Phase formation in the Nd2Ti2O7-SrTiO3 system at 1350°C in the presence of V2O5, CuV2O6, or SrCuO2

Article abstract of DOI:10.1006/jssc.1996.0044

The system Nd₄Ti₄O₁₄ (=Nd₂Ti₂O₇)-SrTiO₃ has been studied by powder X-ray diffraction and electron microscopy. The samples were prepared by normal ceramic methods and contained one of the additives V₂O₅, CuV₂O₆, or SrCuO₂. Each of these additives caused the reaction to take place at greatly increased rates compared to that without additives. The major products were Nd₄Ti₄O₁₄ and SrTiO₃, which were formed in 8 hr at 1350°C. The vanadium in the additives formed NdVO₄ while the other components were incorporated into the product phases. The Nd₄Ti₄O₁₄ was generally well ordered but the SrTiO₃ showed a complex microdomain microstructure which was associated with a variety of changes in the unit cell observed by X-ray diffraction. Heating for several days caused the SrTiO₃ to revert to the normal cubic form, at which point the micro-domain structure Was lost.

Full text of DOI:10.1006/jssc.1996.0044

JOURNAL OF SOLID STATE CHEMISTRY 121, 324–331 (1996)
ARTICLE NO. 0044
Phase Formation in the Nd Ti O –SrTiO System at 1350؇C in the
2
2
7
3
Presence of V O , CuV O , or SrCuO
2
5
2
6
2
J. Sloan and R. J. D. Tilley
School of Engineering, University of Wales, Cardiff, P.O. Box 917, Cardiff CF2 1XH, United Kingdom
Received April 10, 1995, in revised form September 22, 1995; accepted October 3, 1995
and tilting of the BO octahedra. These are responsible
6
The system Nd
4
Ti
4
O
14 (؍Nd
2
Ti
2
O
7
)–SrTiO
3
has been studied for a number of important physical properties exhibited by
by powder X-ray diffraction and electron microscopy. The sam- perovskite-related compounds in general (2). In particular
ples were prepared by normal ceramic methods and contained
one of the additives V , CuV , or SrCuO . Each of these
additives caused the reaction to take place at greatly increased
rates compared to that without additives. The major products
several oxides show high-temperature ferro- and piezoelec-
tric behavior and some, including Nd Ti O , display po-
2
O
5
2
O
6
2
4
4
14
tentially important opto-electronic properties (3, 4). Re-
cently investigations on the similar La Ti O –LaTiO
4
4
14
3
were Nd
350؇C. The vanadium in the additives formed NdVO
the other components were incorporated into the product
phases. The Nd Ti 14 was generally well ordered but the SrTiO
showed a complex microdomain microstructure which was as-
sociated with a variety of changes in the unit cell observed by relations in systems between Nd Ti O and other perov-
4
Ti
4
O14 and SrTiO
3
, which were formed in 8 hr at
system have shown that varied electrical behaviour accom-
panies the structural changes taking place as the stoichiom-
etry is altered (5, 6).
1
4
while
4
4
O
3
Although Nd Ti O is of considerable interest, phase
4
4
14
4
4
14
X-ray diffraction. Heating for several days caused the SrTiO
to revert to the normal cubic form, at which point the micro-
3
skite structure ABO oxides have not been widely studied
3
with the exception of the Nd Ti O –CaTiO system (7–
4
4
14
3
domain structure was lost.
 1996 Academic Press, Inc.
10). Because of the interest in layered perovskites, both
from the point of view of physical properties and crystal
chemistry we have investigated the phases occurring in the
Nd Ti O –SrTiO system. As Nd Ti O and SrTiO are
INTRODUCTION
Crystallographically Nd Ti O belongs to a small group
4
4
14
3
4
4
14
3
both refractory oxides, reactions were expected to proceed
very slowly. In a previous paper we have given details of
the phases found in this system when samples were melted
as part of a preparation procedure designed to overcome
this problem (11). An alternative technique, the incorpora-
tion of small quantities of rate-enhancing additives into
the system is described here. The additives chosen were
V O , CuV O , or SrCuO . All are low-melting com-
2
2
7
of A B O oxides that includes La Ti O , Ca Nb O , and
2
2
7
2
2
7
2
2
7
Sr Nb O (1). The structure of these materials can be
2
2
7
thought of as built up from layers of perovskite-like sheets
which are four BO octahedra in thickness bounded by
6
͕
110͖ planes relative to a cubic perovskite parent. In all of
these materials the larger A cations occupy positions within
the perovskite sheets and at the interface between them.
The structure of Nd Ti O is shown in Fig. 1.
2
5
2
6
2
pounds and would be expected to form liquid phases, which
are known to increase sintering rates in solids. The oxide
V O was chosen because it has well-known spreading and
2
2
7
It has long been recognized that a homologous series of
related oxides can be constructed, at least hypothetically,
by variation of the thicknesses of the perovskite slabs. The
2
5
surface wetting properties. SrCuO was selected because
2
(
unpublished) results in our laboratory had shown that this
n
n
3nϩ
2
formula of the series is A B O , where the subscript n
was a good sintering aid for the formation of ordered
Aurivillius phases. CuV O was chosen to see if the surface
is equal to the thickness of the perovskite-like slabs mea-
sured in octahedra. The oxides of interest here are thus
representatives of the n ϭ 4 member of the series. To
emphasize this structural feature we have found that it is
often preferable to use the formula A B O rather than
2
6
wetting properties combined with the Cu might give supe-
rior results to SrCuO . The results of this study are de-
2
scribed and discussed.
4
4
14
A B O to describe these phases and we follow this prac-
tice here.
2
2
7
EXPERIMENTAL
The real structures of these materials display a variety
The starting materials were Nd O , (99.99%, Alfa Indus-
2
3
of distortions which include A and B cation displacements tries) and CuO, TiO , V O , and SrCO of Specpure qual-
2
2
5
3
3
24
0
022-4596/96 $18.00
Copyright  1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

PHASES IN TITANATES IN THE PRESENCE OF ADDITIVES
325
fashion but subjected to extended heating times. In these
cases the pellets were removed from the furnace after 3
days heating at 1350ЊC, reground, pressed, and heated to
give a total of 7 days at 1350ЊC. This was repeated once
more so that pellets were subjected to a final heating time
of 14 days at 1350ЊC. Some preparations were also made
in the absence of additive. These samples were heated for
7
days without any intermediate treatment.
After reaction each sample was examined by powder
X-ray diffraction using a Guinier–H a¨ gg focusing camera
employing monochromatic CuKͰ radiation and KCl as
1
an internal standard (a ϭ 0.62923 mm at 25ЊC). Lattice
0
parameters were refined using least-squares techniques.
Samples were also examined by scanning and transmission
electron microscopy. For transmission studies, fragments
of the pellets were crushed under n-butanol in an agate
mortar. A drop of the resultant suspension was placed onto
2 2 7 4 4
FIG. 1. Projection of the structure of Nd Ti O (Nd Ti O14) along
[
010]. The circles represent Nd atoms at two different heights and the a copper grid that had previously been coated with a holey
shaded diamonds represent TiO
6
octahedra. In some structure determina-
carbon film. Fragments of the crystal were examined in a
Jeol 200CX transmission electron microscope fitted with
a goniometer stage operating at 200 kV and equipped
tions the a-axis is double that shown here.
with a LaB filament. For scanning electron microscopy,
6
ity (Johnson Matthey). The additive-assisted preparations samples were examined in a Jeol 35CF microscope fitted
used Nd O , SrO, and TiO , and one of the additives V O , with energy and wavelength dispersive analysis facilities.
2
3
2
2
5
CuV O , or SrCuO . The oxide TiO and additive V O
Samples were gold coated to avoid problems associated
2
6
2
2
2
5
were employed in the as-supplied form without further with electrostatic charging.
modification. The oxide SrO, because of its tendency to
RESULTS
absorb CO from the atmosphere, was prepared immedi-
2
ately prior to use by calcining SrCO in an open crucible
3
Powder X-Ray Diffraction
at 1150ЊC for 8 hr. CuV O was prepared by heating to-
2
6
No additives. Samples with an overall composition
gether a 1 : 1 molar ratio of CuO and V O in an open
2
5
crucible to 600ЊC for 8 hr. SrCuO was prepared by arc
n
nϪ
4
n
3nϩ2
Nd Sr Ti O with n ϭ 4, 4.5, and 5 were used for these
2
preliminary experiments. X-ray diffraction showed that no
significant reaction had taken place after 7 days at 1350ЊC
beyond the production of a small amount of Nd TiO .
melting a 1 : 1 molar ratio of SrO, prepared as above, and
CuO, according to standard procedures (11).
The compositions prepared all fell along the Nd Ti O –
2
5
4
4
14
The major materials present were the starting oxides. No
Nd Ti O had formed.
SrTiO line in the phase diagram if the additives are ig-
3
nored and corresponded to the hypothetical oxides
4
4
14
n
nϪ
4
n
3nϩ
2
Nd Sr Ti O with n ϭ 4, 4.5, 5, 5.5, and 6. Additionally
Additive reactions. Irrespective of the starting compo-
preparations at some other compositions were occasionally sition, Nd Ti O and, when appropriate, SrTiO were
4
4
14
3
made to clarify aspects of the reactions taking place. Ap- formed after heating for 8 hr at 1350ЊC. No traces of the
propriate weights of the reactants were mixed with 5 wt% starting oxides or Nd TiO were observed on the X-ray
2
5
of one of the additives. The mixtures were then ground films.
thoroughly and compressed to form pellets 10 mm in diam- The Nd Ti O patterns were in good agreement with
4
4
14
eter and 3–5 mm thick. Stacks of two pellets of each mix- that given in the JCPDF file (entry 33-942). The refined
ture were placed onto a sheet of freshly cleaned platinum lattice parameters for Nd Ti O were typically (from a
4
4
14
foil resting on a ceramic tile inside a chamber furnace. The sample corresponding to the composition Nd Ti O with
4
4
14
pellets were heated at 200ЊC/min to 800ЊC and maintained V O as additive) a ϭ 1.3001 Ϯ 0.0001 nm, b ϭ 0.5460 Ϯ
2
5
0
0
at this temperature for 30 min before heating at 50ЊC/min 0.0001 nm, c ϭ 0.7680 Ϯ 0.0001 nm, and ͱ ϭ 98.62 Ϯ
0
to the reaction temperature of 1350ЊC. The pellets were 0.04Њ. These values compare favorably with the reported
maintained at the reaction temperature for 8 hr and were values a ϭ 1.3008 nm, b ϭ 0.54648 nm, c ϭ 0.7679 nm,
0
0
0
then air quenched to room temperature. After reaction and ͱ ϭ 98.56Њ. No significant variations in the powder
the bottom pellet was discarded as a precaution against pattern of this phase were found across the phase range
platinum contamination.
and no trace of an orthorhombic modification of Nd Ti O
4 4 14
A further set of samples were prepared in an identical was observed.

326
SLOAN AND TILLEY
TABLE 1
X-Ray Phase Analysis
SrTiO
parameters nm
3
Composition, na
Phases
CuV
2
O
6
8 hr
4
Nd
Nd
Nd
Nd
Nd
4
4
4
4
4
Ti
Ti
Ti
Ti
Ti
4
O
4
O
4
O
4
O
4
O
14 , NdVO
14 , NdVO
14 , NdVO
14 , NdVO
14 , NdVO
4
4
4
4
4
, (TiO )
2
4
5
5
6
.5
, SrTiO
, SrTiO
, SrTiO
, SrTiO
3
3
3
3
(cs)
(cs)
(c)
0.7699
0.7711
0.3856
0.7078
.5
(cs)
2 6
CuV O 3, 7, 14 days
5
6
7
.333
Nd
Nd
Nd
4
4
4
Ti
Ti
Ti
4
O
4
O
4
O
14 , NdVO
14 , NdVO
14 , NdVO
4
4
4
, SrTiO
, SrTiO
, SrTiO
3
3
3
(c)
(c)
(c)
0.3861
0.3870
0.3870
V
2
O
5
8 hr
4
4
5
5
6
Nd
Nd
Nd
Nd
Nd
4
4
4
4
4
Ti
Ti
Ti
Ti
Ti
4
O
4
O
4
O
4
O
4
O
14 , NdVO
14 , NdVO
14 , NdVO
14 , NdVO
14 , NdVO
4
4
4
4
4
2
, ((TiO ))
.5
.5
, SrTiO
, SrTiO
, SrTiO
, SrTiO
3
(cs)
(cs)
(c)
0.7700
0.7702
0.3856
0.7708
3
3
3
(cs)
2
SrCuO 8 hr
4
4
5
5
6
Nd
Nd
Nd
Nd
Nd
4
4
4
4
4
Ti
Ti
Ti
Ti
Ti
4
O
4
O
4
O
4
O
4
O
14
.5
.5
14 , SrTiO
3
3
3
3
(t)
(t)
(t)
a ϭ 0.3890b
a ϭ 0.3870, c ϭ 0.3900
a ϭ 0.3878, c ϭ 0.3899
a ϭ 0.3860, c ϭ 0.7816
14 , SrTiO
14 , SrTiO
14 , SrTiO
(st)
2
SrCuO 3, 7, 14 days
4
4
6
6
7
7
.333 (3 days)
.33 (7, 14 days)
(3 days)
(7, 14 days)
(3 days)
Nd
Nd
Nd
Nd
Nd
Nd
4
4
4
4
4
4
Ti
Ti
Ti
Ti
Ti
Ti
4
O
4
O
4
O
4
O
4
O
4
O
14 , SrTiO
14 , SrTiO
14 , SrTiO
14 , SrTiO
14 , SrTiO
14 , SrTiO
3
3
3
3
3
3
(st)
(c)
(st)
(c)
(st)
(c)
a ϭ 0.3849, c ϭ 0.7892
0.3895 (7 days)
a ϭ 0.3853, c ϭ 0.7790
0.3893 (7 days)
a ϭ 0.3845, c ϭ 0.7788
0.3884
(7, 14 days)
a
n refers to the formula (Sr, Nb)
The lines were not intense enough to determine a value for c .
0
n
Ti
n
O
3n
ϩ
2
.
b
The SrTiO powder pattern displayed a degree of varia- (rutile). All phases had patterns which agreed with those
3
tion. In some preparations X-ray powder pattern corres- reported in the literature and the JCPDS files. No lines
ponded closely to the normal cubic form reported in the could be detected in the powder patterns which could be
literature. This modification will be referred to as attributed to CuV O . In the case of the other prepara-
2
6
SrTiO (c). In other preparations a low-angle line was seen tions, Nd Ti O and NdVO were present together with
3
4
4
14
4
which could be incorporated into a cubic pattern if a unit SrTiO . No TiO could be found on these other prepara-
3
2
cell approximately double that of normal SrTiO was as- tions. No differences could be found between the powder
3
sumed. This modification will be called SrTiO (cs). Other patterns of Nd Ti O and NdVO from one film to another
3
4
4
14
4
preparations resulted in a powder pattern which showed apart from the gradual change in intensity expected as
that the phase was tetragonal with a cell of dimensions the phase range was traversed. In the n ϭ 4.5, 5, and 6
similar to those of the normal cubic modification. This will preparations the SrTiO (cs) modification was found. In
3
be referred to as SrTiO (t). The final modification was also the n ϭ 5.5 sample SrTiO (c) was present. The lattice
3
3
tetragonal, but with a double cell similar to the SrTiO (cs) parameter of the SrTiO was slightly smaller than that
3
3
material. This will be referred to as SrTiO (ts). It was also recorded in the literature.
3
found that the unit cell was slightly contracted compared
The X-ray diffraction results for the samples heated for
with that reported in the literature. The X-ray powder the longer times of 3, 7, and 14 days were identical to those
diffraction phase analysis is summarized in Table 1 and of the shorter heating time preparations with the single
considered in detail below.
exception that only SrTiO (c) was observed. This also had
3
a slightly smaller cell than that reported for pure SrTiO3 .
5
wt% CuV O additive. The n ϭ 4 samples heated for
2
6
8
hr contained Nd Ti O , NdVO , and a trace of TiO
5 wt% V O additive. The X-ray powder diffraction
4
4
14
4
2
2
5

PHASES IN TITANATES IN THE PRESENCE OF ADDITIVES
327
results for these samples were identical to the equivalent of Nd Ti O formed in the presence of SrCuO additive
4
4
14
2
CuV O additive preparations. Longer periods of heating was found to be considerably less and so is not illus-
2
6
were not utilized because of this similarity.
trated here.
No order or disorder was found in any of the fragments
examined which could be interpreted as due to slabs of a
perovskite type wider than those characteristic of
5
wt% SrCuO additive. For the n ϭ 4 composition,
2
only Nd Ti O was present. In the other preparations only
two phases were observed, Nd Ti O and SrTiO . In the
n ϭ 4.5, 5, and 5.5 preparations the SrTiO was present as
4
4
14
4
4
14
3
Nd Ti O .
4
4
14
3
SrTiO (t) while in the n ϭ 6 preparation it was SrTiO (st).
SrTiO3 . The SrTiO crystal fragments presented a sim-
3
3
3
The X-ray powder diffraction results for the samples ilar picture regardless of the sample preparation method
heated for 3, 7, and 14 days showed that the SrTiO patterns used. In the samples heated for 8 hr most crystal fragments
3
evolved in a consistent fashion. After 3 days total heating exhibited contrast indicative of a microdomain substruc-
time the X-ray patterns of this phase were characteristic of ture. Similar microdomain structures arose in the (sc), (t),
SrTiO (st) and after 7 days total heating time the material and (st) forms of SrTiO . The microdomains consisted
3
3
conformed to the SrTiO (c) parent structure.
of regions of altered contrast lying along two orthogonal
3
directions. We show two examples in Figs. 3a and 3b. The
relative extent of the microdomains varied from one crystal
fragment to another and the two examples shown are rep-
resentative of many others, each of which is slightly differ-
ent from the others in the extent and arrangement of the
microdomain texture. When crystal projections which were
at an angle to the cubic axes were examined the microdo-
mains appeared as irregular patches in the regular fringe
patterns of the matrix.
Scanning Electron Microscopy
Scanning electron microscopy was mainly used to deter-
mine whether the additives were present after reaction
rather than for morphological studies. Nevertheless it was
found that in all cases well-formed crystals made up the
bulk of the samples. In none of the preparations were
discrete crystals or other phases found which analyzed to
a composition identical to an additive phase. Instead the
elements which comprised the additive for any particular
sample were found to be distributed fairly evenly through-
out the sample. In the cases of CuV O and V O the
overall amounts of Cu and V found were identical to those
originally incorporated into the preparation within the ac-
curacy of the technique. In the case of SrCuO the amount
of Cu remaining after reaction was a little less than that
originally incorporated into the sample. The Sr could not
be determined uniquely because of interference with that
already present as SrTiO3 .
All the electron diffraction patterns showed extra reflec-
tions compared to that expected for cubic SrTiO . As ex-
3
amples, the diffraction patterns corresponding to the im-
ages shown in Fig. 3 are included in Figs. 4a and 4b. The
intensities of the additional spots in the diffraction patterns
varied in a way that was consistent with the amount of
microdomain material present and the underlying symme-
try of the structure as determined from the X-ray pow-
der patterns.
2
6
2
5
2
In all samples diffraction patterns similar to those ex-
pected from normal SrTiO were also found. In such cases
3
the micrographs did not show microdomain contrast. In
the samples which were annealed for long times this was
the normal state of affairs and microdomains, or the rem-
nants of this contrast, were only rarely seen.
Transmission Electron Microscopy
Nd Ti O . A small difference was observed between
4
4
14
fragments of Nd Ti O formed in the presence of CuV O
4
4
14
2
6
and V O as opposed to those produced with SrCuO .
2
5
2
Most of the fragments of Nd Ti O obtained from the
4
4
14
DISCUSSION
CuV O and V O preparations exhibited a measure of
2
6
2
5
disorder along the long axis of the unit cell. This was
One of the principal conclusions of this study was that
n
n
3nϩ
2
revealed by uneven contrast on the micrographs, as shown no A B O phases with n greater than 4 were found.
in Fig. 2a, taken from a preparation using CuV O as the Moreover, no wider spacings were found as defects within
2
6
additive. The disorder was also manifested on the diffrac- the matrix of Nd Ti O , which could be considered to be
4
4
14
tion patterns by diffuse streaking along and parallel to a* evidence for the incipient formation of the n ϭ 5 or higher
and by a doubling of the spot spacing in alternate rows of phases. This result is likely to reflect the temperatures
diffraction maxima running parallel to a*. Inspection of employed, that is, thermodynamic considerations, as we
the contrast on these micrographs revealed that the disor- have found that the phases SrNd Ti O , Nd Ti O , and
4
5
17
5
5
17
der was due to microtwinning with (100) acting as the Nd Ti O form in arc-melted preparations (11, 12). This
6
6
20
composition plane. The twins were sometimes seen to be is supported by the fact that Nd Ti O formed in just 8
4
4
14
aggregated into larger bands, as shown in Fig. 2b, which hr and no other phases formed even after 14 days, strongly
was also taken from a preparation utilizing CuV O . The suggesting that it is not kinetic factors which are a problem
2
6
degree of disorder and twinning observed in the fragments in the synthesis of high-n phases.

328
SLOAN AND TILLEY
4 4 2 6
FIG. 2. Micrographs of Nd Ti O14 (n ϭ 4) prepared using CuV O additive. (a) Preparation conditions: 8 hr at 1350ЊC. The uneven contrast
is due to microtwinning across (100) planes, imaged as dark lines running vertically in the photograph. (b) Preparation conditions: 14 days at 1350ЊC.
The twins are aggregated into bands, imaged as strips of darker contrast running vertically.

PHASES IN TITANATES IN THE PRESENCE OF ADDITIVES
329
FIG. 3. Micrographs showing two different forms of microdomain contrast in SrTiO
b) Preparation conditions: CuV additive, 8 hr at 1350ЊC.
3
. (a) Preparation conditions: V
2
O
5
additive, 8 hr at 1350ЊC.
(
2 6
O

330
SLOAN AND TILLEY
FIG. 4. (a,b). Diffraction patterns from the images shown in Fig. 3a and 3b, respectively. It is seen that the intensity distribution of the superlattice
reflections varies considerably between the two patterns. The indices refer to the SrTiO subcell.
3
A second important finding was that the reactions pro- tives while 7 days in the absence of additives gave virtually
ceeded to form mainly Nd Ti O and SrTiO and that no reaction at all. The fact that Nd Ti O formed readily
4
4
14
3
4
4
14
the presence of all of the additives, V O , CuV O , and in all preparations at 1350ЊC suggests that an even lower
2
5
2
6
SrCuO , increased the rate of the reaction considerably temperature could have been employed in the synthesis.
2
compared to that without additives. As far as the kinetics The lower temperature limit for this reaction was not inves-
of the reaction are concerned, a time of 8 hr gave an tigated, however. Even so, the reaction conditions used
apparently complete reaction in the presence of all addi- compare favorably with those reported in the literature.

PHASES IN TITANATES IN THE PRESENCE OF ADDITIVES
331
For example, Nd Ti O , La Ti O , and Sr Nb O typically the decrease in the lattice parameter of the SrTiO is
2
2
7
2
2
7
2
2
7
3
require temperatures in excess of 1600ЊC for their for- due to substitutions taking place perhaps in conjunction
mation. with vacancy formation. One likely scheme would involve
3
ϩ
2
ϩ
We have no information on the mechanism by which the incorporation of Nd on the A sites and Cu on
the reaction rate was increased so drastically. However, the B sites.
the additives were chosen because of their low melting
The microdomain contrast observed in the SrTiO3
points and it is likely that a transient liquid phase plays a material would arise naturally if small regions of different
role in the kinetics. The importance of gas-phase transport symmetry are coherently grown into the perovskite ma-
is not eliminated, though, and is worthy of further investi- trix. In all of the cases described the normal cubic
gation.
matrix is converted to some extent into material of lower
The additives utilized in this study were not observed symmetry. The alternative ways in which these lower
after reaction as discrete phases. The fact that no second symmetry variants can be disposed in the structure would
phase at all was found when SrCuO was used as the addi- then give rise to the microdomain motifs seen. Similar
2
tive, even for the n ϭ 4 preparations, indicates that this microdomains have been observed in LaTiO in contact
3
must be incorporated into the Nd Ti O structure. Proba- with La Ti O (6). Modeling of these structures is not
4
4
14
5
5
17
bly the Sr enters the Nd sites and the Cu enters the Ti practical until the exact defect structure of each material
sites. As the charges are balanced no other defects are is obtained. Further studies of the microdomain structure
needed. A similar situation is also likely to hold with of SrTiO are needed to clarify the remaining uncertain-
3
SrTiO . SrCuO could readily be incorporated without the ties. These will be reported in the future.
3
2
necessity of forming cation counter defects, if the addi-
tional Sr population enters Sr sites and the Cu occupies
Ti sites. The well-known tolerance of the perovskite struc-
ture for substitution supports this view.
ACKNOWLEDGMENTS
J.S. is indebted to the School of Engineering, University of Cardiff,
for the provision of a grant which made this study possible.
For the V O and CuV O additives, NdVO formed as
2
5
2
6
4
a side-product. The formation of NdVO indicates that
4
REFERENCES
neither V O5 or CuV O could be incorporated into
2
2
6
Nd Ti O or SrTiO readily. This is possibly due to the
4
4
14
3
1
. K. Scheunemann and H. M u¨ ller-Buschbaum, J. Inorg. Nucl. Chem.
7, 2261 (1975).
2. J. K. Brandon and H. D. Megaw, Philos. Mag. 21, 189 (1970).
. M. Kimura, S. Nanamatsu, T. Kawamura, and S. Matsushita, Jpn. J.
Appl. Phys. 13, 1437 (1974).
. J. K. Yamamoto and A. S. Bhalla, J. Appl. Phys. 70, 4469 (1991).
. T. Williams, H. Schmalle, A. Reller, F. Lichtenberg, D. Widmer, and
G. Bednorz, J. Solid State Chem. 93, 534 (1991).
fact that such incorporation would involve substitution on
the Ti sites and the necessity of generating energetically
unfavorable counter defects in the A sites. However, small
amounts of incorporation have not been ruled out by the
present study.
3
3
4
5
The metastable superstructures formed in the SrTiO3
phase present features very similar to those reported in a
study of the SrTiO –La TiO system (13). These authors
6. T. Williams, F. Lichtenberg, D. Widmer, G. Bednorz, and A. Reller,
J. Solid State Chem. 103, 375 (1993).
. M. Nanot, F. Queyroux, J-C. Gilles, R. Portier, and M. Fayard, Mater.
Res. Bull. 10, 313 (1975).
. M. Nanot, F. Queyroux, J-C. Gilles, and R. Chevalier, Acta Crys-
tallogr. Sect. B 32, 1115 (1976).
3
2/3
3
7
8
found powder patterns almost identical with our (sc)
type for samples close to compositions of
3
0%SrTiO : 70%La TiO and interpreted their observa-
3 2/3 3
tions as indicative of vacancy ordering on the A cation
positions. Although this makes a suitable model for the
SrTiO3 formed in our preparations, we do not have
sufficient information on the presence or absence of
additives in our materials to draw further conclusions
9. M. Nanot, F. Queyroux, and J-C. Gilles, Mater. Res. Bull. 12, 915
1977).
0. M. Nanot, F. Queyroux, and J-C. Gilles, J. Solid State Chem. 28,
37 (1979).
1. J. Sloan and R. J. D. Tilley, Eur. J. Solid State Chem. 31, 673 (1994).
2. E. Connolly, J. Sloan, and R. J. D. Tilley, in preparation.
(
1
1
1
1
at present. However, it is reasonable to conclude that 13. T. Y. Tien and F. A. Hummel, Trans. Brit. Ceram. Soc. 44, 233 (1967).