. Bele, M. Cazacu, G. Stiubianu and S. Vlad
A high molecular weight polydimethylsiloxane-a,u-diol, synthesized in the lab, was used as a matrix for
nanocomposites. Barium titanate nanoparticles, either cubic or nanorods, were obtained by
a
hydrothermal procedure and used as a filler. In order to ensure a good compatibility with the matrix, the
filler was surface treated with a commercial surfactant. A highly reactive trifunctional silane was used as
crosslinking agent in the presence of organometallic catalyst. Two other samples, the first consisting of
pure crosslinked polydimethylsiloxane and the second being the polymer matrix filled with commercial
barium titanate, were prepared and used as references to evaluate the influence of the presence of
barium titanate nanoparticles and their shape on some characteristics of the resulting crosslinked
composites: morphology, thermal behavior, moisture sorption, mechanical and dielectric characteristics.
The electromechanical sensitivity and energy output were calculated on the basis of appropriate
experimental data in order to estimate the potential of the composites for future electromechanical
applications.
Received 5th September 2014
Accepted 29th October 2014
DOI: 10.1039/c4ra09903f
www.rsc.org/advances
not as much as the theoretical values because the side organic
groups (methyl groups in the case of PDMS) hinder the dipoles
from getting too close to each other.12,13 In order to increase the
Introduction
There is a growing interest for materials with high electrome-
chanical sensitivity, because such materials are useful as
capacitors that can store a large amount of energy and then
deliver it instantaneously,1–4 as novel potential electromechan-
ical actuators that deform when stimulated by an electrical
eld,5–7 or as sensors used mainly for damage detection, char-
acterization of structures and fatigue studies of materials.5,8
Therefore, many studies have been carried out to develop high
permittivity materials.1 Polymeric materials having a large
dielectric constant and high energy density are promising for
electromechanical energy conversion technologies.3,4,9 Besides
uoropolymers, acrylic polymers, polyurethanes and other
synthetic or natural rubbers, silicones are oen chosen for such
purposes. Due to their structural peculiarities (high exibility of
the Si–O bond, low intermolecular interactions), silicones, with
polydimethylsiloxane (PDMS) as the main representative of this
type of polymer, have high free volume, low glass transition
temperature and unusual rheological/ow properties with
highly elastic behaviour.10,11 The polarizability of the Si–O bond,
a premise for a high dielectric constant, is larger when
compared to organic nonpolar polymers (e.g., polyethylene), but
dielectric constant, the polysiloxanes are chemically modied
by attaching polar groups to the silicon atoms.14–16 Although this
procedure proved to be an efficient pathway, due to incompat-
ibility between dimethylsiloxane segments of the main chain
and polar groups it has the disadvantage that with increasing
content of such groups, a phase separation phenomenon
appears, thus leading to worsening of the mechanical proper-
ties of the resulting material.16,17
Another approach consists in the incorporation of highly
dielectric or conductive nanollers in polymeric matrix to
obtain high permittivity nanocomposites which combine the
useful properties of polymers, such as versatile chemistry,
exibility, light weight and processability, with the superior
electrical properties of adequate nanoller.3,18 High-
performance dielectric materials such as ferroelectric ceramic
particles are oen used in this aim because of their high
dielectric constant and low dielectric loss1,19,20 but used alone
such materials show the disadvantages of limited tunability of
their own dielectric properties, high production cost and poor
processability.18 A well-known ceramic is BaTiO3, which is in the
tetragonal phase in the temperature range between 5 and 120
ꢀC, and in this phase it is a ferroelectric crystalline material
exhibiting spontaneous polarization, with the dipole moment
arising mainly because of the movement of Ti atoms with
respect to the O atoms in the same plane inside the O6 octa-
hedra.21 It has been proved that particles of BaTiO3 suspended
Petru Poni Institute of Macromolecular Chemistry, Aleea Grigore Ghica Voda 41A, Iasi,
700487, Romania. E-mail: mcazacu@icmpp.ro; Fax: +40-232211299
† Electronic supplementary information (ESI) available: Stress relaxation and
permanent set data, DTG and DSC curves for the prepared composites. See DOI:
10.1039/c4ra09903f
58522 | RSC Adv., 2014, 4, 58522–58529
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
in a medium show dielectric relaxation, similar to that observed
Methyltriacetoxysilane (MTAS) has been prepared by a
for polar molecules and the dielectric behaviour is inuenced procedure adapted from the literature33 that consists in a simple
by the size and structure of the aggregates.1 Thus, the capacity condensation reaction between methyltrichlorosilane and ace-
of spherical particles to increase the dielectric constant is small tic anhydride in stoichiometric ratio (1 : 3). Acetic anhydride
at low volume fractions.22 Higher volume fractions lead to was dropwise added over methyltrichlorosilane under stirring
increased dielectric constant, but also to a reduction in elec- at room temperature, aer that a condenser was attached to the
trical breakdown strength and mechanical properties.1,23–26 The reaction vessel and the temperature of the reaction mixture was
ꢀ
particles in a colloidal suspension subjected to a strong electric increased to 110 C in order to remove acetyl chloride (vapour
ꢀ
eld are polarized and tend to form chainlike aggregates temperature 53 C), a byproduct of the condensation reaction.
oriented in the direction of electric eld,27–30 this behavior being Finally a vacuum of 10 mm Hg was applied to collect methyl-
known as the Winslow effect.27,28 This effect is more pronounced triacetoxysilane (yield around 75%, b.p.10 ¼ 95–97 ꢀC, d20
¼
ꢀ
when high aspect ratio llers are used, this being a promising 1.17, freezing point ꢁ40 C).
path for the development of materials with high dielectric
constant and low dielectric loss at a low ller volume fraction,
for use as capacitor and electric eld materials.1
Equipments
In this work, we used barium titanate particles with prismatic Scanning electron microscopy (SEM) images were acquired with
and nanorod shapes, prepared by hydrothermal procedure, as a an electronic microscope type Quanta 200 operating at 30 kV
ller for a well-dened, high molecular weight polydimethylsi- with secondary and backscattering electrons in low or high
loxane-a,u-diol (Mn
triacetoxysilane in presence of dibutyltin dilaurate as catalyst.
¼
438 000) crosslinked with methyl- vacuum mode.
Transmission electron microscopy (TEM) observations were
We have chosen barium titanate due to its high dielectric made with Hitachi High-Tech HT7700 Transmission Electron
permittivity, 30 ꢁ 1600 at room temperature. In addition, the Microscope, operated in high contrast mode at 100 kV acceler-
approached hydrothermal preparation method is simple, ating voltage. The samples were prepared on carbon coated
reproducible and easily scalable one. Because the dielectric copper grids of 300 mesh size. Microdrops of dispersions
permittivity depends on many factors, such as the grain size, previously ultrasonicated in chloroform were placed on the
shape and size of the crystals, impurities, and on processing grids, and then solvent was removed in vacuum. A vacuum of
techniques,31 here we studied the effect of its morphology. One >10ꢂ4 Torr was achieved in the sample holder room before
sample, a pure crosslinked polymer and one consisting of performing the analysis.
polymer lled with commercial barium titanate were prepared
Wide Angle X-rays Diffraction (WAXD) measurements were
and used as reference to evaluate the inuence of the barium performed on a Bruker-AXS D8 ADVANCE diffractometer, with
titanate presence and its morphology on some characteristics Bragg Brentano parafocusing goniometer and Cu anode with
(dispersibility, thermal and moisture sorption) of the resulted Ka1 ¼ 1.5406.
materials and properties of interest for electromechanical
Stress–strain measurements were performed on TIRA test
applications (dielectric and mechanical parameters). The 2161 apparatus, Maschinenbau GmbH Ravenstein, Germany on
presumptive electromechanical sensitivity was estimated on the dumbbell-shaped cut samples with dimensions of 50 ꢃ 8.5 ꢃ 4
basis of mechanical and dielectric measurements.
mm. Measurements were run at an extension rate of 20 mm
minꢂ1, at room temperature. Cyclic tensile stress tests were
performed on the similar samples between 2 and 100% strain.
Five stretch-recovery cycles were registered. The stationary time
at minimum and maximum applied stress was 5 s.
Experimental
Materials
Dielectric spectroscopy was performed using the Novocontrol
The polydimethylsiloxane-a,u-diol, that was used as matrix, was “Concept 40” broadband dielectric spectrometer (Hundsangen,
prepared by bulk polymerization of octamethylcyclotetrasilox- Germany), at room temperature in the frequency domain 1 Hz to
ane catalyzed by H2SO4, at room temperature and atmospheric 1 MHz. Samples in the form of lms having uniform thickness
humidity, according to an already described procedure.32 The in the 0.7–1 mm range were placed between gold plated round
average molecular mass values of the polymer, as estimated electrodes, the upper electrode having a 20 mm diameter.
from the monomodal elution curve traced by gel permeation Temperature was controlled using a nitrogen gas cryostat and
chromatography, with CHCl3 as eluent, aer calibration with the temperature stability of the sample was better than 0.1 ꢀC.
standard polystyrene samples, were Mn ¼ 438 000 and Mw
¼
Water vapor sorption (DVS) capacity of the samples was
determined in dynamic regime, in the relative humidity (RH)
641 000 (polydispersity index, I ¼ 1.46).
Barium titanate, BaTiO3, BT, coded as CO to indicate ller type, range 0–90% by using the fully automated gravimetric analyzer
with particle size <3 mm and Pluronic L-31, HO-poly(ethyleneglycol)- IGAsorp produced by Hiden Analytical, Warrington (UK).
block-poly(propyleneglycol)-block-poly(ethyleneglycol)-OH (M
¼
Dielectric strength measurements were made at PERCRO
25
1100, d ¼ 1.02) were purchased from Fluka AG, barium chloride Laboratory – TeCIP Institute – Scuola Superiore Sant'Anna, Pisa,
25
dihydrate, BaCl2$2H2O from Chimopar S.A. Romania, and tita- Italy, on a home-made installation consisting in high-speed
nium dioxide, Ti-Pure® R-902+ (min. 93 wt% TiO2, a rutile titanium high-voltage power amplier, function generator, and an
dioxide pigment as a ne, dry powder) from Du Pont USA.
oscilloscope. The brass electrodes were applied on the lm
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 58522–58529 | 58523

View Article Online
RSC Advances
Paper
samples and the measurements were performed at 60 Hz, and a polymerization at room temperature) and used as matrix for a
voltage increase rate of 2000 V sꢂ1 at room temperature (25 ꢀC). ceramic ller, i.e., barium titanate to obtain lms with
The samples were previously brought into equilibrium with the increased electromechanical sensitivity. Barium titanate parti-
humidity of the environment in which the measurements were cles with cubic (CU) and nanorod (NR) morphologies were
made. Three samples were analyzed for each composite prepared in this aim by hydrothermal technique. The cubes
formulation and the lowest value was taken into consideration. were obtained with dimensions ranging from 350 nm up to 550
Thermogravimetric (TG) measurements were conducted on a nm on each side, while the nanorods had about 200 nm in
STA 449 F1 Jupiter device (Netzsch, Germany), in nitrogen diameter and 2–3 mm in length as could be estimated based on
atmosphere, in the temperature range 25 ꢀC to 700 ꢀC at a scanning and transmission electron microscopy images
heating rate of 10 ꢀC minꢂ1. DSC measurements were con- (Fig. 1b, c, e and f). For commercial product, although SEM
ducted with a DSC 200 F3 Maia (Netzsch, Germany) at a heating images taken on powder spread as such onto a substrate indi-
rate of 10 C minꢂ1 in nitrogen atmosphere.
cate large prisms with high dispersity in size (between 1–7 mm)
(Fig. 1a), the TEM images taken on the sample deposited on the
grid from chloroform suspension showed much smaller (less
than 50 nm) and very uniform particles (Fig. 1d). Each particle,
as seen in TEM images, is composed of multiple crystalline
grains, which in turn are composed of multiple crystallites. The
sizes of the crystallites for the specic type of particles, calcu-
lated with Debye–Scherrer equation, are: 59 nm (CO), 58 nm
(CU) and 60 nm (NR).
Wide angle X-ray powder diffraction spectra, registered at
room temperature in the range 2q from 20–70 (ꢀ), show for
both our samples (CU and NR) and for commercial sample
(CO) the same patterns specic for crystalline barium tita-
nate.35 The presence of (100) planes shows that ferroelectric
domains have 180ꢀ walls and act as boundaries between
domains with antiparallel polarization and these form as a
result of pressure developed in stainless steel autoclaves at
200 ꢀC.21 Using high concentration solutions of sodium
hydroxide, contributes to increasing the concentration of the
reaction mixture which favors the nucleation rate and the
formation of smaller particles uniformly distributed. This is
visible in Fig. 2, where for cubic (CU) nanoparticles formed in
10 M NaOH solution there is increased peak intensity of
barium titanate.36
ꢀ
Procedure
Preparation of barium titanate with different morphologies,
CU and NR. In order to synthesize cubic barium titanate (CU)
proper amounts of TiO2 (0.25 g) and BaCl2 (0.65 g) were mixed
together in 20 mL of 10 M NaOH aqueous solution.34 The same
amounts of TiO2 and BaCl2 were used to synthesize barium
titanate nanorods (NR) in 20 mL of 5 M NaOH aqueous solution.
The resulting mixtures were transferred in stainless steel auto-
claves coated with Teon and kept 72 h at 200 ꢀC in a Venticell
55 hydrothermal oven. Heating and cooling of samples were
done with a rate of 1ꢀC minꢂ1. Aer cooling, the particles were
washed with distilled water by centrifugation at 6000 rpm for 15
min. This was repeated for 2 more times. Then the particles
were heated at 100 ꢀC over night in a laboratory oven to remove
the remaining water before analyses.
Preparation of the silicone nanocomposites, C–CO, C–CU,
C–NR, C–R. 0.5 g barium titanate, either commercial (CO) or
obtained as in the above described procedure (CU or NR), was
grounded using an agate mortar with pestle and then was mixed
with 0.5 g surfactant by vigorous mechanical stirring followed by
ultrasonication for 5 minutes. The mixture was then transferred
over 10 g PDMS and stirred again. Then 0.25 mL crosslinker
(MTAS) and 0.025 mL catalyst (DBTDL) were added and stirred
thoroughly until the components formed a homogenous mixture.
The resulted mixture was sonicated for 10 minutes and aer-
wards it was poured into a Teon mold (15 ꢃ 5 cm). The molds
were maintained in atmospheric humidity at room temperature
for 24 h so that crosslinking of the silicone matrix could occur.
Aer that the formed lms with thickness between 0.5 and 0.7
mm were easily peeled off from the substrate. The lms labeled as
C–CO (polymer matrix with commercial barium titanate as ller),
C–CU (polymer matrix with cubic barium titanate nanoparticles
as ller) and C–NR (polymer matrix with barium titanate nano-
rods as ller). A similar lm consisting in pure crosslinked PDMS
without any ller has been prepared as a reference sample, C–R.
The lms were then kept in the laboratory environment about two
weeks for aging before characterization by different techniques
(DSC, TGA, dielectric measurements and mechanical testing).
Barium titanate species thus obtained were used as active
llers for PDMS. In order to assure a good compatibility of the
ller with the matrix, the former was previously surface treated
with a surfactant. A commercial surfactant, Pluronic L-31 was
Results and discussions
Fig. 1 Top – SEM images for: (a) commercial barium titanate (CO); (b)
cubic barium titanate (CU); (c) barium titanate nanorods (NR). Bottom
– TEM images for: (d) commercial barium titanate (CO); (e) cubic
A poly(dimethylsiloxane)-a,u-diol of high molecular mass was
prepared by a known procedure (bulk cationic ring-opening barium titanate (CU); (f) barium titanate nanorods (NR).
58524 | RSC Adv., 2014, 4, 58522–58529
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
Fig. 3 SEM images in cryofractured section for the sample: (a) C–CU;
(b) C–NR.
The impact of barium titanate particles, their size and
morphology on some characteristics and properties of silicone
composites thereof was studied by comparing the results with
those obtained on a similar sample prepared without ller. Pure
crosslinked PDMS shows very high value for the strain (2198%)
(Fig. 4a and Table 1). This is due to the high molecular mass of
the used polymer. Being long and exible, and because the
crosslinking is achieved through the ends of the chain (giving
freedom between crosslinking nodes), the polymeric chains are
able to slide over each other in the absence of constraints
imposed by other intermolecular forces. A decreasing of this
value is observed as a result of incorporating the llers, when
the strain drops to about 900% but this still is a good value
compared with literature.13,14,37 This would be partially due to
the plasticized effect of surfactant. The decrease of tensile
strength and elongation in PDMS–BT composites proves the
non-reinforcing effect of barium titanate as has also been found
in another study.37
Young's modulus increases as a result of the incorporation of
prismatic barium titanate (commercial and cubic one), and this
is an expected effect, according to the rule of mixtures. However,
in the case of the composite lled with nanorods, Young's
modulus decreases dramatically. We presume that high
dimensional ratio of this ller and its good dispersion within
matrix contribute to this result
Stress–strain cycles, with strains up to 100% of initial length,
show an viscoelastic behaviour (shown by stress relaxation,
Fig. 1 ESI†) in all cases with a clearly visible hysteresis loop only
Fig. 2 X-ray powder diffraction spectra for barium titanate (BT) with
cubic (CU) and nanorod (NR) morphologies in comparison with that of
a commercial (CO) sample; assignment of diffraction peaks was made
according to ref. 35.
used in this aim at a rate of 100 wt% relative to the mass of ller.
Incorporation of the ller in the matrix was performed by
mechanical stirring and sonication; the last operation was done
in order to remove all the air bubbles trapped in the mixture.
The crosslinking occurred in lms cast on the Teon
substrate by condensation using MTAS as a crosslinker agent
and DBTDL as catalyst, in presence of the atmospheric mois-
ture, at room temperature, according to Scheme 1.
The distribution of the ller within the silicone matrix was
emphasized by SEM images taken on cryofractured section
(Fig. 3). Contrary to expectations, the surfactant aggregates
incorporating ller are visible as white particles, probably due
to large amount of surfactant used (surfactant : ller ¼ 1 : 1
weight ratio). However, the aggregates are relatively uniform
distributed within the matrix. The stick morphology of the ller
is observed in the case of the composite C–NR, where these were
used (inset Fig. 3b).
Scheme 1 Crosslinking pathway applied for poly(dimethylsiloxane)- Fig. 4 Mechanical tests for the crosslinked films: (a) stress–strain
a,u-diol matrix.
curves, (b) mechanical fatigue behavior.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 58522–58529 | 58525

View Article Online
RSC Advances
Paper
Table 1 The main parameters of the mechanical and dielectric tests
Young's
Tensile
Dielectric
Dielectric
modulusa, strength, Elongation at permittivity, 30
permittivity, 30
Dielectric loss,
Ebfb,
Sample MPa
MPa
break, %
(at 10 Hz, 25 ꢀC) (at 5 kHz, 25 ꢀC) 300 (at 10 Hz, 25 ꢀC) (kV mmꢂ1
)
bc, MPaꢂ1
s
d (J)
C–R
0.0958
0.20
0.17
0.13
0.16
2198
1196
894
3.11
3.90
9.04
5.29
3.01
3.72
7.96
4.75
0.21
0.15
0.30
0.07
102.28
21.74
18.28
5.76
31.42
15.97
406.12
27.30
6.37
C–CO
C–NR
C–CU
0.2330
0.0196
0.1740
21.98
22.21
0.61
902
a
Young's modulus calculated at 15% strain. b Dielectric strength. c Electromechanical sensitivity estimated according to ref. 6 as a ratio between
the dielectric permittivity at 5 kHz and Young's modulus calculated at 15% strain. d Estimated energy output.40
at the rst strain-release cycle; in the subsequent strain cycles, 20 vol% BT bers42 or 30 php BT nanorods,37 respectively. There
the siloxane chains are rearranged in the lm and the difference is a sudden drop (the sharper as the value of the permittivity is
between strain and release was smaller than 1% of the stress greater) at about 104 Hz for 30, accompanied with the maximum
value at each point (Fig. 4b). In fact, this behavior might be dielectric loss, 300, suggesting a relaxation process occurring at
associated with Mullins effect (rst-cycles stress-soening38), silicone–ller interface. The decrease in the dielectric permit-
the temporary phenomenon which disappears aer a few cycles tivity is due to the inability of the dipoles, beyond a certain
of solicitation when the elastomer response is completely frequency, to orient and to return to the initial position in
stabilized and reproducible as long as the maximum stretching accord with the oscillating electric eld. This happens when the
level is not exceeded.38,39 Permanent set values (Fig. 1 and Table time taken for the dipoles to return to its original random
1 ESI†) reveal that the four samples have different elastic orientation, known as relaxation time, is larger than the rate of
portions of the stress–strain curve. For sample C–R, the elastic oscillating electric eld. Thus, the polarization cannot follow
portion is up to 20%, with 0.01967 MPa of stress. A decreasing the oscillating frequency resulting in the energy absorption and
in the maximum limit of pure elastic strain at 10, 12.5 and dissipation as heat.42
12.8% occurs when commercial, nanorod and cubic barium
titanate particles, respectively, are incorporated in the silicone chanical actuation or energy harvesting devices, the dielectric
elastomer while the stress values increase. parameters are very important besides the mechanical ones.
For further applications as active elements in electrome-
The dielectric behavior of the obtained composites is a The correlation of these two characteristic types is well
normal one (Fig. 5, Table 1). As expected, the dielectric expressed by electromechanical sensitivity, b, calculated
permittivity of the silicone matrix increases by incorporation of according to established procedure6 as the ratio between the
barium titanate, a dielectric ceramic, the effect being more dielectric permittivity at 5 kHz and the Young's modulus at
obvious when particles of ller have high aspect ratio, i.e., 15% strain (Table 1). The obtained value revealed that the
nanorods. Thus, while pure crosslinked PDMS have 30 ¼ 3.01 at morphology of the ller has a signicant inuence on this
10 Hz, this increases at 3.90, 5.29 and around 8 by incorporating parameter. The highest effect proved to have the ller with
commercial barium titanate, prepared cubic and nanorods one, high aspect ratio
– nanorods. By comparing the three
respectively. These values can be considered good if they are composites containing the same percent of ller but with
compared with those reported in the literature for some PDMS– different morphologies, the sample C–NR shows an electro-
BT composites, for example 30 ꢁ 3.8 (1 Hz) at 5 wt% BT particles mechanical sensitivity more than 12 times as that of pure
of 100 nm,41 or 30 ꢁ 6.5 42 and 30 ꢁ 12,37 but obtained at a load of PDMS. The three sensitivity values are higher as compared
with other reported in literature for PDMS–BaTiO3 composites,
for example that used for comparison in (ref. 6) as having 5.7
MPaꢂ1 (no information about the characteristics of the PDMS
used as a matrix, ller content and applied crosslinking
procedure has been provided). As expected, the dielectric
strength of the polymeric matrix consisting in crosslinked
silicone fell when the ller has been incorporated into it (Table
1). However, values for samples C–CO and C–NR remain
comparable with literature.1,13–15 On the basis of the measured
dielectric strength values and elongation determined from
mechanical tests, the energy gained from a stretching and
relaxing uniaxial deformation cycle, s (J), could be estimated
according to literature.40 While the energy output of the
composite C–CU decreases dramatically (about 10 times) than
the reference sample without any ller, C–R, the composite C–
Fig. 5 Dielectric measurements for tested samples: (a) dielectric
constant; (b) dielectric loss as a function of frequency.
58526 | RSC Adv., 2014, 4, 58522–58529
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
NR has shown a harvesting capacity more than three times
Thermogravimetrical analysis results (Fig. 7 and Table 2)
higher (22.21 J) than sample C–R but only slightly higher than reveal that both the reference samples and those lled with
C–CO. The dielectric strength and higher value of elongation at commercial and cubic barium titanate begin to decompose at
break recorded for sample C–CO are those which contribute to 300–310 ꢀC, the process occurring in two steps. Instead, the
raising it close to the energy output level recorded for sample decomposition of the sample C–NR starts at 276 ꢀC and occurs
C–NR, even if it has a lower dielectric constant.
in three steps. Somewhat surprisingly, the reference sample is
ꢀ
For further applications, i.e., actuators, sensors or genera- broken down almost entirely at 700 C, the residue being 0.72
tors, the elastomer should also ensure a stable operation in percent, although, according to silicon content, it would have
different conditions of moisture and temperature.43 The been expected to be 81%. We assume that the presence in the
moisture sorption–desorption isotherms were registered in sample of the dibutyltindilaurate, used as crosslinking catalyst,
dynamic regime for the crosslinked composite lms (Fig. 6). causes the siloxane bond breaking which leads to the frag-
Polydimethylsiloxanes and derived silicone materials are mentation of the long polysiloxane backbone turning it into
known as being very hydrophobic. An increase in moisture volatile compounds. The samples lled with barium titanate
sorption capacity of composites might have been expected due leaves higher residues due to the formation of metal oxides or
to the hydrophilic nature of the ller (barium titanate particles ceramic materials. As can be seen from the data presented in
are easily dispersed in water). However, only a slight increase Table 2, composite samples based on home-prepared barium
in the value of this characteristic was registered. Thus, the titanate show similar values of thermal decomposition resi-
maximum sorption capacity increases from 0.50% in pure dues, 4.76 and 4.37 wt%, respectively, but lower than in the
crosslinked PDMS (C–R) to 1.18% for the composite with cubic case of the sample containing commercial barium titanate that
BT (C–CU), 1.47% for the composite with commercial BT (C– le a residue of 10.13 wt%. This difference may stem from the
CO) and 3.53% in the composite with BT nanorods (C–NR). different preparation methods underlying the commercial and
This diminished effect could be assigned to low loading ller home-made barium titanate and different purity degrees. DSC
level and the surfactant coating barium titanate particles analysis (Table 2) reveals that all the samples show transitions
combined with the known tendency of the polysiloxane to typical of silicone polymers: glass transition in the range ꢂ120
migrate at the interface material–air that will lead to a surface to (ꢂ140) oC, crystallization around ꢂ70 to (ꢂ71) oC and
that is richer in hydrophobic siloxane component. The higher melting in the range ꢂ38 to (ꢂ42) oC without apparent
value for the sorption capacity of the composite with BT modications due to the presence of the ller. The peak cor-
ꢀ
nanorods as compared with the others could be explained by responding to melting (at ꢂ40 C) is smaller in area than the
their higher specic surface. By their shape, isotherm recorded one corresponding for crystallization (at ꢂ70 ꢀC) and this
for the four samples could be associated with an isotherm of indicates that a small amount of crystalline phase develops
type III according to IUPAC classication – specic for non- during the cooling scan, due to the high degree of cross-
porous hydrophobic materials with weak interactions linking. The DSC curves for the cooling scans overlap for the
between adsorbent–adsorbate. The kinetic curves (not shown) reference samples and the samples with barium titanate
for water vapor sorption show the rate of desorption of water nanoparticles. However, the heating scans do not overimpose,
vapors are generally slower than the sorption rate, leading to a especially the sample with barium titanate nanorods C–NR,
small hysteresis loop.
which shows an additional glass transition at ꢂ70 ꢀC. This can
be attributed to the surfactant which, as can be seen in the
SEM image, in Fig. 2b, is organized into larger aggregates in
this case.
Fig. 6 Sorption–desorption isotherms of the reference sample C–R
and composites C–CO, C–CU and C–NR registered at room
temperature (the meaning of the endings: S
Desorption).
–
Sorption; D
–
Fig.
7 Thermogravimetrical curves registered for the prepared
composites and reference sample.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 58522–58529 | 58527

View Article Online
RSC Advances
Paper
Table 2 Main centralized parameters obtained from TG/DTG and DSC data
DSC data
b
c
d
e
f
g
Sample
C–R
Stagea
Tonset (ꢀC)
Tpeak (ꢀC)
Tendset (ꢀC)
Weight loss, %
Residue, wt% (700 oC)
Tg
Tc
Tm
1
2
1
2
310.35
487.37
313.39
451.67
359.2
570
387.21
600.97
369.51
666.14
52.71
44.98
30.77
57.62
0.72
ꢂ121.5
ꢂ120.6
ꢂ70.7
ꢂ38.16
ꢂ39.16
C–CO
344
10.13
ꢂ71.56
528.2
641.3
339.6
382.1
515.1
339
C–NR
1
2
3
1
2
276.46
355.49
473.78
302.31
444.93
330.7
16.92
48.66
27.76
12.36
80.67
4.76
ꢂ124.2
ꢂ70.01
ꢂ71.56
ꢂ42.39
ꢂ39.16
410.76
567.6
365.59
628.45
ꢂ70.99
C–CU
4.37
ꢂ120.61
551.4
a
b
c
Stage of thermal degradation. The temperature at which the thermal degradation begins. The temperature at which the degradation rate
d
e
f
reaches its maximum value. The temperature at which the process ends. Glass transition on the second heating run. Crystallization
temperature on the rst cooling run. g Melting temperature on the second heating run.
Framework Programme of European Community and co-
nanced by UEFISCDI (Contract 205EU). Support by the COST
Conclusions
Action (MP1003) “European Scientic Network for Articial
Muscles (ESNAM)”, through Short-Term Scientic Mission
COST-STSM-MP1003-14672, and PERCRO Laboratory – TeCIP
Institute – Scuola Superiore Sant'Anna Pisa Italy for technical
support are also gratefully acknowledged. The authors
addressed his thanks to Dr V. Musteata, C. D. Varganici and F.
Doroei, for registration of dielectric spectra, thermal curves,
and SEM images, respectively.
Barium titanate, with cubic and nanorod morphologies
respectively, were obtained by hydrothermal procedure and
then were incorporated in a high molecular weight polydi-
methylsiloxane matrix, followed by crosslinking. Very long
chains and crosslinking system (through the ends of chains)
used have created prerequisites for high elongations that
actually have been acquired (between 894 and 2198%). Incor-
poration of barium titanate into such an array resulted in
composite materials with improved dielectric characteristics.
The most effective was proven to be the ller having particles
with high dimensional ratio (nanorods) which has led to a
signicant increase of dielectric permittivity of resulted
composite (about three times) as compared with the matrix, the
mechanical characteristics remaining in the domain of interest
(Young modulus 0.0196 MPa and elongation around 900%).
The electromechanical sensitivity and gained energy capacity,
parameters of interest for further target applications (active
elements in electromechanical devices), estimated on the basis
of the electrical and mechanical characteristics, proved to be 12
and more than three times higher, respectively for the
composite lled with barium titanate nanorods than for pure
crosslinked PDMS. These effects were obtained at lower load (5
wt%) than those reported in the literature for similar results.
According to our knowledge, this is the rst time when a PDMS
having such high molecular weight is addressed in such
purposes. All composites are stabilized aer their rst loading
cycle. The moisture sorption and thermal stability of the PDMS
changed to slightly increased values by the incorporation of the
ceramic llers, while the thermal phase transitions of the sili-
cone matrix are almost unaffected by their presence.
Notes and references
1 Z. Wang, J. K. Nelson, J. Miao, R. J. Linhardt, L. S. Schadler,
H. Hillborg and S. Zhao, IEEE Trans. Dielectr. Electr. Insul.,
2012, 19, 960–967 and references herein.
2 B. Li, H. Chen and J. Zhou, Composites, Part A, 2013, 52,
55–61.
3 G. Kofod, H. Stoyanov, M. Kollosche, S. Risse, H. Ragusch,
´
D. N. McCarthy, R. Wache, D. Rychkov and
¨
M. Dansachmuller, Molecular level materials design for
improvements of actuation properties of dielectric
elastomer actuators, in Proc. SPIE 7976, Electroactive
Polymer Actuators and Devices (EAPAD) 2011, March 2011,
vol. 79760J, DOI: 10.1117/12.880981.
4 S. Zhang, C. Zou, X. Zhou, D. Anderson, B. Zellers and
Q. Zhang, Polymer lm capacitors with high dielectric
constant, high capacitance density, and high energy
density, in Power Modulator and High Voltage Conference
(IPMHVC), 2010, IEEE International, 2010 May 23–27,
Atlanta, GA, pp. 221–224.
5 X. Li, R. Zhang, W. Yu, K. Wang, J. Wei, D. Wu, A. Cao, Z. Li,
Y. Cheng, Q. Zheng, R. S. Ruoff and H. Zhu, Sci. Rep., 2012, 2,
870.
6 H. Zhao, D. R. Wang, J. W. Zha, J. Zhao and Z. M. Dang, J.
Mater. Chem. A, 2013, 1, 3140–3145.
Acknowledgements
The work presented in this paper is developed in the context of
the project PolyWEC (http://www.polywec.org, prj. ref. 309139),
a FET-Energy project that is partially funded by the 7th
7 G. M. Spinks, G. G. Wallace, L. Liu and D. Zhou, Macromol.
Symp., 2003, 192, 161–170.
58528 | RSC Adv., 2014, 4, 58522–58529
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
8 P. Murugaraj, D. Mainwaring, N. Khelil, J. L. Peng, R. Siegele 25 S. Cho, J. Lee, J. Hyun and K. Paik, Mater. Sci. Eng., B, 2004,
and P. Sawant, Carbon, 2010, 48, 4230–4237. 110, 233–239.
9 C. Naber, C. Tanase, P. Blom, G. Gelinck, A. Marsman, 26 T. Hu, J. Juuti, H. Jantunen and T. Vilkman, J. Eur. Ceram.
F. Touwlager, S. Setayesh and D. de Leeuw, Nat. Mat., 2005,
4, 243–248.
Soc., 2007, 27, 3997–4001.
27 D. Khastgir and K. Adachi, Polymer, 2000, 41, 6403–6413.
10 R. Shankar, T. K. Ghosh and R. J. Spontak, So Matter, 2007, 28 W. M. Winslow, J. Appl. Phys., 1949, 20, 1137–1140.
3, 1116–1129.
29 K. Koyama, K. Minagawa, T. Watanabe, Y. Kumakura and
11 S. Koulouridis, G. Kiziltas, Y. Zhoz, D. J. Hansford and
J. Takimoto, J. Non-Newtonian Fluid Mech., 1995, 58, 195–206.
J. L. Volaki, IEEE Trans. Microwave Theory Tech., 2006, 54, 30 K. Koyama, Int. J. Mod. Phys. B, 1996, 10, 3067–3072.
4202–4208.
31 L. Guo, H. Luo, J. Gao, L. Guo and J. Yang, Microwave
hydrothermal synthesis of barium titanate powders, Mater.
Lett., 2006, 60(24), 3011–3014.
12 M. Andriot, S. H. Chao, A. Colas, S. Cray, F. de Buyl,
J. V. DeGroot, A. Dupont, T. Easton, J. L. Garaud,
E. Gerlach, F. Gubbels, M. Jungk, S. Leadley, J. P. Lecomte, 32 M. Cazacu, M. Antohi, C. Racles, A. Vlad and N. Forna, J.
B. Lenoble, R. Meeks, A. Mountney, G. Shearer, S. Stassen, Compos. Mater., 2009, 43, 2045–2055.
C. Stevens, X. Thomas and A. T. Wolf, in Silicones in 33 K. A. Andrianov, A. A. Zhdanov and A. A. Bogdanova, Dokl.
Industrial Applications, ed. F. Gubbels, R. D. Jaeger and M. Akad. Nauk SSSR, 1954, 94, 697–699.
Gleria, Nova Science Publishers, New York, 2007, pp. 61–161. 34 Z. Lin, Y. Yang, J. M. Wu, Y. Liu, F. Zhang and Z. L. Wang, J.
13 D. M. Opris, M. Molberg, C. Walder, Y. S. Ko, B. Fischer and
Phys. Chem. Lett., 2012, 3, 3599–3604.
¨
F. A. Nuesch, Adv. Funct. Mater., 2011, 21, 3531–3539.
35 Z. Lazarevic, N. Romcevic, M. Vijatovic, N. Paunovic,
M. Romcevic, B. Stojanovic and Z. D. Mitrovic, Acta Phys.
Pol., A, 2009, 115, 808–810.
´
¨
14 S. Risse, B. Kussmaul, H. Kruger, R. Wache and G. Kofod,
DEA material enhancement with dipole graing PDMS
network, in Proc SPIE. 7976, Electroactive Polymer Actuators 36 Z. Shen, W. Zhang, J. Chen and J. Yun, Chin. J. Chem. Eng.,
and Devices (EAPAD) 2011, March 24, 2011, vol. 79760N,
2006, 14(5), 642–648.
DOI: 10.1117/12.881919.
15 S. Risse, B. Kussmaul, H. Kruger and G. Kofod, Adv. Funct.
37 A. Nayak, T. K. Chaki and D. Khastgir, Adv. Mater. Res., 2013,
622–623, 897–900.
¨
Mater., 2012, 22, 3958–3962.
38 L. Mullins, Rubber Chem. Technol., 1969, 42(1), 339–362.
16 C. Racles, M. Cazacu, B. Fischer and D. M. Opris, Smart 39 R. Vertechy, M. Fontana, G. Stiubianu and M. Cazacu, Proc
Mater. Struct., 2013, 22, 104004.
SPIE 9056, Electroactive Polymer Actuators and Devices
(EAPAD) 2014, March 8, 2014, vol. 90561R, DOI: 10.1117/
12.2045053.
¨
17 S. Risse, B. Kussmaul, H. Kruger and G. Kofod, RSC Adv.,
2012, 2, 9029–9035.
18 J. Chon, S. Ye, K. J. Cha, S. C. Lee, Y. S. Koo, J. H. Jung and 40 S. Chiba, M. Waki, R. Kornbluh and R. Pelrine, Smart Mater.
Y. K. Kwon, Chem. Mater., 2010, 22, 5445–5452. Struct., 2011, 20, 124006.
19 Y. Rao, S. Ogitani, P. Kohl and C. P. Wong, J. Appl. Polym. Sci., 41 A. L. Skov, A. Bejenariu, J. Bogelung, M. Benslimane and
2002, 83, 1084–1090.
20 G. Subodh, V. Deepu, P. Mohanan and M. Sebastian, Appl.
Phys. Lett., 2009, 9, 062903.
A. D. Egede, Electroactive Polymer Actuators and Devices
(EAPAD) 2012, SPIE, 2012, vol. 8340, p. 83400M, DOI:
10.1117/12.912114.
21 L. Tu and A. Ilhan A, Cryst. Growth Des., 2001, 1(5), 401–419. 42 Z. Ahmad, Polymer Dielectric Materials, Dielectric Material,
22 A. Sihvola and E. Alanen, IEEE Trans. Geosci. Electron., 2002,
29, 679–687.
ed. M. Silaghi, ISBN: 978-953-51-0764-4, 2012, InTech, DOI:
10.5772/50638, http://www.intechopen.com/books/
23 Y. Bai, Z. Y. Cheng, V. Bharti and H. Xu, Appl. Phys. Lett.,
2000, 76, 3804–3806.
24 K. Cheng, C. M. Lin, S. F. Wang, S. T. Lin and C. F. Yang,
Mater. Lett., 2007, 61, 757–760.
dielectric-material/polymer-dielectric-materials.
43 E. A. Chigorina, T. M. Chigorina, A. A. Arutyunyants and
M. V. Bestaev, Polym. Sci., Ser. D, 2010, 3(4), 228–230.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 58522–58529 | 58529