Challenges and opportunities for future non-volatile memory technology

Abstract

Recent progress made in new non-volatile semiconductor memory materials such as chalcogenide, binary metal oxides, perovskites, ferromagnetics, ferroelectrics, organic materials and carbon based materials is quite exciting and will open new horizon for the future terabit memory era. It is, however, interesting to note that most of such new non-volatile memory operation can be described by a “change of resistance” of materials. For instance, phase change memory depends upon the change of resistivity based upon whether the material is amorphous or crystalline, but the function of resistive switching memory family, often called ReDox, RRAM, ReRAM, CBRAM, is ascribed to motion of either ions or vacancies resulting in creation of nanoscale conductive path inside of the material. Here we can clearly see a new paradigm, as opposed to traditionally accepted notion of semiconductor device operation where only electrons and holes contribute to current and charge storage in structurally stable crystal lattice, is now emerging; and non-volatile memory functionality is contributed by the structural changes of the material itself coupled with motion of electrons and holes or even include such state variables as spin involved. This creates an enormous challenge to our technical community in terms of needs for cross-boundary collaborations of multi-disciplinary expertise and knowledge base.