ORIGINAL PAPER

Polydimethylsiloxane/silica/titania composites preparedby solvent-free sol–gel technique

Mihaela Alexandru • Maria Cazacu • Alexandra Nistor •

Valentina E. Musteata • Iuliana Stoica •

Cristian Grigoras • Bogdan C. Simionescu

Received: 7 May 2010 / Accepted: 10 August 2010 / Published online: 21 August 2010

� Springer Science+Business Media, LLC 2010

Abstract Composites based on polydimethylsiloxane

incorporating silica and titania were prepared by mixing

polydimethylsiloxane with proper oxides precursors, tet-

raethyl-orthosilicate and tetrabutyl-orthotitanate. In the

presence of environmental humidity and in acid catalysis,

hydrolysis/condensation processes take place with forma-

tion of oxides and concomitantly polymer crosslinking.

Partial replacement of SiO2 in a polydimethylsiloxane/

silica composite with titania (both generated in situ by

sol–gel process) affects surface hydrophilicity (evaluated

by dynamic contact angle), water vapor sorption ability

(determined by dynamic vapor sorption) and thermal sta-

bility. The dielectric properties are also controlled by

composition.

Keywords Composites � Sol–gel technique �Polysiloxanes � Titania � Silica

1 Introduction

The obtaining of composite materials by sol–gel technique

has been widely investigated [1–5]. As compared to other

techniques, the sol–gel approach has several advantages:

low costs, low temperature of heat treatment, unique ability

to achieve molecular level uniformity in the synthesis of

organic–inorganic composites, and strong adhesion of the

coating to the substrate [6, 7].

Polydimethylsiloxane/SiO2 composites prepared by the

reaction of polydimethylsiloxane (PDMS) and tetraethyl-

orthosilicate (TEOS) have also been extensively studied

[8–16]. These materials can be considered as ‘‘ceramic

rubbers’’ depending on the TEOS/PDMS molar ratio.

When PDMS concentration is increased, the final material

presents rubber-like properties. For high TEOS concen-

trations, hard composites are obtained. The rubbery prop-

erties are dependent on various reaction parameters, such

as temperature, acid concentration, reaction time, etc.

The incorporation of different inorganic components,

instead of or aside from TEOS, into the hybrid structure is

usually carried out in order to improve the mechanical,

thermal, and optical properties or to obtain new properties

derived from the hybrid nature of the material [17–22].

Due to their interesting properties such as elasticity,

insulating ability and easy processing, silicones are used in

microelectromechanical systems (MEMS) where they play

a structural role as protective layers, encapsulating ele-

ments, valves and diaphragms. However, by using active

fillers, the dielectric properties of silicones can be modified

[23, 24]. Titania has a high dielectric constant (e * 89)

being of real interest for this purpose. In addition, it is an

important inorganic functional material, with good physical

properties, which make it suitable for thin film applica-

tions. Films containing TiO2 have been often used in

microelectronic devices, e.g. in capacitors, or as a dielectric

gate in metal-dielectric-semiconductor devices. Titania

occupies also an important place as a photocatalyst, due to

its high photocatalytic activity, excellent functionality,

high chemical and thermal stability and non-toxicity [25].

The preparation of PDMS/SiO2/TiO2 composites by

the sol–gel method has already been reported in literature

M. Alexandru (&) � M. Cazacu � A. Nistor �V. E. Musteata � I. Stoica � C. Grigoras � B. C. Simionescu

‘‘Petru Poni’’ Institute of Macromolecular Chemistry, Aleea Gr.

Ghica Voda 41 A, Iasi 700487, Romania

e-mail: amihaela@icmpp.ro

B. C. Simionescu

Department of Natural and Synthetic Polymers, ‘‘Gh. Asachi’’

Technical University of Iasi, Iasi 700050, Romania

123

J Sol-Gel Sci Technol (2010) 56:310–319

DOI 10.1007/s10971-010-2307-5

[19, 26, 27] and phase-separated materials in which SiO2

and TiO2 particles are dispersed in the silicone matrix were

generally obtained [28, 29]. As an alternative to literature

approaches, the present paper deals with the obtaining of

PDMS/SiO2/TiO2 composites in the absence of solvents.

Solvent-free systems present environmental advantages

and, on the other hand, this method diminishes the problem

created by the solvent removal at the end of the reaction,

which determines porosity in the material. In addition, this

would be a preferred alternative when a high molecular

mass polydimethylsiloxane is used as polymer matrix,

taking into account that this is insoluble in solvents com-

monly used in sol–gel technique. A series of composites

based on PDMS filled with in situ generated SiO2 and TiO2

was prepared. The influence of TiO2 content on the

dielectric properties was investigated. Other properties of

interest in dielectric applications, namely surface (dynamic

contact angle, dynamic vapor sorption) and thermal prop-

erties were also studied.

2 Experimental

2.1 Materials

Polydimethylsiloxane-a,x-diol (Mv ¼ 48; 000) was pre-

pared according to a previously described procedure [8, 30].

Tetraethyl-orthosilicate (TEOS), purchased from Fluka

(purity [ 98%, b.p. = 163–167 �C, d420 = 0.933) was used

as received. Tetrabutyl-orthotitanate (TBT) (d = 0.966

g/cm3, b.p. = 310–314 �C, m.p. = -55 �C, d420 = 1.486).

Dibuthyltin dilaurate (DBTDL, d420 = 1.055) was received

from Merck-Schuchardt, and was used as received.

2.2 Equipments

Fourier transform infrared (FTIR) spectra were obtained on

a Bruker Vertex 70 FTIR analyzer. The analyses were

performed in transmission mode, in the 400–4,100 cm-1

range, at room temperature with 2 cm-1 resolution and

accumulation of 32 scans. The ground samples were

incorporated in dry KBr and processed as pellets.

Water vapors sorption capacity of the film samples was

measured by using the fully automated gravimetric ana-

lyzer IGAsorp supplied by Hiden Analytical, Warrington

(UK). An ultrasensitive microbalance measures the weight

change as the humidity is modified in the sample chamber

at a constant regulated temperature. The measurement

system is controlled by a user-friendly software package.

Dynamic contact angles (DCA) and contact angle hys-

teresis were measured on films by using a KSV Sigma 700

tensiometer system—a modular high-performance, com-

puter controlled surface tension/contact angle meter. The

measurement parameters were: advancing–receding speed,

5 mm/min; start depth, 0 mm; immersion depth, 8 mm;

number of cycles, 3; the average values were taken into

consideration.

Thermogravimetric measurements (TGA) were per-

formed in the 0–750 �C temperature range at a heating rate

of 10 �C/min in air using a Q-1500D System.

Glass transition and melting processes of PDMS based

composites were followed using a Pyris Diamond DSC

(Perkin Elmer USA) instrument. The samples were cooled

from room temperature to -150 �C, held at this tempera-

ture 2 min, and then heated up to 50 �C at a heating rate of

20 �C/min. Helium gas was purged through the cells at

20 ml/min to assure an inert atmosphere and good thermal

conductivity. Before measurements, the DSC instrument

was calibrated for temperature and energy scale using n-

hexane and pure water as recommended standards for LN2

range of DSC analysis. The glass transition temperature

was calculated as a midpoint of the heat capacity of the

sample.

Novoncontrol setup (Broadband dielectric spectrometer

Concept 40, Germany), integrating an ALPHA frequency

response analyzer and a Quatro temperature control sys-

tem, was used to investigate the dielectric properties of the

polymer composites over a broad frequencies window,

100–106 Hz, in the -140 7 30 �C temperature range. The

bias voltage applied across the sample was 1.0 V. Samples

having uniform thickness in the 0.2–0.9 mm range were

placed between gold plated round electrodes, the upper

electrode having a 20 mm diameter.

The AFM measurements were made on a Scanning

Probe Microscope Solver Pro-M platform (NT-MDT,

Russia), in air, in semi-contact mode, using a rectangular

NSG10/Au cantilever with a nominal elasticity constant

KN = 11.5 N m-1 and a 10 nm radius of curvature of the

tip. A 257.8 kHz oscillation frequency was used. The scan

area was 20 lm 9 20 lm, 256 9 256 scan point size

images being thus obtained. The AFM image processing

and the calculation of the surface texture parameters were

realized by the Nova Software (NT-MDT, Russia).

2.3 Composites preparation

The synthesis of a polydimethylsiloxane/silica/titania com-

posite was carried out as follows. 1.000 g (0.020 mmol) of

PDMS with Mv ¼ 48; 000 was introduced in a Teflon dish

and mixed with 0.500 g (2.400 mmol) TEOS and 0.038 g

(0.111 mmol) TBT. After about 10 min of stirring, 0.003 g

(0.005 mmol) of DBTDL was added and the stirring was

continued for another 10 min. The mixture was poured on a

Teflon foil and vacuumed for 10 min to eliminate the

incorporated air. This sample was labeled as T5. The other

J Sol-Gel Sci Technol (2010) 56:310–319 311

123

composites were obtained following the procedure described

above but using amounts of reactants according to Table 1.

3 Result and discussions

A polydimethylsiloxane-a,x-diol was used as matrix for

preparing composite materials with SiO2 and TiO2 by

adapting a solvent-free sol–gel procedure. We chose this

way because as is known, PDMSs of high molecular mass

(e.g., 48,000) have a low solubility in solvents that are

common for sol–gel technique, such as alcohols, ketones

and dimethylsulfoxide. Instead, they are soluble in non-

polar solvents like hydrocarbons and their halogenated

derivatives, many of them being toxic.

For this purpose, oxide precursors (TEOS and TBT) and

a condensation catalyst (DBTDL) were added in pre-

established amounts to the polymer. After energic stirring

and gas removing, the mixtures were processed as films.

The hydrolysis of the corresponding precursors took place

under the influence of the atmospheric humidity followed

by catalytic condensation with the formation of silicon and

titanium oxides networks [26]. In addition, as a result of

the reactions of the alkoxydes and their hydrolysates with

Si–OH PDMS end groups, a crosslinking of PDMS occu-

red. Obviously, TEOS and TBT also act as crosslinking

agents (Scheme 1).

The method used for the samples preparation in this study

is based on the same principle as the room temperature

vulcanization of the silicone rubber (RTV), in which the

crosslinking of the polydimethylsiloxane-a,x-diol occurs by

polycondensation reactions in the presence of environmental

humidity, and needs a few days for curing. Water must dif-

fuse deep into the film, while low molecular condensation

compounds (i.e. alcohol) migrate outwards [31]. Therefore,

the films were kept in the laboratory environment for about

2 months before investigations, when the weighting dem-

onstrated the mass stabilization. In our case there was a

relative humidity of 40–60% in the laboratory. The obtained

white opaque films (of about 0.2–0.9 mm thickness) were

easily peeled off from the substrate.

The inorganic part of these composites is considered to

be the sum of SiO2 and TiO2, while the organic one is

represented by PDMS. The SiO2/TiO2 ratio was varied, but

their cumulated amounts remained constant. While silica is

introduced as a reinforcing material, the expected effect of

titania is to modify PDMS dielectric properties [32].

3.1 Fourier transform infrared spectroscopy

Figure 1 shows the FT-IR spectra of PDMS based com-

posites obtained by using different TEOS/TBT mass ratios

in the spectral range between 4,100 and 400 cm-1. The

composites show absorption bands around 2,900 and

2,964 cm-1, assigned to stretching vibrations of C–H in

methyl groups. The absorption band around 1,262 cm-1 is

the main characteristic band of methyl groups bonded to

silicon and assigned to the symmetric deformation of C–H.

The absorption bands around 1,020–1,097 cm-1 are

assigned to Si–O–Si stretching vibrations in siloxane

network [33, 34]. The broad peaks of OH bond are

observed at 3,000–3,500 cm-1 for all samples [29]. TBT

Table 1 The recipes used to prepare PDMS/SiO2/TiO2 composites

Sample Inorganic Organic Catalyst

TEOS

(wt. %)

TBT

(wt. %)

PDMS

(wt. %)

DBTDL

(wt. %)

T0 35 0 65 0.165

T5 32.5 2.5 65 0.196

T7.5 31 4 65 0.245

T10 30 5 65 0.222

T15 28 7 65 0.132

T20 26 9 65 0.166

Tm 63 5 32 0.174

Si O

CH3

CH3

Si O Si

CH3

CH3

O Si

O

O

O

Si

Si

Si

Si O

O

O

Si

CH3

CH3

n

O

O

O

O

O Si

Si

O

Si O Si

O

Si

O

Ti

O

O

O

Ti

Si

Si

OC2H5

OC2H5

OC2H5

OC2H5 TiO

OC4H9

OC4H9

OC4H9

C4H9SiOH

CH3

O

CH3

Si O

CH3

CH3

Si

CH3

CH3

OHn

+ +

DBTDLroomtemp.

Scheme 1 Proposed

networking within composites

312 J Sol-Gel Sci Technol (2010) 56:310–319

123

self-condensation reaction gives Ti–O–Ti bonds, while its

copolymerization with TEOS or PDMS gives Ti–O–Si

bonds [27]. Some reports suggest that the Ti–O–Si bonds

formed as a result of the reactions between TBT, TEOS and

PDMS are unstable, disappearing during the aging process

[35]. However, a certain amount of Ti–O–Si bonds can be

found in the final material. To verify the formation of the

Si–O–Ti bond, not visible in our spectra due to the higher

content in PDMS as compared with the inorganic part, a

model sample was prepared based on an inversed mass ratio

between the organic and inorganic components (sample Tm,

Table 1). One can observe, in the FTIR spectrum of the Tm

sample, a shoulder in the 910–960 cm-1 region, shoulder to

be assigned to the Si–O–Ti bond [36]. Due to the fact that in

this region Si–OH bond could be also present [37], the

sample was calcinated at 900 �C (Tm-c), the band centered

around 951 cm-1 and corresponding to Si–O–Ti vibration

thus becoming well defined. According to literature [38],

the intermolecular interaction between titania and silica

in composites with predominant SiO2 content occurs by

replacement of silicon atoms in the SiO4-4 tetrahedra with

titanium atoms, while maintaining the tetrahedral coordi-

nation of titanium with respect to oxygen. The reverse

substitution in the titanium dioxide-enriched composite is

difficult, probably due to the limited coordination ability

[38]. A strong band is developed at 478 cm-1 assigned to

both Si–O–Si and Ti–O–Ti overlaped bonds [39].

3.2 Vapor sorption capacity

Water vapors uptaking capacity for the samples at 25 �C in

the 0–90% relative humidity range (RH) was investigated

by using the IGAsorp equipment. The vapors pressure was

increased in 10% humidity steps, every having a pre-

established equilibrium time between 30 and 40 min

(minimum time and time out, respectively). At each step,

the weight gained was measured by electromagnetic com-

pensation between tare and sample when equilibrium was

reached. An anti-condensation system was available for

vapor pressure very close to saturation. The cycle was

ended by decreasing the vapor pressure in steps to obtain

also the desorption isotherms. The drying of the samples

before sorption measurements was carried out at 25 �C in

flowing nitrogen (250 ml/min) until the weight of the

sample was in equilibrium at RH \ 1%. The sorption/

desorption isotherms registered in these conditions are

presented in Fig. 2.

The partial replacement of SiO2 in a PDMS-silica

composite with TiO2 (both generated in situ by the sol–gel

process) yielded an increase of water vapor sorption

capacity from 0.4 (T0) up to 1.0–1.1 (T5–T20). According

to IUPAC classification, the sorption–desorption curves

can be associated to type III curves. These types of iso-

therms describe sorption on hydrophobic/low hydrophilic

material with weak sorbent–water interactions [40]. GAB

and BET kinetic models applied to the obtained data gave

the values presented in Table 2.

Fig. 1 IR spectra of the PDMS/SiO2/TiO2 composites

Fig. 2 Rapid water vapors sorption isotherms for composites

J Sol-Gel Sci Technol (2010) 56:310–319 313

123

As one can see, the surface area values increase with

TiO2 content although the average pore size remains

almost constant (1.2–1.6 nm). The incompatibility between

the oxide and the polymer, leading to the increasing pore

number, can explain this behavior [41].

3.3 Dynamic contact angle

Film surface wettability was analyzed by measuring the

dynamic contact angle on composites films by using the

tensiometric method (Wilhelmy plate technique). Water

was used as measurement liquid. The DCA runs were

performed on rectangular films. The obtained contact angle

values are given in Table 3.

The water contact angle depends on the polarity of the

surface, i.e. by increasing the polarity, the hydrophilicity

increases. This happened by incorporating TiO2 in the

studied composites. The presence of Ti–O–Ti bonds con-

tributes to the hydrophilicity. A decrease of advancing

contact angles was observed with increasing titania content

in the samples. The differences between the maximum

advancing and minimum receding contact angle values,

known as contact angle hysteresis—a measure of surface

heterogeneity and roughness [42]—were calculated. The

decreasing hysteresis value is determined by the surface

smoothing caused by the increase of TiO2 content in the

sample.

3.4 Thermogravimetric measurements

Results of thermogravimetric analysis of the prepared

composites in the 0–750 �C temperature range in air are

shown in Fig. 3.

Two weight loss stages are observed, below 350 �C and

between 350 and 550 �C, the latter one being a significant

weight loss stage. The weight loss below 350 �C is

attributed to the evaporation of free water, the volatiliza-

tion and the thermal decomposition of the remnant organic

solvents. Between 350 and 550 �C, the weight losses could

be probably ascribed to the further combustion of organic

moieties [43, 44].

The samples containing TiO2 yield a larger amount of

residue during the thermo-oxidative decomposition than

sample T0 that contains only SiO2. It was previously

shown [31, 45, 46] that traces of organometallic catalyst

favors depolymerisation with formation of volatile com-

pounds and lowers the residue amount. According to the

Table 2 Maximum water

vapors sorption values for the

composites

Sample The average

pore size (nm)

Weight

(% d.b.)

BET analysis (5–35%) GAB analysis (5–80%)

Area

(m2/g)

Monolayer

(g/g)

Area

(m2/g)

Monolayer

(g/g)

T0 1.37 0.4552 6.622 0.001886 – –

T5 1.27 1.0891 17.107 0.004873 20.866 0.005943

T7.5 1.39 1.0947 15.773 0.004492 18.806 0.005356

T10 1.59 1.0282 12.996 0.003702 15.207 0.004331

T15 1.25 1.0414 16.675 0.004749 20.703 0.005897

T20 1.44 1.1290 15.679 0.004466 18.422 0.005247

Table 3 The main parameters of the water contact angle

measurements

Sample %TiO2 H2O

A R H

T0 0 109.89 77.82 32.07

T5 5 97.13 74.63 22.5

T7.5 7.5 83.74 75.06 8.68

T10 10 94.42 81.6 12.82

T15 15 84.46 73.70 10.76

T20 20 76.39 77.32 –Fig. 3 The TGA curves of PDMS/SiO2/TiO2 composites

314 J Sol-Gel Sci Technol (2010) 56:310–319

123

here reported data, in the presence of TiO2 this process is

hindered.

3.5 Differential scanning calorimetry

PDMS is well known as a polymer having very low glass

transition and melting temperatures down to 0 �C. Glass

transition processes in PDMS/SiO2/TiO2 systems resulted

from DSC scans are shown in Fig. 4a for all samples,

including the T0 sample. It is noticeable that the T0 sample

has a higher glass transition temperature (-118 �C) as

compared to the composites with TiO2 (T5….T20)

(Table 4) which show glass transition processes around

-123 �C for all compositions (with small, insignificant

differences up to 1 �C). It is reasonable to consider that, by

filling with SiO2 and crosslinking of PDMS, a reduction of

chain mobility occurs, which leads to the increase of Tg.

The incorporation of TiO2 in the PDMS/SiO2 system in

various fractions generally has a contrary effect; the

increasing of the free volume and the increased chain

mobility allows a lower glass transition temperature. The

amorphous phase of the material is affected by both the

chemical linking of SiO2 and by the addition of TiO2 (no

matter the amounts added).

The melting endotherms of PDMS/SiO2/TiO2 compos-

ites are shown in Fig. 4b. Sample T0 clearly shows a

melting process around -45 �C. This sharp endothermic

peak suggests that the PDMS/SiO2 system has a crystalline

morphology consisting of well-defined crystallites that

need a high thermal energy to melt (24.42 J/g). All com-

posites containing TiO2 present lower-energy melting en-

dotherms with enthalpy values decreasing from T5 to T20,

as shown in Table 4. The shape of melting endotherms for

samples which contain TiO2 shows a duality of their

crystalline morphology, which comes from the develop-

ment of two major types of crystallites. The first one is

formed by the PDMS/SiO2 composite and melts at -47 �C,

while the second type is developed by the PDMS/SiO2/

TiO2 composite and has a melting interval around -43 �C.

This duality of morphology is not very clearly separated

because the two types of crystallites have almost the same

thermodynamic stability, and, in addition, at least one

polymer chain that forms a crystallite type may pass

through the other crystallite type. According to Fig. 4b, for

higher contents of TiO2 (higher than 15%) in the PDMS/

SiO2 system the dual morphology becomes well defined.

The morphology developed only by PDMS/SiO2 fraction

seems to be preserved.

3.6 Dielectric measurements

Measurements of the complex dielectric permittivity

e* = e0 - ie00 (where e’ is the storage component and e00 is

the loss component) were carried out at seven fixed fre-

quencies, by sweeping the temperature from -140 �C up to

30 �C with 5 �C/min heating rate. The real and imaginary

parts of complex dielectric permittivity have a direct

physical interpretation. e0 is related to the reversible energy

stored in the material by polarization, whereas e00 is pro-

portional to the energy which is dissipated per cycle,

divided into relaxation and conductivity contribution

(energy required to align dipoles and move ions).Fig. 4 Glass transition (a) and melting endotherms (b) in PDMS/

SiO2/TiO2 composites

Table 4 The main parameters of the DSC curves registered for the

composites

Sample Tg (�C) DCp J/g(�C) Tm (�C) DHm (J/g)

T0 -118.93 0.09 -44.60 24.425

T5 -123.27 0.229 -40.61 19.145

T7.5 -123.52 0.073 -41.89 18.38

T10 -123.32 0.269 -41.23 19.713

T15 -123.36 0.089 -40.84 17.634

T20 -123.66 0.059 -40.58 17.65

J Sol-Gel Sci Technol (2010) 56:310–319 315

123

For the T5 sample the data are represented as the real

part of complex dielectric permittivity (e0) and the imagi-

nary part or dielectric loss (e00) as a function of temperature

and frequency in Fig. 5a and 5b, respectively.

The temperature dependence of dielectric parameters, e0

and e00, exhibits three regions which are associated to the

mobility of the polymer chains. Initially, e0 presents an

increasing step and e00 a peak which corresponds to the

segmental a relaxation associated with the glass transition

of amorphous PDMS. The shifting of these peaks to higher

temperatures by increasing frequency is a characteristic of

the dielectric relaxation. Due to the frequency of the

dielectric measurement, the loss peak appears at a tem-

perature significantly higher (-100 �C) than the calori-

metric Tg (-123 �C). A a relaxation peak follows, this one

being partly superposed by a smaller peak at higher tem-

peratures. The last one could result from crystallization of

the sample during the heating scan. The idea of crystalli-

zation is supported by e0 behavior, which displayed a

decrease at the same temperature; at this temperature the

mobility of the dipoles is reduced due to the immobiliza-

tion and/or constraint of some fractions of the responding

dipoles by increasing bulk crystallinity.

At temperatures higher than -40 �C, an increase in the

permittivity of composites by increasing temperature is

observed, especially at lower frequencies, due to the

Fig. 5 Dielectric permittivity e0 (a) and dielectric loss e00 (b) as a

function of temperature and frequency for T5 sample

Fig. 6 Dielectric permittivity, dielectric loss versus temperature at

1 kHz

Fig. 7 Dielectric permittivity versus TiO2 percent at 25 �C

316 J Sol-Gel Sci Technol (2010) 56:310–319

123

interfacial polarization at electrode/sample or amorphous/

crystalline interface. This process appears at lower fre-

quencies, when the mobile charges have enough time to

migrate between boundaries.

From dielectric measurements, it appears that e00

increases with temperature, especially at low frequencies,

this behavior being attributed to the increased conductivity;

the mobility of the charge carriers increase by increasing

Fig. 8 AFM images of the composites

J Sol-Gel Sci Technol (2010) 56:310–319 317

123

temperature. These effects on e0 and e00 values are more

pronounced in the composite samples containing TiO2.

No significant variation of the a relaxation temperature

with composition is observed, this demonstrating that the

glass transition of polymer matrix is not considerably

influenced by the introduction of TiO2, as also emphasized

by DSC (Fig. 6).

As shown by the vapor sorption study, the samples with

higher TiO2 content have an increased hydrophilicity,

suggesting that it’s possible for those samples to have a

higher content of absorbed atmospheric water. It is known

that the dielectric constant of a material increases by

uptaking water [47] because the concentration of mobile

dipoles becomes higher, leading to the increase of the

permittivity [48, 49]. Therefore, the increase of the

dielectric constant by rising the TiO2 content could have

two causes: the polarity of TiO2 (e = 89) and the polarity

of water (e = 78.5 at 25 �C) [50]. Taking into account that

the amount of absorbed water is smaller (*1.1%) than the

TiO2 content (5–20%), it is presumed that the increase in

dielectric constant is mainly due to the titania amount.

However, the increase in dielectric loss at positive tem-

peratures for the samples containing TiO2 could be due to

the absorbed water molecules.

On the other hand, the slight increasing of the dielectric

constant values of the samples with TiO2 against those

without (Fig. 7) implies that the TiO2 particles are uni-

formly dispersed in the PDMS matrix. Otherwise, the

porosities associated with the agglomerated TiO2 particles

would significantly degrade the dielectric constant.

3.7 Atomic force microscopy

Figure 8 shows AFM images of the T0–T10 composites.

The films have relatively high surface roughness (the root

mean square roughness is about 10 nm). Table 5 presents

the surface particles characteristics and roughness param-

eters of T0–T10 films, with 20 9 20 lm2 scanned areas,

corresponding to the 2D AFM images.

Although the T0 particles are smaller (the average of 10

measurements is 812 nm), roughness parameters have

higher values (Sa is 13.56 nm and Sq is 23.37 nm), and the

particles tend to agglomerate. In the case of T10 sample,

the particles have a larger average diameter, about

1427 nm, but are not crowded, thus determining a lower

film roughness (Sa is 6.03 nm and Sq is 10.40 nm). The

introduction of TiO2 in PDMS/SiO2 composites leads to an

increase of particle size, but one can not differentiate

between TiO2 and SiO2 particles.

4 Conclusions

A series of PDMS/SiO2/TiO2 composites have been pre-

pared by solvent-free sol–gel technique. The amounts of

TiO2 were quite small and didn’t induce spectacular

effects, but higher amounts of TiO2 yield low quality films

and structuration processes. The partial replacement

of SiO2 in a PDMS-silica composite with TiO2 (both

generated in situ by the sol–gel process) increases the

Table 5 Surface particles characteristics and roughness parameters of T0–T10 films, corresponding to the 2D AFM images

Sample AFM scanned

area (lm2)

Particle characteristics Surface roughness parameters

Number of particles

measured

Average particle

length (nm)

Average roughness,

Saa (nm)

Root mean square

roughness, Sqb (nm)

T 0 20 9 20 10 812 13.56 23.37

T 5 20 9 20 10 1169 10.13 13.15

T 7.5 20 9 20 10 922 4.10 5.48

T 10 20 9 20 10 1427 6.03 10.40

a The average roughness parameter, Sa, is the most used surface roughness parameter. It is the arithmetic mean or average of the absolute

distances of the surface points from the mean plane. The digital equation that represents this algorithm is displayed below, where M is the number

of columns in the surface and N is the number of rows in the surface:

Sa ¼ 1MN

PN

j�1

PM

i�1

zj j xi; yj

� �ð1Þ

b The root mean square (RMS) roughness parameter, Sq, is the root mean square of the surface departures from the mean plane within the

sampling area:

Sq ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

1MN

PN

j�1

PM

i�1

z2 xi; yj

� �s

ð2Þ

318 J Sol-Gel Sci Technol (2010) 56:310–319

123

hydrophilicity, porosity, and thermal stability of the com-

posites. While the Tg value is only slightly influenced by

TiO2 addition, the crystallization and melting temperatures

values change as compared to those of the blank sample.

Both e0 and e00 values increase throughout the studied range

of temperatures and frequencies following the addition of

TiO2. The temperature dependence of both curves show

inflections at Tg, Tm, and Tc values close to those detected

by DSC. The possibility to control the dielectric parameters

recommends these materials as elastic dielectric—active

elements for nano-actuation devices.

Acknowledgments This research was financially supported by

European Regional Development Fund, Sectoral Operational Pro-

gramme ‘‘Increase of Economic Competitiveness’’, Priority Axis 2

(SOP IEC-A2-O2.1.2-2009-2, ID 570, COD SMIS-CSNR: 12473,

Contract 129/2010-POLISILMET).

References

1. Schmidt H (1994) J Sol-Gel Sci Technol 1:217–231

2. Lev O, Wu Z, Bharathi S, Glezer V, Modestov A, Gun J, Rabi-

novich L, Sampath S (1997) Chem Mater 9:2354–2375

3. Sanchez C, Julian B, Belleville P, Popall M (2005) J Mater Chem

15:3559–3592

4. Shindou T, Katayama S, Yamada N, Kamiya K (2003) J Sol-Gel

Sci Technol 27:15–21

5. Husing N, Bauer J, Kalss G, Garnweitner G, Kickelbick G (2003)

J Sol-Gel Sci Technol 26:73–76

6. Fabes BD, Uhlmann DR (1990) J Am Ceram Soc 73:978–988

7. Simionescu B, Aflori M, Olaru M (2009) Constr Build Mater

23:3426–3430

8. Alexandru M, Cristea M, Cazacu M, Ioanid A, Simionescu BC

(2009) Polym Compos 30:751–759

9. Hu Y, Chung YJ, Mackenzie JD (1993) J Mater Sci 28:6549–

6554

10. Sun CC, Mark JE (1989) Polymer 30:104–106

11. Mark JE, Pan SJ (1982) Macromol Rapid Commun 3:681–685

12. Rajan GS, Sur GS, Mark JE, Schaefer DW, Beaucage G (2003)

J Polym Sci Polym Phys 41:1897–1901

13. Mark JE, Jiang C-Y, Tang M-Y (1984) Macromolecules

17:2613–2616

14. Ning Y-P, Tang M-Y, Jiang C-Y, Mark JE, Roth WC (1984)

J Appl Polym Sci 29:3209–3212

15. Tang M-Y, Mark JE (1984) Macromolecules 17:2616–2619

16. Ning Y-P, Mark JE (1985) J Appl Polym Sci 30:3519–3522

17. Yamada N, Yoshinaga I, Katayama S (2000) J Sol-Gel Sci

Technol 17:123–130

18. Rojas-Cervantes ML, Lopez-Peinado AJ, Martın-Aranda RM,

Gomez-Serrano V (2003) Carbon 41:79–86

19. Pena-Alonso R, Tellez L, Rubio J, Rubio F (2006) J Sol-Gel Sci

Technol 38:133–145

20. Enescu D, Hamciuc V, Timpu D, Harabagiu V, Simionescu BC

(2008) J Optoelectron Adv M 10:1473–1477

21. Epure V, Hamciuc V, Pricop L, Pinteala M, Airinei A, Harabagiu

V, Simionescu BC, Enescu D, Perichaud A (2007) High Perform

Polym 19:270–282

22. Racles C, Cazacu M, Ioanid A, Vlad A (2008) Macromol Rapid

Commun 29:1527–1531

23. Cazacu M, Ignat M, Vlad A, Alexandru M, Racles C, Zarnescu G

(2009) Polym Int 58:745–751

24. Cazacu M, Ignat M, Vlad A, Alexandru M, Zarnescu G (2010)

Optoelectron Adv Mater Rapid Commun 4:349–351

25. Raileanu M, Crisan M, Dragan N, Crisan D, Galtayries A, Bra-

ileanu A, Ianculescu A, Teodorescu VS, Nitoi I, Anastasescu M

(2009) J Sol-Gel Sci Technol 51:315–329

26. Rubio F, Rubio J, Oteo JL (2000) J Sol-Gel Sci Technol

8:105–113

27. Tellez L, Rubio J, Rubio F, Morales E, Oteo JL (2003) J Mater

Sci 38:1773–1780

28. Lantelme B, Dumon M, Mai C, Pascault JP (1996) J Non-Cryst

Solids 194:63–71

29. Lee B-S, Kang D-J, Kim S-G (2003) J Mater Sci 38:3545–3552

30. Cazacu M, Marcu M (1995) Macromol Rep A 32:1019–1029

31. Alexandru M, Cazacu M, Vlad S, Iacomi F (2009) High Perform

Polym 21:379–392

32. Dire S, Babonneau F, Sanchez C, Livage J (1992) J Mater Chem

2:239–244

33. Pirson A, Mohsine A, Marchot P, Michaux B, Van Cantfortc O,

Pirard JP (1995) J Sol-Gel Sci Technol 4:179–185

34. Brusatin G, Guglielmi M, Innocenzi P, Martucci A, Battaglin G

(1997) J Non-Cryst Solids 220:202–209

35. Babonneau F (1994) Polyhedron 13:1123–1130

36. Gu H, Bao DH, Wang SM, Gao DF, Kuang AX, Li XJ (1996)

Thin Solid Films 283:81–83

37. Noggin M (1985) J Non-Cryst Solids 69:415–423

38. Murashkevich AN, Lavitskaya AS, Barannikova TI, Zharskii IM

(2008) J Appl Spectro 75:730–734

39. Kochkar H, Figueras F (1997) J Catal 171:420–430

40. Ng E-P, Mintova S (2008) Microporous Mesoporous Mater

114:1–26

41. Yang Y, Zhang H, Wang P, Zheng Q, Li J (2007) J Memb Sci

288:231–238

42. Yeh K-Y, Chen L-J, Chang J-Y (2008) Langmuir 24:245–251

43. Que W, Zhou Y, Lam YL, Chan YC, Kam CH (2001) J Sol-Gel

Sci Technol 20:187–195

44. Wang B-L, Hu L-L (2005) Mater Chem Phys 89:417–422

45. Cazacu M, Vlad A, Alexandru M, Budrugeac P, Racles C, Iacomi

F (2010) Polym Bull 64:421–434

46. Chiang C-L, Ma C-CM (2004) Polym Degrad Stab 83:207–214

47. Fraga AN, Frullloni E, de la Osa O, Kenny JM, Va0zquez A

(2006) Polym Testing 25:181–187

48. Li Y, Cordovez M, Karbhari VM (2003) Compos Part B

34:383–390

49. Pethrick RA, Hayward D (2002) Prog Polym Sci 27:1983–2017

50. Weast RC (ed) (1985) CRC handbook of chemistry and physics.

CRC, Boca Raton

J Sol-Gel Sci Technol (2010) 56:310–319 319

123