IBM Demos Atomic-Scale Circuitry

Christopher Pereda sends us a LA Times story about IBM demonstrating atomic scale circuitry. Or see IBM's press release. Who needs Coppermine?

Comments

Re:*The* Nanotech Book?

[jblackman]

Yeah, calling Drexler the world's foremost authority on nanotech is, in my eyes, completely accurate. Remember the ceiling fresco in The Diamond Age? With portraits of Feynman, Merkle and Drexler? As of today, those really are the big three.

Drexler's written a couple major nanotech books, Engines of Creation which is basically the what of nanotech, and the big one, Nanosystems, which is the how.

Nanosystems is a highly technical (and through no fault of its own, hard for me to understand) book, but if you'll spend a little time with it, it'll prove to be absolutely fascinating. It's somewhat expensive (around textbook prices) but certainly well worth it, especially if you have groundings in chemistry and physics. It's *the* nanotech book of today, and likely will be for quite some time to come.

-jay

Binary?

[StarFace]

What I find interesting is that the it seems they are still driving to create a binary system. At least, that is what the media is protraying. Perhaps the research going on is a bit more open minded. It would seem to me that technology such as this would allow us to leap beyond the on/off limitations that we have now. There has been a few efforts to create a trinary system. I believe a research facility in Canada is working on something along those lines.

The problem with anything much more than binary is that you'd have to drastically increase the voltage to get any sort of differentiation between different voltage states, or you would have to delay the reactor to wait until the initial spike levels out, which would defeat any speed enhancement +binary would give you.

It would seem to me that these "nanotechnology" advancements would give us a much more refined control, a much more liberal control, over what is going on electrically, and allow a vast amount of <i>states</i> that the relay could be set at.

Are there any nanotech/quantum/electric buffs out there who could add a little reality insight to this theory? I do not know enough about it to really go beyond wondering.

It's a first step, of many...

[WNight]

This is akin to the invention of wire, for building CPUs. They have a way to pass a signal from one place to another. Albeit, a very fast, energy efficient way.

We've seen STM pictures that show an echo of one atom, but does that mean that the STM is innacurate, and the image is being distorted by the ring of atoms? Or does it mean that the atom actually appears, at an atomic level, to be in both positions?

If it's just an artifact of the STM, then you'd need to use an STM to view this echo, and it won't be of much use. If it's actually an echo of the atom in the new position, then you should be able to detect a change in the atomic property at that location somehow. If it uses electrons as waves, focusing them, like the whisper room does sound, then surely an atom at the other focal point would be effected.

And, if this does effectively change the atomic state at a distance, in a way that can be read without an STM, then you'll only need transistors, and a way to move the atoms and you'll basically have the tools for a CPU. Perhaps, if this works, putting two source atoms in, such that both cast an echo in the same spot, could be used as a transistor, where both need to be in place, and echoing on the output, for the output to change.

But, a circuit needs to change. If you use voltage, the voltage can easily be changed. If you use the position of atoms, how do you move them?

So, you need three things, one of which they might have part of. Signal paths, gates, and changeable states.

They do have a signal path, but not over an arbitrary path, and it would be hard to extend, because you'd need two interlocking rings, one so that the echo of the first created its own echo. But, that would mess up their elipse and probably stop the echo from being created in the first place.

Any ideas on how to overcome these issues?

Re:How do you read the "mirage"

[irongull]

IANAQP (I am not a quantum physicist) but I do have a Nature online account, and I've read the actual scientific paper, so I'll take a stab at it.

This 'echo' is actually an echo of something called the Kondo effect. Basically, when you have a single magnetic molecule (like cobalt) in a non-magnetic metal (like copper) and you lower the temperature, the electrons on the surface of the copper begin to align their spins to cancel the magnetic moment of the cobalt atom. At sufficiently low temp, these shielding electrons enter a many-particle single-spin quantum state that completely masks the cobalt's magnetic moment. This is the Kondo effect. The ellipse on the Cu surface creates a number of possible waveforms (more properly, eigenstates) that can refocus this resonance to the other focus - creating another Kondo effect where there is no cobalt atom. This effect can be measured with a STM (scanning-tunneling electron microscope). IIRC, STM uses a very, very fine molecular 'tip' that is passed above the sample. As the tip moves over an atomic surface, a tunnelling current is generated that is proportional to the distance between the tip and the sample. This is commonly used to generate topographic maps of electron density around single atoms.

Theoretically, this resonance could be used to sample the orientation of a magnetic molecule at a distance. Of course, measuring this moment would disrupt it as per Heisenberg's Uncertaintly Principle, but in this experiment, they are only measuring the presence of the field, not its orientation. Since the effect disappears when the Co atom is moved off the focus of the ellipse, this could also be used sense small movements of atoms at a distance. And it could be used to link the quantum states of two molecules at a distance on a surface, effectively forming a specific quantum 'wire'. They also speculate that ellipsoids could be used to do this in a 3D solid.

Don't get too excited - the effect only happens at 4K (brrrrr) and an electron microscope is a rather impractical sensor, so don't look for it any time soon. But its still cool.

I may very well have butchered some or all of this explanation. I welcome any corrections or clarifications.

Another link with more info ...

[Troed]

(which I submitted yesterday and got rejected - have we heard that one before? ;)

link

How do you read the "mirage"

[konstant]

For the benefit of those who prefer to think on a more graspable scale, IBM is exploiting an interesting property of closed ellipses. Namely, that a disturbance at once of the focii will create a miraged disturbance at the other focus. If, for example, a swimmer dives into an elliptical pool and strikes a focus, a splash will actually appear at both that focus and the one on the other side of the pool. Similarly, IBM sticks a cobalt atom at one focus of an elliptical ring of cobalt atoms. A miraged cobalt atom appears at the other focus, I'm guessing this is because atoms can be expressed as probability waves - which look a bit like the splash from a diver - and the overlap of all these waves causes an elliptical reflection. If somebody understand particle physics fairly well, I'd appreciate a clarification on that point.

Anyway, what interest me far more is how IBM plans to read the state of this "circuit" without causing a sever disruption, per the Heisenberg uncertainty principle. You can turn the device on or off by pushing the odd cobalt atom around, but surely attempting a read operation on the device would cause its state to alter? Does anyone have ideas as to how they would avoid this?

-konstant
Yes! We are all individuals! I'm not!

Re:Not too excited

[Spazmoid]

I disagree, if you shrink the size of the computing device,you increase the power/space ratio. We are currently struggling with processors that run with 64-128 bit datapaths. If you can fit the equivelant of 50 Athalons in the space of one, at the same speed, and parallize them then the increase is much more noticable. You have effectively gone fron a 32 bit bus to a 1600 bit bus or five 320 bit bit buses. Even if the 'processor' were taking data at 500Mhz you would be able to crunch the same amount of data in much less time. Think about a single processor that coule render one frame of video the quality of "Toy Story" or better in a second or less. The implications/applications are endless.

Food for thought.

[lifebouy]

I see a lot of posts as to how lame this is.
Please excuse IBM for proving something we have all suspected for most of our geek lives. Give credit where its due.
It seems to me that at this point its all a matter of arranging these little mirage pools such that the output of two produce enough of a mirage in a third to produce an output in the third.
-^------^-
-^--------^--
-^------^-
something like that. if you can make it so that both inputs are required to produce an output, you have a logic gate. The spacing would determine whether or not it was an AND or OR gate.
NAND and NOR should be a simple matter of reversing the output's connection to the next pool(s) in line.

I really don't see this as far off in the future.
My LCARS may not be too far off after all...

This is nothing new...

[Aurik]

HP Labs demonstrated a molecular logic gate last July. See the article here or news coverage at ABC News.