Czochralski growth of lithium tetraborate single crystals

Czochralski growth of lithium tetraborate single crystals

Volume 11,number 8,9 MATERlALS July 199 LETTERS I Czochralski growth of lithium tetraborate single crystals Tadeusz tukasiewicz Institute ofTec...

279KB Sizes 3 Downloads 127 Views


11,number 8,9


July 199



Czochralski growth of lithium tetraborate single crystals Tadeusz tukasiewicz Institute



I February

and Andrzej Majchrowski WIT, 00-908 Warsaw Poland


Single crystals of Li2B407, substrate material for surface acoustic wave (SAW ) devices, have been grown by the Czochralski technique. The length of the crystals was up to 60 mm and the diameter up to I5 mm. Regions with inverted orientation were found by chemical etching. The density of dislocations, as revealed by etch pits. was 9 x lo3 cmeZ. Properties of some simple SAW devices are presented.

1. Introduction

of its low photon energy dependence [4]

Microwave devices using surface acoustic waves are in common use for IF filters for colour television and radar signal processing elements. New polar materials for SAW devices have been examined so as to obtain single crystals with better properties than the two materials most commonly used nowadays: aquartz and lithium niobate. These two materials have serious shortcomings for certain SAW applications [ I]. For instance, the refativefy poor electromechanical coupling of u-quartz makes it unsuitable for devices with large bandwidth while the poor temperature stability of SAW velocities in LiNb03 makes it unsuitable for devices with high environmental stability. A very promising new acoustic material is tetragonal lithium tetraborate LiZB407. The first single crystals were obtained by Garrett et al. [2]. The quality of their crystals was rather poor. Considerable improvement of the quality was obtained by Robertson et al. [ 3 1. Measurements of L&B.,O, SAW devices revealed their considerable acoustic properties for use as SAW substrates. The material’s electromechanical coupling coefficient for SAW is 1.06%. It is about ten times higher than for u-quartz. The temperature coefficient of the SAW velocity is 3x 10e6/K; the coefficient for YZ-cut lithium niobate is thirty times higher. LiZB407, with some dopants (e.g. Mn), seems to be a promising thermoluminescent material because 0167-577x/91/$

03.50 0 1991 - Eisevier Science Publishers

2. Preparation As starting materials, high-purity lithium bicarbonate Li2C03 and boric acid HsBO,, produced by Polish Chemical Reagents Factory, were used. The phase diagram for the LizO-Bz03 system shows the existence of two congruently melting compounds: monoclinic LiBOz with m.p. 849°C and tetragonal Li,B,O, with m.p. 9 17°C [ 5 ]. The latter is useful as a substrate for SAW devices and growth of this material has been investigated. Melts were prepared with mole ratio 2Bz03 : 1Liz0 in platinum crucibles. First, the boric acid is melted. Before and after the melting process the crucible was weighed. This enabled us to find the real mass of BzOJ in the container and to calculate the appropriate quantity of lithium bicarbonate. Addition of L&CO3 to the melted B,03 protects the platinum against corrosive attack by this reagent. The crucible was heated by rf and the melted material was completely transparent.

3. Experimental results Crystal growth was carried out in the apparatus shown schematically in fig. 1. We used resistance heating with silicon carbide elements placed in a horizontal plane around the crucible walls. It al-

B.V. (North-Holland



Volume 11.number 8,9

July 1991





HIcml Fig. 2. Axial temperature distribution in the furnace.





Schematic diagram of growth apparatus

lowed us to obtain a suitably high temperature gradient above the surface of the melt. A highly stable controller maintained the temperature under the bottom of the crucible to an accuracy of J- 0.1 ‘C. The crucible is placed on a ceramic pedestal. The temperature gradient above the melt was additionally adjusted by the crucible position inside the

heater. The pulling mechanism allowed us to change the pulling speed of the crystal from 0.2 to 10 mm/ h with rotations from a few to eighty r-pm. Seed crystals with [ 0011 and [ IOO] orientations were obtained on platinum wire. We carried out experiments using different temperature gradients at the interface. Crystal rotation rates from 2 to 30 rpm and pulling rates in the range of 0.2 to 3 mm/h were abo investigated, We found that the axial temperature gradient in the vicinity of the surface of the melted material has to be kept above 200 K/cm. With lower axial temperature gradients, single-crystal growth does not occur and a glass is formed. Fig. 2

Fig. 3. [email protected], single crystai. The striations on the surface of the crystal are due ta decreasing the temperature of the melt and do not deteriorate the optical quality of the crystal. The scale is in cm.



11, number



shows the temperature distribution near the interfacial surface and above the melt. The best results were obtained with a seed rotation of 5 rpm and a pulling rate of 0.5 mm/h. A typical pulled crystal is shown in fig. 3. Higher pulling rates caused formation of cores, which had a tendency to increase in width with growing pulling rate. The same effect was observed when the temperature of the melt was decreased too rapidly. When the seed rotation was increased above 10 rpm, the shape of the interface became strongly concave towards the crystal, which caused unstable growth or even disconnection of the growing crystal from the melt. During growth the melt evaporates. Analysis of a powder condensed on cold walls of the furnace revealed higher contents of B203 than in the melt. We have found 25.7% of B and 4.4% of Li (by weight) which corresponds approximately to the compound Liz0.4Bz03, whereas in lithium tetraborate, those values are respectively 25.6% and 8.2%. In connection with changes in stoichiometry of the melt, a diminishing of the diameter of the growing crystal occurred if the temperature of the melt was not strongly decreased. To decrease this effect, we used an excess of Bz03 (up to 2 mol%). The quality of the pulled single crystals was investigated by a chemical etching method. [OOl ] planes were etched in a mixture of glycerine and water, in a volume ratio of 1 : 1, at 25°C for 24 h. The mixture revealed regions with orientation rotated by 180” and which extended along the crystals. This enabled us to select seeds free of these defects. The density of dislocations, as revealed by etch pits, in our single crystals of Li,B,O, was 9x lo3 cm-‘. The lattice parameters of Li,B,O, single crystals were: a=9.477 A, c= 10.286 A. Measurements of SAW properties on Z-propagating X-cut LiZB407 substrates were made. The length of the plate was 30 mm and the width was 10 mm. Simple interdigital transducers worked at 62.5 and 3 13 MHz with losses of 16 and 36 dB respectively. In the range of 65-200 MHz, strong bulk wave formation was observed. The dielectric constant of the material is t= 10, when the SAW coupling factor k’ exceeds 1%. The SAW velocity is equal to 3530 m/ s. The temperature coefficient of delay (TCD) was measured in the range of 25-70’ C and shows a quasiparabolic dependence with a peak at 55 “C. Zero TCD was observed between 45 and 50°C; between 25 and

July I99

















[“Cl Fig. 4. Temperature dependence of the relative change of delay As/r. of SAW waves in X-cut Z-propagating Li2B,0,.

45°C the TCD is equal to 20 ppm/“C

(fig. 4).

4. Conclusion We have determined the optimum conditions for growth of high-quality single crystals of lithium tetraborate by the Czochralski method. Boules of Li,B,O, up to 60 mm in length with circular cross sections up to 15 mm in diameter were obtained (fig. 3). The crystals were transparent and colourless. Some of them contained narrow cores in their central parts. We have found that this kind of defect is due to increasing the pulling rate and decreasing the temperature gradient. Acknowledgement The authors would like to thank Dr. E. Wieteska for chemical analysis quoted in this paper. References [I ] R. Whatmore, J. Crystal Growth 48 ( 1980) 530. [2] J.D. Garrett, M. Natarajan-Iyer and J.E. Greedan, J. Crystal Growth41 (1977) 225. [3] D.S. Robertson and I.M. Young, J. Mater. Sci. 17 ( 1982) 1729. [ 41 B. Chandra and R.C. Bhatt, Nucl. Instr. Methods 184 ( I98 1) 557. [ 51B.S.R. Sastry and F.A. Hummel, J. Am. Ceram. Sot. 42 (1959) 216.