News & Analysis
Comment
bill.shockley
"if the ions can be moved to form and modify the TiO oxygen vacancy pn junction ...
resistion
Yes I would expect so too.
HP, Hynix delay memristor debut
Peter Clarke
9/26/2012 1:42 PM EDT
LONDON – The memristor, a two-terminal, non-volatile memory technology in development by Hewlett-Packard since 2008, won't be ready for commercialization until the end of 2013, according to Stan Williams, and HP Senior Fellow and director of its Cognitive Systems Laboratory.
[Get a 10% discount on ARM TechCon 2012 conference passes by using promo code EDIT. Click here to learn about the show and register.]
Williams previously said HP and Hynix, now SK Hynix, planned to launch a resistive memory as a replacement for flash memory in the summer of 2013. That estimate came during an International Electronics Forum in October 2011, when HP indicated that an early application could be solid state drives. The decision to delay was include a roundtable discussion posted recently on the Web site of the Kavli Foundation (Oxnard, Calif.).
The panel discussion, entitled "How atomic scale devices are transforming electronics," also included: Michelle Simmons, director of the Australian Center of Excellence for Quantum Computation and Communication Technology, University of New South Wales; and Paul Weiss, Kavli Professor at UCLA and Director of the California NanoSystems Institute.
"In terms of commercialization, we'll have something technologically viable by the end of next year," Williams said, adding that the timing of any launch would be influenced by market demand for such memories.
"Our partner, Hynix, is a major producer of flash memory, and memristors will cannibalize its existing business by replacing some flash memory with a different technology," Williams said. "So the way we time the introduction of memristors turns out to be important. There's a lot more money being spent on understanding and modeling the market than on any of the research."
Williams said memristor research was essentially complete, adding: "If you know what you're doing – and there's a lot of intellectual property involved – literally any foundry could make memristors tomorrow."
The terminology used for resistive memories has itself become contentious, perhaps for reasons related to patent language (see HP responds to memristor debate). Williams said "memristor" refers to a two-terminal resistive memory. However, other companies and research institutes also are working on two-terminal memories, often of multilayered metal-oxide composition, referring to the devices as RRAM or ReRAM, for resistive random access memory.
Related links and articles:
How atomic scale devices are transphorming electronics
The Kavli Foundation
HP responds to memristor debate
Memristor 'brouhaha' bubbles under
HP, Hynix plan to launch memristor in 2013
|
|
|
|
|
Navigate to related information
resistion
9/26/2012 3:13 PM EDT
Hynix also studying phase change and spun torque memories, more to wait and see.
Sign in to Reply
R0ckstar
9/27/2012 2:29 PM EDT
I don't know if I agree with the intentional market introduction delay. If they really can build these "tomorrow", then I think if they do build it, they will come. It's not just about cannibalizing existing product, there are also new non competing sockets to consider. It's a completely new technology. Let's get this party started. Quite frankly, HP could use a bump about now after their really good run of bad decision making.
Sign in to Reply
Justin860
9/27/2012 2:43 PM EDT
Well, to refresh everyone's memory, this has been in development longer than 10 years.
Below is an announcement by Stan Williams on Sept.9, 2002. Funny he called Flash memory nitch market. After more than 10 years, not a single chip came out, no spec. and just smoke and more smoke.
No wonder HP is where it is today.
Just google "2002 HP leaps toward next generation memory"
HP scientist R. Stanley Williams, an HP Fellow and director of Quantum Science Research at HP Labs and his group have:
created the highest density electronically addressable memory reported to date.
combined, for the first time, both memory and logic using rewritable, nonvolatile molecular-switch devices; and
fabricated the circuits using a process called nano-imprint lithography
Williams said he would like to see commercialization of such a memory in five to 10 years. "Our most optimistic hope hope for this type of technology is five years," he said.
The first products resulting from the work would likely be aimed at niche memory markets, such as Flash memory and other areas where a need for inexpensive, nonvolatile memory exists.
The first products resulting from the work would likely be aimed at niche memory markets, such as Flash memory and other areas where a need for inexpensive, nonvolatile memory exists.
Such a memory is expected to have a cost advantage over Flash memory when it hits the market because of the manufacturing process.
"We view this as a highly manufacturable and scalable process," Williams said.
Sign in to Reply
GoGoGeek
9/28/2012 3:30 PM EDT
MRAM is in development much earlier. I heard about it in the last decade already.
Sign in to Reply
GoGoGeek
9/28/2012 3:31 PM EDT
Actually it was in the 90's
Sign in to Reply
R G.Neale
10/2/2012 5:52 AM EDT
Perhaps HP and S K Hynix have read the recently published paper: Fundamental Issues and Problems in the Realization of Memristors by Paul Meuffels and Robit Soni.
http://arxiv.org/abs/1207.7319
And decided a delay was in order. I think the authors of the above reference have expressed a clear objection to the oxygen ionic PN junction model.
My view has always been if the ions can be moved to form and modify the TiO oxygen vacancy pn junction by the use of very little energy, then at elevated temperatures thermal diffusion will remove the concentration gradient and the memory contents.
If this cannot be proved by simple experiment then I think a metallic filament radial oxidation-reduction model offers a more likely explanation of the NV memory effect.
Sign in to Reply
resistion
10/2/2012 9:28 AM EDT
Yes, enough heat can erase anything. But as the gradient flattens, this effect should weaken.
Sign in to Reply
R G.Neale
10/2/2012 11:47 AM EDT
Resistion; If you are claiming the junction with the "flattened gradient" will have the same resistance, i.e.one of the data states, as a steep gradient, then your implied observation would be correct. I think you will find the leakage current across the shallow gradient will be much higher.
Sign in to Reply
resistion
10/3/2012 12:09 AM EDT
Yes I would expect so too.
Sign in to Reply
bill.shockley
10/8/2012 2:14 PM EDT
"if the ions can be moved to form and modify the TiO oxygen vacancy pn junction by the use of very little energy"
They are not claiming this at all. Don't you remember the paper that made you think HP had a phase change device, because of the heat involved and the physical evidence of different phase states in the active element? The memristive magic happens at high current densities under very large potentials. Temperatures are around 300C.
Sign in to Reply