Fwd: Aaron Gibby Orals Abstract
Aaron Matthew Gibby
agibby at stanford.edu
Thu Aug 7 16:56:58 PDT 2008
----- Original Message -----
From: "Diane Shankle" <shankle at ee.stanford.edu>
To: <ee-students at lists.stanford.edu>
Sent: Thursday, August 07, 2008 4:20 PM
Subject: Aaron Gibby Orals Abstract
> Title: A Layered Chalcogenide Phase-Change Memory Device
> Name: Aaron Gibby
> Advisor: Prof. S. Simon Wong
> Time: 10:00 am, Monday, August 25th, 2008 (refreshments served at
> 9:45 am)
> Place: CIS-X Auditorium, 420 Via Palou
> Non-volatile memory (NVM) is the fastest growing sector of the
> semiconductor market. With sales growing from $26 billion in 2006 to
> more than $64 billion in 2011, there is a large economic incentive to
> improve on NVM performance. This has lead to aggressive scaling of
> Flash memory, the dominant NVM technology. There are, however,
> limitations to the degree that Flash can be scaled, with failure of
> the technology predicted around the 22 nm node.
> As a result, research has accelerated in alternative NVM
> technologies. Among these, phase-change memory (PCM) shows
> particular promise given its small cell size, non-destructive read,
> direct overwrite ability, large sensing margin, and fast speed.
> Despite these advantages, the requirement for a large (mA - range)
> programming current remains the major obstacle to mainstream
> implementation for PCM.
> To address this issue, we propose a novel layered structure where the
> phase-change material (Ge2Sb2Te5, or GST) is sandwiched between two
> other layers (called GST-x layers). With the correct choice of
> material, the GST-x layers improve the thermal isolation of the GST,
> while reducing the volume of material programmed. This, in turn,
> lowers the energy and current required to operate the device.
> After a thorough analysis of GST-x candidates using X-ray
> Diffraction, X-ray Photoemission Spectroscopy, Thermal Reflectance
> Thermometry, and electrical measurements, GeTe (GT) was chosen to
> function as the GST-x film. Single layer GST and multi-layer
> GT/GST/GT devices were then fabricated and compared. Analysis was
> conducted using a combination of electrical measurements,
> transmission electron microscopy and scanning Auger spectroscopy.
> As a result of this analysis, GT/GST/GT devices showed a modest (~2x)
> improvement in programming current on the first cycle and a
> significant (~40x) improvement on subsequent cycling. Reasons for
> this and reliability concerns are discussed.
> EE students mailing list
> ee-students at lists.stanford.edu
More information about the labmembers