Electrothermal analysis of spin-transfer-torque random access memory arrays

  • Authors:
  • Subho Chatterjee;Sayeef Salahuddin;Satish Kumar;Saibal Mukhopadhyay

  • Affiliations:
  • Georgia Institute of Technology;University of California Berkeley;Georgia Institute of Technology;Georgia Institute of Technology

  • Venue:
  • ACM Journal on Emerging Technologies in Computing Systems (JETC) - Special issue on memory technologies
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

Spin Transfer Torque RAM (STTRAM) is a promising candidate for fast, scalable, high-density, nonvolatile memory in nanometer technology. However, relatively high write current density and small volume of the memory device indicate the possibility of significant self-heating in the STTRAM structure. This article performs a critical analysis of the self-heating induced temperature variations in STTRAM. We perform a 3D finite volume method based study to characterize self-heating effect in a single cell. The analysis is extended for STTRAM arrays by developing a computationally efficient RC compact model based thermal analyzer. The analysis shows that self-heating can results in considerable increase in both steady-state value and transient change in temperature of individual cells. The effect is less pronounced at the array level and depends on the activity level, that is, number of active cells within an array size. The analysis further illustrates that self-heating negatively impacts electrical reliability metrics namely, read margin and detection accuracy; degrades cell performance; and modulates energy dissipation.