Since second harmonic generation was discovered by Franken et al. in 1961, nonlinear optics has been one of the research areas which were studied most intensively. Although the theory of nonlinear optics was well established by the end of 1970s, the applications of nonlinear optics were not fully exploited since nonlinear optical materials of large nonlinear coefficient were not readily available by then. In 1980s a number of nonlinear optical materials of excellent second-order properties such as $KTiOPO_4$(KTP), β-$BaB_2O_4$(β-BBO) and $LiB_3O_5$(LBO) were developed, which triggered the active research on fundamental phenomena and practical applications of nonlinear optics. Among them, KTP has been most studied and utilized for second harmonic generation and mid-infrared generation because of its large nonlinear effect and mechanical stability.
The conditions for phase-matching imposes a practical limit in the application of nonlinear optical crystals. However if we have a periodic structure of refractive index with the period of the coherence length, strong hannonic generation can be obtained by quasi-phase-matching. In ferroelectric crystals such as $LiNbO_3$ and KTP quasi-phase-matching geometry can be established by poling the domains periodically. Low coercive field and low electrical conductivity are very important parameters for effective polings. $LiNbO_3$ has been the most successful material for this purpose, but it shows low laser damage threshold. The periodic poling of KTP was successful recently, but the electrical conductivity is high in KTP because of hopping motions of $K^+$ ions. For these reasons $RbT_1OAsO_4$(RTA) was chosen for the study of crystal growth and dielectric properties to exploit the quasi-phase-matching. It is expected that the electrical conductivity could be reduced significantly by replacing $K^+$ ions by $Rb^+$ ions and $P^{5+}$ ions by $As^{p+}$ ions in KTP.
KTP and RTA single crystals were grown from the flux method. The typical sizes of KTP crystals was 33×17×18㎣ using b-axis seed and 30×19×21㎣ using c-axis seed. And the typical sizes of RTA was 19×41×20㎣ using b-axis seed and 20×35×20㎣ using c-axis seed. The temperature dependence of the lattice constants of KTP and RTA single crystals were examined, and the crystals were found to be stable for the temperature variation.
The UV cutoff wavelength of KTP and RTA single crystals was 350 nm, and the IR cutoff wavelengths of KTP and RTA single crystals were 4.5㎛ and 5.3㎛, respectively. The second-order nonlinear optical coefficients of KTP and RTA single crystals were measured using Maker fringe method. From the refractive index data, the phase matching angles of KTP and RTA single crystals for Nd-YAG laser wavelength were calculated. The most efficient phase matching condition in KTP is type Ⅱ with (θ, φ) = (90°, 26°), and the most efficient phase matching condition in RTA is type Ⅰ with (θ, φ) = (53° 41°) New method of determining electro-optic coefficients using Maker fringe method was developed, and the electro-optic coefficients of KTP single crystal were measured. The values were $r_{13}$ = 8.04 pm/V for λ = 1.064 ㎛, $r_{23}$ = 15.17 pm/V for λ = 1.064 ㎛, and $r_{33}$ = 40.19 pm/V for λ = 0.532 ㎛.
The dielectric constants of KTP and RTA single crystals were investigated using an impedance analyzer. The ferroelectric-paraelectric phase transition temperature of RTA was 1018 K. The thermal hysteresis of the phase transition was observed and the amount of change in the transition temperature was 4 K from the heat run to the cool run. The conductivities of RTA along the three principle axes have the same order of magnitude, which is contrasted with the high anisotropy shown in KTP. The observed small conductivity along the c-axis in RTA may in part reflect the facts that the replacement of $K^+$ ions by heavier $Rb^+$ ions makes hopping motions more difficult and the replacement of phosphorus atoms by larger arsenic atoms contributes to the hindrance of the hopping motions of $Rb^+$ ions. A dielectric relaxation was found just below the paraelectric-ferroelectric phase transition temperature. The detailed analysis indicates that the relaxation may be attributed to the domain freezing and the domain freezing temperature was estimated to be 983 K.
