Search

Stoichiometric SrTiO3 (STO) films doped with Fe or Cr were prepared by r.f. magnetron sputtering technique. The effects of Fe or Cr doping in the SrTiO3 films were studied on the leakage current property which was discussed by defect chemistry. The experimental results can be explained by a model in which oxygen vacancies are the key defects responsible for the leakage current. Acceptor doping, with a small concentration of Fe or Cr, has led to a substantial improvement to 10−9 order in the leakage current density. Above the concentration of 0.01∼0.02 mol% Fe2O3, Cr2O3, however, as the concentration increased, the leakage current increased. These acceptors in Ti4+ site are expected to electrically compensate for donor species such as oxygen vacancies, thereby reducing the concentration of mobile carriers that contribute to electrical conduction. Consequently, acceptor doped STO films have been shown to be superior to undoped films for applications requiring high leakage resistance, such as dynamic random access memory capacitors. © 1998 Chapman & Hall

SrTiO3 thin films were deposited by rf-magnetron sputtering under various sputtering conditions followed by conventional furnace annealing at 600 and 700 °C. The amorphous SrTiO3 thin films crystallized into polycrystals at 600 °C. The leakage current of the SrTiO3 thin films decreased with increasing oxygen partial pressure in the sputtering gas. On the contrary, the dielectric constant increased with increasing the oxygen content in the sputtering gas. The leakage current and dielectric constant increased with increasing substrate temperature and post-annealing temperature. The ratio of Sr∶Ti approached 1∶1 with increasing oxygen content in the sputtering gas and substrate temperature during deposition. The oxygen content in the film decreased with increasing the substrate temperature. The capacitance-voltage (C-V) curves showed that the capacitance was nearly independent of the applied voltage. Scanning electron microscopy (SEM) micrographs showed that interdiffusion between the bottom electrode (Pt) and the buffer layer (Ti) occurred during post-annealing, but that the interface between SrTiO3 and Pt was stable.

The degree of phase formation was investigated by Sb2O3 behaviour in the calcination process, and the change of microstructure and breakdown properties due to the phases formation in ZnO varistor. The samples were calcined at several temperatures for 2 h and were sintered at 1150 to 1300 ° C, respectively, for 1 h. After that,V-I characteristics were investigated. Phase analysis by X-ray diffraction and microstructures observation by SEM were done. As a result, spinel phase was not formed and low nonlinear resistance was shown in the samples without Sb2O3. In the samples containing Sb2O3, it was shown that the pyrochlore and spinel phase are formed at the conventional calcination temperature or even below that temperature. This primary spinel phase and the spinel phase transformed from pyrochlore phase in sintering process inhibit ZnO grain growth, and so nonlinear resistances should be changed. Hence ZnO grain growth in ZnO-based varistor system is strongly dependent on the Sb2O3 behaviour in the calcination process.

PZT thin film has been expected to replace conventional dielectric as a charge storage element in DRAM without use of a sophisticated structure such as a trench, because PZT has a dielectric constant much higher than silicon dioxide [1]. However, PZT thin film has relatively high leakage current [2]. To improve the high leakage current, the mechanism must be studied. Some studies have reported that the leakage current mechanism of PZT thin films on platinum showed Schottky emission [3, 4]. Their interpretation depends on the linearity of l n J versus E 1/2 (J: current density, E: electric field). This method is ,not able to distinguish Schottky emission from Poole-Frenkel emission, which may be one of the main mechanisms in PZT thin films. The equation for Poole-Frenkel emission predicts a straight line graph In J against E 1/2. That is, under most circumstances, experimental results can be plotted as a good straight line on either type of graph. Therefore, the mechanism cannot be separated by this means. To tell them apart, it must be determined that the leakage current results from a bulk-limited process or a barrier limited process, Poole-Frenkel emission being bulk limited and Schottky emission being barrier limited [5]. In this study, we measure J V curves for two different polarities using different electrodes, for example, upper electrode aluminium and bottom electrode platinum. In this case, if the leakage current depends on the barrier limited process, the J V curve will be asymmetrical when electrode polarity is reversed because the work functions of aluminium and platinum are different [6]. If leakage current depends on the bulk limited process, the J V curve shows symmetrical characteristics. Platinum film (250nm) with an intervening titanium (300 nm) layer as an adhesion layer were deposited as the bottom electrode by DC sputtering on thermally surface oxidized Si substrate, and then PZT films were deposited by RF magnetron sputtering, the films having a compositional Zr/Ti ratio of 50/50 and a thickness of 200 nm. The deposition conditions are shown in Table I. The aluminium top electrode of radius 100 ~m was deposited by thermal evaporation. X-ray diffractometry revealed that the film was of (1 1 1)-oriented perovskite phase. Leakage current at various voltages was measured using a HP 4145B semiconductor parameter measurement equipment.