Response Surface Methodology Study of Ferroelectric Memory Devices.

摘要

Fatigue in ferroelectric memory devices was studied. The application of fatigue pulses to a ferroelectric sample was controlled by the RT-66 Ferroelectric Tester, a variation of the Sawyer-Tower circuit. The RT-66 also controlled data collection following the fatiguing process. Seven variables were evaluated for their potential- affect on fatigue. A sequence of fatigue tests, which varied the settings of these variables, was developed using Response Surface Methodology (RSM) experimental designs. Least-squares regression models were developed once a particular RSM set of design experiments was completed. These models were evaluated using RSM tools, and the analysis of these models allowed the iterative development of new RSM designs. Within the experimental operating region it was determined that aging did not affect fatigue, and that differences in ferroelectric materials and electrode materials were the most important factors in determining fatigue. First-order models failed to fit the data for any of the RSM designs. The final model developed had two pure quadratic terms and three first-order terms. However, this design only involved one type of ferroelectric. Further testing is necessary before the findings of this study can be extended to modeling fatigue in other ferroelectric materials.