Slashdot Mirror
Stretched Silicon Speeds Semiconductors
byrd77 writes: "IBM is touting new 'strained' silicon as being up to 35 percent faster while reducing power requirements. Let's hope this is more than just an exercise in straining credulity..." See also their press release.
Why the dubious attitude towards IBM? IBM's discoveries have led to significant breakthroughs in technology. I especially like their GMR technology, and although they DO make money from licensing it (OH NO!!!!!!), it HAS advanced all aspects of magnetic storage substantially. Their SOI technology is also very promising. And look at this, they continue to innovate.
So why this "LOLOL IBM WANTS MONEY!!!" commentary? They're a corporation, yes, but their inventions have a wonderful impact on computers.
1. It'll never work - thats right, if the Ph.ds at the IBM lab had consulted me and my RPG-playing zit-studded freshmen pals, they would have understood how idiotic they were to pursue this work, and to claim that it was a success.
2. Those greedy bastards - how dare they hold this innovation closely and not share it with openwhatever.org?
3. What does this have to do with linux? Philip Greenspun is cool! Craig Mundie sucks dick! RMS and Linus rule too! Steve Case also sucks dick!
Thank you, this is /. bitch, signing off.
My first question is that researchers at IBM's Research Lab have been working on this for IBM for literally years. I remember reading papers from IBM on this exact same subject back six or seven years ago. They've been fabricating devices back in strained Si/SiGe interfaces and has been presenting papers documenting the mobility improvement for some time now. So the thing that I find truly puzzling is what is the breakthrough? It's not that I don't think this is great, but I fail to see what they are doing now that they weren't doing before.
So what is strained silicon - essentially it's a way of using the lattice mismatch of silicon and SiGe (silicon germanium) to create tensile strain near the material interfaces. This strain reduces carrier scattering and thus improves mobility for both electrons and holes in the inversion layer of the transistor channel. So, in less engineering speak, the charge carriers in the transistor move around easier and thus faster which improves transistor performance. This mobility improvement can be as high as 70% faster than 'normal' silicon channels.
It's worth mentioning the downsides of this technique - which I notice have been ignored in all the articles that I have read. Thermal conductivity of strained SiGe is substantially lower than 'normal' silicon - like an order of magnitude less. So the devices will be much hotter. This 'self-heating' of the devices results in reduced mobility of the charge carriers due to increased carrier scattering - so essentially the devices are so much hotter than they greatly reduce the effect that was created in the first place. Another issue is the fact that junction leakage is much higher. And another is that the higher dielectic constant and lower band-gap of SiGe result in higher junction capacitances in the transistors.
The technology is interesting, but I don't see how they managed to address the issues that have held back the technology so far. It's a shame that they didn't mention potential issues and how they worked around them in the press release, but I guess we'll have to wait for the technical papers at this year's conferences.
* Not Speaking for Intel Corp. *
IBM is using Silicon On Insulator to make microchips these days. In an SOI process the silicon you actually make the transistors on is very thin, at most a few microns, and beneath that is an insulator, such as SiO2.
If instead you do epitaxial Si deposition on another insulating crystal, perhaps sapphire, with a lattice slightly larger than Si's, you get the benefits of an SOI process *and* this improvement from a strained lattice.
This is a real technical coup on IBM's part.
As for diffusion, well, IBM can use dopants with lower diffusivities should this be a problem for part longevity.
So, you need to build the silicon on top of a substrate, with a similar crystal surface, but a larger lattice parameter(s). Then grow the silicon on top of it by some technique that maintains atomic level consitancy between the layers.
This is difficult to do - your basically talking something along the lines of silicon deposited by some for of epitaxial growth - and for thick layers that's a timeconsuming process. And thus expensive.
One thing that was not mentioned was the cost of this trick - how does it compare with germanium or gallium arsnide? (Ok, projected to compare?).
I think that, baring some lucky find, this is going to be more expensive than the befefits, for general use.
Not only that, but the interdiffusion coefficents of a strained material are, in general, faster than for the an unstrained material, so this will decrease the lifetime of the devices.
Interesting idea, though.
--
... googly moogly!
This must by the first time in months an AC has been moderated up to +5.
Wow. I'm not the only guy who moderates at -1...
-grendel drago
Laws do not persuade just because they threaten. --Seneca
In an age of pushing the envelope where even "compatible" mobo/CPU combos can be unstable because of manufacturing deficits we now add an additional source of strain to the hardware.
I give up, where's my 486?
That would kind of make overclockers shoot themselves in the foot.
Don't be too discouraged. Copper interconnects in IBM's chips didn't take nearly as long as I saw some people predicting.
Depends.
Sometimes IBM manages to get stuff out fairly quickly(like copper interconnects on chips).
OTOH, I think it's been about a decade since I watched on CNN a demo of IBM voice-recognition software that was supposed to eliminate the need for keyboards, and allow you to easily dictate in documents, where synonyms ("to", "too", "2", et cetra) were corrected based on context and so on.
I couldn't help but notice however, how in the demo the software kept getting ahead of the demonstrator, "correctly predicting" and displaying things like a specific noun in a sentence before it was uttered, evidently based merely on her having begun the sentence with the word "The"...
Probably when they develop a 4-dimensional projector that will allow them to project a solid shadow of a 3-D mask into the inside of a block of silicon, and a teleportation device to get the dopant atoms inside.
I didn't even know Joan Rivers was in the microprocessor industry.
Way to go Joan.
Oh, _silicon_!
These days even process shrinks don't give use the speedups they used to (edge effects, RC delays and all that) - however everylittle bit helps - 35% faster silicon probably means 10% faster chips overall - what we really need is a new way to do interconnect - lasers? room temp superconductors? quantum comunications ["I put a cat in my box I don't know if it's dead why don't you check to see if there's any food missing at your end?" :-]? esp? etc etc
Check out the picture at this article at Yahoo.
Since this guy beat me to the punch on the submission, here's the link to the CNN story that I included: http://www.cnn.com/2001/TECH/ptech/06/08/ibm.silic on.ap/index.html. Just in case anybody wanted it.
--------------------------------------
If Murphy's Law can go wrong, it will.
I wonder if/when IBM and Toshiba would cooperate and combine technologies - or if the two technologies are mutual exclusive. Obviously not a lot of technical data is out for real comparison but does anyone else know? I only have rudiments of Chip Design from my EE classes so I don't claim to:)
Still with the X Architecture claiming 10% better peformance and 20% less power dicipation combined with the 30% gains IBM claims, there is turning out to be a lot more room before we hit the physical limit of Moore's law.
Cool.
Ad in classifieds: Pandora's Box (no box) $5
Real or Windows Media only.
--I assume full responsibility for my actions, except the ones that are someone else's fault.
--I assume full responsibility for my actions, except the ones that are someone else's fault.
Let's hope this is more than just an exercise in straining credulity...
Well, it isn't straining credulity, I read it on Slashdot so it HAS to be true!
Just an observation.
If my Mom was reading their press release and all the drooling media blather on it, she'd be left with the impression that IBM had just invented something new and spectacular.
They haven't. Strained semiconductor production and use has been in the labs for *ages*. Nortel and every other company making lasers and optoelectronics components for fiber-optics systems have been using much more exotic strain-compensated quaternary (compound) semiconductor materials systems for ages. And it was all originally "discovered" in public research institutions and universisites.
All IBM has done is figured out how to work it into their current production line (which is in fact quite an accomplishment).
Let me know when AMD gets strain compensated SiGe into their CPUs. Then I'll get excited.
its interesting, but AMD is already incorporating a new type of silicon into their chips. The price increase it supposedly quite small, and the performance gain is similar to IBMs technology... but is said to have incredible heat conductive properties, bringing the athlons to almost no Heatsink levels of temperatures. Since this technology AMD is putting in their chips is an extremely pure form of silicon, it might not be compatitible with IBMs research, but you never know.