Although non-volatile memory has a vast application potential, its applied research has been slow and largely hindered by the physical properties limitation imposed by the ferroelectric materials ability. The best possible ferroelectrics which can be used in non-volatile memory application are yet lead zirconate titanate ($PbZr_{1-x}Ti_xO_3$ or PZT)$ and strontium bismuth tantalate ($SrBi_2Ta_2O_9$ or SBT). PZT should be more favorable in actual application since its crystal structure is simple tetragonal or rhombohedral at room temperature. Its room temperature crystal structure is obtained after phase transition from the high temperature cubic paraelectric phase. But among the undesired properties, its early fatigue in the polarization - electric field(P-E) hysteresis loop experienced after undergoing more than $10^8$ cyclic sweeps have been often singled out. For this reason, many attempts have been made to improve the fatigue property. One way to improve the fatigue property of PZT thin films has been done by doping. By doping $Nb_2O_5$ slight improvements in fatigue property have been made but they have not been enough. On the other hand more modest improvements in fatigue property have been made after inserting oxide conductors such as $La_{1-x}Sr_xCoO_3(LSCO)$, $RuO_2$ or $IrO_2$ over the platinum electrodes.
To overcome the fatigue problem, new ferroelectric materials are also extensively studied. The lead-base relaxor - $PbTiO_3(PT)$ solid solutions such as $Pb(Mg_{1/3}Nb_{2/3})O_3(PMN)$ - PT and $Pb(Sc_{1/2}Nb_{1/2})O_3(PSN)$ - PT are very attractive ferroelectric materials because they show low crystallographic anisotropy and low thermal expansion. Numerous attempts to realize relaxor-PT materials in application area, however, has suffered from their low transition temperatures (typically less than 200℃. For the required temperature stability in the memory application, the phase transition temperature of the relaxor-PT materials is desired raised. Among the relaxor-PT solid solution systems, the $(1-x)Pb(Yb_{1/2}Nb_{1/2})O_3(PYN)-xPT$ or another word PYNT system has been shown to posses the highest transition temperature (~360℃) near the morphotropic phase boundary (MPB).
Although pure PYN is a B-site ordered antiferroelectric, the substitution of $Ti^{4+}$ ion at B-sites induces sequential phase change from the orthorhombic antiferroelectic to relaxor ferroelectric and finally into the tetragonal ferroelectric. The electromechanical properties of the solid solution display optimized just near the MPB composition.
In this work, the physical properties of ceramic and thin film of the PYN-PT solid solution was investigated. In Part Ⅰ, the MPB of antiferroelectric (AFE) - ferroelectric (FE) (1-x) PYN-$xPbTiO_3$ solid solution has been investigated through analysis of crystal structure and dielectric properties with change as a function of temperature and composition. The PYNT MPB marks the boundary between the pseudocubic and tetragonal phases near x=0.48, which is in fact found to be a two phase coexistent band with a concentration width Δx=0.025. An unusual dielectric behavior change near the MPB effected by 10kV/cm poling has been also described. The phase coexistence near the MPB was looked at by macroscopic and microscopic points of view. This temperature dependent boundary could be calculated using Landau-Dvonsire equation represented by polynomials of the order parameter (spontaneous polarization). In the tetragonal region, the 4th and 6th order coefficient can be determined by the fitting of spontaneous polarization vs. temperature curve. However, the thermodynamic coefficient for the pseudocubic region could not be obtained because it displayed very low anisotropy. The series of possible value for other coefficient was used for finding out the phase boundary. The calculated boundary shows the banded shape as observed in real PYNT solid solution. The relationship between the shape of boundary and the 6th order polynomial coefficient is also demonstrated.
PartⅡ is composed of the preparation process and properties of perovskite PYNT thin films. The rf magnetron sputtering was used to prepare ferroelectric PYNT films on the Pt/Ti/$SiO_2$/Si substrate. To control the composition, three targets of Pb, Ti and $YbNbO_4$ were used. The formation of perovskite phase was strongly affected by the Pt/Ti substrate. Thick Pt layers restrain the formation of perovskite grain, whereas pure perovskite phase is favorable grown when $TiO_2$ buffer layers are inserted between PYNT and Pt layers. Although the $TiO_2$buffer layer may play a very important role for growth of the perovskite PYNT films, it also decreases the overall capacitance because of its low dielectric constant. For easy growth and high electrical properties of PYNT films, conductive oxide buffer layer need.
For the conducted buffer layer, deposition processing of $(La_{0.5}Sr_{0.5}CoO_3(LSCO)$ was investigated. It is demonstrated in this study the fact that high quality LSCO thin films can be also prepared by dc magnetron sputtering instead of so far much fashionable pulsed laser deposition technique. These thin films deposited at relatively low temperature (at 400℃) are founded to be uniform and made up of fine perovskite grains. On the LSCO buffer layer, a PYNT film can be grown more easily and it establishes some crystallographic correlation between the crystal orientation of bothe PYNT and LSCO films.
