Computational study of two nonvolatile recording media: resistive random access memory (RRAM) and graded magnetic media.

摘要

The role of nonvolatile memories in recording devices has become extremely important in modern times. Various nonvolatile memory (NVM) concepts have been researched. Some of them have been successfully applied for memory devices: e.g. magnetic hard disk and flash memory. Other NVM concepts are actively researched for possible applications in the future: ferroelectric random access memory (FeRAM), magnetic RAM (MRAM) and resistance RAM (RRAM). Both mature NVM technologies and early stage NVM concepts have their own challenges. In the mature hard disk industry, it has become increasingly difficult to further increase the recording density. On the other hand, for one promising concept, RRAM, researchers are still searching for the dominant physical cause of experimentally observed electrically switching resistance states. These two problems are the focus of this work.;We used software simulators that incorporate necessary physical interactions to duplicate the studied materials: transitional metal oxide in between metallic electrodes in the RRAM study and perpendicular magnetic grain for the magnetic media study.;Using the RRAM simulator, we identified oxygen vacancies in the interface layer next to the electrode as the main driver of the switching behavior. Using the magnetic media simulator, we found that a perpendicular column with the anisotropy constant varying quadratically with its height and a hard layer on the top has the optimal recording characteristic (high energy barrier and low coercivity) for a high density media.