What is Digg?103 Comments
- verkon, on 04/18/2008, -6/+44with just one atom thick, you can pretty much call it two-dimensional.
- smurf22, on 04/18/2008, -4/+37Stop making stupid jokes like that! I remember when digg was intelligible and had comments that actually had to do with the article, now it's just pun after stupid joke. Now as for the feasibility of using these is rather far off, but like all consumer electronics the more it is adapted the cheaper it becomes. I'll look back into this technology in another 2 or 3 years.
- amdforever, on 04/18/2008, -4/+27The problem is not about shrinking size. Intel and AMD have no problem reducing the size of their silicon to below 35nm.
The problem about shrinking has to do with:
(1) Higher power density on your chip makes cooling difficult
(2) To solve power issue, you reduce VTT voltage. BUT, reducing VTT increases the propagation delays, which in turn lowers your clock speed
(3) You also ended up with higher leakage current
In 10 - 15 years time, we should be using optical electronics. - MikeEnIke, on 04/18/2008, -4/+23Pretty much being the key word.
- Mofassa, on 04/18/2008, -0/+19I feel like taking my hand at explaining this as well, without the quantumness
Essentially by 2 dimensions they mean that to describe it, you only need to specify its length and width.
A 3 dimensional material, diamond for example, the crystal which makes it up is in 3 dimensions. You have to consider the relation of each atom to the other atoms in x,y, and z directions. So to say where atom B is in relation to atom A youd have to say something like B is (1,2,3) away from A (I forget the exact crystal structure of diamond)
A 2 dimensional material, such as graphene, only requires 2 of these dimensions to be identified. You can say that atom B is (2,1) away from atom A. It's relation in the z axis is irrelevent as your coordinant system can be made such that it is constant.
similarily, a 1 dimensional material (a nanowire for example) really only requires distance to be identified. So atom B is 3 units away from atom A. No further specification is required.
So while the material exists in 3 dimensions, for the sake of describing it, only 2 dimensions are required. - UnFriendlyFire, on 04/18/2008, -11/+27...graphene, a two-dimensional material...
A two-dimensional material? I think not. - jdepp, on 04/18/2008, -0/+15dynamically two-dimensional from the point of view of electrons -- they can move easily in the plane of the carbon atoms, but motion in the direction normal to the plane of the carbon atoms takes so much energy that it's frozen out even at room temperature.
- revjustin2, on 04/18/2008, -5/+19I can haz tranziztor?
- tHePeOPle, on 04/18/2008, -6/+20Psh. I'll believe it when I see it.
- Sreta, on 04/18/2008, -1/+13It does in engineering. A lot of what engineers do relies on approximations.
- papsmurf, on 04/18/2008, -3/+15they mean 2 dimensional molecular structure. quit trying to sound smart.
- brstilson, on 04/18/2008, -8/+19The thickness of the atom makes it three-dimensional.
- Hananda, on 04/18/2008, -3/+13Look, if you want something approaching a coherent discussion, I'm sure this same article is up on Slashdot right now. For all the faults of the site itself, some of those Slashdotters actually know what they're talking about. This is Digg. We make jokes about cats and retread the same five or six tired arguments over and over again.
- inactive, on 04/18/2008, -3/+12Aaaaand, you're wrong... You can set up any modeling software to use an arbitrary system of measure where 1 easily visible unit = 1 atom width.
- Bilabrin, on 04/18/2008, -4/+11"Pretty much" flies real good in the world of science.
- Seventus, on 04/18/2008, -1/+8They're finding ways around leakage current problems, but heat is still an issue. Electron tunneling is another issue, especially when gates are so narrow. Intel has done a lot of research into the problems of shrinking transistor. It's a very interesting read for those who care.
http://www.spectrum.ieee.org/oct07/5553 - Farik, on 04/18/2008, -1/+8The Singularity is Near...
- DifferentAngle, on 04/18/2008, -0/+7#1 constant field scaling does not increase power density - power density has been increasing because we havent been scaling VDD with the circuit technology as constant field scaling dictates
#2 reducing Vt *increases* drive current - it's VDD that hasnt been scaled, and it's not due to propogation delays since it would still be faster. We havent scaled operating voltage due to variability and noise issues. In particular, it's currently really hard to make 6T SRAM work properly with a supply voltage less than 1.1v, and everyone is avoiding the design delay penalty of moving to 8T (but they cant for much longer).
We're likely to never use optical transistors since they're so bulky. The next technology nodes are going to move to the FIN-FET transistors, likely in 2010-2012. Or at least that's IBM and Intel's current plans. - gn0stik, on 04/18/2008, -0/+7words?
- xyonz, on 04/18/2008, -1/+6Other low dimensional structures include quantum wells (2-D), quantum wires (1-D) and quantum dots (0-D). It's all about quantum confinement of electron wave functions in 1, 2, and all three dimensions (respectively).
- SiliconRain, on 04/18/2008, -0/+5You mean VDD.
And since when have they no problem in going below 35nm? There are HUGE problems in going below 35nm. Remember - 45nm is the absolute cutting edge of what is possible, and the manufacturing processes required to produce devices at that technology node are so complex and outrageously expensive that it blows the mind. Not to mention the design issues in terms of timing, digital noise, parasitics, interconnect delays etc etc etc that massively problemise design at this scale.
Trust me, heat dissipation is the least of these designers worries. - chuckDontSurf, on 04/18/2008, -0/+5As you wish.
- Scrappy1850, on 04/18/2008, -0/+51010101010101010101010101010101010101010101010101
- shadowspawn, on 04/18/2008, -0/+4A friend of mine used to work in design for what was then Motorola in Austin, TX. They were always scrambling to get decent chips out as soon as a change occurred. When they started shrinking the die sizes, lots of bad chips started appearing "in the wild".
Oh they still functioned properly, but one set of engineers would be bitching about the power consumption and the other set (which she belonged to) would bitch about the first group trying to squeeze blood out of a stone. Would never get resolved, no matter how many trips to china they went to. The end result was that for a year or so there were plenty of Motorola handsets that would simply drain the batteries way before they were supposed to.
On a home computer it doesn't make that much of a difference. The mobile market is watching this stuff closely. - jdepp, on 04/18/2008, -1/+5....except that "optical electronics" is practically constrained to be larger than the wavelength of light by Rayleigh criterion, - green light is ~500nm, UV light is ~350nm so current transistors are already ~ 10 times smaller than the smallest optical device ... silicon is indirect bandgap material so it can't emit light efficiently - hence you're into electrooptic modulation or replacing your whole fab to work on Gallium Arsenide = high cost.
- Stupidumb, on 04/18/2008, -0/+4You better be sorry. You made me smile, you son of a bitch. *****. ***** YOU!
- slundal, on 04/18/2008, -0/+3I'm pretty sure you shouldn't eat it.
- slundal, on 04/18/2008, -0/+3absorbing just 2.3% of the light passing through would make me call it transparent.
- moogaman, on 04/18/2008, -0/+3Wow. arguing over 2D vs 3D on a scale of one atom. A lot of people need to get over it. If you think it's 3D because of the thickness of an atom, go ahead. We really shouldn't care this much about defining terms and thicknesses when we are talking about as thin as physically possible.
- Sverre, on 04/18/2008, -0/+3No, the wafer they have made as a thenth of a millimeter, which is quite small compared to the standard 200mm or 300mm silicon wafers. The transistor size would be much smaller - the size is given in the article as 1 atom wide, however much that is in nm.
- enicholas, on 04/18/2008, -0/+3Nope, just ordinary quantum mechanics.
- tHePeOPle, on 04/18/2008, -0/+3Jeez man, I'm really sorry. I was just trying to be funny and I've gone and taken things way too far. I've obviously crossed the line here, and I'll do my best to not let it happen again.
- InfiniteNothing, on 04/18/2008, -1/+4I would say having it's thickness being a statistical distribution qualifies it as not having a real structural thickness perse.
- Sinai, on 04/18/2008, -1/+3Shrug, it all depends on what failure rate/precision you're shooting for. In labs throughout the world, I guarantee you that ~ is being used right now.
- Stupidumb, on 04/18/2008, -0/+2*****, I guess we shouldn't have wasted it all on pencils.
- gn0stik, on 04/18/2008, -0/+2That's the wavelength. Not peak to peak. However, I do agree that scaling optical electronics would be troublesome, so much so that it might not be worth the effort, in light of something like this. However what amdforever was not considering was that the higher the conductivity the lower the resistance, which means lower operating voltage, temperature at a specific voltage, lower leakage current, which means lower power requirements etc, etc.. Almost every single point he made was completely moot with respect to graphene.
Optical electronics might be suitable for intercomponent communication, at the board level, but never for components themselves.
The other thing he may have overlooked is that to get laser diodes down to the size required for even board level operation would require a very small micron process, and very low power consumption. This fits the bill nicely. In fact, his solution, might be dependent on a solution such as this. - masfenix, on 04/18/2008, -0/+2Buried for adding "*****". Quit life if you can't make a simple point without the use of profanity. How old are you anyways?
- Sverre, on 04/18/2008, -0/+2The transistor was 10 atoms wide, not the material. If/when this material becomes useful for electronics, you will stack billions of these transistors in a single chip.
- jdepp, on 04/18/2008, -0/+2the truly interesting thing about graphene is the electrons behave as if their kinetic energy is proportional to their momentum, rather than their momentum squared, eg. in the sheet, the electrons have zero effective mass.
- jdepp, on 04/18/2008, -0/+2the real question here is how big can they make the sheets of graphene - you can buy silicon wafers the size of a plate for a few thousand dollars in quantity, with basically zero defects on the starting material. These graphene sheets are only a few microns square and found one by one under a microscope then wired up and tested, discarding the bad ones that are two atomic layers thick, etc. If you can figure out a reliable way to make plate sized sheets of graphene you'll be the next Gordon Moore.
- moogaman, on 04/18/2008, -0/+2Weee! Look at all that math GO!
- inactive, on 04/19/2008, -0/+2Graphene Ravine.
- ken0624700, on 04/18/2008, -0/+2Silicon Valley needs a new name!
- jdepp, on 04/18/2008, -0/+297.7% of the light goes right through...
point is that the 2.3% absorbed = pi * q_e^2 / (2 * h * c * eps_0).
ie, the amount of light absorbed is controlled by fundamental constants like plancks constant, the charge of an electron, speed of light, dielectric permittivity of vacuum. etc.
pi = 3.142
q_e = 1.602e-19 Coulombs
h = 6.626e-34 Joules / Coulomb
c = 3.0e8 meters / second
eps_0 = 8.854e-12 Farads/meter - chuckDontSurf, on 04/18/2008, -0/+2Well, I wouldn't say they have *no* problem going below 35nm. Neither one has a 32nm chip in production yet (although Intel's is slated for 2009), so there's no telling what kind of yield they'll get. But more importantly, neither has a solid plan of how to go smaller than that.
- mobislink, on 04/18/2008, -0/+2I can't wait until we get annoyingly tiny cell phones. This may make the iPod Flea possible.
- dimplemonkey, on 04/18/2008, -0/+2yeah, yeah. I've heard it all before. One day "invention/discovery" will make our lives easier and it will reduce the costs by "x%". yawn....
So where is my iPod invisio? - stealth210, on 04/18/2008, -0/+2Ok, I stand corrected.
- gn0stik, on 04/18/2008, -0/+2I agree. The higher conductivity means it will run much cooler. Which translates to much quieter. Not to mention the incredible size reduction. Imagine a smart phone with a PC class(or better) processor. Imagine clear displays embedded in contact lenses. The efficiency increase in PV cells. Cheap heads up displays in cars.
The singularity is coming. - enicholas, on 04/18/2008, -0/+2It's not like the silicon they use in computer chips is just melted sand. The silicon wafers have to be unbelievably pure with incredibly regular crystal structures, and they are very expensive to manufacture. The cost of the silicon itself is but a tiny, tiny fraction of the cost of a finished wafer.
Graphene is just carbon. The raw material is in abundant supply, and unlike diamond it's relatively easy to get carbon to transform into graphene. An ordinary candle can do it. We don't know how to make large graphene sheets yet, but when we do I assure you that the cost of the raw material will not be a significant contributor to the overall cost of the finished graphene sheets. -
Show 51 - 100 of 103 discussions
Browsing Digg on your phone just got easier with our enhancements to the 
© Digg Inc. 2009
— Content created and posted by Digg users is