Slashdot Mirror
IBM Pushes Beyond 7 Nanometers, Uses Graphene To Place Nanomaterials on Wafers (ieee.org)
An anonymous reader shares a report: Four years ago, IBM announced that it was investing $3 billion over the next five years into the future of nanoelectronics with a broad project it dubbed "7nm and Beyond." With at least one major chipmaker, GlobalFoundries, hitting the wall at the 7-nm node, IBM is forging ahead, using graphene to deposit nanomaterials in predefined locations without chemical contamination. In research described in the journal Nature Communications, the IBM researchers for the first time electrified graphene so that it helps to deposit nanomaterials with 97% accuracy.
"As this method works for a wide variety of nanomaterials, we envision integrated devices with functionalities that represent the unique physical properties of the nanomaterial," said Mathias Steiner, manager at IBM Research-Brazil. "We also can envision on-chip light detectors and emitters operating within a distinct wavelength range determined by the optical properties of the nanomaterial." As an example, Steiner explained that if you wanted to modify the spectral performance of an optoelectronic device, you could simply replace the nanomaterial while keeping the manufacturing process flow the same. If you take the method one step further, you could assemble different nanomaterials in different places doing multiple passes of assembly to create on-chip light detectors operating in different detection windows at the same time.
"As this method works for a wide variety of nanomaterials, we envision integrated devices with functionalities that represent the unique physical properties of the nanomaterial," said Mathias Steiner, manager at IBM Research-Brazil. "We also can envision on-chip light detectors and emitters operating within a distinct wavelength range determined by the optical properties of the nanomaterial." As an example, Steiner explained that if you wanted to modify the spectral performance of an optoelectronic device, you could simply replace the nanomaterial while keeping the manufacturing process flow the same. If you take the method one step further, you could assemble different nanomaterials in different places doing multiple passes of assembly to create on-chip light detectors operating in different detection windows at the same time.
The wafers are unique in that they are "off-the-shelf" units manufactured by Nabisco and can be found in quantity on aisle 13.
It little behooves the best of us to comment on the rest of us.
GF reportedly hits a wall at 7nm, but not a word on Intel's Great Wall of 10nm..?
Interesting. I would have thought the graphene was a chemical contamination.
The summary is conflating two related, but distinct things here: 1. the ability to place features on a wafer at dimensions less than 7 nm, and 2. the ability to manufacture large quantities of highly reliable transistors with features less than 7 nm.
While 1. is a necessary condition for 2., it is far from sufficient.
The article itself doesn't discuss GF for good reason, since GF dropping out of the "end of Moore's law" race is irrelevant. However, the article wondering why IBM is investing here while it is de-vesting from semiconductor manufacturing misses the point. This is about development of new tools and technologies, not to squeeze one more node out of Moore's law.
These nanotechnologies can be highly useful in lots of areas distinct from chipmaking, for example, the article talks about light sense which could be very important in continuing advances in neuroscience and physical chemistry (to name too examples). There is REAL MONEY in healthcare and these type of new sensors could potentially revolutionize science and practice in many bio-facing areas.
IBM has gotten out of the chipmaking game and this announcement in no way implies it is getting back in.
Bet I'll still have to wait for this, and wait for that, and reboot when the other thing goes bonkers, but aeroglass will work flawlessly.
Might be worth mentioning in this context that scientists have recently published results of experiments showing that graphene is able to turn alternating electric currents in the GHz range into electric currents in the THz range: https://phys.org/news/2018-09-... Thus, instead of using graphene just as some structured base material, it may make a lot of sense to actually build the electronic circuit itself from graphene.
You can just say B. MM is used to designate million because backwards countries see an M and think thousand (mille) before they think million. When they see a B or "billion" they wonder what happened to "milliard", but they don't get confused further and start thinking of some other B number. They're just lost on the long scale from the first issue.
It's hexadecimal for 59 billion. Coincidentally that may have been the final cost for global foundries 7nm process after cost overruns (jk)
This process by IBM sounds pretty expensive, hence it may be useful for special components (microwave transistors, e.g.), but that is it. I think we need to expect no real advancement beyond 7nm for the foreseeable future. Fine by me, maybe then software can start to catch up again instead of crappy coding just relying on more CPU speed.
Most ACs are not even worth the keystrokes to insult them. Be generically insulted by this and ignored otherwise.
Hey, hey, don't diss milliard, long scale is awesome, pity English speaking world no longer uses it.
Their mainframe division is alive and well. Lots of interesting technology, you should check it out.
Only the State obtains its revenue by coercion. - Murray Rothbard
I actually heard they are planning on just using ARM chips for everything... Not sure how much faith I put in that rumor but if true it would be really weird... It would have to be some extreme upgraded ARM to offer any reasonable performance in a desktop machine?
It would have to be some extreme upgraded ARM to offer any reasonable performance in a desktop machine?
You mean like, with superscalar, shadow registers, multi-level cache, branch prediction, that kind of thing? Recent ARM has all of it, differences with X86 microarch have been steadily shrinking. SMT is one of the few x86 characteristic optimizations still missing from ARM. Not clear why ARM hasn't dropped the other shoe on that one, maybe transistor budget. But I would not be surprised at all to see an ARM SMT reveal by this time next year.
When all you have is a hammer, every problem starts to look like a thumb.
The metric system: 3G$US fully specified, 3G$ short.
Thus, instead of using graphene just as some structured base material, it may make a lot of sense to actually build the electronic circuit itself from graphene.
That's great. Wake me when we actually make something useful with graphene that I can actually buy or use instead of just talking about what a miracle material it is. We keep seeing all sorts of articles about how great it is and yet nobody seems to how to actually do anything useful with it outside of a laboratory.