1 Click Subscribe to the RSS feed
1 click ADD to your Google HomepageNantero's Progress in Nanotube NRAM memory
From an interview with Nantero CEO Greg Schmergel on Feb 3rd 2006, Nantero's NRAM, is ready for a production roll out in 2007. http://www.tgdaily.com/2006/02/03/nantero_cnt_memory/
This is a pretty telling statement regarding progress made towards development of Nantero's NRAM memory. Something is happening, and it sounds good.
The NRAM technology is a nanotube based MEMS storage element that is formed from a "mems crosspoint" of 2 opposed matts or bundles of nanotubes, 1 suspended as a flexible (MEMS) Nanotube "strip", above the other fixed anchored carbon nanotube bundle below.
Electrostatic forces can cause the upper Nanotube bundle to move into electrical contact with the lower Nanotube bundle, in a non volatile state, but electrically reversible.
( shades of nonvolatile FLASH, but faster than flash, lower power and cheap as a technology node advance )
CLICK READ MORE... for rest of article.
While the device implementation is not as elegant at the seminal single wire pair Science publication concept by Nantero founder Rouckes, then of Harvard's Lieber Lab, in which was predicted NRAM feasability based on single pair, crossed nanotubes; apparently the implemented NRAM device of crossed BUNDLES of nanotubes, works very well and yields well.
YIELD is the key... (as always)
Long ago when I read of the Nantero start-up, I had some pretty emphatic emails with Jennifer Fonstad of DFJ regarding the near impossibility of single pair, crossed wire NRAM devices, to be able to be made with any yield. ( which still holds today btw AFAIK ).
Just a stubborn principled response from this grizzly process engineer, who has seen many claims before re paper feasability, become well, just paper, due to unwarranted optimism.
While the Rueckes paper in Science implied the feasability of single wire pair NEMS memory element, it was trivially apparent that the state of technology of catalytic growth of Carbon Nanotubes could in no way support successful fabrication of single nanotube crossed wire pair memory elements at ANY acceptable yield (which likely remains true to this very day).
Key is that tensioning of nanotubes in anchored tensioned pairs via catalytic wire growth ,was very much impossible, or not quite in the realm of feasability YET!(at the time of nanteros startup)...among other technical issues regarding encapsulation of single wire crossed pairs of nanotubes and so on...
The Science paper's device concept was very much not trivial to build at successful yield, apparent to someone who both had extensive microfabrication expertise, and decent knowledge of state of the art of catalytic nanotube growth - artifacts and all.
Single Wall Carbon Nanotube processing is fascinating, SWNT growth being basically a CVD process comparable in materials science fundamentals to the very old VLS - Vapor Liquid Solid growth, of silicon nanowires dating back to the late 1950's.
BUT as transpired, it appears that Nantero began to understand this limitation to the original vision of the single wire, crossed pair memory element's "commercial" viability, and struck out addressing the challenges needed to overcome to build a viable Nanotube based MEMS ( ok NEMS ) memory cell, without being fixated on Rueckes original concept. GOOD !
Despite the Science paper showing an ideal artist's rendering of the single wire, crossed pair memory element concept, it was very evident that catalytic Carbon Nanotube growth had difficulty to grow single tubes, point to point, under near neutral tension (critical for tight NEMS geometry control and to avoid weird buckling effects of the upper moving tube).
Additionally obvious was that any practical MEMS memory element, had to be encapsulated while not damaging the memory element. which for a single wire crossed pair was not particularly hopeful. (nor avoiding entrapping particulate in the encapsulation either)
Unsealed devices would never survive assembly singulation (chip dicing/saw), and it was not trivial to encapsulate single wire crossed pair memory elements....without damaging them in the process or making the device sensitive to sealed in particles. hmm.. food for thought.
Dropping the ideal vision of Rueckes' Science paper, was key to find a workable manufacturable device concept.
Presently the Crossed Nanotube Matts are polymer spin deposited (NOT GROWN) by a unique polymer formulation provided by Brewer Scientific - where the nanotubes are dispensed onto prepatterned IC circuit wafers in a Brewer Scientific polymer formulation, using pretty conventional PhotoResist wafer track equipment, to great success. At an appropriate time in the process, the polymer is removed (via dry plasma) - resulting in a crossed pair of nanotubes with a controlled gap, devoid of the original dispensing polymer.
(Hats off to Nantero and Brewer for this elegant practical technical advance).
2-3 lithographies seem to be all that is necessary for the cell formation of the crossed nanotube matts, and the multiwire MEMS cells seem to work very well. ( else production would be hard to contemplate )
Key in this "hybrid" nanotube NEMS memory technology, is that the bundles of nanotubes do not appear to have huge limits on the device performance or yield despite the nanoscale "cosmetic" appearance of the bundles.
All my respect goes to The Nantero Team of Rueckes, Schmergel and Segal, and Jennifer Fonstad of Draper Fisher Jurvetson, for sticking with this, until a working device concept could be found and implemented.
CONGRATULATIONS and BEST OF SUCCESS !
Ref - "Carbon Nanotube Based Non-Volatile RAM for Molecular Computing"; Rueckes, Kim, Joselevich, Tseng, Cheung and Lieber; Science Vol. 289, p94, July 7th 2000
p.s. now that I think of it, NRAM's concept actually seems modestly similar to the Iridigm iMOD device acquired by QualComm MEMS used for display applications - but where the iMOD is made from monolithic thin films for a reversible buckling light modulator, versus Nantero's nanotube based memory element. Both Nantero and Iridigm come out of the Boston area, with Iridigm having started back in the mid 90s by Mark Miles of MIT, targetting low power reflective displays at a larger scale of device structure, and not using nanotubes obviously.
And on a similar theme there is Prof. Bloom's Silicon Light Machines which was again a comparable concept of a suspended mems bridge for projection displays and light modulators, with his later follow on firm LightConnect in Newark making telecom mems optical modulators for dynamic gain adjustment / levelling in DWDM. PLus there is the contemporary Cavendish Kinetics MEMs embedded RAM that will have a much smaller Market than nRAM, mostly because of a less agressive market thrust and seemingly lower technology target than nRAM...
All closely related in device concept of a mems /nems crossed bridge structure - but for different applications and materials and size scales, with Nantero aggressively targetting the mainstream high density memory market, where the potential for large revenue is the greatest . Curious connections ...
wendman nano nanotech carbon+nanotube DFJ DFJ+Fonstad Fonstad+Jennifer Fonstad+Jen non+volatile+memory NVM windows+vista NVM+windows+vista windows+vista+fast+boot windows+fast+boot low+power+memory DRAM DDR DDR2 CRV Charles+River+Ventures Sand+Hill+Road Venture+Capital Leiber+Charles Rueckes nanotechnology nram nanotube+memory
This is a pretty telling statement regarding progress made towards development of Nantero's NRAM memory. Something is happening, and it sounds good.
The NRAM technology is a nanotube based MEMS storage element that is formed from a "mems crosspoint" of 2 opposed matts or bundles of nanotubes, 1 suspended as a flexible (MEMS) Nanotube "strip", above the other fixed anchored carbon nanotube bundle below.
Electrostatic forces can cause the upper Nanotube bundle to move into electrical contact with the lower Nanotube bundle, in a non volatile state, but electrically reversible.
( shades of nonvolatile FLASH, but faster than flash, lower power and cheap as a technology node advance )
CLICK READ MORE... for rest of article.
While the device implementation is not as elegant at the seminal single wire pair Science publication concept by Nantero founder Rouckes, then of Harvard's Lieber Lab, in which was predicted NRAM feasability based on single pair, crossed nanotubes; apparently the implemented NRAM device of crossed BUNDLES of nanotubes, works very well and yields well.
YIELD is the key... (as always)
Long ago when I read of the Nantero start-up, I had some pretty emphatic emails with Jennifer Fonstad of DFJ regarding the near impossibility of single pair, crossed wire NRAM devices, to be able to be made with any yield. ( which still holds today btw AFAIK ).
Just a stubborn principled response from this grizzly process engineer, who has seen many claims before re paper feasability, become well, just paper, due to unwarranted optimism.
While the Rueckes paper in Science implied the feasability of single wire pair NEMS memory element, it was trivially apparent that the state of technology of catalytic growth of Carbon Nanotubes could in no way support successful fabrication of single nanotube crossed wire pair memory elements at ANY acceptable yield (which likely remains true to this very day).
Key is that tensioning of nanotubes in anchored tensioned pairs via catalytic wire growth ,was very much impossible, or not quite in the realm of feasability YET!(at the time of nanteros startup)...among other technical issues regarding encapsulation of single wire crossed pairs of nanotubes and so on...
The Science paper's device concept was very much not trivial to build at successful yield, apparent to someone who both had extensive microfabrication expertise, and decent knowledge of state of the art of catalytic nanotube growth - artifacts and all.
Single Wall Carbon Nanotube processing is fascinating, SWNT growth being basically a CVD process comparable in materials science fundamentals to the very old VLS - Vapor Liquid Solid growth, of silicon nanowires dating back to the late 1950's.
BUT as transpired, it appears that Nantero began to understand this limitation to the original vision of the single wire, crossed pair memory element's "commercial" viability, and struck out addressing the challenges needed to overcome to build a viable Nanotube based MEMS ( ok NEMS ) memory cell, without being fixated on Rueckes original concept. GOOD !
Despite the Science paper showing an ideal artist's rendering of the single wire, crossed pair memory element concept, it was very evident that catalytic Carbon Nanotube growth had difficulty to grow single tubes, point to point, under near neutral tension (critical for tight NEMS geometry control and to avoid weird buckling effects of the upper moving tube).
Additionally obvious was that any practical MEMS memory element, had to be encapsulated while not damaging the memory element. which for a single wire crossed pair was not particularly hopeful. (nor avoiding entrapping particulate in the encapsulation either)
Unsealed devices would never survive assembly singulation (chip dicing/saw), and it was not trivial to encapsulate single wire crossed pair memory elements....without damaging them in the process or making the device sensitive to sealed in particles. hmm.. food for thought.
Dropping the ideal vision of Rueckes' Science paper, was key to find a workable manufacturable device concept.
Presently the Crossed Nanotube Matts are polymer spin deposited (NOT GROWN) by a unique polymer formulation provided by Brewer Scientific - where the nanotubes are dispensed onto prepatterned IC circuit wafers in a Brewer Scientific polymer formulation, using pretty conventional PhotoResist wafer track equipment, to great success. At an appropriate time in the process, the polymer is removed (via dry plasma) - resulting in a crossed pair of nanotubes with a controlled gap, devoid of the original dispensing polymer.
(Hats off to Nantero and Brewer for this elegant practical technical advance).
2-3 lithographies seem to be all that is necessary for the cell formation of the crossed nanotube matts, and the multiwire MEMS cells seem to work very well. ( else production would be hard to contemplate )
Key in this "hybrid" nanotube NEMS memory technology, is that the bundles of nanotubes do not appear to have huge limits on the device performance or yield despite the nanoscale "cosmetic" appearance of the bundles.
All my respect goes to The Nantero Team of Rueckes, Schmergel and Segal, and Jennifer Fonstad of Draper Fisher Jurvetson, for sticking with this, until a working device concept could be found and implemented.
CONGRATULATIONS and BEST OF SUCCESS !
Ref - "Carbon Nanotube Based Non-Volatile RAM for Molecular Computing"; Rueckes, Kim, Joselevich, Tseng, Cheung and Lieber; Science Vol. 289, p94, July 7th 2000
p.s. now that I think of it, NRAM's concept actually seems modestly similar to the Iridigm iMOD device acquired by QualComm MEMS used for display applications - but where the iMOD is made from monolithic thin films for a reversible buckling light modulator, versus Nantero's nanotube based memory element. Both Nantero and Iridigm come out of the Boston area, with Iridigm having started back in the mid 90s by Mark Miles of MIT, targetting low power reflective displays at a larger scale of device structure, and not using nanotubes obviously.
And on a similar theme there is Prof. Bloom's Silicon Light Machines which was again a comparable concept of a suspended mems bridge for projection displays and light modulators, with his later follow on firm LightConnect in Newark making telecom mems optical modulators for dynamic gain adjustment / levelling in DWDM. PLus there is the contemporary Cavendish Kinetics MEMs embedded RAM that will have a much smaller Market than nRAM, mostly because of a less agressive market thrust and seemingly lower technology target than nRAM...
All closely related in device concept of a mems /nems crossed bridge structure - but for different applications and materials and size scales, with Nantero aggressively targetting the mainstream high density memory market, where the potential for large revenue is the greatest . Curious connections ...
wendman nano nanotech carbon+nanotube DFJ DFJ+Fonstad Fonstad+Jennifer Fonstad+Jen non+volatile+memory NVM windows+vista NVM+windows+vista windows+vista+fast+boot windows+fast+boot low+power+memory DRAM DDR DDR2 CRV Charles+River+Ventures Sand+Hill+Road Venture+Capital Leiber+Charles Rueckes nanotechnology nram nanotube+memory
posted by Mark Wendman at 6:08 AM
0 Comments:
Post a Comment
Links to this post:
Create a Link
<< Home