Preparation, structural characterization and ferroelectric properties of compensated Te-doped n = 2 Aurivillius oxides ceramics

Article abstract of DOI:10.1016/S0022-3697(00)00017-2

New solid solutions Bi_2-xTe_xSrNb_2-xB_xO₉ (B = Zr, Hf; 0≤x≤0.5) and Bi_2-yTe_ySr_1-yA_yNb₂O₉ (A = K; 0≤y≤0.25) have been synthesized and characterized by powder X-ray diffraction methods, thermal analysis and dielectric measurements. Evolution of the unit-cell parameters at room temperature shows that doping the A or B perovskite sites dominates the size of (001) plane in the structure, whereas the c value is largely influenced by the cations which occupy the bismuth position in its layer. Thermogravimetric studies show that these materials are stable up to 800 °C for x, y = 0.25 and 750 °C for x = 0.5. Differential thermal analysis curves of x = 0.25 compounds exhibit one exothermic peak at 400 and 500 °C for zirconium and hafnium phases, respectively. This effect is related to the expansion of their network, without changing the symmetry or space group. All compositions are ferroelectric at room temperature, and exhibit a non-zero d₃₃ value after poling. Measurements of the dielectric permittivity as a function of the temperature reveal signs of at least two possible phase transitions. The disappearance of the induced spontaneous polarization when heating samples above the lower temperature anomaly in the dielectric permittivity clarifies the T_c value of the compounds. Assuming they are displacive-type ferroelectrics, it can be shown that the variation of T_c with the composition is an effect of the change of the unit cell size. This result offers the possibility of tailoring a ferroelectric material with the desired thermal working range.

Full text of DOI:10.1016/S0022-3697(00)00017-2

Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
www.elsevier.nl/locate/jpcs
Preparation, structural characterization and ferroelectric properties
of compensated Te-doped n ˆ 2 Aurivillius oxides ceramics
a,
b
b
b
*
P. Dur a´ n-Mart ´ı n , B. Jim e´ nez , P. Mill a´ n , A. Castro
a
Ceramics Laboratory, Materials Department, EPFL Swiss Federal Institute of Technology, CH-1015 Lausanne, Switzerland
b
Instituto de Ciencia de Materiales de Madrid, CSIC, Cantoblanco, 28049 Madrid, Spain
Received 10 August 1999; accepted 7 January 2000
Abstract
New solid solutions Bi2ϪxTe SrNb B O (B ˆ Zr, Hf; 0 Յ x Յ 0:5† and Bi Te Sr A Nb O (A ˆ K; 0 Յ y Յ 0:25†
x
2Ϫx
x
9
2Ϫy
y
1Ϫy
y
2
9
have been synthesized and characterized by powder X-ray diffraction methods, thermal analysis and dielectric measurements.
Evolution of the unit-cell parameters at room temperature shows that doping the A or B perovskite sites dominates the size of
(001) plane in the structure, whereas the c value is largely influenced by the cations which occupy the bismuth position in its
layer. Thermogravimetric studies show that these materials are stable up to 800ЊC for x, y ˆ 0:25 and 750ЊC for x ˆ 0:5:
Differential thermal analysis curves of x ˆ 0:25 compounds exhibit one exothermic peak at 400 and 500ЊC for zirconium and
hafnium phases, respectively. This effect is related to the expansion of their network, without changing the symmetry or space
group. All compositions are ferroelectric at room temperature, and exhibit a non-zero d33 value after poling. Measurements of
the dielectric permittivity as a function of the temperature reveal signs of at least two possible phase transitions. The disap-
pearance of the induced spontaneous polarization when heating samples above the lower temperature anomaly in the dielectric
permittivity clarifies the T value of the compounds. Assuming they are displacive-type ferroelectrics, it can be shown that the
c
c
variation of T with the composition is an effect of the change of the unit cell size. This result offers the possibility of tailoring a
ferroelectric material with the desired thermal working range. ᭧ 2000 Elsevier Science Ltd. All rights reserved.
Keywords: A. Oxides; D. Ferroelectricity; D. Phase transitions
1. Introduction
The double-sided Bi O sheets are composed of square
2 2
pyramidal BiO
carried out in our laboratory have demonstrated that Bi
4
groups sharing their basal edges. Studies
3ϩ
In recent years there has been an increasing interest in the
family of bismuth layered Aurivillius oxides because of
their exciting electrical properties and promise as ferro-
piezoelectric materials for high temperature applications
in this layer can be replaced by other p-cations which posses
a stereoactive lone pair of electrons, allowing preservation
of the basal network [5–7]. Moreover, it was shown that it is
possible to modify the charge balance compensation
[1–4].
2
ϩ
2Ϫ
The structure of these compounds consist of a stacking of
perovskite-like units of nominal composition
between layers, i.e. [Bi
2
O
2
]
2 7
[SrNb O ] , by means of
n
different cationic doping into both the perovskite and
bismuth layers [8–11]. Electrical characterization of these
materials, carried out on highly dense ceramics sintered
using uniaxial hot-pressing technique, has shown that
[
AnϪ1B O ] between [Bi O ] layers. The A-site of the
n 3nϩ1 2 2
perovskite slab can be occupied by several cations accepting
a cubooctahedral coordination, such as Bi, Ba, Sr, K, Ca,
Na, Pb or a combination of them, while smaller ions, such
as Ti, Nb, Ta, Fe, W, Mo and Ga, can occupy the octahedral
B-site.
3
ϩ
increasing the amount of Bi substitution results in higher
ferroelectric transition temperature values [12–14].
In view of the relationship between composition and
properties observed in these Aurivillius-substituted materi-
als, we have undertaken the study of new
*
Corresponding author. Tel.: ϩ 41-21-693-4952; fax: ϩ 41-21-
93-5810.
E-mail address: pedro.duran@epfl.ch (P. Dur a´ n-Mart ´ı n).
6
Bi2ϪxTe
x
SrNb2Ϫx
B
x
O
9
(B ˆ Zr, Hf) and Bi2ϪyTe
y
S-
r
1Ϫy
A
y
Nb
2 9
O
(A ˆ K) bismuth layered solid solutions, in
0
022-3697/00/$ - see front matter ᭧ 2000 Elsevier Science Ltd. All rights reserved.
PII: S0022-3697(00)00017-2

1
424
P. Dur a´ n-Mart ´ı n et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
order to attempt the improvement of their ferroelectric
characteristics. The structural characterization by X-ray
diffraction methods, thermal behavior and electrical proper-
ties of these new materials are reported, and the results are
compared with those previously reported of B ˆ Ti and
A ˆ Na substituted phases.
Redcroft STA-781. Approximately 50 mg of the sample
was employed in each measurement with standard platinum
crucibles, using Al O as the reference material. DTA and
2 3
TG curves were collected simultaneously at a heating rate of
10Њ/min in air.
Dielectric measurements as a function of the temperature
from 20 up to 800ЊC were performed on the different
samples, previously electroded on their largest faces with
platinum paint, mounted in a alumina holder and then
placed into a high temperature furnace. A Hewlett Packard
Impedance Analyzer 4194A was used for AC measure-
ments, in the frequency range from 100 Hz to 15 MHz.
The ferroelectric character of these compositions at room
temperature was confirmed by poling the samples. The
poling process consisted of introducing the samples into a
silicone oil bath and heating them up to 200ЊC. Then a
100 kV/cm DC electric field was applied to the sample at
this temperature and keeping it applied during cooling to
room temperature. The resultant polarization yield was
2. Experimental procedure
Three series of polycrystalline
samples
of
Bi2ϪxTe
Bi2ϪyTe
x
y
SrNb2Ϫx (B ˆ Zr, Hf) with 0 Յ 0:5 and
B
x
O
9
Sr1ϪyK Nb O with 0 Յ y Յ 0:25; were prepared
y 2 9
by solid state reaction from appropriate amounts of stoichio-
metric mixtures of Bi O , TeO , SrCO , Nb O , ZrO , HfO
2
2
3
2
3
2
5
2
2 3
and K CO . The reactants were thoroughly mixed in an
agate mortar. These mixtures were processed in alumina
crucibles, with the temperature sequences: 600, 700, 750,
8
00, 810 and 820ЊC for 24 h and 820ЊC for 12 h. After each
thermal treatment, the samples were quenched in air, and
then weighed, reground and examined by X-ray powder
diffraction methods.
followed by measuring the piezoelectric coefficient, d33,
with a Berlicourt-meter, at room temperature.
Ceramics of these compositions were obtained from the
synthesized powders by uniaxial hot-pressing techniques.
Sintering was performed at temperatures between 800 and
3
. Results and discussion
As reported elsewhere [8,16], the increase of the positive
8
50ЊC, depending upon the composition, and with applied
4
ϩ
3ϩ
charge generated by the substitution of Te by Bi , into its
layer, must be balanced by increasing the negative charge of
the perovskite sheets by means of the simultaneous substitu-
tion of A or B cations with appropriate oxidation state and
ionic radius [17]. In this way, single phases
pressures of 270 Pa, during one hour. Resultant ceramic
materials presented a bulk density of 98% compared to
theoretical.
Room temperature X-ray diffraction data were collected
on a Siemens Kristalloflex 810 generator, a D-501 goni-
ometer, with a graphite monochromator, between 5 and
(2ϩx)ϩ
(2ϩx)Ϫ
[Bi2ϪxTe O ]
[SrNb B O ]
(B ˆ Zr, Hf), 0 Յ
x
2
2Ϫx
x
7
(
2ϩy)ϩ
(2ϩy)Ϫ
x Յ 0:5; and [Bi2ϪyTe
y
2
O ]
[Sr1Ϫy
K
y
Nb
2
O
7
]
, 0 Յ
1
10Њ (2u) with increments of 0.05Њ (2u) and counting time
x Յ 0:25; have been isolated. In contrast with the results
obtained for the equivalent Na doped compounds [8], for
y Ͼ 0:25 not single phases can be isolated but a mixture of
two different n ˆ 2 Aurivillius-type phases, probably due to
the presence of the not substituted Bi SrNb O compound
of 4 s per step. A Phillips PW 1310 diffractometer was
employed to record the XRD patterns at increasing tempera-
tures, between 5 and 66Њ (2u) with increments of 0.02Њ (2u)
and counting time of 2 s per step. In both experiments, Cu
2
2
9
ꢀ
Ka radiation ꢀl ˆ 1:5418 A† was employed. The X-ray
and the lower substituted y ˆ 0:25 one. This fact can be
diffraction profiles were analyzed by a pattern matching
procedure using the Fullprof program [15]. The line shape
of the diffraction peaks was generated with a pseudo-Voigt
function. Fifteen different parameters were refined, includ-
ing six background coefficients, zero-point, half-width,
pseudo-Voigt and asymmetry parameters for the peak
shape, as well as the unit-cell parameters. This refinement
was performed for a variety of symmetries and space groups
and the best agreement was obtained for the B2cb (No. 41),
as it was previously reported for other similar compounds
understood taking into account the ionic radii of the different
2ϩ
ϩ
ꢀ
ꢀ
A-cations [17] ꢀrꢀSr † ˆ 1:44 A; rꢀNa † ˆ 1:39 A;
ϩ
ϩ
ꢀ
ϾrꢀK † ˆ 1:64 A†; because the higher doping of K
involves the increase of the averaged A radius and then
the tolerance factor for the perovskite layer [18], giving
rise to the instability of the phase.
Fig. 1 shows, for example, the X-ray diffraction pattern
and the pattern matching agreement of the representative
Bi1.75Te0.25SrNb1.75Hf0.25O oxide. Table 1 reports the lattice
parameters for all the known single phases
Bi2ϪxTe
9
[8]. Standard X-ray diffraction patterns at room temperature
x
SrNb2Ϫx
B
x
O
9
(B ˆ Ti, Zr, Hf) and Bi2ϪyTe
y
S-
were also collected for the ceramics and no differences
between those and the powder ones were observed. That
rules out the possibility of any second phase appearances
during the sintering process, and insures that this process
keeps the integrity of the synthesized phase.
r1ϪyA Nb O (A ˆ Na, K) together with the averaged A
y
2
9
and B ionic radii for these compounds, according to the
following expressions:
1
2
rꢁ B ˆ ꢀꢀ2 Ϫ x†rNb ϩ xrB†
ꢀ1†
ꢀ2†
Possible phase transition and stability of selected samples
were studied using a thermal analysis instrument Stanton
rꢁ A ˆ ꢀ1 Ϫ y†rSr ϩ yrA

P. Dur a´ n-Mart ´ı n et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
1425
Fig. 1. X-ray powder diffraction profile of Bi1.75Te0.25SrNb1.75Hf0.25
9
O . Crosses are the raw data points, the solid line is the best calculated profile
and the difference plot appears at the bottom. Thin marks indicate the position of the allowed Bragg reflections.
3
ϩ
4ϩ
2
where r
A
and r
B
correspond to the ionic radii of the substi-
behavior between Bi and Te ns lone pair of electrons
[19] and their size. Therefore it seems that the character of
cations forming the perovskite layer dominates the size of
(001) plane in the structure whereas the c value largely
influences from the occurrence of different cations in the
bismuth slab.
tuting cations, and xand y are the doping amounts for each
case.
In general, the a and b unit-cell parameters show a conti-
nuos evolution as a function of the averaged ionic radii of
both A and B cations ꢀ rꢁ and rꢁ †; whereas c parameter is
A
B
more affected by the degree of substitution, x or y. In fact,
since [001] direction is perpendicular to perovskite and
bismuth layers building up such structures, the doping in
the perovskite sites only can have a slightly influence in
the c value, which is particularly governed by the different
This fact could be related with the recent discovery from
the electronic structure of another n ˆ 2 Aurivillius phase,
that the Bi–O layer controls the electronic response of these
materials and the ferroelectric response originates mainly
from the perovskite block [20].
Table 1
Lattice constants and ionic radii of A and B cations in the Bi2ϪxTe
Na, K) solid solutions
x
SrNb2Ϫx
B
x
O
9
(B ˆ Ti, Zr, Hf) and Bi2ϪyTe
y
Sr1Ϫy
A
y
Nb
2
O
9
(A ˆ
Reference
3
˚
˚
˚
˚
˚
˚
V (A )
x
y
A
B
r
A
(A)
r
B
(A)
a (A)
b (A)
c (A)
0
0
0
0
0
0
0
–
–
–
0
–
–
–
–
–
–
0.25
0.5
0.25
Sr
Sr
Sr
Sr
Sr
Sr
Sr
Sr/Na
Sr/Na
Sr/K
Nb
1.44
1.44
1.44
1.44
1.44
1.44
1.44
1.428
1.415
1.490
0.64
5.5147(3)
5.5112(6)
5.5017(6)
5.512(1)
5.514(2)
5.512(1)
5.513(1)
5.4926(6)
5.4386(8)
5.510(2)
5.5122(3)
5.5008(3)
5.4832(4)
5.5132(9)
5.517(2)
5.5153(8)
5.516(1)
5.4911(3)
5.4916(7)
5.5086(8)
25.0775(8)
25.066(1)
25.078(2)
25.034(4)
25.038(3)
25.047(3)
25.047(4)
24.974(2)
24.972(3)
25.047(4)
762.3(2)
759.9(3)
756.5(4)
760.8(8)
762(1)
761.4(7)
761.7(8)
753.2(4)
752.0(6)
760(1)
[5]
[8]
[8]
.25
.5
.25
.5
.25
.5
Nb/Ti
Nb/Ti
Nb/Zr
Nb/Zr
Nb/Hf
Nb/Hf
Nb
0.636
0.631
0.650
0.660
0.649
0.658
0.64
Present work
Present work
Present work
Present work
[8]
Nb
Nb
0.64
0.64
[8]
Present work

1
426
P. Dur a´ n-Mart ´ı n et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
9
Fig. 2. Thermogravimetric (TG) and differential thermal analysis (DTA) curves for Bi1.75Te0.25SrNb1.75Zr0.25O .
The differential thermal analysis (DTA) data from
(B ˆ Zr, Hf, 0 Յ x Յ 0:5† samples
show a very wide effect at the temperature range between
between room temperature and 700ЊC. Fig. 3 shows the
evolution of the unit-cell parameters versus temperature
for this oxide (B ˆ Zr, x ˆ 0:25†: Only an expansion of
the network is observed and any important structural transi-
tion has been detected.
x x 9
Bi2ϪxTe SrNb2ϪxB O
1
00 and 750ЊC. Furthermore, the x ˆ 0:25 oxides exhibit
one well defined exothermic peak at 400 and 500ЊC for
zirconium and hafnium compounds, respectively. Fig. 2
shows the thermogravimetric (TG) and DTA curves for
From the dielectric measurements, the real part of the
0
permittivity, e ; has been represented as a function of the
r
the representative Bi1.75Te0.25SrNb1.75Zr0.25
O
9
phase. The
exhibit,
temperature for the different compositions, Figs. 4 and 5,
grouped as a function of the cationic substitution amount,
y ˆ 0:25; 0.5. As shown, for all samples there are three
anomalies during the heating run. The first of them is rever-
sible, which indicates that it should be related to a structural
ferro–ferro or ferro-paraelectric type phase transition. The
character of this anomaly indicates a diffuse phase transition
occurs at this temperature, because of the thermal hysteresis
between heating and cooling runs, the fat and broad instead
of narrow shape, and also its frequency dependence. The
other two anomalies that appear while heating become
only one anomaly in the cooling run, taking place at an
intermediate temperature between the above two. This
suggests that indeed they are not associated to a real phase
transition. All these temperatures are listed in Table 2.
DTA curve behavior for Bi1.75Te0.25Sr0.75 0.25Nb O
K
2 9
like the above phases, only the very large effect between 100
and 800ЊC. On the other hand, the TG curves, performed in
air conditions, show that these materials are stables up to
8
00ЊC for x, y ˆ 0:25 and 750ЊC for x ˆ 0:5: At those
temperatures the decomposition of the compounds take
place with the volatilization of tellurium as TeO and the
formation of several oxides, like Bi SrNb and BO
2
2
2
2
O
9
(
B ˆ Zr, Hf) or K O, as it has been verify by X-ray
2
diffraction methods. It is interesting to point out as the
decomposition temperatures decrease with increasing tell-
urium doping, while the occurrence of the exothermic peak
in B-doped phases is rather independent of tellurium
content.
X-ray diffraction patterns at higher temperatures have
been recorded for the representative compound
c
To find out the T of the materials, samples were poled.
All of them presented a non-zero d33 value after poling
Bi1.75Te0.25SrNb1.75Zr0.25
O
9
in order to study the possible
(Table 2) which confirms them as ferroelectric materials at
structural transformation joined to the observed exothermic
effect. In this case, as it has been previously reported for Ti
and Na doped materials [8], the viewed evolution involves
only a displacement of diffraction lines to small angles,
together with some intensity variations. The X-ray diffrac-
tion profiles always must be analyzed retaining the symme-
try and space group. The a, b and c parameters have been
calculated for seven increasing temperatures in the range
room temperature. Then, each sample was heated up to the
0
different temperatures of the e anomalies above mentioned,
r
and cooling down to the room temperature to measure again
the d33 value. All samples showed a d33 value equal to
zero, which means the disappearance of the ferroelectric
polarization induced during the poling process, after heating
them above the temperature of their first anomaly. This
strategy provided us the information of the character of

P. Dur a´ n-Mart ´ı n et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
1427
3
˚
˚
˚
˚
Fig. 3. Evolution of the unit-cell parameters (B—a (A), X—b (A), O—c (A) and A—V (A )) as a function of temperature in
Bi1.75Te0.25SrNb1.75Zr0.25
9
O . Represented error bars are the double of the real values, to make them visible.
these anomalies, indicating that this is the first one that is
related to a ferro-paraelectric phase transition.
agreement with results shown in above mentioned Subbar-
ao’s studies. The same family of compounds has a lower T
c
ꢀ
The other higher temperature peaks of the dielectric
constant as a function of the temperature could be due to
changes in the electric conductivity paths that take place
during the heating and the cooling runs. It is shown in
literature that it is possible to obtain very frequency depen-
dent maximum of the dielectric constant as a function of the
temperature, by applying a mean field theory to heteroge-
neous materials [21,22]. In this case, the anisotropic conduc-
tivity of the material would be the parameter to be
considered, and this possibility is under study now.
when B-site is occupied by Tantalum ꢀr ˆ r ˆ 0:64 A†
B
Ta
ꢀ
than when is occupied by Niobium ꢀr ˆ r ˆ 0:64 A†: In
B
Nb
these case, only two groups of compounds with two different
values for the ionic radius r were studied. Now we come
B
with a more extended study that confirms that r
B
is not the
parameter that explains the change of T . One characteristic
c
that changes among the B-cations typically used in Aurivil-
lius’s compounds is their electronegativity. Subbarao [23]
already commented the possible influence of this parameter.
Let us calculate an averaged electronegativity value for
cations in B-sites as:
From the previously reported works of Subbarao [23,24],
it is well-known that, for n ˆ 2 structure compounds with
different A-sites cations, the lower rꢁ value, the higher is T
1
2
eꢁ B ˆ ꢀꢀ2 Ϫ x†eNb ϩ xeB†
ꢀ3†
A
c
of the resulting material. The authors of the present work
have already reported elsewhere [25] that the compounds
with cationic substitutions in the perovskite A-sites that is
Bi2ϪyTe Sr A Nb O (A ˆ Na, K) are in agreement with
where eNb ˆ 1:6 and e ꢀe ˆ 1:5; e ˆ 1:4; e ˆ 1:3† are
B
Ti
Zr
Hf
the Pauling’s electronegativity values for former, Nb, and
doping B-site cations, and x represents the substitution
amount. In Fig. 6, T as a function of this eꢁ ; is represented
y
1Ϫy
y
2
9
c
B
the above-mentioned dependence of T
c
versus the averaged
for those phases doped in B-sites. We can effectively see a
ionic radii, rꢁ : However, if we represent the evolution of T
continuous increase of T with eꢁ ; almost linear for T values
A
c
c
B
c
as a function of the averaged ionic radii, rꢁ ; for the samples
under study in the present work with cationic substitutions
in the B-sites of the perovskite, we will not observe a clear
measured during cooling. Moreover, there are two samples
with very different chemical composition, Bi Te SrNb
1.5-
B
1
.5
0.5
Ti O , and Bi1.75Te0.25SrNb1.75Zr0.25O that have the same
0
.5
9
9
dependence between both of them. T does not decrease with
decreasing value but changes randomly. This is also in
eꢁ ˆ 1:575; and, as shown, the same T measured during the
c
B
c
cooling run. This confirms the hypothesis to consider the

1
428
P. Dur a´ n-Mart ´ı n et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
Fig. 4. Relative dielectric constant as a function of temperature for
Fig. 5. Relative dielectric constant as a function of temperature for
samples with cationic substitution amount of x ˆ 0:25 (A—
samples with cationic substitution amount of x ˆ 0:5 (A—
Bi1.75Te0.25SrNb1.75Ti0.25
O
9
,
W—Bi1.75Te0.25SrNb1.75Zr0.25
O
9
and
Bi1.5Te0.5SrNb1.5Ti0.5
O
9
,
W—Bi1.75Te0.5SrNb1.5Zr0.5
O
9
and K—
K—Bi1.75Te0.25SrNb1.75Hf0.25
9
O ).
9
Bi1.5Te0.5SrNb1.5Hf0.5O ).
electronegativity of the B-site cation as the main parameter
that justifies changes in T when doping in that crystallo-
c
including the composition with A-site substitution,
mentioned above.
graphic position.
Nevertheless, it is curious that two different parameters
are needed to explain changes in T promoted when doping
or when making cationic substitutions. In deed, if we regard
structural parameters like axb or the unit-cell volume, the rꢁ B
parameter becomes no longer necessary. Reordering the
Most of the perovskite-type ferroelectric oxides have a
large Curie–Weiss constant and present a displacive type
ferroelectric phase transition [26]. In these materials occur-
rence of spontaneous polarization can be explained as a
problem of electrostatic equilibrium, and the phase transi-
tion can be associated to structural distortion which
promotes the relative cationic displacement, allowing the
c
c
samples as a function of T (Table 2), we can realize that
the greater the volume the lower the T of the compound,
appearance of spontaneous polarization, P [27]. Thus, this
c
s
Table 2
Several ferroelectric parameters of the compounds under study, ordered as a function of its T
c
˚
Composition
T
c
(up/down) f ˆ 10 kHz
g (A)
d33 (pC/N)
Bi1.5Te0.5SrNb1.5Hf0.5
O
9
230/222
250/250
285/275
300/290
300/295
310/300
330/325
366/346
460/390
9.1326
9.1338
9.1314
9.1290
9.1254
9.1258
9.1118
9.0985
9.0937
5
5
9
9
9
13
7
11
8
Bi1.5Te0.5SrNb1.5Zr0.5
Bi1.75Te0.25SrNb1.75Hf0.25
Bi1.75Te0.25SrNb1.75Zr0.25
Bi1.75Te0.25SrNb1.75Ti0.25
Bi1.75Te0.25Sr0.75 0.25Nb
Bi1.5Te0.5SrNb1.5Ti0.5
Bi1.75Te0.25Sr0.75Na0.25Nb
Bi1.5Te0.5Sr0.5Na0.5Nb O
9
O
9
O
9
O
9
O
9
K
2
O
9
O
9
2
O
9
2

P. Dur a´ n-Mart ´ı n et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
1429
Fig. 6. Variation of the T
c
B
(X—heating, B—cooling) with the averaged electronegativity value of cations in B-crystallographic sites, eꢁ ; for the
different compounds under study.
P
s
is usually estimated as a vectorial sum of the different
g, for all the compounds under study. In other words, it is
confirmed that the cationic substitutions promote a change
in the volume of the unit cells that provides a reciprocal
dipoles contributions inside the unit cell that can be
expressed as:
X
c
change in the T of the compound.
In conclusion, we would remark first that a new solid
1
P ˆ
qiri
ꢀ4†
V
i
solution starting from the well-known n ˆ 2 Bi SrNb O
2
2
9
where V is unit cell volume, r
cation from its position in the network at the paraelectric
state, and q the electric charge of the i-cation.
One of the best-studied compounds belonging to this
group is BaTiO . In BaTiO , the hydrostatic pressure effect
i
the displacement of the i-
have been successfully synthesized and characterized by
different analytical techniques. It has also demonstrated
that all of them are ferroelectric compounds at room
temperature.
i
From the DTA, TG and dielectric analysis data we
have induced that these phases do not show a first order,
or even close to, ferroelectric to paraelectric phase transi-
tion, but a diffuse one. By means of depoling experiences
3
3
on the T of the compound is extensively studied [28,29],
c
which decreases linearly with increasing pressures [30].
Also its solid solutions present a monotonic decrease with
increasing lattice parameters. Finally, it has been recently
c
we found out the T of these materials vary between
shown that change of T
c
can be attributed to a volume effect
230ЊC (Bi1.5Te0.5SrNb1.5Hf0.5O ) and 460ЊC (Bi1.5Te0.5Sr0.5
9
of the unit cell whether the change is produced by a change
in pressure or a change in the composition [31]. In such
Na0.5Nb O ), which corresponds to temperature of the first
dielectric constant anomalies.
2 9
systems, it is usual to select g as structural parameter,
Probably, the most outstanding fact of the present work is
that we have shown that changes in the volume of the unit
cell in these structures, by means of cationic substitutions,
p
3
being g ˆ ꢀa ء b ء c†:
The Aurivillius-type structure is also defined as an alter-
nating stacking of perovskite unit blocks, interconnected by
are responsible for the change of its T
meters that mainly affect this change: for A-site substitu-
tions, the bigger rꢁ A the higher T
. For B-sites substitutions,
increases when increasing the eꢁ B value. The interesting
c
. There are two para-
2ϩ
[Bi
2 2
O ]
layers. They posses a high c value [32], and many
studies present in the literature intend to explain the P
s
c
existence by different structural shifts taking place across
the phase transformation [33,34], and estimate its value by
using expression (4). Then, assuming our materials as
displacive-type ferroelectric compounds, we have calcu-
lated g for them as it is shown in Fig. 7. From this graph,
T
c
point of this result is that it offers the feasibility to tailor by
applying the suitable partial substitution. Those substitu-
tions should take into account that, first, the evolution of
the unit-cell parameters at room temperature shows that
doping in A or B perovskite sites dominates the size of
c
it is clear that T increases with decreasing lattice parameter

1
430
P. Dur a´ n-Mart ´ı n et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
c
Fig. 7. Evolution of T (X—heating, B—cooling) as a function of the g parameter, for the different compounds under study.
(
110) plane in the structure, while the c value is mostly
Tuttle, Qualitative model for the fatigue free behaviour of
SrBi Ta , Appl. Phys. Lett. 68 (1996) 690–692.
[5] P. Millan, A. Castro, J.B. Torrance, The first doping of
lead2 ϩ into the bismuth oxide layers of the Aurivillius
oxides, Mater. Res. Bull. 28 (1993) 117.
2
2 9
O
influenced by the cations occupying the bismuth position
in its [Bi
schmidt’s tolerance factor as well. The comparative study of
the ferroelectric properties of these phases and the role of
the cationic substitutions on their improvement still
remains, and will be the focus of future works.
2
ϩ
2
O
2
]
layer, and second, the limits of the Gold-
[
[
[
[
6] A. Castro, P. Millan, R. Enjalbert, Structural evolution of the
Aurivillius framework in the solid solutions Bi WO
Sb WO , Mater. Res. Bull. 30 (1995) 871–882.
7] P. Millan, A. Ramirez, A. Castro, Substitutions of smaller
2
6
–
2
6
3
ϩ
2ϩ
3ϩ
Sb and Sn cations for Bi in Aurivillius-like phases, J.
Mater. Sci. Lett. 14 (1995) 1657–1660.
Acknowledgements
8] A. Ramirez, P. Millan, A. Castro, J.B. Torrance, Compensated
doping between layers ub-Aurivillius oxides, Eur. J. Solid
State Inorg. Chem. 31 (1994) 173–185.
The authors express their warmly gratitude to Dr A.
Ramirez for the preparation of the powdered samples. The
present work has been supported by the CICYT project,
Mat97-0711.
9] T. Rentschler, M. Karus, A. Wellm, A. Reller, Synthesis
and
characterization of
the
Aurivillius
phases
Bi(2Ϫx)Pb
x
Sr(1Ϫx)Nd
x
2 9
Nb O , Solid State Ionics 90 (1996) 49–
55.
[
10] T. Rentschler, Substitution of lead into the bismuth oxide
layers of the n ˆ 2 and n ˆ 3 Aurivillius phases, Mater.
Res. Bull. 32 (1997) 351–369.
References
[
[
[
1] S. Ikegami, I. Ueda, Pieyoelectricity in ceramics of ferroelec-
tric bismuth compound with layer structure, Jpn. J. Appl.
Phys. 13 (1974) 1572–1577.
2] J.P. Mercurio, A. Souirti, M. Manier, B. Frit, Phase transition
and dielectric properties in some compounds with bismuth
oxide structure, Mater. Res. Bull. 27 (1992) 123–128.
3] O. Auciello, A.R. Krauss, J. Im, D.M. Gruen, E.A. Irene,
R.P.H. Chang, G.E. McGuire, Studies of film growth
processes and surface structural characterization of ferroelec-
[11] P. Millan, M.T. Montero, P. Duran-Martin, B. Jimenez, A.
Castro, Material solid solutions of lead-doped bismuth layer
of Aurivillius n ˆ 2 and 3 oxides: structural and dielectric
characterization, Res. Bull. 33 (7) (1998) 1103–1115.
[12] B. Jimenez, P. Duran-Martin, A. Castro, P. Millan, Obtention
and caracterization of modified Bi
2 2 9
SrNb O Aurivillius-type
ceramics, Ferroelectrics 186 (1996) 93–96.
[13] P. Duran-Martin, B. Jimenez, A. Castro, P. Millan, Influence
of cationic substitutions on the ferroelectric properties of new
aurivillius-type ceramics, Proc. Electroceramics V 1 (1996)
535–538.
[14] P. Duran-Martin, A. Castro, P. Millan, B. Jimenez, submitted
for publication.
tric memory-compatible SrBi
situ, real-time ion-beam analysis, Appl. Phys. Lett. 69 (18)
1996) 2671–2673.
4] H.N. Al-Shareef, D. Dimas, T.J. Boyle, W.L. Warren, B.A.
2 2 9
Ta O layered perovskites via in
(
[

P. Dur a´ n-Mart ´ı n et al. / Journal of Physics and Chemistry of Solids 61 (2000) 1423–1431
1431
[
15] J. Rodriguez-Carvajal, Fullprof program for Rietveld
refinement and pattern matching analysis, XV Congress
International Union Cryst, Toulouse, France, 1990.
[25] P. Duran-Martin, A. Castro, P. Millan, B. Jimenez, Influence
of Bi-site substitution on the ferroelectricity of the Aurivillius
compound Bi SrNb O , J. Mater. Res. 13 (9) (1998) 2565–
2
2
9
[
16] A. Castro, P. Millan, M.J. Martinez-Lope, J.B. Torrance,
2571.
3
ϩ
2ϩ
Substitution for Bi into (Bi
O
2 2
)
layers of the Aurivillius
[26] T. Mitsui, I. Tatsuzaki, E. Nakamura, An introduction to the
physics of ferroelectrics, Gordon and Breach, New York,
1976.
[27] B.A. Strukov, A.P. Levanyuk, Ferroelectric phenomena in
crystals, Springer, Berlin, 1998.
(
(
Bi
1993) 897–903.
2 2 n
O ) (AnϪ1B O3nϩ1) oxides, Solid State Ionics 63–65
[
[
[
17] R.D. Shannon, Revised effective ionic radii and systematic
studies of interatomic distances in halides and chalcogenides,
Acta Cryst. A32 (1976) 751–767.
18] V.M. Goldschmidt, Geochemische verteilungsgesetze der
elemente, VII, Skirft Norske Vidensk. Akad. Ns. (2 and 8)
[
[
[
[
28] W.J. Merz, The effect of hydrostatic pressure of the Curie
point of barium titanate single crystals, Phys. Rev. 78
(1950) 52–54.
(1926).
29] G.A. Smolenskii, K.I. Rozgachev, Seignettoelectric properties
of solid solutions in the system barium titanate–strontium
titanate, Zh. Tekhn. Fiz. 24 (1954) 1751–1760.
19] J. Galy, G. Meunier, S. Andersson, A. Strom, St e´ r e´ ochimie
des El e´ ments Comportant des Paires Non Li e´ es: Ge(II),
As(III), Se(IV), Br(V), Sn(II), Sb(III), Te(IV), I(V), Xe(VI),
Tl(I), Pb(II) et Bi(III) (Oxides, Fluorures et Oxyfluorures), J.
Solid State Chem. 13 (1975) 142–159.
30] G.A. Samara, Pressure and temperature dependences of the
3 3
dielectric properties of the perovskites BaTiO and SrTiO ,
Phys. Rev. 151 (1966) 378–386.
[
[
[
20] J. Robertson, C.W. Chen, W.L. Warren, C.D. Gutleben, Elec-
31] L. Zhang, W.L. Zhong, Y.G. Wang, P.L. Zhang, The cell
volume effect in barium strontium titanate, Solid State
Commun. 104 (5) (1997) 263–266.
tronic structure of the ferroelectric layered perovskite
SrBi Nb O , Appl. Phys. Lett. 69 (1996) 1704–1706.
2 2 9
21] R. Jimenez, Application of the P.E.M. theory to solid electro-
lyte pressed powder pellets: electrically isotropic solid elec-
trolyte particles, Solid State Ionics 90 (1996) 230–244.
22] R. Jimenez, Application of the Percolation Effective Medium
Theory (PEMT) to solid electrolyte pressed powder pellets. II,
Solid State Ionics 92 (1996) 225–241.
23] E.C. Subbarao, Ferroelectricity in mixed bismuth oxides with
layer-type structure, J. Chem. Phys. 34 (1961) 695–696.
24] E.C. Subbarao, family of ferroelectric bismuth compounds, J.
Phys. Chem. Solids 23 (1962) 665–676.
[
32] Landolt-B o¨ rnstein. Vol 16, Ferroelectric and Related
Substances, Subvol. A: “Oxides”, 1981.
[33] A.D. Rae, J.G. Thompson, R.L. Withers, A.C. Willis,
Structure refinement of commensurated modulated bismuth
4 3
titanate, Bi Ti O12, Acta Cryst. B46 (1990) 474–487.
[34] A.D. Rae, J.G. Thompson, R.L. Withers, Structure refinement
[
of commensurately modulated bismuth strontium tantalate,
Bi SrTa O , Acta Cryst. B48 (1992) 418–428.
2 2 9
[