Processing, characterization, and dielectric studies on K(Ta1-xNbx)O3 for use at cryogenic temperatures

Article abstract of DOI:10.1111/j.1151-2916.2001.tb01051.x

Development of the next generation space telescope, to supplement the Hubble Space Telescope, has heightened the need for cryogenic actuators. High performance of electromechanical materials are needed for use from 10 to 77 K. The use of piezoelectric materials to satisfy this need involves major property trade-offs; consequently, alternative materials are of interest - the electrostrictive ceramics. The electrostrictors operate above their Curie (transition) temperature, and have electromechanical properties that can be superior to conventional piezoelectric materials. Because work on the electrostrictive materials is limited, the K(Ta,Nb)O₃ solid solution system has been chosen as a viable material for use at cryogenic temperatures. These solid solutions have predicted transition temperatures in the range of 4 to 150 K. Two separate KTaO₃ powders were synthesized with 10 and 17.5 mol% Nb₂O₅ doping levels. The powders were synthesized via a conventional mixed oxide route and materials were fabricated using a tape cast and laminating technique. Powders and samples were characterized using laser scattering, scanning electron microscopy, and surface area analysis to determine particle size distribution and morphology, powder X-ray diffractometry for phase composition, and dielectric response as a function of frequency and temperature. Tape casting and lamination produced final samples that were fabricated to ～88% of theoretical density. Dielectric measurements resulted in Curie temperatures of 248 and 123 K, and weak-field permittivity values of ～2700 and 4000 for 17.5 and 10 mol% additions of Nb₂O₅ to KTaO₃.

Full text of DOI:10.1111/j.1151-2916.2001.tb01051.x

J. Am. Ceram. Soc., 84 [11] 2547–52 (2001)
journal
Processing, Characterization, and Dielectric Studies on
K(Ta1؊xNb )O for Use at Cryogenic Temperatures
x
3
Cristopher B. DiAntonio* and Stephen M. Pilgrim*
Laboratory of Electronic Ceramics, New York State College of Ceramics at Alfred University,
Alfred, New York 14802
Development of the next generation space telescope, to
supplement the Hubble Space Telescope, has heightened the
need for cryogenic actuators. High performance electrome-
chanical materials are needed for use from 10 to 77 K. The
use of piezoelectric materials to satisfy this need involves
major property trade-offs; consequently, alternative mate-
rials are of interest—the electrostrictive ceramics. The
electrostrictors operate above their “Curie” (transition)
temperature, and have electromechanical properties that
can be superior to conventional piezoelectric materials.
Because work on the electrostrictive materials is limited, the
K(Ta,Nb)O3 solid solution system has been chosen as a
viable material for use at cryogenic temperatures. These
solid solutions have predicted transition temperatures in the
range of 4 to 150 K. Two separate KTaO3 powders were
synthesized with 10 and 17.5 mol% Nb O doping levels.
The powders were synthesized via a conventional mixed
oxide route and materials were fabricated using a tape cast
and laminating technique. Powders and samples were char-
acterized using laser scattering, scanning electron micros-
copy, and surface area analysis to determine particle size
distribution and morphology, powder X-ray diffractometry
for phase composition, and dielectric response as a function
of frequency and temperature. Tape casting and lamination
produced final samples that were fabricated to ϳ88% of
theoretical density. Dielectric measurements resulted in
Curie temperatures of 248 and 123 K, and weak-field
permittivity values of ϳ2700 and 4000 for 17.5 and 10 mol%
phase is cubic and the lattice parameter is almost identical to
that of KNbO in its centrosymmetric phase, Ͼ700 K. Despite
3
its similarity, the Curie temperature of KTaO is believed to be
3
ϳ13 K, one of the two lowest ferroelectric transition tempera-
tures known to date (the other is that of lithium thallium tartrate
monohydrate, LiTlC H O ⅐H O, 10 K). The temperature depen-
4
4
6
2
dence of the dielectric constant has been measured down to 1.3
2
K by Hulm et al. The results show that the dielectric constant
obeys the Curie–Weiss law only down to 52 K, below which it
still increases, but less rapidly than required by that law. A flat
maximum is reached at the Curie point, below which the curve
levels off in a smooth fashion. To date, no conclusive informa-
tion is available about the symmetry of the polar phase. The
deviation from the Curie–Weiss law at low temperatures is a
phenomenon characteristic of ferroelectrics with very low Curie
2
5
2
3
points. Barrett showed that quantum effects play an important
role in these low temperature regions, and that these effects can
explain the deviation from the Curie–Weiss law observed
experimentally.
The K(Ta1ϪxNb )O (KTN) system is a ferroelectric material
x 3
for x Ն 0.05, and exhibits the same three ferroelectric phases
as BaTiO3 and KNbO3 (tetragonal–orthorhombic–rhom-
bohedral). It has a
ϳ4 ϫ 10 °C⅐(m ⅐K) . Its Curie temperature can vary with
high pyroelectric coefficient of
Ϫ6
2
Ϫ1
the composition from 13 K for KTaO , to 700 K for KNbO ,
3
3
and it is stable in air. The phase diagram and variation in Curie
temperature for this solid solution can be seen in Fig. 1. The
additions of Nb O to KTaO .
2
5
3
system of solid solutions K(Ta,Nb)O was first investigated by
3
3
,4
Reisman et al. with thermal and XRD measurements. Trieb-
wasser conducted a detailed study of this system more recently
5
I. Introduction
with particular attention to its dielectric properties. The mea-
surements were made on single crystals of solid solutions with
different compositions. Single-crystal KTN and hot-pressed
polycrystalline KTN, prepared from the conventional mixed
ECENTLY, a need has developed for actuator and transducer
R
systems for applications that operate at cryogenic tempera-
tures. Examples of these applications include flow controllers for
liquid fuels, micropositioners for high-temperature superconduc-
tors, precision positioning devices, active dampers for spacecraft,
and transducers for ultrasonic monitoring of cryogenic liquids or
structures. One group of materials that may be able to address this
need is electrostrictive ceramics. These materials commonly op-
erate above their transition temperature and have superior electro-
6
–9
oxide method, have been studied by other investigators.
The
results show that the Curie temperature varies almost linearly
between the two end-members of the system with a cubic–
tetragonal transition in the region of room temperature when
x Ϸ 0.37 (see Fig. 1). The predicted transition temperatures for
the synthesized compositions are 298 and 173 K for Nb O
2
5
mechanical properties. The KTaO system seems to show potential
dopant levels of 17.5 and 10 mol%, respectively. These pre-
dicted values have been determined from Fig. 1.
3
as an electrostrictive ceramic, and will also be able to operate at
cryogenic temperatures.
1
Matthias reputed the ferroelectric properties of the com-
pound KTaO in 1949. The symmetry of the room temperature
3
II. Experimental Procedure
Potassium carbonate (K CO , 99% purity, standard grade, EM
2
3
Science, Gibbstown, NJ), tantalum oxide (Ta O , 99% purity,
W. Huebner—contributing editor
2 5
ceramic grade, H. C. Stark, Inc., New York), and niobium oxide
Nb O , 99% purity, ceramic grade, H. C. Stark, Inc.), were used
(
2
5
as raw materials. Thermogravimetric (TGA) and differential ther-
mal analysis (DTA) were performed before batching to determine
the amount of volatiles in each of the raw materials, and to
quantify the hygroscopic nature of the K CO .
Manuscript No. 188508. Received June 5, 2000; approved October 16, 2000.
Presented at the 100th Annual Meeting of the American Ceramic Society,
Cincinnati, OH, 1998.
*
Member, American Ceramic Society.
2
3
2
547

2
548
Journal of the American Ceramic Society—DiAntonio and Pilgrim
Vol. 84, No. 11
Table I. Heat Treatment Schedule used for
Binder Burnout and Sintering of the KTN
Laminated Stacks
Temperature ramp
°C/h)
Dwell temperature
(°C)
Dwell time
(h)
(
2
3
4
0
0
0
350
450
4
4
2
2
600
2
00
1300
milled using stabilized zirconia. The powders were first dry-milled
to break down agglomerates and promote mixing. After dry-
milling for 2 h, ethanol was added and the powders were
wet-milled for 24 h. Ethanol was used because the K CO not only
2
3
absorbs water, but also is soluble in it. It is assumed that no
dissolution of K CO into the ethanol occurred. The suspension
2
3
was then stir-dried. The powder was calcined in a sealed alumina
crucible (calciner furnace controlled by a control unit, Eurotherm
Corp., Reston, VA) to 850°C and held there for 4 h. After
calcination, the powder was dry-ball-milled for 4 h to reduce the
particle size and pulverize the larger agglomerates that formed
during calcining. The calcination/milling procedure was then
repeated to assure the reaction had completed.
Final test samples were fabricated using a tape cast and
laminating technique. Figure 2 is a flow chart that shows the
general steps taken to prepare a sample from laminating tape layers
produced by tape casting. The slurry prepared for casting was
composed of 65% KTN powder, 29% binder (No. B73305 vinyl
binder system, Ferro Corp., Cleveland, OH) and 6% toluene
5
,6
Fig. 1. Phase diagram of the KNbO –KTaO system.
3
3
For clarity purposes, the base powder is considered KTaO and
the dopant is Nb O . This results in a formulation equation of the
3
2
5
†
type
(solvent) by weight. This particular binder is formulated with all
necessary solvents, plasticizers, surfactants, and release additives.
The slurry was then mixed for 24 h using zirconia milling media.
It is important not to ball-mix for longer periods than 24 h because
of evaporation of organic substances and mechanical degradation
of the vinyl binder. The degradation of the binder can occur
because of the milling action of the media, which is added only to
promote mixing and homogeneity. The slurry is de-aired, using a
vacuum pump and dessiccator, and cast, using a doctor blade
casting unit. The tape was dried for 45 min inside a vented
chamber. The tape was cut, stacked and laminated using a
laboratory press (Carver, Inc., Wabash, IN) that is equipped with
a heating system (Model CN 800, Omega Engineering, Inc.,
Stamford, CT). This heating system controls the temperature of the
press plates, both above and below the die. The stacks were
laminated at 75°C, 5000 psi (20.7 MPa), and held for 10 min.
The final samples were sintered to 1300°C for 2 h (alumina tube
furnace, Lemont Scientific, State College, PA, controlled by a
Eurotherm control unit). They were fired in a sealed alumina
crucible and the atmosphere was maintained by packing the
1
2
1
K CO –H O ϩ ͑1-x͒ Ta O
5
2
3
2
͓ ͔
2
2
1
͑2
1
1
ϩ
x Nb O 3 K(Ta Nb )O ϩ CO ϩ H O
2 5 1-x x 3 2 2
2 2
͒
Two separate compositions were prepared, one with 10 mol%
Nb O , and the other with 17.5 mol% Nb O . For both composi-
2
5
2 5
tions, the K CO , Nb O , and Ta O were dry- and wet-ball-
2
3
2
5
2 5
†
2 2
The evolved CO and absorbed H O must be accounted for.
Fig. 2. Flowchart showing the general steps for fabricating samples from
laminated tape-cast layers.
Fig. 3. TGA profile for K CO , tested at 10°C/min up to 1000°C.
2
3

November 2001
K(Ta1ϪxNb )O for Use at Cryogenic Temperatures
2549
x
3
Fig. 4. Laser-scattering particle-size distribution analysis.
samples in a sacrificial powder of the same composition. The
sintering schedule for the laminates has several steps because of
the high percentage of organics that need to be burned out. The
heat treatment schedule shown in Table I was used for binder
burnout and sintering of the KTN laminated stacks.
The sintered samples were polished using 320 and 600 grit SiC
polishing paper. Polished sample surfaces were sputtered with a
layer of 60% gold–40% palladium and then electroded using a
silver, fire-on, conductor composition (silver conductor composi-
tion No. 7095, Dupont, Wilmington, DE). The silver paste was
fired on at 550°C for 15 min.
1
1
Fig. 6. XRD pattern of KTN (x ϭ 0.20).
any decomposition products or volatiles. The TGA for potassium
carbonate is shown in Fig. 3. The analysis was performed using a
heating rate of 10°C/min to 1000°C. The initial weight loss at
00°C is caused by the volatilization of absorbed water. At 860°C,
the carbonate decomposes and carbon dioxide volatilizes, leaving
a pure form of potassium oxide (K O). The TGA results show that
1
2
K CO contains ϳ16.6% by weight of absorbed H O. This was
taken into consideration in the batch formulations.
To fully characterize the powders, a number of powder charac-
terization techniques were used. TGA was used to quantify the
amount of volatiles in the raw materials, particularly K CO . The
2
3
2
2
3
particle size distribution and morphology were measured using a
combination of SEM, laser scattering analyzer, and surface area
analysis (BET). Phase purity and phase development of the final
powders was examined using powder XRD. The density of the
powder was determined using a helium pycnometer, and the
specific surface area was measured using BET analysis.
(2) Powder Properties
Powders with Nb O dopant levels of 0.17 [K(Ta0.65Nb0.35)O3]
2
5
and 0.10 [K(Ta0.80Nb0.20)O ] were synthesized using conventional
3
mixed oxide techniques. The particle size distribution (Model
LA-920, Laser Scattering Particle Size Distribution Analyzer,
Horiba Instruments, Inc., Irvine, CA) and a SEM image (Model
1
810, Amray, Inc., Bedford, MA) of the calcined powder can be
seen in Figs. 4 and 5. From the SEM image and particle size
distribution results, the average particle size was determined as
2.90 Ϯ 0.10 ␮m, assuming spherical particles. This value was
consistent for both powders (x ϭ 0.35 and x ϭ 0.20) prepared in
this system. The image also shows the spherical-like morphology
of the particles. BET analysis measured the specific surface area
III. Results and Discussion
(
1) TGA of Raw Materials
TGA was performed on all raw materials used for batching
KTN. The tantalum and niobium oxide showed no indication of
(Model Gemini III 2375 Surface Area Analyzer, Micromeretics,
Norcross, GA) of the calcined powders at a value of 2.5 Ϯ 0.35
2
m /g.
(3) Phase Formation
Final powders were examined for phase formation and purity
using powder XRD techniques (Model D5000, Siemens,
Karlsruhe, Germany; JADE™ pattern processing software, Mate-
rials Data, Inc., Livermore, CA). Figure 6 shows a typical XRD
pattern of the x ϭ 0.20 KTN composition (all patterns were taken
over a 2␪ range of 20° to 70° and a step size of 0.02° at a count
time of 5 s). The major phase in both compositions was the cubic
perovskite KTaO , Pm3m. The KTN, x ϭ 0.20 and x ϭ 0.35,
3
patterns are nearly identical to each other. The only difference, as
seen in X-ray analysis, is a small shift in the 2␪ values and a slight
broadening at the base of the peaks. These changes are caused by
the small variation in unit cell size from the further substitution of
niobium for tantalum at x ϭ 0.35. The pattern was compared with
both end compositions KTaO3 and KNbO3 (Refs. 11 and 12,
respectively) because a reliable solid solution pattern doesn’t exist
in the powder diffraction file database. KNbO , at room temper-
3
ature, has an orthorhombic crystal structure (Cm2m(38)). No
unidentified peaks were present and the pattern indexes as cubic
with the (hkl) values, as noted in Fig. 6.
The XRD analysis confirmed phase formation with no indica-
tion of contamination. The pattern matched that of the cubic
perovskite KTaO (Ref. 11) and no KNbO peaks were apparent in
Fig. 5. SEM image of final KTN (x ϭ 0.20) powder at 4000ϫ
magnification.
3
3

2
550
Journal of the American Ceramic Society—DiAntonio and Pilgrim
Vol. 84, No. 11
†
Table II. Physical Properties of the Laminated Pellets of KTN (x ؍ 0.35 and x ؍ 0.20)
Percent of theoretical density
(%)
3
Bulk density (g/cm )
Applied pressure (psi)
Apparent porosity (%)
x ϭ 0.35
x ϭ 0.20
x ϭ 0.35
x ϭ 0.20
2
3
4
5
500 (17 MPa)
500 (24 MPa)
000 (28 MPa)
000 (34 MPa)
12 Ϯ 1
6.7 Ϯ 4
5.3 Ϯ 2
5.0 Ϯ 3
3
5.0 Ϯ 0.30
5.3 Ϯ 0.25
5.5 Ϯ 0.35
5.7 Ϯ 0.40
5.3 Ϯ 0.20
5.5 Ϯ 0.30
5.6 Ϯ 0.25
5.8 Ϯ 0.20
80.9 Ϯ 3
85.8 Ϯ 3
89.0 Ϯ 1
92.2 Ϯ 1
81.0 Ϯ 5
84.0 Ϯ 3
85.6 Ϯ 2
88.7 Ϯ 1
†␳
(theoretical density) calculated as 6.18 g/cm for x ϭ 0.35 and 6.54 g/cm for x ϭ 0.20 (XRD).
3
T
the pattern, indicating the formation of a solid solution of
K(Ta,Nb)O . The only indication of the added Nb O dopant was
(4) Sintering Behavior
The percent of theoretical density, apparent porosity, and bulk
density, as a function of pressing pressure, was measured for the
3
2
5
a slight shift of the diffraction peaks at the 17.5 mol% niobium
oxide composition caused by the change in lattice spacings with
5ϩ
1
3
KTN samples following a previously outlined procedure. Several
laminating pressures were tested to find the ideal conditions for the
sample fabrication, and those results are listed in Table II for both
the x ϭ 0.35 and x ϭ 0.20 sintered samples. Density values of
ϳ88% of theoretical density were the best obtained because of the
difficulty in sample preparation. The KTN solid solution appears
to sinter at high temperatures (1300°C) and maximum densifica-
tion. At higher temperatures, density values remained the same and
eventually led to melting of the samples. The optimal laminating
condition for these samples is an applied pressure of 5000 psi at
5
ϩ
5ϩ
increasing Nb substitution for Ta in the unit cell. The Nb
5
ϩ
and Ta occupy the same site in the perovskite structure, and the
difference in their ionic radius is only ϳ0.02 Å, therefore, the
cubic KTaO X-ray pattern appears to dominate at dopant levels of
3
0
.20 and smaller.
7
5°C. These conditions provided consistent results and well-
fabricated samples for testing. The average volume shrinkage,
from green to fired state, was ϳ25 vol% after sintering to 1300°C.
(5) Microstructure
The microstructure of the final samples was examined using
SEM. A fractured surface of a pellet of KTN x ϭ 0.20 is shown in
Fig. 7. Both KTN compositions possessed the same microstruc-
tural characteristics. Visual inspection of these images indicates
that grain sizes are on the order of 3–5 ␮m and undergo
intergranular fracture. Also evident in these images are the large
volumes of porosity contained in these samples.
(6) Dielectric Measurements
Dielectric measurements were performed as a function of
frequency and temperature (Model 4284A, precision LCR meter,
Model 3488A, automatic switch scanner, Hewlett-Packard Co.,
Rockaway, NJ, and a test chamber, Delta Design, San Diego, CA,
controlled by a computer running LabView software, National
Instruments, Woburn, MA). The test chamber was capable of
operating in the temperature range of 90 to 475 K (using liquid
nitrogen to cool), thus allowing measurements over a broad
temperature range. True four-point contacts were made at the
sample (as shown in Fig. 8), rather than at the back end of the
probe. The chamber was heated to 445 K and held for 30 min to
allow moisture to volatilize and fully de-age the samples. Samples
that are not properly de-aged will show lower permittivity values,
Fig. 8. Four-terminal measurements largely eliminate the effects of lead
resistance by separating the current-sourcing and voltage-sensing func-
tions.
Fig. 7. SEM images of fractured surfaces of KTN (x ϭ 0.20).

November 2001
K(Ta1ϪxNb )O for Use at Cryogenic Temperatures
2551
x
3
Fig. 9. Weak-field permittivity measurements of KTN (x ϭ 0.35) as a
function of frequency and temperature.
Fig. 11. Weak-field permittivity measurements of KTN (x ϭ 0.20) as a
function of frequency and temperature.
shifted transition temperatures, and relaxation frequencies. Data
were then collected from 445 to 90 K with a Ϫ1 K/min cooling
rate and liquid nitrogen purging the oven.
The average weak-field permittivity and dielectric loss as a
function of temperature and frequency for KTN, x ϭ 0.35 and x ϭ
IV. Conclusions
These results provide a clearer view of the powder properties,
microstructural characterization, and dielectric response of two
separate solid solutions in the KTN system. Although both
compositions show similar physical properties, their dielectric
properties are significantly different. TGA allowed quantification
of volatiles in the potassium carbonate where 16.6% by weight of
0
.20, is shown in Figs. 9–12. The measurements were taken over
a temperature range of 445 to 90 K for frequencies of 1, 10, 100,
and 1000 kHz at 1 Vrms applied voltage. The permittivity values
were collected on the cooling portion of the temperature cycle. The
average transition temperature (Tmax) and average permittivity at
that Tmax as a function of frequency for the x ϭ 0.35 and x ϭ 0.20
samples, are listed in Tables III and IV. Figure 13 shows a
comparison of the weak-field permittivity for KTN (x ϭ 0.20 and
x ϭ 0.35) as a function of temperature at a frequency of 1 kHz. It
is easy to see the difference in the magnitude and overall
diffuseness between the two samples.
the material is adsorbed H O. The tape casting and laminating
2
technique produced samples with densities greater than 88% of
theoretical.
The dielectric results show broad transition curves and a
frequency dispersion that is dependent on the temperature of the
measured variables. This is characteristic of relaxor ferroelec-
tric behavior. The temperatures of maximum dielectric con-
stant, Tmax, also exhibit relaxor-type frequency dependence.
Relaxor ferroelectrics undergo a diffuse phase transition be-
tween the low-temperature ferroelectric phase and the high-
temperature paraelectric phase over a wide temperature range.
This behavior is apparent for the KTN material. The decrease in
The diffuseness of the transition for x ϭ 0.35 and x ϭ 0.20 has
1
0
been calculated using the Gaussian distribution relation:
2
2
g
Ermax ϭ E exp͓͑T Ϫ T ͒ /2␦ ͔
r
i
0
the Nb O doping level, from 17.5 to 10 mol%, resulted in a
where Ermax is the maximum relative dielectric constant, E the
2 5
r
decrease of the Tmax from 250 to 120 K. These values differed
significantly from those predicted by Fig. 1. Also evident was
an increase in the average weak-field permittivity for all
frequencies tested. The difference of maximum permittivity
relative dielectric constant, and ␦ the Gaussian diffuseness. The
g
calculated values are included in Tables III and IV. A method of
limiting the temperature range to that corresponding to two-thirds
of the maximum relative dielectric constant was used to calculate
1
0
values and transition temperatures between KTa0.65 Nb0.35O3
the diffuseness. One of the possible reasons for diffuse phase
transitions in ferroelectric solid solutions is that fluctuations in
composition occur. The diffuseness values characterize the breadth
of the transition peak and allow for another way of comparing the
homogeneity of the compositions to each other. The diffuseness
value for the x ϭ 0.35 is higher than those for the x ϭ 0.20
composition. The higher values may indicate that more fluctua-
tions in the composition are present for a larger Nb O doping
and KTa0.80 Nb0.20O suggests that the random occupation of
3
the same site with different cations result in local variations of
composition and the local transition temperature. However,
further work is needed at lower dopant levels and lower
temperatures to confirm the dielectric nature of KTN for
cryogenic use.
The KTN system shows potential as a base material for low-
temperature electroactive dielectrics. The system exhibits typical
2
5
level.
Fig. 10. Dielectric loss for KTN (x ϭ 0.35) as a function of temperature
Fig. 12. Dielectric loss for KTN (x ϭ 0.20) as a function of temperature
and frequency.
and frequency.

2
552
Journal of the American Ceramic Society—DiAntonio and Pilgrim
Vol. 84, No. 11
Table III. Average Curie Temperature (Tmax), Average
Permittivity, and Diffuseness as a Function of Frequency for
relaxor ferroelectric behavior with a strong dependence on tem-
perature and frequency. Extensive characterization is essential in
determining and reporting the electroactive responses of a device.
Often, powder and final sample features are excluded or not given
the level of importance they deserve. Past and present research
indicates that variations in average grain size, percent porosity,
phase purity, etc., can have profound effects on the dielectric
properties of a device. The potential they have to vary properties
of the material and device is enormous, making it essential to
examine the characteristics of a powder system. However, to fully
quantify the potential of the KTN system, the electromechanical
response of these devices needs to be investigated. This will be a
better indicator of the compositions viability in the area of
electroactive devices and is the goal of future work.
KTN (x ؍ 0.35)
Frequency
Average
(max)
Average
permittivity
Diffuseness
(K)
(
kHz)
T
c
1
242 Ϯ 3.0 K
246 Ϯ 1.8 K
248 Ϯ 3.2 K
250 Ϯ 2.5 K
2850 Ϯ 19.0
2750 Ϯ 15.0
2660 Ϯ 15.0
2570 Ϯ 20.0
55.6 Ϯ 1.5
50.2 Ϯ 2.0
51.2 Ϯ 3.5
52.2 Ϯ 4.0
1
0
1
00
000
1
Table IV. Average Curie Temperature (Tmax), Average
Permittivity, and Diffuseness as a Function of Frequency for
(x ؍ 0.20)
Frequency
Average
(max)
Average
Diffuseness
(K)
References
(
kHz)
T
c
permittivity
1
B. T. Matthias, “New Ferroelectric Crystals,” Phys. Rev., 75 [11] 1771 (1949).
J. K Hulm, B. T. Matthias, and E. A. Long, “A Ferromagnetic Curie Point in
KTaO at Very Low Temperatures,” Phys. Rev., 79 [5] 885 (1950).
J. H. Barrett, “Dielectric Constant in Perovskite Type Crystals,” Phys. Rev., 86 [1]
2
1
121 Ϯ 1.0 K
123 Ϯ 2.8 K
124 Ϯ 2.2 K
126 Ϯ 3.5 K
4260 Ϯ 2.5
4120 Ϯ 3.0
3900 Ϯ 6.0
3470 Ϯ 8.5
23.3 Ϯ 2.0
21.5 Ϯ 1.5
22.3 Ϯ 3.0
24.3 Ϯ 4.5
1
0
3
3
1
00
000
1
18 (1952).
1
4
A. Reisman and E. Banks, “Reactions of the Group VB Pentoxides. VIII. Thermal,
Density and X-Ray, Studies of the Systems KNbO -NaNbO and KTaO -KNbO ,”
3
3
3
3
J. Am. Chem. Soc., 80 [8] 1877 (1958).
5
S. Triebwasser, “Study of Ferroelectric Transitions of Solid-Solution Single
Crystals of KNbO –KTaO ,” Phys. Rev., 114 [1] 63 (1959).
3
3
6
A. Reisman, S. Triebwasser, and F. Holtzberger, “Phase Diagram of the System
KNbO –KTaO by the Methods of Differential Thermal and Resistance Analysis,”
3
3
J. Am. Chem. Soc., 77 [16] 4228 (1955).
7
J. D. Siegwarth and J. C. Holste, “Polarization and Dielectric Constant of
Potassium Tantalate,” J. Appl. Phys., 47 [11] 4791–93 (1976).
8
P. E. Debely, P. Gunter, and H. Arend, “Preparation and Electro-optic Properties
of Hot-Pressed Potassium Niobate-Tantalate Ceramic,” Am. Ceram. Soc. Bull., 58 [6]
6
06–609 (1979).
9
A. L. Gentile and F. H. Andres, “A Constant Temperature Method for the Growth
of KTN Single Crystals,” Mater. Res. Bull., 2 [4] 353 (1967).
1
0
S. M. Pilgrim, S. R. Sutherland, and A. E. Winzer, “Diffuseness as a Useful
Parameter for Relaxor Ceramics,” J. Am. Ceram. Soc., 73 [10] 3122–25 (1990).
1
1
Powder Diffraction File Card No. 38-1470. International Centre for Diffraction
Data, Newton Square, PA, 1989.
1
1
2
3
Powder Diffraction File, Card No. 32-0822; see Ref. 11, 1983.
Standard Test Methods for Liquid Absorption, Apparent Porosity, Apparent
“
Specific Density, and Bulk Density by Vacuum Pressure,” ASTM Designation C
830-88, American Society for Testing and Materials, West Conshohocken, PA, 1988.
Fig. 13. Comparison of the weak-field permittivity for KTN (x ϭ 0.20)
and KTN (x ϭ 0.35) as a function of temperature for a frequency of 1 kHz.
Ⅺ