POLYMORPHISM IN LiGeBO4
P. Becker, L. Bohatý and J. Liebertz
Institute of Crystallography, University of
Cologne, Zülpicher Str. 49b, D-50674 Köln, Germany
e-mail: beckerp@kri.uni-koeln.de
Keywords: LiGeBO4,
phase transition
Lithium germanate borate, LiGeBO4, has been considered as a potential material for nonlinear optics because of its non-centrosymmetry. Erroneously described by Ihara [1] to be orthorhombic (pseudotetragonal), its real symmetry at room temperature is I4_, as proposed by Liebertz and Stähr [2] and, in a more recent paper, verified by Parise and Gier [3] by means of single crystal X-ray structure determination.
Large single crystals of LiGeBO4 were grown from highly viscous stoichiometric melts using a modified Czochralski technique through slow cooling from about 1175 to 1125 K in a period of 3-4 weeks without crystal rotation. The crystals obtain sizes of up to 45 mm (diameter) and 25 mm length, but are always partially cloudy because of inclusions and cracks. The strategy of crystal growth from highly viscous melts has been described by Liebertz [4].
Crystals grown with this technique always show a lamellar twinning which can easily be observed with the polarizing microscope (see Fig.1). The question of the origin of these twins, which do not occur in the very small crystals grown hydrothermally at about 873K [3], led us to a further investigation of LiGeBO4.
Fig.1: Polarization microsopic image of twin domains in LiGeBO4, viewing along [001] of the major twin.
By means of single crystal X-ray investigations and observation of thin crystal plates (100), (001) and (110) with the polarizing microscope, Liebertz and Stähr found the twinning law to be {112} [2], which is now confirmed by our recent reinvestigation. The composition plane (domain wall) also is {112}. The twinned crystals consist mainly of large twin domains oriented with their direction [001] along the pulling direction applied during crystal growth process. These are intersected by thin twin domains of about 5-8 mm thickness.
DTA measurements on single crystals and on powdered samples as well as heating stage polarizing microscope observations indicate a phase transition at about 1164K. At this phase transition the twin domains vanish and the crystalsremain optically uniaxial positive. Only 2-3 K higher, the crystals decompose peritectically into a melt and a crystalline remnant, thus making an investigation of the high temperature modification rather uncomfortable.
Due to the observed optical behaviour of the high temperature phase and a consideration of the features of the cristobalite-type structure of LiGeBO4 and its possible variations, the symmetry of the high temperature modification is likely to be 4_m2. A further investigation of the high temperature phase and the nature of the phase transition is in progress.