An acquaintance recently brought to my attention an article in R&D Magazine about some pioneering research on phase-change memories or PCM. The researchers’ findings hold a lot of promise. (R&D Magazine’s article is based upon an original paper in the journal Science.)
A team led by Ritesh Agarwal, associate professor at the University of Pennsylvania, was trying to develop a better understanding of the mechanism behind the phase changes in PCM. The team found that existing programming algorithms that involve melting the material could be replaced with pulses of electrical current that not only would program the cell without heat, but provided an “On” to “Off” resistance ratio of 2-3 orders of magnitude, which renders the cell significantly easier to read, especially in the presence of noise. This effectively makes memory chip design easier than with current methods.
Furthermore the means of programming the cell and the researchers’ understanding of it indicates that current phase change materials may not be the best to use for PCM, and that new classes of materials may produce better devices.
The Memory Guy hasn’t yet heard back from Dr. Agarwal in response to an e-mail I sent him a week ago, but I hope to establish contact to hear more details of this finding. When I do I will update this post and let readers know whether the team has found a similar way of re-crystallizing the bit to return it t o a high-conductance state, or if that must be done through existing methods. I will also report back on anything I learn about the programming speed of the new approach.
When people ask me which technology is likely to replace NAND flash once it reaches its scaling limit I always say that it’s too early to tell. Research like this really underscores that fact – PCM may gain a significant advantage due to the University of Pennsylvania’s research.
There is a free downloadable white paper about Phase Change Memory on the Objective Analysis website.
I think there’s a critical question left unanswered. When we last left PCM, ambient thermal annealing of the storage state was a big problem rising inversely with feature-size reduction. From your description above, I might conclude that the “other materials” might be needed to overcome this problem but in truth, there’s no mention of it so I simply don’t know. However, I suspect that the problem doesn’t go away just because the researchers have found a way to switch between amorphous and crystalline states using trains of short programming pulses.
There are still lots of questions around this, and I would guess that many years will pass before this research finds its way to manufacturing. Still, I would guess that a significant materials change, as is suggested by the R&D Magazine article, could push us past the difficulties you mention.
In 1970 I did a cinematic study of the switching mechanism of PCM in chalcogenide at Iowa State University. (http://www.youtube.com/watch?v=0bgVsOk17vw) At 2:57 of the video, it visiually showed the “electric wind” effect you described in your publication. It is basically a voltage driven (V square over R) ionization wave that set the material between the electrodes to a more amorphous state starting from the center and propagating outward. Once the material is “amorphsized” then the crystalline filament can propagate from the positive electrode to the negative electrode. As shown in the video the erase of the crystalline filament can be very fast if driven from a current source. I believe the polarity effect and the electric wind will contribute toward the understanding of the failure mode of the device. I also believe this study on a macro scale is applicable to the behavior of the device on a micro scale.
“Memory Devices Using Bistable Resistivity in Amorphous As-Te-Ge Films” C. H. Sie, PhD dissertation, Iowa State University, 1969, order #69-20670, University Microfilm
“Electric-Field Induced Filament Formation in As-Te-Ge Semiconductor” C.H. Sie, R. Uttecht, H. Stevenson, J. D. Griener and K. Raghavan , Journal of Non-Crystalline Solids, 4, 358-370,1970
Charles,
It’s an honor to know that someone of your caliber has read my blog!
Thanks for sharing the video. It is simply amazing!
It’s interesting that you found something in 1970 that appeared to be news in 2012. I hope that researchers aren’t wasting time repeating experiments that were already performed a long time ago.
Jim