Jun 16, 2011
Resistive memory: how small can you go?
Memory technologies continue to obtain favourable projections from the Semiconductor Roadmap under aggressive feature scaling. Currently, resistive memory (ReRAM) is regarded as a very promising emerging technology. There are different types of the resistive memories that operate by forming and breaking conduction pathways via different physical mechanisms. The central question addressed in this study was: what is the smallest volume of matter needed for resistive memory?
To answer the question, researchers from Semiconductor Research Corporation and Micron Technology, both in the US, conducted several Gedanken or thought experiments for solid-state physical systems consisting of a small number of atoms.
Fundamental physical principals were used to analyse how the properties of the system vary by changing the position(s) of one or a few atoms. The two tasks explored in this study were: (i) change in resistance due to the addition or removal of a few atoms; and (ii) stability of a few-atom system.
Prospectus for tiny memory
These thought experiments offered insights on the effects of ultimate scaling on a number of parameters such as density, speed and retention. It was found that the resistive memory cell can, in principle, be scaled down to the volume of 3–4 nm3, which is considerably less than all memory devices based on the storing of electron charge.
The result shows that resistive memories are an attractive option providing that a number of materials and process issues are overcome.
Full details can be found in the journal Nanotechnology.
About the author
Roy Meade and Gurtej Sandhu are with the Process R&D department of Micron Technology, Inc. and Victor Zhirnov and Ralph Cavin are staff members at Semiconductor Research Corporation. (SRC). The genesis of the study can be traced to a 2009 SRC forum on emerging research memories in which a paper was presented on the demonstrable advantages of resistive memories from the standpoint of an energy-space-time metric. These early positive results led the authors to collaborate to explore the scaling limits of resistive memories and ultimately to this paper.