Cryst. Res. Technol. 34 1999 4 519–525
The crystals of KNbO3 have been grown by the micro-pulling-down method. Colorless, transparent, and crack-free crystals were produced from the melts containing excess of K2O as a flux. Growth of relatively large size (up to 2 mm in diameter and up to 30 mm in length) single crystals was found is possible using the crucibles with corresponding nozzle size (up to 2.0 mm in outer diameter). Second harmonic generation was observed in the crystals irradiated by fundamental beam with wavelength about 860 nm.
Keywords: potassium, niobate, fibers, flux, growth, pulling-down
1. Introduction
Potassium niobate (KNbO3 or KN) is a well known ferroelectric material for electro-optic, nonlinear optic, and photorefractive applications (FUKUDA; FLÜCKIGER). It is efficient material for doubling the frequency of near-infrared (Ga, Al)As diode lasers used for recording and reading data from optical compact discs. The information density of optical systems arranged with KN crystal is expected to be four times greater than those of nonarranged ones. However, it is difficult to grow these crystals because KNbO3 melts incongruently at temperature above 1000°C (IMAI; IRLE; REISMAN; ROTH). Therefore the crystals have to be grown from a K2O rich, non-stoichiometric melts. Moreover KNbO3 is known to exist in three phases. High temperature phase crystallizes in the cubic perovskite structure. Within the temperature range 225-435°C KNbO3 has tetragonal structure. At room temperature it is isostructural with the distorted perovskite form of BaTiO3 and has an orthorhombic structure with two formula units per unit cell. Therefore it is also difficult to obtain high quality crystals because of structural reordering that occurs during the crystal cooling. The flux growth technique is widely used to grow KNbO3 crystals from the melts containing K2O excess (FUKUDA; FLÜCKIGER). However reproducibility of growth results is difficult to control because of non-stoichiometry of the starting mixtures and easy volatilization of K2O from the melt and crystal surface (FLÜCKIGER; IMAI). It was reported also (IMAI), that single crystal fibers of K(Ta,Nb)O3 solid solutions were grown by the laser heated pedestal growth method (LHPM). The source rods were enriched with K2O to prevent formation of K4Nb6O17 phase which is stable in the KNbO3 stoichiometric melt without excess of K2O (IRLE; REISMAN; ROTH). The temperature gradient in this method is steep enough to prevent constitutional supercooling and therefore to avoid spontaneous nucleation on the liquid-solid interface. In the case of K(Ta,Nb)O3
The crystals of KNbO3 have been grown by the micro-pulling-down method. Colorless, transparent, and crack-free crystals were produced from the melts containing excess of K2O as a flux. Growth of relatively large size (up to 2 mm in diameter and up to 30 mm in length) single crystals was found is possible using the crucibles with corresponding nozzle size (up to 2.0 mm in outer diameter). Second harmonic generation was observed in the crystals irradiated by fundamental beam with wavelength about 860 nm.
Keywords: potassium, niobate, fibers, flux, growth, pulling-down
1. Introduction
Potassium niobate (KNbO3 or KN) is a well known ferroelectric material for electro-optic, nonlinear optic, and photorefractive applications (FUKUDA; FLÜCKIGER). It is efficient material for doubling the frequency of near-infrared (Ga, Al)As diode lasers used for recording and reading data from optical compact discs. The information density of optical systems arranged with KN crystal is expected to be four times greater than those of nonarranged ones. However, it is difficult to grow these crystals because KNbO3 melts incongruently at temperature above 1000°C (IMAI; IRLE; REISMAN; ROTH). Therefore the crystals have to be grown from a K2O rich, non-stoichiometric melts. Moreover KNbO3 is known to exist in three phases. High temperature phase crystallizes in the cubic perovskite structure. Within the temperature range 225-435°C KNbO3 has tetragonal structure. At room temperature it is isostructural with the distorted perovskite form of BaTiO3 and has an orthorhombic structure with two formula units per unit cell. Therefore it is also difficult to obtain high quality crystals because of structural reordering that occurs during the crystal cooling. The flux growth technique is widely used to grow KNbO3 crystals from the melts containing K2O excess (FUKUDA; FLÜCKIGER). However reproducibility of growth results is difficult to control because of non-stoichiometry of the starting mixtures and easy volatilization of K2O from the melt and crystal surface (FLÜCKIGER; IMAI). It was reported also (IMAI), that single crystal fibers of K(Ta,Nb)O3 solid solutions were grown by the laser heated pedestal growth method (LHPM). The source rods were enriched with K2O to prevent formation of K4Nb6O17 phase which is stable in the KNbO3 stoichiometric melt without excess of K2O (IRLE; REISMAN; ROTH). The temperature gradient in this method is steep enough to prevent constitutional supercooling and therefore to avoid spontaneous nucleation on the liquid-solid interface. In the case of K(Ta,Nb)O3
V.I. CHANI*, K. SHIMAMURA, T. FUKUDA
*
Center for Interdisciplinary Research, Tohoku University, Sendai, 980-8578, Japan
Institute for Materials Research, Tohoku University, Sendai, 980-8577, Japan
Flux Growth of KNbO3 Crystals by Pulling-Down
Method
520 V. I. CHANI et al.: Flux Growth of KNbO3 Fiber Crystals
mixed crystals (IMAI) the addition of a gas blower to the apparatus was necessary to increase
and to control the gradient in the vicinity of the molten zone.
Micro-pulling-down (m-PD) technique (YOON) is reported in the present paper to be a
versatile method of preparation of high quality KNbO3 fiber crystals starting from the melts
of non-stoichiometric composition. We discuss the experimental procedure and the growth
parameters which allow us to produce relatively large KNbO3 crystals.
2. Growth procedure
The starting materials with various compositions were made using K2CO3 (Rare Metallic
Co.) and Nb2O5 (High Purity Chem. Lab.), both of 99.99% purity. Desired quantities of
compounds were carefully weighed and mixed by grinding in ethanol with an agate mortar
and pestle, and dried at 100°C during 5-10 hr. Special attention was paid to prepare waterfree
K2CO3. Therefore preliminary annealing of starting K2CO3 at temperature 350-400°C
during 5-10 hr was necessary. The compositions of the melts used in our experiments are
summarized in Table 1.
No. K2O Nb2O5 Nozzle Seed Pulling
rate
Length
12-2 40 60 1.2x1.0 Pt wire 0.35 35
1-2 50 50 1.2x1.1 Pt wire 0.11 32
4-2 50 50 1.2x1.0 Pt wire 0.10 8
3-3 52 48 1.2x1.0 Pt wire 0.08 10
7-1 54 46 1.2x1.0 Pt tube 0.30 50
7-2 54 46 1.4x1.3 No. 7-1 0.15 34
6-1 55 45 1.2x1.0 Pt wire 0.13 25
8-1 56 44 1.2x1.0 Pt wire 0.09 22
8-2 56 44 1.2x1.0 Pt tube 0.20 45
8-3 56 44 1.4x1.3 Pt tube 0.12 30
8-4 56 44 1.4x1.3 KTN
(flux)
0.13 37
15-1 57 43 1.7x1.6 Pt tube 0.10 30
15-3 57 43 2.0x1.9 Pt tube 0.10 19
11-1 58 42 1.2x1.0 Pt wire 0.11 13
11-2 58 42 1.2x1.0 Pt tube 0.16 27
11-4 58 42 1.2x1.0 Pt tube 0.30 70
11-5 58 42 2.0x1.9 Pt tube 0.10 20
13-1 62 38 1.2x1.0 Pt tube 0.15 18
Table 1.:Crystal growth conditions: melts composition (mol.%) outer and inner diameters of the
crucible nozzle (mm), seed material, pulling rate (mm/min), and crystal length (mm)
Schematic diagram of the m-PD system and the details of the experimental technique are
described in the above-mentioned paper (YOON). The crystals were grown under air
atmosphere. The melt was contained in a crucible, which was made of Pt plate of 0.1 mm
thickness and Pt pipe of 1.0-2.0 mm in outer diameter and a wall thickness of 0.05-0.01 mm,
as it is shown in Fig. 1. We will call the modified arrangement (large nozzle diameter) as
pulling-down (PD) technique to separate it from the conventional (m-PD) system.
In the main two variations of the seeding technique were used in the experiments
described here; solidification of the melt was started on Pt wire of 0.5 mm in diameter or Pt
pipe of 0.4 mm in outer diameter and a wall thickens of 0.05 mm. Using of KNbO3 single
Cryst. Res. Technol. 34 (1999) 4 521
crystal seed was also possible, but it was difficult because of cracking which occurred
during heating or seeding. Extremely high temperature gradient under the crucible nozzle
and phase transitions mentioned above are considered to be main causes of the cracking.
K(Nb,Ta)O3 seed crystals grown by conventional flux technique were used also to prevent
these disadvantages. The crystal grown on K(Nb,Ta)O3 seed (sample No. 8-4 of Tables 1
and 2) was crack-free.
Fig. 1. Schematic diagram of seeding procedure
using Pt pipe as a seed.
In case of seeding on Pt wire one crystallographic direction of high growth rate was
developed by pulling down rate of 0.50-1.00 mm/min and necking procedure. Further
orienting of the crystal was made during the growth process by manipulations of micro X-Y
stage. It was possible due to faceting of the crystals observed in situ by optical microscope.
No. W (mg) D (mm) Color Crystal Remain Melt Cracks
12-2 150 1.0 colorless K4Nb6O17 K4Nb6O17 Yes
1-2 - 1.0 colorless K4Nb6O17 K4Nb6O17 -
4-2 - 1.1 blue KNbO3 - -
3-3 39 1.1 colorless KNbO3 K4Nb6O17 Yes
7-1 206 1.1 colorless KNbO3 - Yes
7-2 171 1.3 dark blue KNbO3 - Yes
6-1 94 1.1 colorless KNbO3 K4Nb6O17 No
8-1 77 1.1 light blue KNbO3 K4Nb6O17 No
8-2 207 1.1 colorless KNbO3 - No
8-3 177 1.3 colorless KNbO3 - No
8-4 236 1.4 light blue KNbO3 - No
15-1 246 1.8 colorless KNbO3 K4Nb6O17 No
15-3 239 2.1 light blue KNbO3 K4Nb6O17 No
11-1 52 1.2 light blue KNbO3 K4Nb6O17 Yes
11-2 93 1.1 colorless KNbO3 KNbO3 No
11-4 255 1.1 light blue KNbO3 - No
11-5 231 2.1 colorless KNbO3 - No
13-1 57 1.0 blue KNbO3 KNbO3 Yes
Table 2. : Crystal growth results: weight (W), diameter (D), color and phase of the crystals grown and remain
melts
Best results were achieved in the runs where Pt tube was used as a seed similar to that of
described earlier (KIMURA) for the crystal growth by Czochralski method. Schematic
522 V. I. CHANI et al.: Flux Growth of KNbO3 Fiber Crystals
illustration of the seeding technique used is illustrated by Fig. 1. As a first step the tube was
inserted into the crucible nozzle and kept there about 1 min. At that time overheating of the
crucible was necessary to prevent solidification of the melt inside the nozzle because of high
thermoconductivity of platinum. Thereafter pulling down was started with a rate close to
that of used for crystal growth as it is given in Table 1.
All growth processes were stopped after observation of any kind of crystal imperfection.
After that the crystals were disconnected from the molten zone and pulled down with the
rate corresponding to cooling rate of about 30°C/min. Thereafter the crystals were removed
from the seed holder. As a second stage of each experiment the remain melt was removed
from the crucible using the same seeding material with a pulling rate of about 0.50 mm/min.
Deposition of small drops of a flux (K2O) was sometimes observed on the surface of the
crystals. Therefore the crystals were washed in warm water.
3. Growth results and discussion.
Fig. 2 shows the as grown KNbO3 crystals. The fibers grown had a habit corresponding to
published data (FUKUDA). The crystals showed simple crystallographic {100} faces because
of presence of flux. In the main the rod-like crystals had four-fold symmetry corresponding
to [100] orientation of pseudo-cubic high temperature phase (ZENG). Similar to the crystals
grown by top-seeded solution growth, the ones reported here had very flat cubic faces
because the progressive nucleation on cubic planes is quite difficult (HULLIGER).
Fig. 2. View of KNbO3 crystals (sample No. 11-2
above and sample No. 15-1 below) grown by PD
technique (scale in mm).
The crystals were blue and colorless depending on melt composition and pulling rate. In
general optimization of crystal growth conditions was necessary for all of the melts reported
here, because at least light blue coloration following from presence of some amount of
oxygen vacancies (FUKUDA; IMAI) was observed in all crystals grown at relatively high
pulling rate. Almost all crystals were transparent, as shown in Fig. 2. The typical size of
crystals was about 1-2 mm in cross-section depending on diameter of the nozzle and few
centimeters in length. In the main about 70-80 vol.% of the melt was crystallized into
KNbO3 single crystals. Maximum yield achieved (crystal/melt volume ratio) was about 90
Cryst. Res. Technol. 34 (1999) 4 523
vol.% for the melts containing insignificant excess of K2O.
It was also possible to grow the KNbO3 single crystals with an extremely high pulling rate of
about 1-2 mm/min. In such a case the crystals also were single phase and had typical fourfold
symmetry. However these crystals were dark blue in color.
Phase homogeneity of the crystals grown and the melts remain after growths were
studied by X-ray powder diffraction analysis. In the main the crystals grown were KNbO3
single phase (JSPDS data card No. 32-822) as it is given in Table 2. The remain melts were
found crystallized as either KNbO3 or K4Nb6O17 depending on composition of starting
mixture.
In the melts corresponding to the vicinity of stoichiometric composition of KNbO3
crystallization of the second phase was often observed. It was assumed that the phase is
K4Nb6O17. However the phase identification was difficult because X-ray diffraction data for
K4Nb6O17 compound found in JSPDS data cards are very different (cards No. 14-287, 21-
1295, 31-1063, and 31-1064). Therefore it was necessary to prepare this material by
ourselves. The K4Nb6O17 compound was produced by solid state reaction technique.
Moreover the K4Nb6O17 single crystals were grown by the PD method from stoichiometric
melt of the above composition using the procedure similar to that of described above. The
K4Nb6O17 crystals were transparent, colorless, and were well developed in shape. X-ray
diffraction pictures of the poly- and single-crystalline samples were very similar. X-ray
diffraction data are given in Table 3. The results of Table 3 show relatively high correlation
with the data found in JSPDS data card No. 14-287.
2q dobs I/I0
9.15 9.66 100
13.95 6.35 20
16.58 5.35 16
19.63 4.52 13
21.20 4.19 13
23.00 3.87 16
27.58 3.23 46
29.90 2.99 33
31.55 2.84 77
35.95 2.50 17
38.10 2.36 15
40.28 2.24 21
45.30 2.00 16
46.20 1.96 26
51.20 1.78 15
58.60 1.58 16
Table 3: X-ray powder pattern of the single crystal grown
from K4Nb6O17 stoichiometric melt in the range of 2q = 6-60°
(sample No. 12-2 of Tables 1 and 2)
The KN crystals were cut and polished. Second harmonic generation (SHG) was observed in
the samples with fundamental laser beam irradiated along the growth axis (a-axis).
4. PD and related growth methods
Comparison of most common features of PD and related growth techniques is given in
Table 4. Two most important advantages of the m-PD technique modified by increasing of
524 V. I. CHANI et al.: Flux Growth of KNbO3 Fiber Crystals
the diameter of capillary channel (PD) are discussed below briefly.
Methods m-PD and LHPM* PD Flux Growth
Crystal Diameter ~ 0.5 mm ³ 2.0 mm ³ 10 mm
Segregation
Coefficients
K » 1 K» 1 K» 1
Flux Growth Difficult because
K » 1
Possible Possible
Growth Rate
(mm/min)
Very high (0.1-1.0) Very high (0.1-1.0) Very low (< 0.01)
Growth control Easy Easy Difficult
LHPM* - laser heating pedestal method
Table 4: Comparison of PD and related techniques
4.1. Flux growth
One of the most important result of this study seems to be related with presence of
considerable amount of flux in the starting melts. For example, recalculation of the melts
used for the growth of the crystals No. 11-5 and No. 13-1 (Table 1) results KNbO3 : K2O =
84 : 16 and KNbO3 : K2O = 76 : 24 molar ratios, respectively.
In the case of conventional m-PD growth reported earlier (YOON) the crucibles were
fabricated with a nozzle diameter less than 1 mm. In such a case the segregation phenomena
was not observed because of low mass transport inside the narrow capillary channel.
Therefore intensity of the cations exchange between the liquid and solid phases was very
low. Same phenomena is usually observed in the LHPM crystal growth (IMAI). In both these
methods the segregation coefficients reported were close to unity: K » 1. However, in the
modified PD arrangement reported here the diameter of capillary channel has been increased
considerably up to 2 mm that is close to size of the crucible (about 10 x 5 x 2 mm, as it is
shown in Fig. 1). This way, the rate of natural convection has been increased also, and the
segregation on the liquid/solid interface was observed become possible (K » 1). Thus the
modification discussed here results unusual possibility of flux growth with high pulling rate
(0.1-1.0 mm/min).
4.2. Growth of macro-crystals
Another important result is related with examination of macro-limitations of the m-PD
system. In all previous reports concerning oxide crystal growth by m-PD technique the
diameter (or cross-section) of the fibers reported did not exceed 1 mm. The maximal size of
the KNbO3 crystals grown in this study was greater than 2 mm. Therefore fields of
application of the PD technique and the crystals reported is assumed increase considerably.
5. Conclusions
KNbO3 single crystals were grown by the pulling-down technique. The crystals were SHG
active. It was found that combination of the conventional pulling-down system with the
crucible arranged with large capillary channel can be used to produce relatively large macrocrystals
with diameter greater than 2 mm by flux method.
Cryst. Res. Technol. 34 (1999) 4 525
Acknowledgments
We thank Dr. K. Imai (NGK Insulators, Ltd.) for preliminary results related with measurements of
SHG in the KN crystals reported.
References
FLÜCKIGER, U., AREND, H., J.: Cryst. Growth 43 (1978) 406
FUKUDA, T., UEMATSU, Y.: Jpn. J. Appl. Phys. 11 (1972) 163
HULLIGER, J., GUTMANN, R., WÜEST, H.: J. Cryst. Growth 128 (1993) 897
IMAI, T., YAGI, S., SUGIYAMA, Y., HATAKEYAMA, I.: J. Cryst. Growth 147 (1995) 350
IRLE, E., BLACHNIK, R., GATHER, B., Thermochim. Acta 179 (1991) 157
KIMURA, H., NUMAZAVA, T., SATO, M., J. Cryst. Growth 165 (1996) 408
REISMAN, A., HOLTZBERG, F., J. Am. Chem. Soc., 77 (1955) 2117
ROTH, R. S., Progr. Solid State Chem, 13 (1980) 159
YOON, D.H., YONENAGA, I., FUKUDA, T., OHNISHI, N., J. Cryst. Growth 142 (1994) 339
ZENG, H. C., J.Cryst.Growth 173 (1997) 446
(received October 7, 1998; accepted October 30, 1998)
Author's address:
Dr. V.I.CHANI
Center for Interdisciplinary Research
Tohoku University
Aramaki aza Aoba, Aoba-ku
Sendai, 980-8578
Japan
Fax: +81 (022) 215-2104
E-mail: chani@lexus.imr.tohoku.ac.jp
*
Center for Interdisciplinary Research, Tohoku University, Sendai, 980-8578, Japan
Institute for Materials Research, Tohoku University, Sendai, 980-8577, Japan
Flux Growth of KNbO3 Crystals by Pulling-Down
Method
520 V. I. CHANI et al.: Flux Growth of KNbO3 Fiber Crystals
mixed crystals (IMAI) the addition of a gas blower to the apparatus was necessary to increase
and to control the gradient in the vicinity of the molten zone.
Micro-pulling-down (m-PD) technique (YOON) is reported in the present paper to be a
versatile method of preparation of high quality KNbO3 fiber crystals starting from the melts
of non-stoichiometric composition. We discuss the experimental procedure and the growth
parameters which allow us to produce relatively large KNbO3 crystals.
2. Growth procedure
The starting materials with various compositions were made using K2CO3 (Rare Metallic
Co.) and Nb2O5 (High Purity Chem. Lab.), both of 99.99% purity. Desired quantities of
compounds were carefully weighed and mixed by grinding in ethanol with an agate mortar
and pestle, and dried at 100°C during 5-10 hr. Special attention was paid to prepare waterfree
K2CO3. Therefore preliminary annealing of starting K2CO3 at temperature 350-400°C
during 5-10 hr was necessary. The compositions of the melts used in our experiments are
summarized in Table 1.
No. K2O Nb2O5 Nozzle Seed Pulling
rate
Length
12-2 40 60 1.2x1.0 Pt wire 0.35 35
1-2 50 50 1.2x1.1 Pt wire 0.11 32
4-2 50 50 1.2x1.0 Pt wire 0.10 8
3-3 52 48 1.2x1.0 Pt wire 0.08 10
7-1 54 46 1.2x1.0 Pt tube 0.30 50
7-2 54 46 1.4x1.3 No. 7-1 0.15 34
6-1 55 45 1.2x1.0 Pt wire 0.13 25
8-1 56 44 1.2x1.0 Pt wire 0.09 22
8-2 56 44 1.2x1.0 Pt tube 0.20 45
8-3 56 44 1.4x1.3 Pt tube 0.12 30
8-4 56 44 1.4x1.3 KTN
(flux)
0.13 37
15-1 57 43 1.7x1.6 Pt tube 0.10 30
15-3 57 43 2.0x1.9 Pt tube 0.10 19
11-1 58 42 1.2x1.0 Pt wire 0.11 13
11-2 58 42 1.2x1.0 Pt tube 0.16 27
11-4 58 42 1.2x1.0 Pt tube 0.30 70
11-5 58 42 2.0x1.9 Pt tube 0.10 20
13-1 62 38 1.2x1.0 Pt tube 0.15 18
Table 1.:Crystal growth conditions: melts composition (mol.%) outer and inner diameters of the
crucible nozzle (mm), seed material, pulling rate (mm/min), and crystal length (mm)
Schematic diagram of the m-PD system and the details of the experimental technique are
described in the above-mentioned paper (YOON). The crystals were grown under air
atmosphere. The melt was contained in a crucible, which was made of Pt plate of 0.1 mm
thickness and Pt pipe of 1.0-2.0 mm in outer diameter and a wall thickness of 0.05-0.01 mm,
as it is shown in Fig. 1. We will call the modified arrangement (large nozzle diameter) as
pulling-down (PD) technique to separate it from the conventional (m-PD) system.
In the main two variations of the seeding technique were used in the experiments
described here; solidification of the melt was started on Pt wire of 0.5 mm in diameter or Pt
pipe of 0.4 mm in outer diameter and a wall thickens of 0.05 mm. Using of KNbO3 single
Cryst. Res. Technol. 34 (1999) 4 521
crystal seed was also possible, but it was difficult because of cracking which occurred
during heating or seeding. Extremely high temperature gradient under the crucible nozzle
and phase transitions mentioned above are considered to be main causes of the cracking.
K(Nb,Ta)O3 seed crystals grown by conventional flux technique were used also to prevent
these disadvantages. The crystal grown on K(Nb,Ta)O3 seed (sample No. 8-4 of Tables 1
and 2) was crack-free.
Fig. 1. Schematic diagram of seeding procedure
using Pt pipe as a seed.
In case of seeding on Pt wire one crystallographic direction of high growth rate was
developed by pulling down rate of 0.50-1.00 mm/min and necking procedure. Further
orienting of the crystal was made during the growth process by manipulations of micro X-Y
stage. It was possible due to faceting of the crystals observed in situ by optical microscope.
No. W (mg) D (mm) Color Crystal Remain Melt Cracks
12-2 150 1.0 colorless K4Nb6O17 K4Nb6O17 Yes
1-2 - 1.0 colorless K4Nb6O17 K4Nb6O17 -
4-2 - 1.1 blue KNbO3 - -
3-3 39 1.1 colorless KNbO3 K4Nb6O17 Yes
7-1 206 1.1 colorless KNbO3 - Yes
7-2 171 1.3 dark blue KNbO3 - Yes
6-1 94 1.1 colorless KNbO3 K4Nb6O17 No
8-1 77 1.1 light blue KNbO3 K4Nb6O17 No
8-2 207 1.1 colorless KNbO3 - No
8-3 177 1.3 colorless KNbO3 - No
8-4 236 1.4 light blue KNbO3 - No
15-1 246 1.8 colorless KNbO3 K4Nb6O17 No
15-3 239 2.1 light blue KNbO3 K4Nb6O17 No
11-1 52 1.2 light blue KNbO3 K4Nb6O17 Yes
11-2 93 1.1 colorless KNbO3 KNbO3 No
11-4 255 1.1 light blue KNbO3 - No
11-5 231 2.1 colorless KNbO3 - No
13-1 57 1.0 blue KNbO3 KNbO3 Yes
Table 2. : Crystal growth results: weight (W), diameter (D), color and phase of the crystals grown and remain
melts
Best results were achieved in the runs where Pt tube was used as a seed similar to that of
described earlier (KIMURA) for the crystal growth by Czochralski method. Schematic
522 V. I. CHANI et al.: Flux Growth of KNbO3 Fiber Crystals
illustration of the seeding technique used is illustrated by Fig. 1. As a first step the tube was
inserted into the crucible nozzle and kept there about 1 min. At that time overheating of the
crucible was necessary to prevent solidification of the melt inside the nozzle because of high
thermoconductivity of platinum. Thereafter pulling down was started with a rate close to
that of used for crystal growth as it is given in Table 1.
All growth processes were stopped after observation of any kind of crystal imperfection.
After that the crystals were disconnected from the molten zone and pulled down with the
rate corresponding to cooling rate of about 30°C/min. Thereafter the crystals were removed
from the seed holder. As a second stage of each experiment the remain melt was removed
from the crucible using the same seeding material with a pulling rate of about 0.50 mm/min.
Deposition of small drops of a flux (K2O) was sometimes observed on the surface of the
crystals. Therefore the crystals were washed in warm water.
3. Growth results and discussion.
Fig. 2 shows the as grown KNbO3 crystals. The fibers grown had a habit corresponding to
published data (FUKUDA). The crystals showed simple crystallographic {100} faces because
of presence of flux. In the main the rod-like crystals had four-fold symmetry corresponding
to [100] orientation of pseudo-cubic high temperature phase (ZENG). Similar to the crystals
grown by top-seeded solution growth, the ones reported here had very flat cubic faces
because the progressive nucleation on cubic planes is quite difficult (HULLIGER).
Fig. 2. View of KNbO3 crystals (sample No. 11-2
above and sample No. 15-1 below) grown by PD
technique (scale in mm).
The crystals were blue and colorless depending on melt composition and pulling rate. In
general optimization of crystal growth conditions was necessary for all of the melts reported
here, because at least light blue coloration following from presence of some amount of
oxygen vacancies (FUKUDA; IMAI) was observed in all crystals grown at relatively high
pulling rate. Almost all crystals were transparent, as shown in Fig. 2. The typical size of
crystals was about 1-2 mm in cross-section depending on diameter of the nozzle and few
centimeters in length. In the main about 70-80 vol.% of the melt was crystallized into
KNbO3 single crystals. Maximum yield achieved (crystal/melt volume ratio) was about 90
Cryst. Res. Technol. 34 (1999) 4 523
vol.% for the melts containing insignificant excess of K2O.
It was also possible to grow the KNbO3 single crystals with an extremely high pulling rate of
about 1-2 mm/min. In such a case the crystals also were single phase and had typical fourfold
symmetry. However these crystals were dark blue in color.
Phase homogeneity of the crystals grown and the melts remain after growths were
studied by X-ray powder diffraction analysis. In the main the crystals grown were KNbO3
single phase (JSPDS data card No. 32-822) as it is given in Table 2. The remain melts were
found crystallized as either KNbO3 or K4Nb6O17 depending on composition of starting
mixture.
In the melts corresponding to the vicinity of stoichiometric composition of KNbO3
crystallization of the second phase was often observed. It was assumed that the phase is
K4Nb6O17. However the phase identification was difficult because X-ray diffraction data for
K4Nb6O17 compound found in JSPDS data cards are very different (cards No. 14-287, 21-
1295, 31-1063, and 31-1064). Therefore it was necessary to prepare this material by
ourselves. The K4Nb6O17 compound was produced by solid state reaction technique.
Moreover the K4Nb6O17 single crystals were grown by the PD method from stoichiometric
melt of the above composition using the procedure similar to that of described above. The
K4Nb6O17 crystals were transparent, colorless, and were well developed in shape. X-ray
diffraction pictures of the poly- and single-crystalline samples were very similar. X-ray
diffraction data are given in Table 3. The results of Table 3 show relatively high correlation
with the data found in JSPDS data card No. 14-287.
2q dobs I/I0
9.15 9.66 100
13.95 6.35 20
16.58 5.35 16
19.63 4.52 13
21.20 4.19 13
23.00 3.87 16
27.58 3.23 46
29.90 2.99 33
31.55 2.84 77
35.95 2.50 17
38.10 2.36 15
40.28 2.24 21
45.30 2.00 16
46.20 1.96 26
51.20 1.78 15
58.60 1.58 16
Table 3: X-ray powder pattern of the single crystal grown
from K4Nb6O17 stoichiometric melt in the range of 2q = 6-60°
(sample No. 12-2 of Tables 1 and 2)
The KN crystals were cut and polished. Second harmonic generation (SHG) was observed in
the samples with fundamental laser beam irradiated along the growth axis (a-axis).
4. PD and related growth methods
Comparison of most common features of PD and related growth techniques is given in
Table 4. Two most important advantages of the m-PD technique modified by increasing of
524 V. I. CHANI et al.: Flux Growth of KNbO3 Fiber Crystals
the diameter of capillary channel (PD) are discussed below briefly.
Methods m-PD and LHPM* PD Flux Growth
Crystal Diameter ~ 0.5 mm ³ 2.0 mm ³ 10 mm
Segregation
Coefficients
K » 1 K» 1 K» 1
Flux Growth Difficult because
K » 1
Possible Possible
Growth Rate
(mm/min)
Very high (0.1-1.0) Very high (0.1-1.0) Very low (< 0.01)
Growth control Easy Easy Difficult
LHPM* - laser heating pedestal method
Table 4: Comparison of PD and related techniques
4.1. Flux growth
One of the most important result of this study seems to be related with presence of
considerable amount of flux in the starting melts. For example, recalculation of the melts
used for the growth of the crystals No. 11-5 and No. 13-1 (Table 1) results KNbO3 : K2O =
84 : 16 and KNbO3 : K2O = 76 : 24 molar ratios, respectively.
In the case of conventional m-PD growth reported earlier (YOON) the crucibles were
fabricated with a nozzle diameter less than 1 mm. In such a case the segregation phenomena
was not observed because of low mass transport inside the narrow capillary channel.
Therefore intensity of the cations exchange between the liquid and solid phases was very
low. Same phenomena is usually observed in the LHPM crystal growth (IMAI). In both these
methods the segregation coefficients reported were close to unity: K » 1. However, in the
modified PD arrangement reported here the diameter of capillary channel has been increased
considerably up to 2 mm that is close to size of the crucible (about 10 x 5 x 2 mm, as it is
shown in Fig. 1). This way, the rate of natural convection has been increased also, and the
segregation on the liquid/solid interface was observed become possible (K » 1). Thus the
modification discussed here results unusual possibility of flux growth with high pulling rate
(0.1-1.0 mm/min).
4.2. Growth of macro-crystals
Another important result is related with examination of macro-limitations of the m-PD
system. In all previous reports concerning oxide crystal growth by m-PD technique the
diameter (or cross-section) of the fibers reported did not exceed 1 mm. The maximal size of
the KNbO3 crystals grown in this study was greater than 2 mm. Therefore fields of
application of the PD technique and the crystals reported is assumed increase considerably.
5. Conclusions
KNbO3 single crystals were grown by the pulling-down technique. The crystals were SHG
active. It was found that combination of the conventional pulling-down system with the
crucible arranged with large capillary channel can be used to produce relatively large macrocrystals
with diameter greater than 2 mm by flux method.
Cryst. Res. Technol. 34 (1999) 4 525
Acknowledgments
We thank Dr. K. Imai (NGK Insulators, Ltd.) for preliminary results related with measurements of
SHG in the KN crystals reported.
References
FLÜCKIGER, U., AREND, H., J.: Cryst. Growth 43 (1978) 406
FUKUDA, T., UEMATSU, Y.: Jpn. J. Appl. Phys. 11 (1972) 163
HULLIGER, J., GUTMANN, R., WÜEST, H.: J. Cryst. Growth 128 (1993) 897
IMAI, T., YAGI, S., SUGIYAMA, Y., HATAKEYAMA, I.: J. Cryst. Growth 147 (1995) 350
IRLE, E., BLACHNIK, R., GATHER, B., Thermochim. Acta 179 (1991) 157
KIMURA, H., NUMAZAVA, T., SATO, M., J. Cryst. Growth 165 (1996) 408
REISMAN, A., HOLTZBERG, F., J. Am. Chem. Soc., 77 (1955) 2117
ROTH, R. S., Progr. Solid State Chem, 13 (1980) 159
YOON, D.H., YONENAGA, I., FUKUDA, T., OHNISHI, N., J. Cryst. Growth 142 (1994) 339
ZENG, H. C., J.Cryst.Growth 173 (1997) 446
(received October 7, 1998; accepted October 30, 1998)
Author's address:
Dr. V.I.CHANI
Center for Interdisciplinary Research
Tohoku University
Aramaki aza Aoba, Aoba-ku
Sendai, 980-8578
Japan
Fax: +81 (022) 215-2104
E-mail: chani@lexus.imr.tohoku.ac.jp
