Phase relations in the Li2O – Nb2O5 binary system
The first phase diagram of the system Li2O – Nb 2O5 was published by Reisman and Holtzberg  describing two incongruently and two congruently melting compounds: lithium meta-niobate (LiNbO3, melting point 1253 °C) and lithium ortho-niobate (Li3NbO4, melting point 1408 °C). Lerner et al.  made a substantial improvement, showing that LiNbO3 has a fairly large solid solubility range, and the maximum melting temperature does not occur at the stoichiometric composition but between 48 and 49 mol% Li2O. The congruent composition is reported to be between 48.35 and 48.60 mol% Li2O [16, 20, 21, 22, 23, 24] depending on the purity of the starting materials, growth conditions, measuring methods and also on the direction of the crystal growth [21, 22]. In the homogeneity range (~ 47 – 50 mol% Li 2O ) the physical parameters are sensitively changing (e. g. lattice constant, density, ferroelectric Curie temperature, refractive index, phase matching temperature of the 2nd harmonic generation, the position of the ultraviolet absorption edge, position and shape of the OH-vibration spectrum). Several calibration methods use these variations for the precise determination of the compositions of the crystals as will be shown later. A slight shift in the composition causes a large change in the physical parameters in this composition range; therefore the exact knowledge of the Li2O content of the crystal is very important.
Phase relations in the K2O – Li2O – Nb2O5 ternary system
The investigations of the K2O – Li2O – Nb2O5 ternary system started in the 1960’s. Potassium lithium niobate (K3Li2Nb5O15, KLN) with good non-linear optical and electrooptical properties was grown from this system [26, 27]. A new interest was given to the subject when the composition of lithium niobate (LN) grown from K2O containing starting material was found to have better properties (sharper NMR and EPR lines) than those grown from congruent melt . Based on these first results in-depth investigations of the ternary oxide system were started. It was established that lithium niobate bulk crystals with the compositions closest to stoichiometric can be grown from this system (Polgár et al. ). Within the confines of the lithium niobate existence field the phase boundaries of the stoichiometric lithium niobate existence field were determined as shown in Fig. 1.4.
High temperature top seeded solution growth (HTTSSG) method
The starting materials (the solvent and the solute) are mixed and melted in a crucible and kept at high temperature for homogenization. An oriented crystal seed (with the same composition as that of the crystal to be grown) is brought into connection with the surface of the mixture. After reaching thermal equilibrium between the seed and the solution close to the saturation temperature, with the slow lowering of the temperature the crystal begins to grow on the seed with corresponding orientation. During the procedure the seed is rotated and pulled (fig. 1.5). The temperature programme of the growth is chosen in a way that the diameter of the crystal increases continuously from the few mm diameter of the seed crystal to the maximal size required and maintains this constant diameter till the end of the growth. If the crystal reaches the expected length, it is detached from the flux.
Solid state reactions, thermoanalytical investigations and phase characterization
In order to understand the role of the alkali metal oxide solvents in the growth of sLN crystals several methods have been combined. Series of X2O – Li2O – Nb2O5 (X = Na, Rb, and Cs) polycrystalline samples were prepared by solid state reaction for thermal analytical, x-ray powder diffraction and crystal growth investigations.
The raw materials used were Na2CO3 (Merck, analytical grade), K2CO3 (Merck, analytical grade), Rb2CO3 (99,9 % (Johnson-Matthey), Cs2CO3 (Merck, extra pure), Li2CO3 (Merck, Suprapur) and Nb2O5 (Starck, LN grade). The sample preparation was carried out in a high temperature electric furnace. In all cases the starting materials were dried (200 °C for 12 hours), weighed with 0.1 mg precision, mixed and reacted in solid phase (800 °C for 3 hours) then cooled to room temperature and reground. The release of the whole amount of CO2 was controlled by weighing. The samples used for thermal analyses and for crystal growth were prepared by a second reaction step: melted at 1200°C or sintered just below the melting temperature (1100 oC), respectively. The constituent phases were assessed by X-ray phase analysis with a Philips PW 1710 diffractometer using Cu K radiation in the 2 range of 0-80o for the measurements described in chapter 3.1.1. For the results in chapter 3.1.2. an INEL MPD CPS 120 diffractometer was used.
DSC measurements were used to determine the phase transition temperatures of X2O – Li2O – Nb2O5 (X = Na, Rb or Cs) polycrystalline samples in the composition range of [X] = 10-16 mol% and [Li]/[Nb] = 1. The DSC curves were recorded with a PL Thermal Sciences 1500 differential scanning calorimeter in Ar atmosphere with Al2O3 as the reference material. The calibration of the system was done by using 6N pure standard metals (In, Sn, Pb, Zn, Al, Ag, Au, Si). The calorimeter was heated up to 1370oC at a rate of 10oC/min and then cooled down at the same rate to 300oC. For the determination of the phase transition temperatures the heating part of the curves was taken into account in order to eliminate the mistakes from supercooling. The compositions of the polycrystalline samples are represented in a ternary diagram shown in Fig. 2.1.
Crystal growth and refinement of the phase diagram
Crystal growth was done by the HTTSSG method from fluxes along the line in the ternary diagram joining to the LiNbO3-X2O join of the X2O – Li 2O – Nb 2O5 (X = Na, Cs or Rb) ternary systems from the starting composition of X2O = 10 mol% and [Li2O] / [Nb2O5]=1.
For the growth experiments <00.1> oriented LiNbO3 seeds were used. The crystals were pulled at rates of 0.3-0.5 mm/h and rotated with 8-10 rpm (rotation / minute).
If it is assumed that the system shows no solid solubility for the X ion (it does not enter into the LN lattice) the compositional changes which occur during the solidification may be simply traced. In a first approximation, assuming a stoichiometric composition for the crystallized solid phase, the actual alkali oxide concentration of the liquid can be calculated from the pulled amount of the crystal. The real solid composition (the Li2O content) of the pure LN crystals can be determined by ultraviolet/visible (UV/Vis) spectroscopic measurements. In the previous studies K2O based fluxes proved to be ideal solvents  therefore a similar behaviour was expected also for the other alkali ions.
The crystallization temperature of the lithium niobate crystal at a given composition can be deduced from the DTA measurements of the polycrystalline samples with the same composition as the actual flux from which the crystal is growing. The composition of different parts of the crystals was determined from slices cut perpendicular to the growth axes of the pulled-out LiNbO3 crystals. The knowledge of these two sets of data lets us determine the corresponding liquidus and solidus points (tie-lines) (Fig. 2.3).
Table of contents :
1. Physics and chemistry of crystalline LiNbO3
1.1. The LiNbO3 crystal structure
1.2. Phase relations in the Li2O – Nb2O5 binary system
1.3. Phase relations in the K2O – Li2O – Nb2O5 ternary system
1.4. Crystal growth processes
1.4.1. High temperature top seeded solution growth (HTTSSG) method
1.5. Crystal composition and characterization methods
2. Experimental methods and instruments
2.1. Solid state reactions, thermoanalytical investigations and phase characterization
2.2. Crystal growth and refinement of the phase diagram
3.1. Phase identification and phase diagram determined on polycrystalline samples
3.1.1. Thermal analysis and X-ray phase identification in the X2O – Li2O – Nb2O5 ternary systems
3.1.2. Phase diagram of the Cs2O – Li2O – Nb2O5 ternary system
3.2. Single crystal growth and characterization
3.2.1. Crystal growth experiments in the X2O – Li2O – Nb2O5 ternary systems
3.2.2. Solid composition determination by UV/Vis spectroscopic measurements.
3.2.3. Phase relations along the vertical section of LiNbO3-X2O systems
3.3. Spectroscopic characterization and composition determination of the crystals grown from alkali metal oxide fluxes
3.3.1. IR absorption measurements
126.96.36.199. OH- vibrational spectra of crystals grown from Rb2O and Cs2O containing fluxes
188.8.131.52. OH- spectra of crystals grown from Na2O containing flux
3.3.2. Raman spectroscopic properties of the crystals
184.108.40.206. Composition calibration for lithium niobate crystals based on Raman experiments
220.127.116.11. Raman spectra of crystals grown from Rb2O and Cs2O containing fluxes
18.104.22.168. Raman spectra of crystals grown from Na2O containing flux
5. New results