Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

0-3 Connectivity

Download as pdf or txt
Download as pdf or txt
You are on page 1of 29

706 IEEE Transactions on Dielectrics and Electrical Insulation Vol. 3 No.

5, October 1996

Inorganic Ceramic/ Polymer Ferroelectric


Composite Electrets
C. J. Dias

Universidade Nova de Lisboa, Portugal

and D. K. Das-Gupta

School of ElectronicEngineering and Computer Systems, University of Wales, Bangor, United Kingdom

ABSTRACT
Ferroelectric composites are now an established alternative to conventional ferro-
electric ceramic materials and to the more recently discovered ferroelectric polymers.
These materials due to their unique blending of polymeric properties of mechani-
cal flexibility, formability and low cost with high electro-active properties have been
been suggested to be a viable alternative both in piezoelectric and pyroelectrictrans-
ducer applications. This review is devoted to the piezoelectricand pyroelectricprop-
erties exhibited by these type of composites with a special reference to those made of
ceramic particles embedded in a polymer matrix (i.e. 0-3 connectivity type compos-
ite). A review of models predicting the electro-active properties of 0-3 composites
is presented together with a proposal for a new mixed connectivity cubes model to
be applicable to the case of high ceramic loading and/or when the ceramic grain size
incorporated in the polymer matrix is comparable to the thickness of the sample. A
review of the experimental results of the piezo- and pyroelectric properties of vari-
ous ferroelectric composite materials, reported by several workers, is also presented
in this paper. Special reference is made to composites made from calcium modified
lead titanate and poly(viny1idenefluoride-trifluorethylene) emphasizing their ad-
vantages in the poling process which is a critical phase in the process of obtaining
successful electro-active 0-3 composite electrets.

1. INTRODUCTION a wide selection of permittivity (E" M 100 to 2400) and low di-
electric and mechanical losses albeit with a large acoustic charac-
I N modern usage, composite materials are made from a filler,
either particles, flakes or fibers, embedded in a matrix made of
polymer, metal or glass. Nevertheless, according to the broadest
teristic impedance resulting in a more difficult acoustic match to
soft media such as organic tissues and water.
definition, a composite is any material consisting of two or more Thus range of electro-active materials have, more recently,
distinct phases. So, in this latter sense most ferroelectrics may be been enlarged with ferroelectric polymers such as PVDF and
included in the composite category. P(VDF/TrFE) [6]which have a low acoustic impedance but also
Combining a piezoelectricceramic and a polymer host to form a lower electromechanical coupling factor and higher dielectric
a flexible ferroelectriccomposite has been pursued in recent years and mechanical losses. These properties are coupled to a low per-
in view of the greater flexibility allowed by these materials to mittivity, wluch can be advantageous in certain situations, such
suit particular properties such as mechanical, electrical, thermal as in pyroelectric detectors.
and/or a coupling between these properties. Composites can be viewed as intermediate materials between
Conventional piezoelectric materials such as lead zirconate- these two extremes 171 combining such attributes as flexibility
titanate [l,21, lead metaniobate and more recently modified lead and formability, to a whole range of piezoelectric and pyroelec-
titanates [3-51 are nowadays the most popular choices in appli- tric properties which themselves depend both on the ceramic-
cations which use either piezo- or pyroelectric properties. These polymer mixture and on the processing employed in its manu-
materials offer high electromechanicalcoupling (kt M 0.4 to 0.5), facture.
1070-9878/96/ $3.00 @ 1996 IEEE
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 3 No. 5, October 2996 707

The identification of the relevant variables controlling the sol-gel route [16]. Concern however, must be exercised when us-
property behavior is thus of utmost importance for the manufac- ing fine powders as properties vary with the ceramic grain as
ture of a successful composite. The most obvious one is that the was mentioned above. Presently however, commoln composites
ceramic material employed should exhibit high pyro- and piezo- are N 5 to 10 p m thick while the ceramic grains embedded in
electric activity. It was recognized however in the work of Newn- them are typically > 0.8 to 1 p m and thus acoustic frequencies
ham and coworkers [S-101 that the connectivity of the phases is
also a very important parameter and t h s finding prompted re-
-
up to 500 MHz can be used without resonance effects.
Other effects such as interfacial, percolation and porosity ef-
search on composites consisting of embedding rods of a piezo- fects can also play a role in composite properties [17]. Interfa-
electric material in a polymer host thus forming a 2D parallel net- cial effects can occur between the ceramic grains and the poly-
work which exhibited better properties than the simpler powder mer matrix leading to large dielectric relaxations normally at low
dispersed composites. Other ceramic-polymer configurations frequencies otherwise known as Maxwell-Wagner relaxations,
have been investigated such as using ceramic fibers and coral while in other cases interesting Schottky barrier ph.enomena can
ceramic templates to make a 3D self connected ceramic phase. also occur. Space-charge compensation of dipolar polarization
Nevertheless, the powder in a matrix approach has never been is also known to occur at the interface boundaries. Percolation
quite abandoned mainly because of the attraction of mass pro- of ceramic grains inside a composite is an effect which can be of
duction of cheap piezoelectricand pyroelectric materials that can great help in aiding its polarization by establishing a continuos
impart in reliable way special properties to various structures by electrical flux through the ferroelectric grains.
a coating process. Large and thin film of electro-activearrays for Porosity, when not planned, can have adverse effects. Howev-
ultrasonic and infrared detection are also part of an agenda that er, it has been intentionally induced in PZT up to 70%in volume
still awaits its fulfillment. [18,19] so that air was the filler material with very compliant
In this latter type of composites, of so-called 0-3 connectivi- properties. Decrease of the permittivity, the Young modulus, the
ty, the characteristics of the grains of the ceramic powder is also transverse coefficients and of the mechanical Qm were observed
of practical relevance. T h s is because the properties of ceram- with increasing porosity whilst d33 remained stable and g33 in-
ics depend not only on the composition and their crystal struc- creased. Such materials could be thus very attractive for longi-
ture but also on the microstructure morphology including grain tudinal mode operation although some pressure dependence of
size, grain boundaries, pores, crystallinity, microcracks etc. For properties should be expected.
instance, fine grained ceramic of lead titanate (PT) and PZT have These 0-3 composites can be poled to form pijezo and pyro
been reported to display a reduction in the polarization and an in- electrets. In this review an introduction to the composite mod-
crease in the room temperature permittivity which was assigned els will be outlined with an emphasis on those taking into ac-
to a lower tetragonality in the structure [ll]. The effect of me- count the connectivity of the composite phases in such electrets.
chanical grinding, lapping and polishng has also been assessed Ths will be followed by Sections covering piezoellectric and py-
in PLZT ceramics discs [12] whch showed a decrease of polar- roelectric properties, where their measurements techniques are
ization and permittivity in mechanically treated samples when described and the composite properties so far obtained are re-
compared to those treated chemically. This effect has been attrib- viewed.
uted to the existence of a layer over the surface of the ceramic
grain characterized by an absence of ferroelectricity and a corre-
2. CONNECTIVITY
sponding low dielectric constant. Making composite materials, by a combination of a ferroelec-
tric ceramic and a polymer of suitable properties, means not only
The matching of electrical [13] and mechanical properties [14]
choosing the right materials processed in a particular way, but
is another area which should be given due consideration. The also coupling them with the best possible design structure. This
importance of electrically matching the ceramic to the polymer concept of connectivity, first established by Newnham et al. [9],
phase stems mainly from the difficulty of orienting the polariza- to describe the interspatial relationships in a multiphase material
tion of randomly distributed ceramic grains of low resistivity and is of the utmost importance because it controls the mechanical,
high permittivity embedded in a high resistivity, low permittivity electrical and thermal fluxes between the phases.
polymer matrix, by means of an external field without triggering
In a composite each phase may be spatially self-connected in
an electrical breakdown.
either one, two or three dimensions. In a diphasic or two phase
The periodicity and scale in composites regarding both the system there are ten different combinations of phase connectivity
filler and the matrix are important factors in determining, for in- which are usually indicated using two digits, dentoting the con-
stance, the upper frequency limit where resonance and interfer- nectivity of the filIer and that of the matrix respectively in t h s
ence occur. In order to get higher frequencies as well as other ad- order. These are the: 0-0, 0-1, 0-2, 0-3,l-1,l-2,l-3, 2-2, 2-3,
vantages such as miniaturization, nanocomposites consisting of and 3-3 connectivities represented in Figure l(a) using a cube as
particles in the nanometer scale dispersed in a matrix, have been a building block. For example, the commonly usecl composite of
recently been experimented [15] which have been brought about dispersed particles embedded in a matrix would then be indicat-
by synergies in the production of fine ceramic powders using the ed as a 0-3 composite while another common system consisting
708 Dias et al.: Inorganic CeramiyTolymer Ferroelectric Composite Electrets

of fibers implanted on a matrix and oriented across one of the temperature with the right proportions of the resin, adhesive and
dimensions of the sample would be known as a 1-3 composite. ceramic powder in a recipient.
T h s classification, does not however specify the directions up- In the case of a thermoplastic polymer the blending process
on which the connectivity takes place nor does it say, in the case may be done in two ways. In one method the polymer is first
of a ferroelectric composite, in what direction the polarization heated in a hot rolling machine to a temperature between its soft-
vector points to. This information must obviously be asserted or ening and melting temperature, after which the ceramic powder
be self evident for the case in hand. For example, for the two is gradually added while mixing the components until a reason-
dimensional case, both the so-called series and parallel connec- able blend is obtained. In another method the polymer is dis-
tivities shown in Figure l(b), would be categorized using this solved in a suitable solvent agent and then the ceramic powder
classificationunder the 2-2 connectivity pattern. They render ob- is added to the fluid. This fluid is then dispersed in a plate, im-
viously quite different properties in all respects as will be demon- parted with release properties, and the solvent allowed to evap-
strated later. orate by a curing process at a suitable temperature. The material
The definition, in a diphasic system, of what is the filler and thus obtained may be further processed in order to get composite
what is the matrix can be confusing when in the case of certain films with the desired thckness either by pressing again at the
composites the volume fraction occupied by the filler is higher softening temperature of the polymer or by means of polishmg
than that of the matrix. The connectivity criteria can be used in and grinding operations.
this case so that the matrix is the phase of highest connectivity in Some problems may be faced during the mixing operation
the composite. which are linked to a poor distribution of the polymer, poor ad-
hesion of component phases or to air entrapment. Inadequate
2.1 . 0-3 CONNECTlVlTY distribution of the polymer is commonly coupled to an agglom-
COMPOSITES AND THEIR eration of the ceramic grains which may be overcome by a lower-
FA BRICAT IQN ing of the polymer viscosity via heating, during the mixing pro-
cess. Air entrapment during mixing also is a common problem
The most commonly studied composites are the 0-3 and 1-3
and vacuum degassing procedures may have to be undertaken
configurations, although for different reasons. The 0-3 connec-
to prevent t h s effect.
tivity owes its popularity to the easy fabrication procedure which
allows for mass production at a relatively low cost. To impart piezoelectric and pyroelectric activities, these mate-
rials must be initially subjected to an external electric field (i.e.
The preparation of 0-3 ceramic polymer composites involves
poling) for a certain time at a suitable temperature, to orient the
proper consideration regarding the materials to be used. The
spontaneous polarization in the ceramic phase of the electret. For
ceramic has to be transformed into a powder form from its usu-
this poling action to be effective, the electric field across each of
al block shape obtained after the pressing, firing and sintering
the grains must be greater than the so-called coercive field of the
stages [20]. This form can be aclveved either by mechanically
ceramic. Two methods of poling are usually in use (1)the con-
grinding the ceramic until the ceramic grains have the desired di-
ventional method where the sample electroded on both sides,
mensions or by heating the ceramic block and then quenchng it
is immersed in an inert insulating fluid customarily silicone oil
in a water bath kept at room temperature, fracturing the ceramic
kept at a constant temperature while an electric field is applied
along the cleavage planes arising from the thermal stresses thus
for a period and (2) the corona method where a one-side elec-
developed [21]. It is clear that the grinding method provides a
troded sample is charged on the unelectroded face by means of a
better control of the size of the ceramic grain powder than the
corona current produced at a corona point sometimes controlled
quenching process where a wider distribution of grain sizes is
by a grid placed in between the corona point and the sample
obtained. The quenched powder can be sieved to separate the
whilst the electroded face of the film rests on a grounded plate
grains into a narrower size range, however there is the disad-
(Figure 2).
vantage that this is inevitably a costly procedure if a too strin-
gent range is to be attained. Mechanical processes such as grind- 2.2. 1-3 AND 3-3 CONNECTIVITY
ing and polishmg have however been reported to induce anoma-
lous dielectric properties and reduced polarization in PLZT ce- The 1-3 connectivity composites show in general a higher per-
ramics which were attributed to the existence of surface layers formance in terms of piezoelectric properties albeit at a lugher
produced during these operations [12], whch could be atten- manufacturing cost [23,24]. T h s usually consists of pillars reg-
uated through chemical etching and/or thermal annealing. A ularly spaced within a polymer matrix oriented with their long
more recent approach for producing ceramic powders uses the axis perpendicular to the surfaces of the film.
sol-gel route where a very fine particle grain size is obtained [22] A problem in 1-3 composites arises as a result of the period-
through the direct calcination of the gel powders. icity of the lattice rods which cause spurious resonances and de-
The polymer to be used determines the way in which the ma- grades the performance of ultrasonic transducers [25,26]. These
terials will be mixed. If it is an epoxy (i.e. a thermoset polymer) resonances are caused by Bragg reflections of Lamb waves prop-
the mixing stage can be performed very simply by hand at room agating laterally in a periodic array of rods. The frequency of
IEEE Transactionson Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 709

0-0

a , 53
parallel series
connectivity conne'ctivity
........
....
..
........
....
...... .........
.......
........
........
.........
... ...............
.......
1-1

polarization
vector

(a) 2-3 3-3 (b)


Figure 1. (a) Connectivity patterns in a diphasic composite system [9]. (b) Parallel and series connectivities with their respective orientation
of the polarization vector.

To overcome this limitation two approaches have been pro-

avo
mixing and hot rolling
- -

.
hot pressing aluminum
deposition posed which are as follows (1)a carbon fiber is woven into the
coating desired structure by textile techniques and then the carbon stmc-
ture replicated with a piezoelectric ceramic [27], (2) in another
approach a 'lost wax' method [28] is used to produce a comple-
mentary structure in plastic; a ceramic slip is injected into this
corona yling ,,&
-1lkV conventional
paling (ac or dc )
mold and fired; the plastic mold burns away during the ceramic
firing and a polymer is cast back into its place. Such methods
allow for the production of large area and fine scales at relatively
low cost.

An alternative and simpler procedure widely in use today,


called the dice-and-fill techruque, was devised by Savakus et al.
[24] where a prepolarized ceramic disk, as commercially avail-
Figure 2. Experimental procedure for obtaining 0-3 ferroelectric
composites films. able from the manufacturers, is mounted on a diamond saw. A
series of parallel grooves of a suitable thickness and pitch are
made on the disk after whch it is rotated by 90°C and a second
these Iateral modes are inversely proportional to the rod spacing similar series of parallel cuts made. The grooves are then filled
while the thckness mode resonance depends on the plate thick- with a polymer and the resulting composite disk sliced off the ce-
ness. To overcome this problem the spacing between the pillars ramic base. Practical minimum spatial scale is in thle 50 pm range
can be made finer, which pushes the lateral resonances to higher where the ceramic rods become increasingly fragik. Finer scales
frequencies and allowing the composite to behave as an effective can, however, be attained by using a laser to cut the grooves ei-
homogeneous medium.
ther by laser ablation or laser-induced chemical etching. A wide
range of volume fractions and pillar thickness to length ratio can
The original method [23,26] involves an extrusion of fine be achieved by changing either the saw cut thckness and/or the
piezoceramic pillars which are fired prior to their careful posi- pitch, More recently it has been proposed that the grooves could
tioning in a mold. The complete assembly is then potted with a be made using an ultrasonic cutter [29].
suitable epoxy polymer and a section of the required thickness
sliced from the composite block thus obtained. Electroding and More recently, it has been suggested a new design of compos-
poling of the ceramic is then performed to obtain a workable ites made of tubular piezoelectric ceramic embedded in a poly-
piezoelectric composite sample. The method is effective for rod mer matrix [30]. These composites could be tuned so that the
diameters of 2 200 pm while being too cumbersome for finer effective piezoelectric coefficient in the radial direction be posi-
spatial scales due to the fragility of the ceramic rods. tive, zero or negative by varying the ratio of the outer radius to
710 Dias et al.: Inorganic Ceramiqh'olymer Ferroelectric Composite Electrets

the inner radius of the tube. They could also provide compos- composite was shown to be given by
ites with exceptionally high values of hydrostatic piezoelectric
coefficients in the range of 5 to 10 pPa- I.
3-3 connectivity binary composites [lo], where the two phas-
where 7 is a parameter dependent on the shape of the ellipsoidal
es connect in the three dimensions, have been produced by the
particles [32] and their orientation in relation to the surface of the
replamine form process. This consists of shaping a suitably cho-
composite film, while 4' is the volume fraction of the ceramic.
sen coral as a template for the ceramic material. The coral is first
The other symbols E, E' and E P , refer to the magnitude of the
vacuum impregnated with a casting wax after whch the coral
permittivity in the composite, ceramic and polymer respectively.
is leached away in hydrochloric acid leaving a negative of the
Throughout this Section superscripts c and p will refer to the
coral template. The positive of the template is afterwards filled
ceramic and polymer properties respectively, while an absence
with a solution containing the ceramic and the wax is burned off
of a superscript will usually denote a composite property.
at 300°C leaving only the ceramic structure whch is sintered at
high temperature (N 1300°C). The three dimensional structure Composite properties:E. d

thus formed is then impregnated with a polymer of a suitable ceramic dispersoid EC dC


viscosity to form a 3-3 composite. ceramlc volume fraction: ,$C
These composites of 3-3 connectivity have been successful
[lo], due to the existence of electrical ceramic paths across the polymer dispersoid EP,dP
thickness of the composite, enabling an efficient polarization of
the ceramic grains, even at low ceramic volume fractions and
thus with the added advantage of a low permittivity and densi- Figure 3. Binary system consisting of piezoelectric ellipsoidalpar-
ticles dispersed in a continuos medium.
ty.
Assuming that the polymer material used in the matrix does
3. MODELSOF not show piezoelectric activity an expression for the piezoelectric
C ERAMIC/POLYMER constant d in the binary composite is given by
FERROELECTRIC COMPOSITE
ELECTRETS d = atj5'Gd" (2)
where a is the ceramic poling ratio (one for fully polarized and
3.1. INTRODUCTION zero for a total absence of polarization in the ceramic), G the local
field coefficient and d" the piezoelectric constant of the piezoelec-
T h s Section concerns the prediction of composite properties tric dispersoid (i.e. the ceramic). The local field coefficient using
made with electro-activeceramic powders embedded in a poly- the shape parameter 7 is given by
mer host. Tlus particular task is a small part of ongoing work
G= 7&
across the scientific community to model the performance be-
havior of mixed phase materials. Various models have been pro- 7" + (EC - E )
(3)

posed which deal specifically for example with electrical prop- where the permittivity of the composite is given by Equation (1).
erties, namely the resistivity and the permittivity and a large lit- In that work [31] a shape parameter 7 = 8.5 was found to fit
erature exists on this subject alone, while the same goes for the these theoretical equations to the experimental results obtained
mechanical properties. In this short review, however, we will for PZT-PVDF composites with different ceramic volume frac-
be concerned with predictions on the coupling coefficients be- tions. This value of the shape parameter 7 demanded that the
tween the electrical and the mechanical properties through the ceramic ellipsoid axes ratio was 2.8 with the long axis perpendic-
piezoelectric effect and with the coupling coefficient between the ular to the surface of the composite film. This was difficult to ac-
thermal and electrical properties through the pyroelectric effect. cept in view of the procedure used to obtain the composite, which
involved a h g h temperature mixing of the approximately ellip-
3.2. YAMADA MODEL FOR 0-3 soidal ceramic grains inside the polymer followed by a pressing
COMPOSITES stage to obtain the composite films thus forcing the longer axis
of the ellipsoidal particles to be parallel to the surface of the film
Yamada and his coworkers [31] have studied the binary sys- and not perpendicular. An alternative interpretation put forward
tem consisting of a PZT powder embedded in a polymer matrix by Banno [21] points out that the ceramic grains could be approx-
of PVDF, with respect to their dielectric, piezoelectric and elastic imately spherical in shape but, if one compared their dimensions
properties. In the same work, a model was proposed to explain with those of the composite sample in whch they are embedded
the behavior of the properties of the composite using the proper- it would appear to be bigger in the thckness direction than in
ties of its constituents materials. Considering the binary system the other two perpendicular directions of the film and as a re-
schematically shown in figure 3, composed of ellipsoidal parti- sult, the shape factor 7 representing the spheres would appear to
cles dispersed in a continuous medium, the permittivity of the have been stretched in the thickness direction. In conclusion, the
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 711

parameter 11 is not an absolute measure of shape of the ceramic where the symbol c stands for the Young modulus of the materi-
grains but takes also into account the dimensions of the films into als concerned. Expressions for the piezoelectric constants d, e, g
which they are embedded. and h whch have also been derived, were given by
Based on the same model and assuming that the ceramic is the d = 4"LxLEd" (7)
only pyroelectric phase in the composite its magnitude p may be e = 4"L2LEeC (8)
expressed as
g = d'"LXLDg" (9)
p = aq5'Gp'
(4) h = 4"L,LDh" (10)
where p c is the pyroelectric coefficient of the ceramic whilst the where L x , L,, L~ and L E are the local fields coefficientswith
other quantities already have been defined above. Using this respect to X (stress), z (strain), D (electric displacement) and E
expression, good agreement could be found between the experi- (electric field) respectively. The meaning of the loc,il fields coef-
mentally observed data and the calculated values in composites ficients can be understood by the case of L E (local electric field
made of PZT powder embedded in a PVDF matrix 132,331. coefficient) for example, whch is the ratio of the field applied to
the composite as a &hole to the local electric field priduced in
3.3. FURUKAWA MODEL FOR 0-3 the ceramic inclusion.
COMPOSITES
It will be instructive to follow the derivation of one of the

a matrix (i.e.polymer) (Figure 4) which in turn is covered with a


homogeneous medium whose properties approximate the aver-
age composite properties.
d= [a] E
(11)

When a stress X is applied to the composite, it produces the


local stress X , in the ceramic inclusions
X " = LXX (12)
but as the stress and the electric displacement in the ceramic in-
clusion are related by Equation (ll), we have
D" = d c X c= d C L x X (13)
The apparent electric displacement D in the composite due to
the local D" is given by [34]
D= LED" = 6'LxLgd"X (14)
Referring to Equation (14) it may be observed, that ihe composite
behaves as a piezoelectric material whose piezoelectric d is given
by Equation (7). The local field coefficientshave been derived by
Furukawa in terms of the properties of the constituent materi-
als of the two-phase system [35] and are given by the following
Figure 4. Representation of a composite system consisting of equations
spherical ceramic inclusions embedded in a polymer matrix sphere
5C"
which in turn is covered with a homogeneous medium whose prop- Lx =
erties approximate the average composite properties. 3(1 - @ ) c P + (2 + 3 4 " ) ~ " (15)

In t h s work, expressions for the dielectric, elastic, and piezo-


electric constants were given and compared with experimental
results. The permittivity of the composite was given, using its
values for the constituents materials, by [34]and references there-
in
2EP + 24C(EP
E" - - &") Ep
E =
2EP + +
E" @(EP - E")
(5)
Using this formalism the ceramic/ polymer composite systems
while the elastic constant, which was deduced on the assumption of PZT-epoxy, PZT-PVDF, PZT-PE and PZT-PVA with ceramic
that the composite was incompressible, is given by volume fractions < 21 % were analyzed for their dielectric, elas-
tic and piezoelectric properties. The largest piezoelectric activity
3cp + 2cc 3$"(CP - "
-
amongst these composites was obtained in the PZT-PVDF system
IC. (6)
= 3cp + 2 c " + 24"(cP - CC) in which case the expressions given above, provide reasonable
712 Dias et al.: Inorganic CeramicfPolymer Ferroelectric Composite Electrets

approximations for volume fractions of the ceramic < 10%.For


a PZT volume fraction in the 20% range the measured values of
d were found to be 2 to 3x larger than predicted while similarly
the experimental dielectric constant was found to be larger than
that given by Equation (5) by a factor of two. It was suggest- large grain size small grain size
ed that the cause for the discrepancy between observation and Figure 5. The effect on the connectivity of large grain size and
prediction was a consequence of some hgher order composite volume loading of the ceramic phase in a composite film
effect like aggregation of the ceramic inclusions whch was not
accounted for in the derivation of the model. et al. [9]these structures are classified as having a 2-2 connectivi-
ty pattern and will form the building blocks of our understanding
In order to account for deviations of the 0-3 connectivity be-
of 0-3 composites which have either a high concentration of ce-
havior, an approach was conceived by Pardo and coworkers in
ramic or a large ceramic grain compared with the dimensions of
which the composite was perceived as having a mixed connec-
the sample. The evaluation of properties of the parallel structure
tivity pattern [36]. Considering 4 to be the total volume fraction
have been the object of numerous works notably by Newnham
of the ceramic in the composite, it was assumed that part of it
et al. [9], Smith [14,38] and others [39-451. The elastic sij, piezo-
say 4 0 - 3 , would have a 0-3 connectivity whilst the rest of the
electric d i j , dielectric F and pyroelectric p and linear expansion
ceramic in the composite i.e. $1-3 = 4 - 4 0 - 3 , would con-
a property coefficients of a parallel diphasic structure in terms
nect to both surfaces of the composite film exhibiting thus 1-3
of the properties of the constituents materials are
connectivity. On measuring the ceramic grain size g s and the
composite film thickness t their ratio (i.e. g s / t ) was computed
enabling then through a Monte Carlo simulation, the calculation
of the magnitudes of the 0-3 and 1-3 volume fractions, 4 0 - 3 and
4 1 - 3 respectively. Upon doing this calculation the piezoelectric
constant d33 was calculated from
d33 (41-3 + 40-3aLE)LXd& (19)
where a is the poling ratio for the fraction of the ceramic which
has 0-3 connectivity while the other symbols have the mean-
ing stated above. A reasonable agreement was found between
Equation (19) and the experimental results in the transition re-
gion from the 0-3 to the 1-3 connectivity when the ceramic grain
size was of the same order of magnitude as that of the thickness
of the sample.
Zewdie and Brouers [37] have proposed a calculation proce-
dure similar to that of Pardo and coworkers 1361 using modified
local field coefficientswhich were claimed to include the shape of
the dispersoid through an ellipsoid index. Those local fields were
found to be appropriate for intermediate ceramic concentrations.
3.4. CONNECTIVITY APPROACH a 4CaC p a p
~~

533
-
-
553
+- 4 3
(28)
In the previous models an assumption was made that the ce-
ramic grains were sufficiently small so that a quasi-homogeneous Superscripts c and p represent properties of the polymer and ce-
medium was assumed to exist around the ceramic grains. How- ramic phases respectively while an absence of superscripts de-
ever, that can no longer be held true if a high concentration of the notes composite properties. Some of the composite properties,
ceramic dispersoid is present and/or when the ceramic grains such as d33 for example, are given using the properties of the
are of a similar size as the dimensions of the composite film (Fig- constituent materials as well as the calculated elastic compliance
ure 5). Nevertheless, it has been found experimentally that most of the composite 533.
of the useful ferroelectric composites fall in one of the latter cas- It was assumed for reasons that will be apparent later on and
es which prompted the need to build up models that could take without losing generality that both phases constituting the com-
those facts into better account. posite are ferroelectric.
A central concept in the theoretical modeling of composite In order to derive these expressions some assumptions have
behavior is that of connectivity [9] described in Section 2. For been made whch are (1)the strain and the electric field in the 3
two phase structures, those whch we are interested in, the sim- (i.e. 2 ) direction as well as the stress in the transverse directions
plest cases to treat are the series and the parallel two-dimensional 1and 2 (i.e. 1c and y respectively) are equal in both phases (2) the
structures as represented in Figure l(b). According to Newnham stress and the charge in the 3 direction as well as the strain in the
IEEE Transactionson Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 713

transverse directions 1and 2 are distributed between the phases


according to their volume fractions (see Smith [14,38] for a full
discussion of these assumptions).
Properties in composites can be advantageously classified with
reference to those of the constituent phases as: sum, combination
and product properties [9,46]. A sum property, such as the case These expressions do not include the product terms except in the
of the density, is one which derives its value from an equation case of the s33 coefficient (Equation (29) where that contribution
whch involves an averaging of properties of the phases so that was found to be significant. For the other propert:ies that term
the composite property value is bounded by those property val- was found to be small or negligible as for instancle in the case
ues of the constituent phases. of the permittivity where the additional term contributed with
< 2 x lop5% to the total value of that property in the composite.
A figure of merit for a given application often involves two
Comparing the expressions of the permittivity between the
or more material properties. This is an example of a combination
parallel and series connectivity diphasic structures one can ob-
property whch requires an averaging of more than one property.
serve that their sum property contributions are nothng more
The pyroelectric figure of merit p / & depends both on the pyro-
than the rules for calculating the paraIIe1and series values of ca-
electric coefficient and on the permittivity and due to their differ-
pacitance respectively applied to the permittivity of the compos-
ent mixing rules the pyroelectric FOM, can be hgher or smaller
ites. The same type of reasoning is applicable to the composite
than any of its constituent phases. Product properties are the
elastic constants whilst in the case of the piezoelectric and pyro-
most complex ones as they originate due to a coupling between
electric constants which involve coupling between different sets
the two properties whch do not match in the constituent phases
of properties (i.e.charge/stress and charge/temperature respec-
and thus produce an entirely different property. As an example,
tively) that reasoning is not quite as straightforward.
the coupling of the piezoelectric effect with the linear expansion
of a material produces a contribution to the pyroelectric effect in It is useful to compare the piezoelectric and pyroelectric con-
Equation (27). stants calculated for a composite based on the parallel and series
connectivities. In Figure 6 we have calculated the pyroelectric
More generally it is often the case that a property in the com- figure of merit FOM,) defined asp/&and the hydrostatic piezo-
posite will have all these three different types of contributions. electric FOMh which is dhgh, for both types of connectivitiesin a
The properties of a series connectivity diphasic composite are PTCa-PVDF composite. Comparing these graphs wie can observe
listed as functions of both the properties the constituent materials that the parallel connectivity structure shows superior piezoelec-
and their respective volume fractions. tric and pyroelectric properties than its series countlerpart for the
same ceramic volume fraction. In fact, that was found to be the
case for other choices of ceramic/polymer composites. T h s is an
important consideration in the design of 0-3 composites as will
be demonstrated later.
In order to simulate theoretically 0-3 type of colmposites, an
approach to the properties of diphasic materials kniown as cubes
model, was first proposed by Pauer [47]. This model consists
of a cube with unit dimensions made of the host material (i.e.
polymer) where another smaller cube made of the dispersoid (i.e.
ceramic) is embedded (see Figure 7(a)). The aim is then to find
out the dimensions of the smaller cube such that the new-found
equivalent cube should have similar properties as; those of the
composite material.
Ths model was later modified and generalized by Banno
[21,48] on the so-called 'modified cubes' model. As can be seen
in Figure 7(b) the ceramic block inside the cube was given an ad-
ditional degree of freedom with its height n independent of its
length and width, which remained equal to each other m. Thus,
on varying the parameters n and m of the ceramic block either
a pure series or parallel connectivity could be obtained. Other
modifications [49] were proposed to thus model notably by Gar-
ner et al. [50] to improve on the isotropy of the modified cubes
model.
A difficulty in Banno's modified cubes model is8encountered
when the height of the ceramic block inside the equivalent cube
714 Dias et al.: Inorganic Ceramic/PolymerFerroelectric Composite Electrets

2500
-
7

'Q 2000
n

2 1500
Y

4
3a1000
-c

~jj500

0
0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1
(4 (pc (ceramic volume fraction) (b) $ c (ceramic volume fraction)
Figure 6. PyroelectricFOM (pi.) and hydrostatic piezoelectric FOM (dhgh) for parallel and series connectivity composites.

white = polymer
shaded = ceramic
Computing scheme for
connectivity the mixed connectivity
cube model

1 1

(a)
Figure 7. Schematic of the various proposed cubes models. (a) simple cube, (b) Banno's modified cube and (c) mixed connectivity cubes.

approaches unity where the dielectric, piezoelectric and pyro- branches: a pure polymer, a pure ceramic together with a branch
electric properties all change very fast for small increments of wluch has a series connection of a ceramic and a polymer part.
the height n of the ceramic thus making it difficult to find the The n factor is thus a measure of the degree of 1-3 connectivity
proper equivalent cube. Experimentally that is the case when a of the composite such that a value of unity means a pure 1-3 con-
large ceramic grain size is used in the composite causing a sig- nectivity sample while lower values translate into a higher series
nificant ceramic volume fraction to connect to both electrodes character.
hence imparting an appreciable degree of parallel connectivity
to the composite. This is reflected in the equivalent cube model
through the fact that the embedded ceramic parallelepiped al-
most touches both electrodes in a region of n values where the
gradient of the properties is very high.
In order to overcome this problem we propose a different
unit cell model called mixed connectivity cube as schematically
shown in Figure 7(c) [51,52]. It starts as in the previous models The calculation proceeds by computing first the properties of
with a host polymer unit cube. However, the ceramic part con- the series branch using Equations (29) to (36) for the series con-
sists in this case of a cube of size m in which a fraction of it n, nectivity structure and then calculating the total expressions for
in form of a parallelepiped, connects to both electrodes so that the properties of the cube using the parallel connectivity equa-
a part of the total ceramic fraction dispersed in the composite tions as given in Equations (20) to (28). The latter equations for
will be connected in series whilst the rest will be connected in a diphasic parallel configuration need yet be generalized for a
parallel. T h s is emphasized in the way of computing the prop- triphasic parallel structure, wlvch is easily done in the follow-
erties of this mixed connectivity cubes, outlined schematically in ing way due to the inherent symmetry of those equations. In the
its equivalent model (Figure 7(c)).This consists of three parallel case of the permittivity (the other properties generalize similarly)
IEEE Transactionson Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 715

Table 1. Volume fractions of each of the phases for the series and One can see from the charts (Figure 8) that tlus particular com-
parallel branches. posite will not have good pyroelectric properties, although its
dC I dP I os hydrostatic piezoelectric properties can be of some interest. One
series m 1-m can thus use these charts to find out which combination of ce-
parallel n2m2 1 - m2 m y 1 - n2) ramic and polymer to use in order to achieve a given value of a
Total m3 + n2m2(1- m) target property.
These charts, however, are given in terms of the n and m pa-
Equation (26) transforms in a triphasic system into rameters which characterize the equivalent mixed connectivity
E = @EC + ( Y E P + @ E S cube. These parameters are controlled via the volume fraction
and the grain size of the ceramic. Unfortunately there is no sim-
- ( 4 3 - 43)25334c4p- ( 4 3 - 43)2s334C4S
ple rule to relate these variables and one must resort to an edu-
sz343 G3G3 (37) cated trial-and-error procedure. A possible route is to choose a
- ( 4 3-43)2s334sP particular volume fraction and if possible change the grain size
5kSp33
gradually until the composite is located in the region of interest
(i.e.has the correct degree of 1-3 connectivity character). Anoth-
where the s superscript denotes properties of the series branch. er complication whch may arise is dealt in the following Section
As can be observed from this Equation the sum contribution for
and concerns the assumption that the ceramic inside the compos-
the permittivity contains an additional term coming from the se-
ite is supposed to be fully polarized.
ries element whlst the product contribution has two extra terms
to account for the interaction of the series branch with each of 3.5. CONSIDERATIONS REGARDING
the other phases (i.e.the polymer and the ceramic phases). By in- POLING IN COMPOSITES
spection the same argument holds for the following properties:
811, s12, d31 and p . For the piezoelectric constant d33 and the A major restraint in simulating the electro-active properties
elastic constants 533 and s i 3 only one additional term is needed arises from the fact that the piezoelectric and pyroelectric prop-
as no product contributions exist. Simplifications in the expres- erties of the ceramic depend on the efficiency of poling which is
sions could be carried out now to take into consideration the lack the process the ferroelectric ceramic undergoes when its polar-
of intrinsic electro-activity in the polymer phase (i.e. d:j = 0, ization switches irreversibly to the same direction as that of the
p=0). applied field. This process is usually referred to as a field acti-
vated one, because it is effective only when the field across the
To calculate the properties of the mixed connectivity model
ceramic is higher than the so-called coercive field E,.
one needs to know the volume fractions of each of the phases, as a
function of the dimensions n and m, to use in the expressions for The importance of the connectivity of phases in the efficien-
the series and parallel connectivities. These are listed in Table 1 cy of poling can now be appreciated. Whilst the ceramic in the
and should be read as follows: the calculation of the series branch parallel branch would certainly be poled as long as the applied
requires the value of its ceramic and polymer volume fractions voltage ensured the field to be hgher than the coercive field (usu-
in relation to the series branch whilst, for the parallel calculation ally 2 MV/m), in the case of the series branch, the field across
N

the volume ratios of the ceramic, polymer as well as that of the the ceramic depends on the electrical properties of both the ce-
series branch in relation to the total volume (i.e. unity) needs to ramic and polymer in that branch. The expression found for the
be found. The ceramic total volume fraction inside the unit cube field across the ceramic together with its limiting values for low
cell is found by summing up the ceramic volume in the series and high frequencies are Equations (38) to (40).
branch to that in the parallel one.
In Figure 8 are shown the pyroelectric coefficient p , and d h g h
charts obtained through tlus calculation procedure for a PZT5-
epoxy (Stycast) composite using the property values of Smith
[14]. The dashed lines represent constant volume fraction curves. for all frequencies, and
It means that along these lines the n and m values give a con-
stant value for 4'. The solid lines represent iso-value curves for ,.
v,= v, mPc
mpc + (1 - m)pP
the property concerned calculated, using the mixed connectivity (39)
model. One way of exploring these charts is to choose a particu- 1
lar volume fraction and then progress from n = 0 to n = 1 (i.e. f<-2 7 r P € P
from a simple cube composite of Figure 7(a), to a 1-3 connectivity mEP
composite). It is pointed out that along the line n = 1 the com- v,= v,~A

E +P (1 - m)Ec
posite properties are the same as those predicted by Smith [14]
for 1-3 composites whilst in between we have a combination of
series and parallel properties.
716 Dias et al.: Inorganic Ceramiflolymer Ferroelectric Composite Electrets

------- --
0.9 -------- 0.9

-----------
0.8 0.8
m m
0.7 0.7

0.6 0.6

0.2 0.4 0.6 0.8 (b) 0.2 0.4 0.6 0.8


(a)
Figure 8. Pyroelectric coefficient and d h g h product charts for a PZT5-Stycast epoxy 0-3 composites. The solid lines represent the constant
property lines while the dashed lines with values in underline style represent constant volume fraction curves. (a) Pyroelectric coefficient, p
( , d / m 2 K),(b) d h g h piezo-FOM (pPaC').

The general expression has two regimes namely a h g h fre- copolymer P(VDF/TrFE) is a ferroelectric material whose Curie
quency one in which the voltage across the ceramic is roughly temperature T, is in the 70 to 150°C range depending on the mo-
proportional to the ratio @/eC which can be quite small when lar content of VDF and as a consequence, the permittivity can
using a h g h permittivity ceramic. In the low frequency regime increase to 60 [54]in the vicinity of its T,, bringing the permittiv-
the voltage across the ceramic is proportional to the ratio pc/pP ity mismatch down to just a factor of 3 in a PTCa-P(VDF/TrFE)
which can also be a small quantity due to the high resistivity of composite.
the polymer relative to the ceramic. Using the mixed Connectivity model one can find the yroper-
Various researchers tried to improve the degree of poling by ties of the composite as a function of the degree of poling a, of
increasing the ceramic or decreasing the polymer resistivities. Sa- that part of the ceramic whch is in the series branch while assum-
Gong et al. [53] obtained a decrease of the polymer resistivity ing that the parallel connectivity ceramic is always fully poled.
by doping it with a conducting powder such as carbon, silver, TIUS was done by assuming that both the piezoelectric constant
silicon or germanium in PZT/polymer and PT/polymer com- and the pyroelectric coefficients used in the series connectivity
posites and as a result they found that these materials could be branch are proportional to a parameter a measuring the efficiency
poled effectively under a field of 3 to 4 MV/m for 5 min at 100°C. of poling. In Figure 9 is shown both the pyroelectric FOM, @ / E )
T h s approach can have, however, some adverse effects due to and kt for a PTCa-PVDF [65:35vol%]composite as function of n
an increase of the dielectric loss of the composite and thus in its (i.e. the degree of 1-3 connectivity character) for various poling
signal-to-noise ( S I N )ratio. The decrease of the ceramic resistiv- efficiencies. It is apparent that the FOM, is significantly depen-
ity is usually not an option because the ceramics employed are dent on the degree of poling of the series connectivity branch in
in general optimized to have the hghest electro-activeproperties contrast with the electromechanical coupling factor whch is not
with low dielectric loss. Two other approaches [13] have been re- affected as much. This implies that the series connectivity branch
ported whch are based on the hgher temperature dependence makes a higher contribution to the overall FOM, than for the ICi,
of the resistivity of the polymer and on the hgher diffusion coef- the latter being mostly dependent on the ceramic connected in
ficients of moisture on a polymer, thus effectively decreasing the parallel.
polymer resistivity relative to that of the ceramic. A successful
exploitation of these routes requires the optimization of poling 3.6.OTHER APPROACHES
conditions such as the temperature and the humidity of the at- Other theoretical models akin to the connectivity approach
mosphere in which case high piezoelectric sensitivities can be have been reported. Hashimoto and Yamaguchi [55]have ob-
obtained. tained similar results to those mentioned earlier for the one-
Another approach is the use of host polar polymers, whch due dimensional series and parallel connectivity models using a ma-
to their hgher permittivity, provide a better balance of permit- trix formulation of the piezoelectric equations. In their calcula-
tivities, while employing low permittivity ferroelectricceramics tions it was assumed that for a parallel connectivity the equiva-
such as modified lead titanates as dispersoids. Polymers such lent composite properties of stiffness and permittivity are char-
as PVDF and its copolymer with TrFE can be useful alternatives acterized by the corresponding ensemble average raw materi-
when the permittivity are in the range of 5 to 12. Moreover the al properties whlst for a series connectivity composite it is the
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 717

1.6
PTCa:PVDF [65:35vol%]

-0.00
PTCa:PVDF [65:35vol%l - - - - - 0.50
- - _ - _ _ _ _0.75
_
. .. . . . . . .
ceramic FOM pk = 1.7e-6 c/m*)c, 1.oo

0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1


(4 n (b) n
Figure 9. Pyroelectric FOM @ / E ) and .kt dependence on the poling ratio LY of the series branch in a mixed connectivity model.
equivalent compliance and inverse permittivity which are char- 3. the way in which they are connected.
acterized by the average of the compliances and inverse permit- A fourth factor emulating the microgeometry of the inclusions
tivities of the constituents materials. can be also of considerable importance in 0-3 composites.
PARALLEL CONNECTIVITY COMPOSITE
The importance of the dielectric properties of composites has
been recognized long ago through the establishment of the so-
1 called mixing rules of wluch the the most notable ones are the
parallel, the series, the Maxwell and the Lichtenecker rules [20].
2
(a)
iso-strain 1 .D Other rules have also been proposed [37,58-601 t'ogether with
those mentioned above in the context of composite electro-active
property models.
ceramic
0 polymer
The wider field of dielectric composites is a large one and it
encompasses applications from low to very high frequencies and
(b)
non-uniform
strain
from the cement industry applications to microwave absorbers.
Here however, we will be interested in the dielectric properties
Figure 10. Schematic plot of the deformation profiles (dashed of composites which find their use as electro-active materials.
lines) of a 2-2 composite under the assumptions (a) the strain is the
same in both phases, (b) non-uniform deformation

A more rigorous treatment of the one dimensional parallel - -8- composite (experimental)
connectivity (i.e. 2-2 connectivity composites) has been report- - _ _ _ _ .

ed recently [56,57] where the longitudinal strain in the 3 direc-


series connectivitv
- parallel connectivity ,
tion is no longer assumed uniform as in the work of Smith [14]
but rather to be continuos across the phases namely at the ce-
ramic/polymer interface. In this case the interfacial bonding ac-
counts for the continuos nature of the strain providing the mech-
anism for the stress transfer from the more compliant polymer to
the stiffer ceramic (see Figure 10).
0 20 40 60 80 100
PTCa(q) vol fraction
4. DIELECTRIC PROPERTIES
Figure 11. Permittivity of composites of PTCa (quenched pow-
4.1. RELATIVE PERMITTIVITY OF der) ceramic in a matrix of P(VDF/TrFE) as a function of the ceramic
COMPOSITES AND THEIR volume fraction
MIXING RULES The volume fraction dependence has been studied at a fre-
quency of 1 kHz and 30°C for PTCa:P(VDF/TrFE:) composites
The dielectric properties of composites depend as other prop- as shown in Figure 11. In the same Figure are drawn the curves
erties on three factors: corresponding to a pure 2-2 parallel and series Connectivity com-
1. the properties of the constituent phases, posite. It can be observed that the experimental points describe
2. their intervening volume fractions and a curve that lies between these two limiting connectivity types.
718 Dias et al.: Inorganic CeramiflolymerFerroelectric Composite Electrets

A similar type of outcome can be observed to apply to the imag-


inary part of the permittivity. That is what should be expected
if the rules which apply to the real part of the permittivity are
also applicable to its imaginary part. This is an interesting point
because the derivation of the formulas in the composite models
should still be valid if the variable properties have complex val-
ues, thus enabling a prediction of complex valued properties of
composites.
Figure 12 [31] shows the experimental values for permittivity
of a PZTPVDF composite at various volume fractions together
with a fitting performed using the Yamada model. In that work
a shape parameter q = 8.5 was found to be the best fit for the
experimental data whch as was commented above could be ex-
plained by the high content of ceramic in the composite (see Sec-
tion 3.2 for a full explanation).

w
250 1 PZT-PVDF
n
E(PVDF)=S.O
.?
._
200
11
._
c

8.5
150
~

- -3
Q _ _ - 1
experimental points
100

50 -
Das-Gupta and Abdullah [61] measured the dielectric dis-
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 persion of PZT/PVDF, PZT/P(VDF/TrFE), PZT/PP and Piezel
ceramic volume fraction. 9‘ (Daikin Industries) in the frequency range 10 Hz to 100 kHz at
Figure 12. Theoretical plot of the permittivity for three values of 363 K and the temperature dependence of E” at 1 kHz. Dielec-
the Yamada shape factor in a PZTPVDF composite together with
the experimental data for this composite [31].
-
tric dispersion with a broad maximum at 350 K was observed
for PZT/PVDF and Piezel composites and was assigned to the ac
In Figure 13 is shown a contour chart for the permittivity of relaxation of PVDF. In the low temperature region at 270 K, N

PTCa:P(VDF/TrFE)using the mixed connectivity model. For t h s an absorption ascribed to molecular motions in the amorphous
chart the experimental values of the constituent material have phase of the polymer was observed. The imaginary part of the
been used. The constant volume fraction curves have also been permittivity E / / vs. the frequency is shown in Figure 14 for vari-
ous composites of the same volume fraction of ceramic (i.e.50%).
drawn in the chart as dashed lines.

- -
A decrease in the dielectric losses was observed up to 500 Hz
The location of each of the points in a certain property chart is and at 5 kHz a broad maximum of E” was found for compos-
found by using simultaneously the volume fraction of the com- ites having the PVDF matrix. Thus, tlus absorption process can
posite and the experimental value of that property. For instance, be assigned to the relaxation in PVDF polymer.
in the case of the 60 vol% PTCa composite with a permittivity of Sinha et al. [62] measured dielectric dispersion of PZT/PVDF
49, one has to locate first the line of 60% volume fraction in the composites in the frequency range 10 Hz to 10 MHz at room tem-
permittivity chart and then find the property line closer to the perature. They observed a minimum in E” at the frequency of
value of 49. This locates the point in the chart and enables one to -
find an estimate for the values m and n characterizing the equiv- ceramics.
10 kHz and assigned it to the domain wall motion in PZT

alent mixed connectivity cube of the composite. In the present


- -
case m 0.85 and n 0.45 and thus 23% of the 60% total ce-
ramic loading of the composite touches both the upper and lower
From these studies it can be concluded that dielectric relax-
ations in composites reflect those of the constituent materials d-
though new relaxations arising from the interactions between the
electrodes whilst the rest is connected in series with the polymer.
phases should also be expected. These new relaxations should
In t h s chart it can be noticed that the experimental points have possibly be located in the low frequency range as was found
an n value which is in the range of 0.4 to 0.5 and thus about half in a composite of Nb-modified (Pb,Ba)(Zr,Ti)O, dispersed in
of the ceramic cube touches both electrodes, which in turn means P(VDF/TrFE) copolymer in the 350 to 420 K temperature range
that almost a quarter of the volume of the cube is represented by and in a frequency region < 100 Hz [63]. This was a small re-
ceramic exhibiting 1-3 connectivity. laxation which has been ascribed to a relaxation of defects with
IEEE Transactions on Dielectrics and Electdeal Insulation Vol. 3 No. 5, October 1996 719

an activation energy of 0.7 eV and has been observed to decrease determination of the frequency response of an ultrasonic trans-
with thermal annealing. ducer as well as its time response through a Fourier transforma-
In the h g h frequency region above 5 MHz a considerable in- tion. The key to t h s calculation is a matrix of impedances first
crease of E" accompanied by a small decrease in the E' value for derived by Bui et al. [66,69,70] for a lossy piezoelectric material,
a PZT/rubber composite has been observed. The origin of this relating the voltage and mechanical forces to the electric current
dispersion is still unknown and needs further investigation. and displacement velocities (see Figure 15 and Equation (42):
I 1 I
1
E"
16
vs
fo = 5
(42)
8 2, = A z , =
~ Apu,a

a=&+i- Qf
0 fo

Figure 14. Dielectric absorption at 363 K for 0-3 composites of (1) where fo is the half wave resonance frequency, A the area of the
PZT/PVDF, (2) PZT/P(VDF/TrFE), (3) Piezel and (4) PZT/PP. The sample, 1 the thickness, p the density, 'U, the velocity of sound, z ,
ceramic volume fractionis 50% [61,64].
N the acoustic characteristic impedance of the composite, $ the me-
chanical loss and h33 the piezoelectric voltage coefficient whch
5. PIEZOELECTRIC PROPERTIES is related to the electromechanical coupling factor /ct by
AND APPLICATIONS
5.1. INTRODUCTION AND (43)
MEASUREMENT METHODS
In contrast with inorganic piezo materials, which have a very A knowledge of the backing acoustic impedance Zb = Azb
high quality factor Qm and thus low mechanical losses, for or- allows one to relate the force Fl to the displacemeint velocity VI
ganic materials like PVDF and piezoelectric composites, the fa- at the back of the transducer through. This makes it possible to
miliar methods of analyzing the frequency dependence of their reduce the above matrix (Equation (42) to a two-port network
electrical impedance give rise to large errors in the determina- in which the input quantities (i.e. the voltage and current) are
tion of the piezoelectricconstants. Various procedures to analyze related to the output ones (i.e.the force and the displacement
the impedance spectra of lossy materials have been proposed velocity at the front surface of the transducer) through a 2x2

q;-lfiJ q-p&
amongst others by Brown [65], Bui [66] and Ohigash [67l and matrix.
the last one will be described here. In this method [67],both
the permittivity under constant strain E; and the elastic stiffness
constant under constant displacement cy3 for a lossy material, Zb
backing 2- matrix medium
are taken to be complex quantities

mechanical - + mechanical
ABCD "C ABCD
matrix matrix

where 4 = tan 6,is the dielectric loss and $ = Q;l = tan(6,) Figure 15. Network representation of an ultrasonic transducer
is the mechanical loss. The total measured impedance of a piezo-
electric sample resonating in its transverse mode can then be de-
duced to be a sum of two terms i.e. the electric impedance 2,( U ) input
of the sample capacitance, and an acoustic term Z,,(w) [67].
Following the analogy between the mechanical quantities (i.e.
force and displacement velocity in an acoustic transmission line), Figure 16. Network representation of a mechanical layer.
and the electrical ones (i.e.voltage and current), the electrical net-
work theory developed over the years to deal with electrical lossy Furthermore according to this formalism (the so-called ABCD
transmission lines can be used readily for the study of acoustic matrices) the behavior of a non-piezoelectric mechanical layer
transmission lines encountered in the design of ultrasonic trans- can be described also by a matrix which relates the input force
ducers. This scheme which has been used earlier [68], allows a Fz and displacement Vzto the output force F3 and displacement
720 Dias et al.: Inorganic Ceramiyl'olymer Ferroelectric Composite Electrets

V3 as follows resulting output voltage compared with that of a calibrated mi-


crophone from which the voltage sensitivity g h is then obtained.
The d h coefficient is computed by multiplying the g h constant
by the constant stress permittivity. A useful figure of merit for
hydrophone materials is the product of the hydrostatic strain co-
To calculate the overall matrix relating the voltage/current to
efficient d h and the hydrostatic voltage coefficient Q h whch may
the force/displacement in the output medium one then requires
thus be computed. The d 3 3 stress coefficient is usually measured
the determination of the product of all the matrices describing
using a Berlincourt meter at 100 kHz.
each of the elements composing the transducer. It is necessary
however, to define both the mechanical and the electrical quanti-
ties appropriately, so that the Newton's third law, the continuity
of displacement velocity and the Kirchoff laws are obeyed, while
allowing for the chain rule of the ABCD matrices to apply (see
Figure 15 and 16).

(45)

Finally, to calculate the frequency response we must know the


acoustic impedance of the front output medium Z f = Azj . The
equation, coupling the force and the velocity Pout = Z f Voulto-
Figure 17. Schematic setup for measuring the g h coefficient.
gether with the matrix equation (Equation (45))relating the input
to the output quantities in the transducer allows us to determine 5.2. CO M POSITE PIEZOELECTRIC
the ratio of the output force to the input voltage FoUt/Vc which PROPERTIES
is the transfer function of the emitting transducer.
When the transducer rather than a transmitter works as a re- Good review papers have been published on the piezoelectric
ceiver unit, the input quantities (i.e. voltage and current ) will composites for ultrasonic applications by Gururaja [71], Smith
then be the output quantities and vice versa for the case of force [26] and Ting [72]. Here we will review some of the results re-
and velocity at the front surface of the transducer. In t h s case it ported for 0-3 composites.
turns out that the order of the elements in the matrix describing PZT- epoxy composites are amongst the earliest reported com-
the receiver will be transformed to DBCA. posites for piezoelectric applications [13,47], namely by Fu-
rukawa and coworkers [35] who prepared films 200 p m in thck-
The determination of the piezoelectric properties of the com-
ness, using ceramic grains with diameter range between 0.2 and
posite have been acheved from the measurement of the electri-
cal impedance of a sample vibrating as a free resonator. Thus
2 pm up to a volume loading of 23%. In Figure 18 are shown the
poling characteristics of these composites in relation to the piezo-
FI = Fz = 0 and it can be shown that the total measured imped-
electric stress coefficient where a maximum field of 12 MV/m
ance of a piezoelectric sample resonating in its tlckness mode is
was applied. Using the Furukawa model described in Section 3.3
equal to a sum of two terms: the electric impedance Z,(w) of the
it was found that the experimental d-constants were 3/4 of the
N
sample capacitance and an acoustic term Z,,(w) [67].
predictions leading to the conclusion that these composites were
+
VC = Z , ( w ) Z,,(W) = 7 -
- + 2i (;)qX) imperfectly poled. Pardo and coworkers [36] also have report-
ed on these composites in w l c h the grain size to thickness ratio
IC IWC Z, 2afo
(46) ( g s l l ) was varied from 0.5 to 3.8. They found that a mixed con-
nectivity composite was obtained if the ceramic grain size was-
The experimentally determined impedance can thus be fitted larger than the thickness, regardless of the volume fraction of the
by this equation which upon optimizing the values of the piezo- ceramic. The d33 coefficient was found to be higher for the hgh-
electric voltage coefficient h, which relates to kt in accordance to est ceramic volume fraction and grain to thickness ratio used (i.e.
Equation (43), the mechanical loss $ and the resonant frequen- 4 = 43% and g s / l = 3.44) and equal to d33 = 165 pC/N whlst
cy fo, allows the determination of the piezoelectric properties of the 9 3 3 coefficient peaked at a lower volume fraction of 27% and
the materials when its mechanical losses are not negligible. Once gs/E ratio of 0.8. Espinosa et al. [73]have reported on a manufac-
these parameters are found and the density is known, other ma- turing procedure which consists in multifracturing a commercial
terial properties (i.e., the velocity of sound u s , the characteris- piezoceramic plate by using different techniques and then pour-
tic acoustic impedance z,, the stiffness constant c 3 3 ) can also be ing the plastic phase between the ceramic elements with the re-
computed. sult that the mechanical losses were increased whilst the sensitiv-
The g h and d h hydrostatic piezo coefficients can be measured ity based on the electromechanical coupling factor (- 50 to 60%)
using a low frequency acoustic method (see Figure 17). The sam- was similar to those obtained on 1-3 composites despite being
ple is placed in a box where a sound pressure is generated and the easier to fabricate.
IEEE Transactions on Dielectn'cs and Electrical Insulation Vol. 3 No. 5, October 1996 721

loading showed that 50% was a good compromise when opti-


mizing the d h g h product. The piezoelectric constants d h , gh and
0.5 ............. ...............+.. 0.23
i d h g h , obtained using a poling field of 20 MV/m, was 22.5 pC/N,
80 mV.m/N and 1 . 8 ~ Pae1 respectively. Shaulov [80] has
-
?
0.4 ............ ................................ ......... ............. ........ ...............
also reported good results on a 1-3 composite made of the same
ceramic and two different polymer hosts one being stiffer (Sty-
8- 0.3 ............. : ......... ................
cast epoxy sll = 108 x lo-'' Pa-') than the other (Spurrs
U- , - . epoxy s l l = 332 x Pa-'). Better piezoproperties were
0.2 ........................
obtained for the composite using the stiffer host where d h , gh
0.1
and the d h g h , product were equal to 32 pC/N, 66 rnV.m/N and
2.1x10-~' Pap1 respectively.
0 Research on the practical way of preparing a piezoelectric
0 2 4 6 8 1 0 1 2 1 4 paint has been done by Hanner and coworkers [83,]. Two poly-
Poling field [MV/m]
Figure 18. Poling characteristics of PZT-epoxy composites [34].
mers, an acrylic and a polyurethane were utilized in their study
and loaded with 60 to 70% of PZT or co-precipitated lead ti-
A series of 0-3 type composites called piezo mbbers (PR- tanate (PT) ceramic. Both polyurethane and the aciylic polymer
303/PR-307) have been developed more recently by NTK Tech- were dissolved in a suitable dispersing agent to which the pow-
nical Ceramics Division of the NGK Spark Plugs Corporation in der was then added. The films were prepared by casting the
Japan, which consist essentially of fine lead titanate particles dis- mixtures onto brass plates and cured in a vacuum oven. Air
persed in a chloroprene polymer matrix and whose main charac- dry silver paint was used as an electrode for the final dry films
teristic was that h g h d h g h products in the range of 1to 5 x 10-l' while the brass plate acted as the second electrode, with result-
Pa- were aclueved and good transmitter/receiver pulse re- ing thicknesses of 200 to 500 pm. Composites of the combi-
sponses were obtained without ringing [21,74,75]. nation of PT and acrylic gave the best results in terms of their
A new method of preparing 3-3 connectivity composites was piezoelectric properties whle the lowest value was that of the
reported by Gururaja [75] which consisted of preparing a ceram- PZT/ polyurethane composite. The piezoelectricproperties of the
ic (PbTiO3) using a co-precipitation method, and then form pel- PT-acrylic composite were d h = 32 pC/N, gh = 67 mV.m/N
lets using a PVA binder which was burnt out during a firing op- and d h g h = 2.15 x 10-l' Pa-'.
eration. The voids were subsequently filled with an epoxy to Ferroelectric polymers such as PVDF and its copolymer
form the composites whose ceramic loading was N 70%. The P(VDF/TrFE) have also been used to make composites. Their
composites were poled both conventionally and by the corona main features, apart of being pyro and piezoelectric, are their
method which provided similar results. The piezoelectric hydro- h g h permittivity due to their polar nature and h g h stiffness rel-
static F O M h was 1750x Pa-' whist the electromechanical ative to other thermoplastics. The composites, loaded with PZT
coupling factor was 8%. A measurement of the 002 and 200 X- to 21% [35], were obtained through the hot-rolling technique and
ray diffraction peaks before and after poling showed an apparent pressed to 200 pm in thickness and gold evaporated on both
saturation of the poling. sides. Poling performed at 120°C during 30 min was studied for
A more uniform composite was claimed to be prepared by Han the dependence of the electric field (Figure 19) which shows that
and coworkers 176,771 using a colloidal processing in the pro- poor piezoelectric activity is obtained probably due to insuffi-
duction of composites of (Pbo.5,Bio.5) (Tio.5, (Fel - Mn,)o. 5)O3 cient poling field together with a low ceramic loading. It is clear
(z = 0.00 to 0.02) in an epoxy matrix so that poling fields equal that with the low poling fields applied it is not possible to orient
to 15 MV/m were possible to be sustained without breakdown the polymer phase whose coercive field is > 80 MTJ/m.
of the sample. The values of the piezoelectriccoefficients, d33, d h 5
and gh were as h g h as 65, 41 pC/N and 145 mV-m/N respec-
tively and thus the d h g h = 5.95 x lo-'' Pa-'. 4
Calcium modified lead titanate composites in an epoxy ma-
trix has been extensively studied by Pardo 1781, Chilton [79],
Garner [50] and Shaulov [80]. Pardo has reported on the de-
pendence of the piezoelectric activity on the grain size of the
ceramic particles relative to the thickness of the sample where
he found that similarly to the PZT ceramic composites case a 1
higher grain size to thickness ratio was beneficial to its proper-
ties. Garner [50] reported on the effect of obtaining the ceramic
powders through two different routes namely by ball milling and 2 4 6 8 10 12
Poling field [MV/m]
quenching with the result that the quenched material was found
to perform better. Investigations into the appropriate volume Figure 19. Poling behavior of a PZT-PVDF composite [35].
722 Dias et al.: Inorganic Ceramiflolymer Ferroelectric Composite Electrets

Yamada [31] also has reported on PZT-PVDF composites with 8 MV/m at 85°C and 15 min of poling and a remnant polariza-
concentrations of ceramic up to 70%. Improved properties were tion of 0.13 to 0.14 C/m2 was measured whle the coercive field
found whch for a 67% PZT content composite resulted in ~ 3 =
150, d33 = 50 pC/N, and Young modulus of 3 GPa. The de-
3 was - 0.6 MV/m. It should be pointed however, that a satu-
ration of poling does not imply 100% poling of the ceramic but
pendence of the permittivity and the piezoelectric constant d33 rather that a plateau in the target property has been reached such
on the PZT volume fraction is shown in Figure 20 in whch the that increasing the poling field, the poling time or temperature,
simulated curves are those obtained according to Equation (2) any further, does not lead to an improvement in that property.
with the parameter 7 = 8.5 used in the calculation, representing Measurements of the properties against the grain size showed
a ceramic elongated ellipsoid particle with its long axis arranged a sharp increase of the piezoelectric constant d33 for grain sizes
perpendicular to the surface of the film (see discussion of the Ya- larger than 20 pm whch was ascribed to a transition from single
mada model, Section 3.2). domain to multi-domain behavior for ceramic grains > 200 pm.
160 80 Piezoelectric 0-3 composites using P(VDF/TrFE), also have
been extensively investigated at our laboratory using different
140 70 ceramics (e.g.PZT and PTCa) and powder processingroutes (e.g.
120 quenching and milling) as schematically shown in Figure 21.
60
100

80
50

40
jlPICal/ pl
milled quenched milled
60

40

20
30

20

10
I
I 'I P(VDFfTrFE) I I
Figure 21 . Compositesevaluated for their piezoelectric properties
D "
n at our laboratory.
0 0.10.2 0.3 0.4 0.5 0.6 0.7
The fabrication route of these composites was the hot-rolling
ceramic vol. frac. ["%] techruque as outlined in Figure 2. Measurements of the elec-
Figure 20. Dependence of the permittivity and piezoelectric c o n
tromechanical coupling factor were performed by the resonance
stant on the PZT volume fraction [31]. method whlst the hydrostatic measurements were performed
The ceramic particle size dependence has also been investigat- using the setup of Figure 18. The piezoelectric property mea-
ed by Lee [49] with a barium titanate-phenolic resin composite. surements were always done at room temperature and pressure
The ceramic particles size range was 2 to 130 pm whle the ce- and for the electromechanical related properties they are strictly
ramic volume fraction was kept constant at 60%.They were able valid around the resonant frequency range.
to explain the experimental results based on the existence of a In Table 2 are listed the piezoelectric properties found for these
-
surface layer on BaTi03 [82]grains 1.59 pm thick whose di- composites namely the electromechanical coupling factor kt, the
mensions become significant for the smaller grains of the ceram- quality factor Qm, the acoustic impedance 2,and the hydrostat-
ic and whose characteristics were their low permittivity ( w 105) ic piezo properties d h and the product d h g h . We have listed in
and non-ferroelectricbehavior. This effect disappeared when the the same Table the properties of some 0-3 composites reviewed
ceramic grains were > 100 pm. It was also pointed out, how- above as well as the properties of PTCa and P(VDF/TrFE).
ever, that the thickness and properties of the surface layer may No results for the PZT-P(VDF/TrFE)composite regarding its
depend on the previous processing of the ceramic powders. electromechanicalcoupling factor and its related properties were
Lead titanate PTP(VDF/TrFE) composites were investigated reported because it was not possible to detect any resonance peak
by Ngoma et al. [83]. They found that the ceramic inclusions in the impedance spectra in any of its polarized samples. This
do not modify the nature of copolymer crystalline phase with fact together with the low d h and FOMh coefficients leads us to
respect to its dielectric and DSC behavior. Due to the low fields suggest that this material has a low piezoelectric activity.
applied whch were in the range of 9 MV/m, they also found that Regarding the PTCa composite a reasonable electromechani-
a poor piezoelectric activity was present in these composites. cal coupling factor was observed with the poling field together
Poling studies have been evaluated by changing the poling with a low mechanical quality factor. The latter has a low val-
conditions and measuring the resulting piezoelectric properties ue between 5.5 and 8 probably due to the very dispersive nature
and XRD diffractograms, on a composite of 70% volume frac- of the composite media in which the ceramic grains act as scat-
tion co-precipitated PT in a gel polymer (Eccogel 1365-0) [84]. ters leading to an attenuation of sound waves. There is no clear
In these studies which used ceramic grains in the 12 to 1100 nm dependence of this property either on the volume fraction or on
size range, a saturation of poling was found for poling fields of the size of the grains. One should employ some caution as the
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 723

thickness variation which is 5%, could lead to a smoothing of


N the polarization does not exhibit a decay phenomenon. This lat-
the resonant peak which in turn would artificially decrease the ter finding further indicates that although the host polymer is of
estimated quality factor. a ferroelectric nature it behaves in this composite as a passive
In Figure 22 is shown a plot of the impedance measurements carrier of the electro-active (i.e. ceramic) material.
made on a 50% quenched PTCa composite. The experimental The hydrostatic properties of these composites are directly re-
curves were fitted to Equation (46) using the electromechanical lated to the ceramic materials properties from which they are
coupling factor, the quality factor and the resonance frequency produced. In t h s respect the PTCa ceramic has a much higher
as fitting parameters. As can be seen a good match between the gh than the PZT ceramic and t h s is in agreement with experi-
experimental and the theoretical values is achieved in the case of mentally observed value of g h for the PTCa composite which is
these composites. higher than that of the PZT composite. In many applications,
1000 viz. in receiver applications however, g h acts as a FOMh and in
PTCa:P(VDFRrFE) 50%
relation to tlus constant the composite is a better material than
the pure ceramic. In Table 2 we notice that the 60% quenched
ceramic composite have higher g h coefficients than that of the
milled ceramic probably as a result of the bigger grain size of the
former.
In Figure 23 are shown the poling field characteristics of the
quenched PTCa 65% composite in terms of the dZL3coefficient.
imaginary pari of impedance An alternating poling field with 0.05 Hz frequency with a vari-
calculated cuwes
able amplitude was applied to the sample and the ~1~~ measured
after a complete poling cycle. Although an indication of a sat-
uration seems to appear for poling field amplitudes approach-
ing 25 MV/m it is not certain what the saturation polarization
was in this experiment as the maximum electrical breakdown
strength of the composite did not allow for higher values of the
field. However, for a polarization method which involved ap-
plying a step field of 25 MV/m during 1h we found a value of
d33 = 60 pC/N which can thus be estimated to be the saturation
value for this property using this poling field.

55 -
/
PTCa:P(VDFTTrFE)
50 -
e 45 -
00
-3 40 L
U
35

30 c

25 1 1 1 1 1 1 1 I l l I l l I l l I l l 'Ilk

10 12 14 16 18 20
2!4 26 22
Poling field [MVlm]
Figure 23. Poling field characteristics of the PTCa-F'(VDF/TrFE)
composite.
In Figure 24 is shown a mixed connectivity chart for the d33 for
a PTCa-P(VDF/TrFE)composite where the experimental points
for the quenched ceramic powder of PTCa composite have also
been plotted. The determination of the location of the points was
achieved through a similar procedure as that used for the permit-
tivity chart. It can be observed that the n values for t h s property
are in general a bit lower than those of the permittivity chart of
724 Dias et al.: Inorganic Ceramic/Polymer Fevvoelectric Composite Electrets

Table 2. Piezoelectric properties of selected materials.


~ ~

~ ~

%vel host Qm za [Ref.]


___ _____
MRayl
57 epoxy 36
50-90 Araldite 0.6-0.5 5-9 15-26 73
? rubber 17-44 1-5 74
70 epoxy 0.08 25 1.8 75
35 epoxy 41 6.0 77
50 epoxy 22 1.8 50
55 epoxy 45* 2.4+ 78
25 Stycast 32 2.1 80
70 acrylic 32 2.2 81
PZT 67 PVDF 48" 1 . 7 ~ 31
PTCa 30 P (VDF/ TrFE) 25" 1.7* 83
PZT 50 P(VDF/ TrFE) 10 0.12 69
PTCa m. 60 P(VDF/ TrFE) 0.06 12 14.7 8.5 0.1
PTCa q. 60 P(VDF/ TrFE) 0.11 7.4 14 12 0.2
PTCa q. 65 P(VDF/TrFE) 0.24 4.3 17 28 1.3
PTCa 100 0.47 1200 30 62 2.1
P(VDF/TrFE)
\ . I -
100 0.30 20 4.5 9 1.2 67

Figure 13.There is a tendency however, as with that chart for the Kumar et al. [87]where it was shown, using a Cole-Cole dia-
hgher volume fraction composites to show an increase in their gram, that PZT-epoxy resin exhibited typical Debye-type piezo-
1-3connectivity character. electric dispersion. These relaxations was claimed to be due to
the frequency behavior of heterogeneity and trapped charges.
6. PYROELECTRIC PROPERTIES
0.9 6.1. STATIC AND DYNAMIC
MEASUREMENTS METHODS
m
The pyroelectric constant p of a material is determined from
0.8 the change of the total charge AQ at the electrodes due to a
change in sample temperature A T

(47)
0.7
In the above equation the approximation is only valid when
the area A of the sample does not vary significantly in the tem-
perature range of measurement. In that case it must be pointed
out that a relevant practical quantity is still obtained using the
0.6
simplified equation if one envisages using the assessed material
for making pyroelectric detectors.
In order to measure this coefficient the most commonly used
way, the so-called direct method [SS], is to heat and/or cool the
0.2 0.4 0.6 n 0.8 sample at a constant rate p in a cryostat while the current I =
Figure 24. d 3 3 piezoelectric coefficient contour chart for a PT- AQ/At is monitored [32,88]. The pyroelectric coefficient is then
Ca:P(VDF/TrFE)composite using the mixed connectivity cubes. given by
I
p=-
Due to the nature of the piezoelectric phenomena which por- PA
trays a coupling between the dielectric polarization and the elas- Because this procedure is the same as for the thermally stim-
tic properties of a material one can expect that the piezoelectric ulated current (TSC) measurement, one has to ensure that this
coefficients show frequency dispersion. The piezoelectric relax- kind of contribution to the total current is negligible. Using the
ation behavior in composites has been reviewed in the work of property that the pyroelectric current is a reversible current wlule
IEEE Transactionson Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 725

the TSC is an irreversibleone, t h s is accomplished by heating the mismatch (see Section 3.4 on the composite model's) of the com-
sample several times until a reproducible current is measured or ponent phases. The ferroelectricPZT material usecl was of a for-
by performing an annealing of the sample at a temperature high- mulation designed to give lugh piezoelectric d coefficients and
er than the highest temperature of measurement for a extended thus was not the best one suited for pyroelectric applications.
period of time until the level of current subsides to negligible However, it had the advantage of being able to reproduce con-
values relative to those which are to be measured. sistent electrical properties. The epoxy material, SPURRS was a
In the radiative dynamic method of determining the pyroelec- multicomponent system designed for casting and replication due
tric coefficient a light source is used to deliver an ac thermal pow- to its very low viscosity and, therefore suited for the preparation
er W . Assuming that the sample heats up uniformly the pyro- of the 3-3 connectivity composites. The ceramic loading used
electric coefficient is given by [89] was 40%. Pyroelectric and dielectric measurements demon-
N

strated that the pure 0-3 composite behaved as a dilute PZT sys-
r c tem whose pyroelectric coefficient and permittivj ty were both
P=A= (49)
reduced by a factor of ten in such a way that its ratio p / (i.e.~
where c is the volume specific heat and w is the angular frequen- composite p / = ~ 0.35 pC/m2K) was roughly equal to that of
cy. When the incident radiation is chopped a lock-in amplifier the pure ceramic (PZTp/&= 0.27 pC/m2K). In the 3-3 compos-
must be used to monitor the current generated at the chopping ites made, whose structural unit size was 100 to 150 pm, it was
frequency. observed that critical size effects appeared for tlucltnesseslarger
Another method [90] based on oscillating the temperature of than 6 to 10 structural units (> 1 mm). The pyroelectric coeffi-
the sample uses Peltier elements whose current is controlled to cient exlbited under 45°C a sign opposite (i.e.positive) to that of
generate a low-frequency temperature wave T = Tocos(wt). the ceramic changing to the same sign for l g h e r temperatures.
As the pyroelectric current is proportional to the time derivative This effect was related to a change of the relative contributions of
of the temperature of the sample the magnitude of the current, the primary pyroelectricity, which is due to the intrinsic pyroelec-
which is 90" out of phase relative to the temperature is related to tric properties of the ceramic, and secondary pyrolelectricity, the
the pyroelectric coefficient through latter showing a stronger contribution for thicker samples and
lower temperatures due to the stiffer nature of the polymer host
I 1 in these conditions.
p= -~
A WT,
BaTi03-rubber 0-3 composites have been synthesised by Amin
The current in phase with the temperature, should be discard- and coworkers [92,93]. The powdered ceramic was introduced
ed because its origin is not pyroelectric. Care should be exercised in a butadiene acrylonitrile rubber after which thle whole mix-
so that the exciting frequency is low enough so that the temper- ture was vulcanized in stainless steel molds to obtain thin films.
ature across the sample is uniform, i.e. Poling of the composites was performed at 130 and 170°C (i.e.
K above the Curie temperature of the ceramic) under poling fields
W < < F ranging from 0.1 to 5 MV/m. A peculiar behavior was found:
the pyroelectric coefficient was found to decrease with increas-
where K is the thermal diffusivityand E the thickness of the sam- ing poling field when poling at 130°C. This effect was ascribed
ple; otherwise currents due to temperature gradients in the sam- to space charge formation which also contributed to the pyroelec-
ple will also add to the measured current. tric current measured by the direct method. At 170"C, however,
6.2. PYROELECTRIC PROPERTIES the sample was found to be free of the space charge and the true
pyroelectric coefficient was found to be M 60 pC/m2K for a 30%
Various combinations of ceramic/ polymer 0-3 composites BaTiOs volume loading. The permittivity was found to be 17 N

have been reported in view of optimizing the pyroelectric co- and thus the pyroelectric FOM, was 3.5 pC/m2K which is quite
efficient and/or the pyroelectric figure of merit FOM, which is a h g h value.
p / & [33]. These reports have also included studies of the effect Tripathi et al. [94] have produced 0-3 composites of BaTiO3-
of the volume loading and in some instances the ceramic grain PVDF in which the BaTiOBpowder, 1pm in size, was obtained
size effect. Although the amount of publications on this subject by means of the sol-gel route. The permittivity of the composites
does not compare with that performed on composite piezoelec- was 20 and good pyroelectric properties relative to the pure
N

tric properties there is ground to suppose that successful com- BaTi03 were found; however no absolute values for the pyro-
mercial pyroelectric devices can be made out of these materials. electric coefficients were reported
Here we will review some of this work. TGS-PVDF 0-3 composites have also been evaluated for their
Bhalla and coworkers [91] have reported on the pyroelectric pyroelectric properties [33,95]. TGS crystals grounded to par-
coefficient of PZT-epoxy composites of both 0-3 and 3-3 connec- ticle sizes of 45 to 75 pm were mixed in different proportions
tivity. It was anticipated that improved pyroelectric properties up to 80% volume loading with PVDF dissolved in an organ-
would be brought about as a result of the enhancement of the ic solvent and then poured over glass plates to obtain films 50 to
secondary pyroelectric effect arising from the thermal expansion 100 pm thck. The polarization procedure was performed at 70 to
726 Dias et al.: Inorganic CeramicyTolymerFerroelectric Composite Electrets

80°C for 2 to 3 h whilst applying a field of 0.4to 0.6 MV/m. The curve increases for ceramics loading > 35% and thus no opti-
pyroelectric properties increased with the ceramic loading and mum volume fraction may exist. In their work it was suggested,

withplc = 3.3 pC/m2K.


-
at 80% the pyroelectric coefficient obtained was 90 pC/m2K according to a composite model developed therein, that polymer
hosts with higher permittivities would be desirable in order to
60
improve on the magnitude of the pyroelectric coefficient.
PZT-PVDF and PZT-P(VDF/TrFE) together with a composite
50
w
Piezel made of PZT and a polymer matrix were evaluated by
p 40 Abdullah and Das-Gupta [97,98]. The composites were manu-
c factured by means of the hot-rolling technique using a fine pow-
E 30
- dered ceramic using a 50% volume loading. The pyroelectric co-
20 efficient at room temperature was found using the direct method
10
to be 10 pC/m2K increasing to 140 pC/m2K at 70°C.The corre-
spondingplc ratio wasp/& = 0.35pC/m2K and 1.5pC/m2K
0
10 20 30 40 50 60 70"
respectively at room temperature and 70"C,demonstrating a
PT volume fraction ["/I high dependence of the pyroelectric properties on temperature
Figure 25. Permittivity and pyroelectric coefficient for a PT-PVDF as u7as also found by Bhalla et al. [91],thus at 70°C their FOM,
composite as a function of the PT content [96]. was higher than that of the pure ceramic. Results on the py-
roelectric coefficient dependence on the poling time, field and
Experiments with PVDF and PE (polyethylene) as hosts poly- temperature have been also undertaken (see Figure 27) on these
mers have been performed by Yamazaki and Kitayama [96] using composites where it can be observed that saturation of polariza-
as dispersoids ceramic grains of PZT and PT in the 0.5to 3 pm tion is attained for poling times in excess of 5 h [61]. However,
size range. The 0-3composites produced utilizing the hot-rolling increasing the poling temperature and field results in a monot-
techruque described above, were then pressed to obtain thm films onous increase of the pyroelectric coefficient and no saturation
in the 30 to 70 pm thickness range. Pyroelectric measurements pattern is apparent.
were then performed using the radiative dynamic method. It was
found that for a similar ceramic volume fraction of 62% the hgh- p (C.mT', K") x lo4
est pyroelectric voltage responsivity was that of PT-PVDF wluch
was six times higher than PZT-PVDF and 50 x that of PE-PT.
N

The pyroelectric coefficient for the PT-PVDF 62% volume fraction


composite was found to be 130 pC/m2K whilst its permittivity
was N 54 (see Figure 25) resulting in a pyroelectric FOM, equal
to 2.4 pC/m2K. Ths value compared well with that of the pure
ceramic wluch was reported to have a p / & = 3 pC/m2K.

2.5

P 2
(U

.
E
3 1.5
0 2 4 6
2"
9 1 E, 1 Vn"l x 106
Figure 27. The dependence of pyroelectric coefficient on poling
0.5 parameters at 343 Kin a PZT-P(VDF/TrFE)50% composite [98].
0-3composites using the ferroelectriccopolymer P(VDF/TrFE)
0 [75:25 mol%] as a host material have been also fabricated
0 10 20 30 40 50 60 70 [99,100]. This material has the attractive property of having a
0 , (ceramic volume fraction) high permittivity relative to other polymers which, as was point-
ed out earlier enhances the poling efficiency of the ceramic phase.
Figure 26. FOM, as a function of the content of the ferroelectric The ceramics used as dispersoids included PLZT, PZT and calci-
powder [96].
um modified PT (PTCa). The latter ceramic powder of PTCa was
In Figure 26 is shown the pyroelectric FOM, as a function of obtained either by a grinding or quenchmg process.
the PT content where it can be observed that the slope of the The pyroelectric coefficient for composites made with PZT,
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 727

PLZT and PTCa is shown in Figure 28. These are 50% ceramic the voltage applied to the grid, whch for a low resistivity sam-
volume fraction composites poled by the conventional method ple may not be the case. For instance, in the case of the 60% vol-
(Figure 28(a))at 100°C for 3 h while applying an electric field
N ume fraction PTCa composites (i.e.lower resistivity materials)
of 20 MV/m. It may be observed that the pyroelectric coefficient we have measured lower pyroelectric coefficients for the corona
of PTCa ceramics composites are 50 and 150%hgher than PZT
N poled samples than for the conventional poled ones whch can
and PLZT composites respectively at room temperature. Howev- only be explained if the conduction current through the compos-
er, as the temperature increases, the PZT and PLZT composites ite sample restricts the surface voltage to a much lower value
perform better relative to the PTCa composites due to the low than that set by the corona grid.
temperature dependence of the latter. We note however that a
Figure 29(a) shows the pyroelectric coefficient for three calci-
low dependence on the temperature is a useful feature to assure
um modified lead titanate composites using two ldifferent vol-
a steady working performance of a pyroelectric detector system.
ume fraction, 50 and 60% and two different powdler processing
The PLZT ceramics have a much higher resistivity (lo1' a m ) methods either milled and quenched with ceramic grain sizes in
than any of the other ceramics for which the resistivity is of the range of 1pm and > 20 p m respectively, all of them conven-
the order of 10' Om. This would improve the poling effi- tionally poled.
ciency by having a higher fraction of the poling voltage across The difference in the behavior of the quenched and milled
the ceramic grains for long poling times. As can be seen in composites can be explained partly by the larger grain sizes pro-
the Table 3 and Figure 28(a), the pyroelectric coefficient for the duced by the quenching procedure. T h s has the efflect of increas-
PLZT/P(VDF/TrFE) is not as great as that of PTCa composites ing the 1-3 connectivity which helps both in the poling and in
and is even lower than that of the PZT-P(VDF/TrFE) composite. the enhancing the active properties. It should be noted however,
Thus, we can conclude that the resistivity matching is not a criti- that the grain size effect does not seem to explain certain fea-
cal factor in determining the final pyroelectric response and fur- tures such as a slightly lugher permittivity whch ir; found in the
thermore it appears that the permittivity mismatching is a more milled ceramic composites compared with the quenched ones.
important factor which, as explained above, can affect both the This contradicts the 1-3 connectivity argument, whch predicts
poling efficiencyand the pyroelectric coefficient once the ceram- a simultaneous rise in the permittivity for an increase in the 1-3
ic is fully poled. From the data measured for these composites it connectivity, and thus requires one to admit that the grains ob-
can be beneficial in pyroelectric terms that the ceramic that goes tained through the milling procedure are in some way altered by
into the polymer matrix should not have a very high permittiv- the mechanical forces drawn upon them during the processing
ity. The permittivity of the composite will be consequently low, stage. It is thus possible that a mixture of both hypothesis can
whilst the pyroelectric coefficient can be high thus improving its explain their different behavior.
figure of mepit.
We do not know of any published work on the effects of the
Figure 28(b) shows the direct pyroelectric coefficient for the grain size on the properties of PTCa ceramic. However, it is
same composites which were corona poled. Here again, PTCa known that for BaTi03 [101,102] the permittivity i:ncreaseswith
composites have a higher pyroelectric coefficientat room temper- decreasing grain size, reaching a peak value at 1pm and decreas-
ature than the other ceramic composites. It is 30 and 70% higher ing subsequently with any further decrease of the g a i n size. It is
than the corresponding PZT and PLZT composites respective- possible that a similar behavior arising from the considerations
ly. The temperature dependence of these composites is the same of the effects of grain boundaries, internal stress and field and
as that exhibited by the conventionally poled samples, the PTCa domain wall area, may also be expected for PTCa.
composite showing the lowest dependence whereas the PZT and
PLZT have a similar variation with temperature. Figure 29(b) shows a plot of the direct pyroelectric coefficient
in the 20 to 70°C temperature range in composites made with
It is apparent from a comparison of Figure 28(a) and (b) for the quenched PTCa powder using volume fractions of 30,50,60, and
conventional and corona poled samples respectively that these 65%. The pyro coefficient increases by 33 % when the ceram-
N

50% volume fraction composites are more efficiently poled by ic volume fraction rises from 50 to 60% (i.e.16%),wluch is well
the corona poling method than by the conventional one. This in excess of a linear increment. The higher volume fraction will
can be due to a higher poling field applied in the corona pol- increase the pyroelectric coefficient because of two combining ef-
ing than in the conventional method without serious breakdown fects namely: the a presence of a higher proportion of an electro-
problems. From the comparison of both poling methods it is active material and an increased likelihood of ceramic paths con-
seen that the improvement in the pyroelectric response in rela- necting the upper and lower electrodes.
tive terms is higher for the PZT and PLZT coinyosites than in the
PTCa which is an indication of a higher degree of poling for the Because of the restrictions due to electrical breakdown a mod-
PTCa composites when using the conventional poling method. erate poling field in the range of 15 to 25 MV/m was used in this
work. As a result, the polymer does not contribute to the over-
It must be pointed out, that it is usually assumed that the sur- all pyroelectricity of the composite, because its coercive field is in
face voltage of the sample during corona poling is the same as the 50 to 80 MV/m range. Furthermore, the annealing procedure,
728 Dias et al.: Inorganic Ceramiflolymer Ferroelectric Composite Electrets

1t
r
--nventional
""1
+ding -
50:50 vol% composites
_Q_

_e_
PLZT
PZT
PTCa

1 I
20 30

which consisted of an short-circuit aging at 90°C for 12 h, also re- In order to test the mixed connectivity model, the experimen-
duces its contribution. Thus, the polymer although of a ferroelec- tal points taken from the quenched PTCa composite were plotted
tric nature, does not contribute with its electro-active properties in the mixed connectivity cubes charts of Figure 30. In principle
to those of the composite and acts only has a high permittivity, we expect that the location of the points in the nm plane diagram
low stiffness host to the electro-active ceramic powder. to be coincident in the charts for the permittivity (Figure 13), the
piezoelectric constant d33 (Figure 24) and the pyroelectric coeffi-
cient (Figure 30).
Table 3 gives the values of the pyroelectric coefficients of the
composites at 30°C. The values of the permittivity and loss at The experimental behavior of these composites in the theo-
1 kHz and room temperature for the materials tested, are also retical charts seems to be consistent revealing an increase in 1-3
included in Table 3. From these data one can calculate the figure connectivity character and thus better electro-active properties
of merit p/e, which is roughly proportional to the responsivity for hgher volume fraction composites. T h s assumption is re-
of a detector made with that particular sample. This parameter inforced both in the permittivity and the pyroelectric coefficient
stresses the fact that it is important to have a high pyroelectric
coefficient whle keeping the permittivity low.
-
charts where the n value is 0.4 for the 30% composite rising
to z 0.55 in the 65% volume fraction composite. It is clear that
IEEE Transactions on Dielectrics and Electrical lnsulation Vol. 3 No. 5, October 1996 729

Table 3. Pyroelectric coefficient of selected composites.


__
~
__
~
-
__
ceramic %vel host E P(RT) FOM, [Ref.]
__ __ pC/m2K pC/m2K __
PZT 40epoxy SPURRS 110 40 91 0.35
BaTiOa 30rubber 17 60 92 3.5
BaTiOs 20PVDF 20 94
TGS 80PVDF 12 90 3.3 95
PT 62PVDF 54 130 2.4 96
PZT 50PVDF 90 10 0.11 97
PZT 50P(VDF/ TrFE) 118 39 0.33 99
PLZT 50P(VDF/ TrFE) 80 17 0.21 99
PTCa m 50P(VDF/ TrFE) 56 50 0.90 99
PTCa m 60P(VDF/ TrFE) 66 60 0.91
PTCa q 30 P(VDF/TrFE) 28 29 1.03
PTCa q 50 P(VDF/TrFE) 40 44 1.10
PTCa q 60 P(VDF/TrFE) 49 95 1.93
PTCa aI 65 P(VDF/ TrFE) 67 130 1.94
the m and q suffixes stand for mill Iand quenched powder respectively.

Measurements using the dynamic method [loo] have been also


performed in the quenched PTCa composites (Figure 31) as de-
scribed in Section 6.1 on the measurement techniques. One can
observe that the dynamic techruque can indeed be an accurate
method for the measurement of the pyroelectric Coefficient, even
in the presence of other sources of current such as those of depo-
larization currents. However, one has to take into account only
the portion of the oscillating current which is 90" out of phase
with the temperature in order to get a true value (of pyroelectric
activity.

1000
.pb.~.03~p(vDF.~.~.E)[6~~40.voi;ji
~.
.
...................... ...+ ...................... ; ..../.....; .................... ..;.......................
P
(U
,E
0
c
._
m
g
0

e 100
g
a2 a4 a6 n a8 Qe
Figure 30. Mixed connectivitycharts for the permittivity and py- =-.
CL

roelectric constant in PTCa-P(VDF/TrFE) composites. ............................ AFTER annealing: 4 and 5th.runs ......... .........................
.........
for a given composite similar, although not equal values, of the
n and m values are found in the permittivity and pyroelectric
10
coefficient charts. In fact one must regard these charts as a semi-
10 20 30 40 50 60
quantitative estimation of the properties of the composites and Temperature PC]
one must recognize that a multitude of effects can be present in Figure 31 . Dynamic and direct (i.e. quasi-static)pyroelectricmea-
such a material whch were not taken into consideration, notably surement of a PTCa-P(VDF/TrFE)60:40 vol%composite [loo].
the interface effects at the ceramic-polymerboundaries.
In conclusionit may be said that the mixed connectivity mod- In order to investigate the influence of the frequency of the
el seems to help to portray the general behavior of a composite poling field in the electro-active properties of the composites a
for certain conditions, but care should be taken in ensuring that sample was first polarized with a +25 MV/m dc step voltage for
the premises such as 100% poling and consistent raw material 1h after which the pyroelectric coefficient was measured. Then
properties are met. the sample was turned to the other side and a 0.05 Hz ac field
730 Dias et a1.I Inorganic Ceramicyl'olymer Ferroelectric Composite Electrets

coupling to water and a more adjusted buoyancy than the hgher


density ceramics used for hydrophones. Also, compliant materi-
als have better resistance to mechanical shock than conventional
ceramic transducers and a larger compliance also leads to h g h
damping whch is desirable in passive devices.
Biomedical acoustic imaging is an important diagnostic tool as
it provides images of the internal organs without the use of ion-
PTC:P(VDFTTrFE) [65:35vol%]
300 - izing radiation. A piezoelectric transducer generates an acoustic
pulse whch is then detected as weak echoes. The requirements
250 L poling time: 1 h
/
/
on the transducer are:
F ; poling temperature: 100°C I
1. a h g h electromechanical energy conversion efficiency,
2. a good transmission/reception of acoustic energy between
the transducer and the tissue implying an acoustic imped-
ance close to that of the human tissue,
3. low dielectric losses (tan S < O . l ) ,
4. formability into curved and other convenient shapes for
focusing and beam steering.
These requirements are met by 1-3 connectivity composites and
thus much effort is being made to this end.
Biomedical Pulse-echo annular arrays [lo31 consist of a cen-
tral disk surrounded by a number of concentric rings, acoustical
and electrically isolated. Ths enables by adjusting the time de-
lays between the elements, a variation of the focal point along
the transducer axis either on transmission or on reception. Such
a transducer thus has the advantage of the characteristics of a
composite which can easily form into complex shapes.
Hydrophones [104]. PTCa ceramic, embedded in a Stycast
epoxy polymer was used as a 1-3 connectivity composite trans-
ducer material. These materials showed no pressure dependence
to 20 MPa whle the free-field hydrophone characteristics were
constant with frequency to 6 kHz.
Piezo-rubber composites using PZT ceramic as the electro-
active material have also been reported as a transducer mate-
rial for hydrophones [74],and for single pulse transmitlreceive
units. They showed excellent hydrostatic sensitivity with d h g h
products in the 1to 5 pPa-' range although there is some pres-
sure dependence of these parameters. The single pulse unit re-
sponses showed a time response where the ringing effect was
almost completely absent in contrast with similar transducer de-
signs using PZT, PZT-polymer 1-3 composite or PVDF electro-
active materials. Other efforts have also been dedicated to 0-3
composite as an hydrophone material (1051.
Piezoelectric paints. The idea of developing a piezoelec-
tric material [81,106,107] which could be used as a paint to
cover large areas has been demonstrated to be a feasible one.
One attempt used PZT and PbTiOB having acrylic as well as
polyurethane as host materials. They showed that the piezoelec-
tric d33 coefficient could be as high as 39 pC/N for a 70% PT
ceramic/acrylic copolymer composite.
0-3 composites have also been prepared from water sus-
pensions containing PZT or co-precipitated PT ceramics, a
methacrylic copolymer and other rheological and surfactants
IEEE Transactionson Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 732

agents [107]. The filled emulsion was spread onto an appropriate [5] B. Jimenez, J. Mendiola, C. Alemany, L. d. Olmo, L. Pardo, E. Mau-
surface, dried and the resulting film electroded and poled. The rer, M. L. Calzada, J. d. Frutos, A. M. Gonzalez and M. C. Fandiii
0, "Contributions to the Knowledge of Calcium Modified Lead Ti-
highest piezoelectric d33 coefficients were found to be 35 pC/N
tanate Ceramics", Ferroelectrics, Vol. 87, pp. 97-108,1988.
for a 70% PT loaded composite whose permittivity was 138.
[6] A. Lovinger, "Ferroelectric Polymers", Science, Vol. 220, pp. 1115-
Vibration and noise control. Piezoelectric paint has been de- 1121,1983.
veloped as a sensor material for built-in vibration sensor appli- [7] H. Takeuchi, S. Jyomura and C. Nakaya, "New Piezoelectric Ma-
cations 11061. A composite made of PZT and an epoxy resin was terials for Ultrasonic Transducers", Japan. J. Appll. Phys., Vol. 24
shown to be a promising candidate for the detection of the nat- (Suppl. 2), pp. 36-40,1985.
ural frequencies and mode shape of structural materials [106] as [8] R. E. Newnham, and G. R. Ruschau, "ElectromechanicalProperties
well as for acoustic emission applications o n laminate structures of Smart Materials", J. Intelligent Mater. Systems and Structures,
[lo81 (see Figure 33). Vol. 4, pp. 289-294,1993.
[9] R. E. Newnham, D.P. Skinner and L. E. Cross, "Connectivity and
Piezoelectric-pyroelectricComposites", Mat. Res. Eiull., Vol. 13, pp.
>-- composite transducer 525-536,1978.
[lo] D. P. Skinner, R. E. Newnham and L. E. Cross, "Flexible Compos-
ites Transducers", Mat. Res. Bull., Vol. 13, pp. 599-607,1978,
[Ill T. Ota, J. Takahashi and I. Yamai, "Effect of Microstructure on the
' Prepreg plies Dielectric Property of Ceramics", in Electronic Ceramic Materials,
edited by J. Nowotny, (Trans-Tech Pub. Ltd., Zurich, 1992), pp.
Figure 33. Exploded view of a laminate structure containing a
composite transducer [108]. 185-246,1992.
[12] Q. Y. Jiang, S. B. Krupanidhi and L. E. Cross, "Effects of the Lapped
Active control of sound. Active control of sound using a 0-3 Surface Layer and Surface Stress on the Dielectric Properties of
composite of a hgh g h value has been proposed by Salloway and PLZT Ceramics", Ceramic Dielectrics: composition, processing
coworkers[l09]. The device fabricated was built in a test tube and properties, Vol. 8, pp. 375-383,1990.
around a fast feedback control unit which could either reduce [13] J. A. Chilton, "Electroactive Composites", GEC Review, Vol. 6(3),
the energy in the active noise reduction mode by a factor of 0.5 pp. 156-164,1991.
or increase to 0.17 i n the active noise enhancement mode, i n the 1141 W. A. Smith, "Modeling 1-3 Composite Piezoelectrics: Hydrostatic
frequency range from 1.5 to 8.5 kHz, i n a test tube. With an air Response", IEEE Ultrason. Ferroelec. Freq. Contr., Vol. 40(1), pp.
backed noise control device, the echo reduction could be further 41-48,1993.
reduced by up to 12 dB from 1to 8 kHz. [15] R. E. Newnham and S. E. Trolier-McKinstry, "Structure-property
Relationships in Ferroic Nanocomposites", Ceramic Dielectrics:
Acceleration sensor. PZT-silicone rubber 1-3 composite have composition, processing and properties, Vol. 8, pp. 235-252,1990,
been used [29] as acceleration sensors wluch had twice as m u c h
[16] R. Igreja, M. P. Wenger, C. J. Dias, D.K. Das-Gupta and J. N. Marat-
output voltage as that of a PZT transducer of a similar design. Mendes, 'Tyro- and Piezoelectricity in Sol-gel Derived Ceram-
T h s was observed in a wide range of frequency, acceleration and ic/ polymer Composites", Intern Symposium on Electrets, Paris,
added load. IEEE, pp. 725-730,1994,
The direct and inverse piezoelectric effects enable such elec- [17] R. E. Newnham, "Ferroelectric Composites", Japan. J. Appl. Phys.,
trets to be classified as intelligent (or smart) materials. In this 24 (suPP1. 2)t PP. l6-l7? 1985.
respect the ferroelectric ceramic/polymer electrets are attractive [I81 K. H i k h K. Yamada, M. Nishioka and M. Ono, "'Effectof l"-ous
for the 21th century in micro- and nanoelectroficappli- Structure to Piezoelectric Properties of PZT Ceramics", Japan. J.
Appl. Phys., Vol. 22 (Suppl. 2), pp. 64-66,1983.
cations.
[19] H. G. Lee and H. G. Kim, "Influence of Microstructure on the Di-
electric and Piezoelectric Properties of Lead Zirconate Titanate-
REFERENCES polymer Composites", J. Am. Ceram. Soc., Vol. 74.(6), pp. 938-942,
B. Jaffe, R.S. Roth and S.Marzullo, "Properties of Piezoelectric Ce- 1989.
ramics in the Solid Solution Series Lead Titanate-lead Oxide: Tin [20] A. J. Moulson and J. M. Herbert, Electroceramics:materials, prop-
Oxide and Lead Titanate-lead Hafnate", J. Res. National Bureau of erties and applications, 1990, Cambridge: Chapman & Hall.
Standards, Vol. 55, pp. 239-254,1955. [21] H. Banno and S.Saito, "Piezoelectric and Dielectric Properties of
S.L. Swartz, "Topics in ElectronicsCeramics", IEEE Trans. on Elec- Synthetic Rubber and PbTi03 or PZT", Japan. J. .4ppl. Phys., Vol.
tr. Insul., Vol. 25, pp. 935-987,1990. 22 (SUppl. 2), pp. 67-69,1983.
I. Ueda, and S.Ikegami, "Piezoelectric Properties of Modified [22] Y. Xu and J. D.MacKenzie, "Ferroelectric Thin Films Prepared
PbTiOa Ceramics", J. Japan. Appl. Phys., Vol. 7, pp. 236-242,1968, by Sol-gel Processing", Integrated Ferroelectrics, Vol. 1,pp. 17-42,
Y.Yamasluta, K. Yokoyama, H. Honda and T. Takahashi, "(Pb, 1992.
Ca)((Col/2W1 p),Ti)OS Piezoelectric Ceramics and their Applica- [23] K. A. Klicker, Piezoelectric Composites with 3-2 Connectivity for
tions", Proc. of the FMA-3, Japan. J. Appl. Phys., Vol. 20 (Suppl. 4), Transducer Applications, PhD thesis, Pennsylvania State Universi-
pp. 183-187,1981. ty, 1980.
732 Dias et al.: Inorganic Ceramiflolymer Fenoelectric Composite Electrets

[24] H. P. Savakus, K. A. Klicker and R. E. Newham, ”PZT-epoxy [45] J. Hossack, Modeling Techmques for 1-3 Composite Transducers,
Piezoelectric Transducers: a Simplified Fabrication Process”, University of Strathclyde, 1990.
Mater. Res. Bull., Vol. 16, pp. 677-680,1980, [46] C. -W. Nan, ”Product Property between Thermal Expansion and
[25] B. A. Auld and Y.Wang, ”Acoustic Wave Vibrations in Periodic Piezoelectricity in Piezoelectric Composites: Pyroelectricity”, J.
Composites Plates”, IEEE Ultrasonics Symposium, Texas (USA), Mater. Sci. Lett., Vol. 13, pp. 1392-1394,1994.
pp. 528-532,1984. [47]L. A. Pauer, in IEEE Intern. Cord. Rec., 1973.
[26] W. A. Smith, ”The Role of Piezocomposites in Ultrasonic Transduc- [48] H. Banno, ”Theoretical Equations for the Dielectric and Piezo-
ers”, IEEE Ultrasonics Symposium, pp. 755-766,1989. electric Properties of Ferroelectric Composites Based on Modified
[27] R. J. Card, ”Preparation of Hollow Ceramic Fibers”, Advanced Ce- Cubes Model”, Japan.J. Appl. Phys., Vol. 24(Suppl. 2), pp. 445-447,
ramic Materials, Vol. 3, pp. 29-31,1988. 1985.
[28] W. Wersing, ”Composite Piezoelectricfor Ultrasonic Transducers”, I491 H. -G. Lee and H.-G. Kim, “Ceramic Particle Size Dependence of
pp. 212-223, IEEE 6th Int. Symp. on Appl. Ferroelectrics (ISAF‘86), Dielectric and Piezoelectric Properties of Piezoelectric Ceramic-
Pennsylvania (USA), 1986. polymer Composites”, J. Appl. Phys., Vol. 67(4), pp. 2024-2028,
[29] Y. Ohara, M. Miyayama, K. Koumoto and H. Yanagida, ”PZT- 1990.
polymer Piezoelectric Composites: a Design for an Acceleration [SO] G. M. Garner, N. M. Shorrocks, R. W. Whatmore, M. T. Goosey,
Sensor”, Sensors and Actuators, Vol. A 36, pp. 121-126,1993. P. Seth and E W.Ainger, ”0-3 Piezoelectric Composites for Large
[30] Q. M. Zhang, H. Wane and L. E. Cross, “Piezoelectric Tubes and Area Hydrophones”, Ferroelectrics,Vol. 93, pp. 169-176,1989.
Tubular Composites for Actuator and Sensor Applications”, J. [51] C. J. Dias and D. K. Das-Gupta, ”Piezo- and Pyroelectricity in
Mater. Sci., Vol. 28, pp. 3962-3968,1993. Ferroelectric Ceramic-polymer Composites”, in Ferroelectric poly-
[31] T. Yamada, T. Ueda and T.Kitayama, ”Piezoelectricity of a High mers and ceramic-po1,vmer composites, pp. 217-248, edited by D.
Content Lead Zirconate Titanate/polymer Composite”, J. Phys. K. Das-Gupta, Trans-Tech Publications, 1994.
Phys., Vol. 53(4), pp. 4328-4332,1982. [52] C. J. Dias, Ferroelectric Composites for Pyro- and Piezoelectric Ap-
[32] D. K. Das-Gupta, ”Pyroelectricity in Polymers”, Ferroelectrics,Vol. plications, University of Wales (Bangor-SEECS),1994.
118, pp. 165-189,1991,
[53] G. Sa-Gong, A. Safari and R. E. Newham, ”Poling Study of
[33j U. Wang, W. Zhong and I? Zhang, ”Pyroelectric Properties of PbTiO3-polymer Composites”, in IEEE 6th Int. Symp. on Appl.
Ferroelectric-polymer Composites”,-J. Appl. Phys., vol. 74(1), pp. Ferroelectrics (ISAF86), pp. 281-284, Pennsylvania (USA), 1986.
512-524,1993.
[54] T. Furukawa and G. E. Johnson, “Dielectric Relaxations in a
[34] T. Furukawa, K. Fujino and E. Fukada, “ElectromechanicalProper-
Copolymer of Vinylidene Fluoride and Trifluoroethylene”,J. Appl.
ties in the Composites of Epoxy Resin and PZT Ceramics”, Japan. Phys., Vol. 52(2), pp. 940-943,1981.
J. Appl. Phys., Vol. 15(11), pp. 2119-2129,1976.
[55] K. Y.Hashimoto and M.Yamaguchi, ”Elastic, Piezoelectric and Di-
[35] T. Furukawa, K. Ishida and E. Fukada, “Piezoelectric Properties in
electric Properties of Composite Materials”, in IEEE Ultrasonics
the Composite Systems of Polymer and PZT Ceramics”, J. Appl.
Symposium, pp. 697-702,1986,
Phys., Vol. 50(7),pp. 4904-4912,1979.
[56] W. Cao, Q. M. Zhang and L. E. Cross, ”Theoretical Study on the
[36] L. Pardo, J. Mendiola and C. Alemany, ”Theoretical Treatment of
Ferroelectric Composites using Monte Carlo Calculations”, J. Appl. Static Performance of Piezoelectric Ceramic/ polymer Composites
with 2-2 Connectivity”, IEEE Ultrason. Ferroelec. Freq. Contr., Vol.
Phys., Vol. 64(10), pp. 5092-5097,1988.
40(2), pp. 103-109,1993.
[37] H. Zewdie and E Brouers, ”Theory of Ferroelectric Polymer-
[57] W. Cao, Q. M. Zhang and L. E. Cross, ”Theoretical Study on the
ceramic Composites”, J. Appl. Phys., Vol. 68(2), pp. 713-718,1990.
Static Performance of Piezoelectric Ceramic-polymer Composites
[38] W.A. Smith, ”Modeling 1-3 Composite Piezoelectrics: Thickness-
with 1-3 Connectivity”, J. Appl. Phys., Vol. 72(12), pp. 5814-5821,
mode Oscillations”, IEEE Ultrason. Ferroelec. Freq. Contr., Vol.
1992.
38(1),pp. 40-47,1991.
[39] M. Takeuchi, Y. Miyamoto and H. Nagasaka, ”EffectivePermittivi- (581 E Brouers, ”Percolation Threshold and Conductivity in Metal-
ty of BaTi03-PVDFSystems”, Japan. J. Appl. Phys., Vol. 24 (Suppl. insulator Composite Mean-field Theories”, J. Phys. C, Vol. C 19,
2), pp. 451-453,1985. pp. 7183-7193,1986.
[40] H.L. W.Chan and I. L. Guy, ”PiezoelectricCeramic/polymer Com- [59] D. A. G. Bruggemann, Annal. Phys. LPZ, Vol. 24, pp. 636,1935.
posites for High Frequency Applications”, in Ferroelectric poly- [60] N. Jayasundere and B. V. Smith, ”Permittivity for Binary Piezo-
mers and ceramic-polymer composites, pp. 275-299, edited by D. electric 0-3 Composites”, J. Appl. Phys., Vol. 73(5),pp. 2462-2466,
K. Das-Gupta, Trans-Tech Pub., Switzerland, 1994. 1993.
[41] L. Li and N.R. Sotos, ”Improving Hydrostatic Performance of 1-3 1611 D. I<. Das-Gupta and M. J. Abdullali, ”Electroactive Properties of
Piezocomposites”, J. Appl. Phys., Vol. 77(9), pp. 4595-4603,1995. Polymer-ceramic Composites”, Ferroelectrics, Vol. 87, pp. 213-228,
[42] B.-C.Shin, ”Computation of Hydrostatic Piezoelectric Coefficients 1988.
for 1-3 Composites by the Finite-element Method”, Sensors and [62] D. Sinha, C. Muralidar and P. K. C. Pillai, Proc. 2nd Intern. C o d .
Actuators, Vol. 40, pp. 191-194,1994. on Conduction and Breakdown in Solid Dielectrics, Erlangen, IEEE
[43] Y.Beneviste, ”Exact Results Concerning the Local Fields and Effec- Service Center, p. 227, Piscataway NY, 1986.
tive Properties in Piezoelectric Composites”, Trans. of ASME, Vol. [63] J. Wolak, ”Dielectric Behavior of 0-3 Type Piezoelectric Compos-
116, pp. 260-267,1994. ites”, IEEE Trans. Electr. Insul., Vol. 28(1), pp. 116-121,1993.
[44] J. A. Hossack and R. L. Bedi, “Design of Composite Piezoelec- 1641 J. Malecki and B. Hilczer, ”Dielectric Behavior of Polymer and
tric Transducers“, in Ferroelectric polymers and ceramic-polymer composites”, in Ferroelectricpolymers and ceramic-polymer com-
composites, pp. 301-322, edited by D.K. Das-Gupta, Trans-Tech posites, pp. 181-216, edited by D. K. Das-Gupta, Trans-Tech Publi-
Pub., Switzerland, 1994. cations, 1994.
IEEE Transactions on Dielectrics and Electrical Insulation Vol. 3 No. 5, October 1996 733

1651 L. Brown and D. L. Carlson, "Ultrasound Transducer Models for [84]D. Waller and A. Safari, "Corona Poling of PZT Ceramics and Flex-
Piezoelectric Polymer Films", IEEE Transactions on Ultrasonics ible Piezoelectric Composites", Ferroelectrics, Vol. 87, pp. 187-95,
Ferroelectricsand Frequency Control, Vol. 36(3),pp. 313-318,1989. 1988.
[66] L. N.Bui, H. J. Shaw and L. T. Zitelli, "Study of Acoustic Wave [85] J. M. Pelmore, "Acoustic Impedance of Compo:site Materials",
Resonance in Piezoelectric PVF2 Film", Vol. SU-24(5), pp. 331-336, Acoust. Lett., Vol. 3(3),pp. 65-68,1979.
1977.
[86] M. G. Crewe, T. R. Gururaja, T. R. Shrout and R. Newnham,
[67] H. Ohigashi, "Ultrasonic Transducers in the Megahertz Range", in "Acoustic Properties of Particle/polymer Composites for Ultra-
The applications of ferroelectric polymers, pp. 237-273, edited by sonics Transducer Backing Applications", IEEE Trans. Ultrason.
T. T. Wang, Blackie, Glasgow, 1988. Ferroelec. Freq. Control, Vol. 37(6), pp. 506-514,1990.
[68] M. Platte, "PVDF Ultrasonic Transducers for Non-destructive Test- [87] G. S. Kumar and G. Prasad, "Piezoelectric Relaxation in Polymer
ing", Ferroelectrics,Vol. 115, pp. 229-246,1991.
and Ferroelectric Composites", J. Mater. Sci., Vol.. 28, pp. 2545-
[69] C. Dias and D.K. Das-Gupta, "Piezoelectric Properties of 0-3 2550,1993.
Ceramic-polar Polymer Composites", in Materials Research Soci-
ety, Symposium Proc., Vol. 276, pp. 25-29, Smart Materials Fabri- [88] R. L. Byer and C. B. Roundy, "Pyroelectric Coefficient Direct Mea-
cation and Materials for Micro-Electro-Mechancial Systems, 1992. surement Technique and Application to a ns Response Time Detec-
tor", Ferroelectrics,Vol. 3, pp. 333-38,1972.
[70] C. J. Dias and D.K. Das-Gupta, "Polymer/ceramic Composites for
Piezoelectric Sensors", Sensors and Actuators, Vol. A 37-38, pp. [89] T. Furukawa and T. T. Wang, "Measurements and Properties of Fer-
343-347,1993. roelectric Properties of FerroelectricPolymers", in 'The applications
[71] T. R. Gururaja, W. A. Schulze, L. E. Cross, R. E. Newnham, B. A. offerroelectricpolymers, pp. 6-20, edited by T. T. Wang, J. M. Her-
Auld and a. Y. J. Wang, "Piezoelectric Composite Materials for U1- bert, and A. M. Glass, Blackie & Son, Glasgow, London, 1988).
trasonic Transducer Applications. Part I: Resonant Modes of Vibra- [90] L. E. Garn and E. J. Sharp, "Use of Low-frequency Sinusoidal
tion of PZT Rod-polymer Composites", IEEE Trans. Son. Ultrason., Temperature Waves to Separate Pyroelectric Currents from Non-
Vol. SU-32(4), pp. 481-498,1985. pyroelectric Currents. Part I. Theory", J. Appl. Phys., Vol. 53(12),
[72] R. Ting, "The Hydroacoustic Behavior of Piezoelectric Composite pp. 8974-8979,1982.
Materials", Ferroelectrics,Vol. 102, pp. 215-224,1990. [91] A. S. Bhalla, R. E. Newnham, L. E. Cross and W. A. Schulze, "Pyro-
[73] F. R. M. Espinosa, V. Pavia, J. A. Gallego-Juarez and M. Pappalardo, electric PZT-polymer Composites", Ferroelectrics,Vol. 33, pp. 139-
"Fractured Piezoelectric Ceramics for Broadband Ultrasonic Com- 146,1981.
posite Transducers", in IEEE Ultrasonics Symposium, pp. 691-696, [92] M. Amin, L. S. Balloomal, K. A. Darwish, H. Osman and B. Kamal,
1986. "Pyroelectricity in Rubber Composite Films", Ferroelectrics, Vol.
[74]H. Banno, K. Ogura, H. Sobue and K. Ohya, "Piezoelectric and 81, pp. 381-386,1988.
Acoustic Properties of PiezoelectricFlexible Composites", Japan. J. [93] M. Amin, H. Osman, L. Balloomal, K. A. Darwish and B. Kamal,
Appl. Phys., Vol. 26 (SUPPI.1), pp. 153-155,1987. "Electrical Properties of Acrylonitrile-butadiene Rubber-barium
[75] T. R. Gururaja, Q. C. Xu, A. R. Ramachandran, A. Halliyal and R. E. Titanate Composites", Ferroelectrics, Vol. 81, pp. 3887-392,1988.
Newnham, "Preparation and Piezoelectric Properties of Fired 0-3
[94] A. K. Tripathi, T. C. Goel and P. K. C. Pillai, "IPyroelectric and
Composites", in IEEE Ultrasonics Symposium, pp. 703-708,1986,
Piezoelectric Properties of Sol-gel Derived BaTiOa Polymer Com-
[76] K. H. Han, R. E. Riman and A. Safari,"(Pbn 5 , Bin s)(Tio 5 , Fen 5)03 posites", pp. 415-420,7th Int. symp. on Electrets, Berlin, IEEE Di-
Powder Prepared by Chemically Precipitated Method for 0-3 Ce- electrics and Electr. Insul. Soc., 1991.
ramic/polymer Composites", Ceramic Dielectrics : composition,
processing and properties, Vol. 8, pp. 227-232,1990. [95] C. Fang, M. Wang and H. Zhou, "Pyroelectric Properties of a
New Composite Material PVDF-TGS Film", pp.507-511,7th Intern.
[77]K. Han, A. Safari and R. E. Riman, "Colloidal Processing for Im- Symp. on Electrets, Berlin, IEEE Dielectrics and 13lectricalInsula-
proved Piezoelectric Properties of Flexible 0-3 Ceramic-polymer
tion Society, 1991.
Composites", J. Am. Ceram. Soc., Vol. 74(7), pp. 1699-1702,1991.
[78] L. Pardo, J. Mendiola and C. Alemany, "Theoretical Study of Fer- [96] H. Yamazaki and T. Kitayama, "Pyroelectric Properties of Polymer-
roelectric Composites from Ca-modified Lead Titanate Ceramics", ferroelectric Composites", Ferroelectrics, Vol. 33, pp. 147-153,
Ferroelectrics, Vol. 93, pp. 183-188,1989. 1981.
[79] J. A. Chilton, G. M. Garner, R. W. Whatmore and F. W. Ainger, "0-3 [97] M. J. Abdullah, A Study of Electro-active Prcperties of Poly-
Composite Sensitivity", Ferroelectrics,Vol. 109, pp. 217-222,1990, mer/ceramic Composites, University of Wales - 'SEECS (Bangor),
[80] A. A. Shaulov, W. A. Smith and R. Ting, "Modified-lead- 1989.
titanate/ polymer Composites for Hydrophone Applications", Fer- [98] M. J. Abdullah and D. K. Das-Gupta, "Electrical Froperties of Ce-
roelectrics,Vol. 93, pp. 177-182,1989, ramic/polymer Composites", IEEE Trans. Electr. Insul., Vol. 25(3),
[Sl] K. A. Harmer, A. Safari, R. E. Newnham and J. Runt, "Thin Film 0-3 pp. 605-610,1990.
Polymer/ piezoelectric Ceramic Composites: Piezoelectric Paints", [99] C. Dias and D.K. Das-Gupta, "Ferroelectric Ceramic/ polar Poly-
Ferroelectrics,Vol. 100, pp. 255-260,1989, mer Composite Films for Pyroelectric Sensors", p ! ~393-396,
. in 6th
[82] W. J. Merz, "Switching Time in FerroelectricBaTiO3 and its Depen- Int. Cod. on Dielectric Materials, Measurements and Applications,
dence on Crystal Tluckness", J. Appl. Phys., Vol. 27(8),pp. 938-943, DMMA6, Manchester, IEE, 1992.
1956. [loo] C. Dias, M. Simon, R. Quad and D.K. Das-Gupt,a,"Measurement
[83] J. B. Ngoma, J. Y. Cavaille, J. Paletto and J. Perez, "Dielectric of the Pyroelectric Coefficient in Composites using a Temperature-
and Piezoelectric Properties of Copolymer-ferroelectric Compos- modulated Excitation", J. Phys. D: Appl. Phys., Vol. 26, pp. 106-
ite'', Ferroelectrics,Vol. 109, pp. 205-210,1990. 110,1993.
734 Dias et al.: Inorganic CerarnicyPolymer Ferroelectric Composite Elecfrefs

[lo11 A. A. Shaikh, R. W. Vest and G. M. Vest, "Dielectric Proper- J. Mater. Sci., Vol. 28, pp, 1667-1672,1993,
ties of Ultrafine Grained BaTiOs", pp. 126-129, in IEEE 1986 6th [107] K. A. Klein, A. Safari, R. E. N e w h a m and J. Runt, "Composite
Int. Symposium on Applications of Ferroelectrics, Pennsylvania Piezoelectric Paints", pp. 285-287, in IEEE 6th Int. Symp. on Appl.
(USA), 1986. Ferroelectrics (ISAF'86), Pennsylvania (USA), 1986.
[102] A. Shaikh, R. W. Vest and G. M. Vest, "Dielectric Properties of [lo81 M. P. Wenger, P.Blanas, R. J. Shuford and D.K. Das-Gupta,
Ultrafine Grined BaTiOj", IEEE Trans. 'Ciltrasonics, Ferroelectrics "Acoustic Emission Signal Detection by Ceramic/polymer Com-
and Freq. Control, Vol. 36(4), pp. 407-413,1989. posite Piezoelectric Embedded in Glass-epoxy Laminates", in 4th
[103] W. A. Smith and A. A. Shaulov, "Composite Piezoelectrics: Basic International Conference on Polymer characterization, Univ. of
Research to a Practical Device", Ferroelectrics,Vol. 87, pp. 307-320, North Texas, Denton (USA), 10-14 January 1996.
1988. [lo91 A. J. Salloway, R. C. Twiney R. W. Whatmore and R. Lane, "The
[104] R. Y.Ting, A. A. Shaulov and W. A. Smith, "Piezoelectric Prop- Active Control of Sound Reflection/ transmission Coefficients us-
erties of 1-3 Composites of a Calcium Modified Lead Titanate in ing Piezoelectric Composite Materials", Ferroelectrics,Vol. 134, pp.
Epoxy Resins", pp. 707-710, in IEEE Ultrasonics Symposium, 1990. 89-91,1992.
[lo51 R. l? Tandon, D.R. Chaubey, R. Singh and N.C. Soni, "Dielec-
tric, Piezoelectric and Acoustical Properties of High Performance Manuscript was received on 15 December 1995, in final form 3 Septem-
Piezorubber Composite Hydrophone", J. Mater. Sci. Lett., Vol. 12, ber 1996.
pp. 1182-1184,1993.
[lo61 S. Egusa and N. Iwasawa, "Piezoelectric Paints: Preparation and
Applications as Built-in Vibration Sensors of Structural Materials",

You might also like