Ferroelectric Thin Films For Nonvolatile Memory Applications

IT IS WELL KNOWN THAT FATIGUE OF FERROELECTRIC THIN FILMS WAS REDUCED SIGNIFICANTLY SINCE CERAMIC ELECTRODES SUCH AS RuO(x) WERE DEVELOPED. IT IS, HOWEVER, REALIZED THAT HIGH LEAKAGE CURRENT WITH EARLIER ELECTRICAL DEGRADATION AND POOR RETENTION PROPERTIES ARE STILL PROBLEMS THAT NEED TO BE SOLVED FOR COMMERCIAL MEMORY APPLICATIONS. A FAST POLARIZATION DECAY OCCURS AT THE INITIAL STAGE OF RETENTION (OR AGEING), WHICH IS CALLED IMPRINT. IN PHASE I WORK, THE IMPRINT PHENOMENON WILL BE MODELED QUANTITATIVELY BASED ON THE BEHAVIOR OF MOBILE CHARGES IN CONJUNCTION WITH DOMAIN BOUNDARIES, GRAIN BOUNDARIES, AND FERROELECTRIC-ELECTRODE INTERFACES. AGEING WILL ALSO BE CORRELATED WITH THE IMPRINT. THE IMPRINT PHENOMENON WILL BE REDUCED IN TWO WAYS: 1) IMPROVEMENT OF EXISTING LOW-FATIGUE FERROELECTRIC-ELECTRODE CONFIGURATIONS SUCH AS RuO/PZT/RuO(x) AND 2) USE OF NEW FERROELECTRIC MATERIALS SUCH AS YMnO(3). SPACE CHARGE CONCENTRATION, THE KEY ROLE OF THE IMPRINT PHENOMENON, DUE MAINLY TO OXYGEN VACACIES WILL BE MINIMIZED BY USING DOPANTS SUCH AS La OR Nb IN PZT FILMS. YMnO(3) IS BELIEVED TO HAVE BETTER AGEING AND IMPRINT PROPERTIES BECAUSE OF ITS LOW DEFECT CONCENTRATION SUCH AS OXYGEN VACANCIES. SIMPLE NONVOLATILE MEMORY DEVICES WILL BE FABRICATED AND TESTED USING MOCVD DOPED PZT OR YMnO(3) FILMS DURING PHASE II.