Carbon Nanotube Semiconductor Possibilites

hin writes "At the recent ISSCC in San Francisco, a review paper was presented at the conference in regard to the feasibility of carbon nanotubes for use in semiconductors. " As most people know, put nano in something, and you'll get my attention-but these engineers are talking about using carbon nanotubes that would be hexagonally shaped (thoughts of benzene rings run through my head), and can be as small as 1.4 nm, and as long as 10 microns. There are still a significant number of hurdles to jump before having this even practical, but as the artile states: "It's becoming an experimental field of research rather than a theoretical one."

Make your own [Nanotubes for Sale!!!]

[Sync]

If you would really want to, you can make them at home. The process isn't to difficult. The tubes are grown using a metal catalyst and a carbon containing gas (CO or easier, methane).
The catalyst is made by impregnating a carrier (e.g silica) with a Ni salt. Heat it up and send a mix of N2/He/Ar + CO/CH4 over it.
Maybe a bit dangerous to di it at home, but you could ask your chemistry teacher :-) (he should also be able to find some articles with a more correct way of making them %)

Bas

Overly optimistic

[Artemisia]

This article was definitely overly optimistic. There are several major problems with nanotubes. Currently we have no way of choosing whether we make metallic or semiconducting tubes.

And we have no way to know whether they are metallic or semiconducting without hooking up traditional wires to them.

And we have no way to arrange them as we like on a surface.

And if we did, we couldn't make devices out of them, since we have no way to make decent electrical contacts between them.

It is not that I don't respect the people who are working on tubes. I work with a bunch of them, and they are very clever people. But they would be the first to tell you that they aren't likely to be able to make computers out of these things.

Making buckyballs

[jwilloug]

They were originally discovered by vaporizing carbon in a vacuum, then shooting it through a mass spectrometer. The current (last I heard) method for making buckyballs (not tubes) is still very similar. Suspend two graphite rods in a bell jar, with their tips a fraction of a centimeter apart. Run a current through the rods so you get sparking across the gap, which vaporizes the graphite. Soot will collect on the inside of the bell jar, and you can seperate out the fullerenes by washing it in benzene.

Presumably you do get some miniscule amount in any sooty flame, but it's really not measurable. A little hydrogen in the environment goes a long to impede fullerene formation. There are measurable amounts in some types of lightening-struck rock, though. The only natural occurance of stuff.

maybe the next great thing, maybe...

[Anonymous Coward]

These nanotubes are really interesting. They have intriguiging (sp?) optical properties, they are the strongest fibers yet known, they have exceptional thermal conductivity, and they might have great electrical properties. But there are a few significant things that need to be overcome before these are at all ptactical.

First, you have to be able mass produce these tings. In the paper I saw, they said that they were using atomic force microscopes to move these one at a time. We're not talkin' pentiums here (thank god).

Second, the electrical properties they displayed are not very desirable. The contacts between these tubes and metal has about 500 kOhms resistance, that's a lot. This will seriously slow anything down on these things. Now, they said that they think they can reduce that by a factor of 10, but wait untill they do.

The traditional devices they showed (tubeFET) didn't really have a usefull aplication. The output resistance was very low (this time you want to be able increase it), therefore the gain was low. Digital as well as analog curcuits using these will not work (the gain is actually 1).

One of the more interesting devices has not even been built yet, that is the single electron transistor (SET). Theswe look nice because they are small and fast, but we have yet to see then. An other issue with these is that we may not be able to build good analog cuircits with these. If you think we're lining in a digital world, then how does the 24 bit music on your super-hi-fi sterio get into that format? You need an analog to digital converter. For 24 bit, you need to be able to resolve signals to 1 part in 16 million. with 1 electron, hmmmm, that's tough. But, there may bw architectures that can do this, maybe oversampled ADCs will be able to use these. The thing is, this has yet to be seen. We may have a future where analog circuts are still on silicon, and digital circuits are in 0.01 um, 5000 GHz nanotubes. I guess it wouldn't be so bad.

That said, I really like carbon nanotubes. I think that they may be the next great thing. They certainly have more promise than any of the other new technologis to come along since silicon. I just think that there are a lot of hurdles to get over first. It's tempting to get excited at the first sign of hope (I know, I've done it several times), just realize that this could still fall apart like a Chinese motorcycle.

Matt

More about buckytubes and buckyballs

[WillWare]

The hexagonal shape refers to the arrangement of atoms along the surface of the tube. This is because both buckytubes and buckyballs are rolled-up sheets of graphite (rolled into cylinders and spheres respectively). Graphite is a hexagonally-tiled arrangement of sp2-hybridized carbons. Normally this hybridization has two single bonds and one double bond, so you could think of the graphite as a hexagonal tiling with a carbon atom at each vertex, and all the edges in one direction being double bonds, and all the remaining edges being single bonds.

Actually, X ray diffraction studies show that the bond lengths are all identical, so it's really more like each is a 1.666-order bond. This phenomenon, called resonance, also appears in benzene. It is thought to be a quantum superposition of all the possible different ways of arranging the double bonds. (What we call a double bond is actually a superposition of a sigma orbital and a pi orbital.)

The pi orbitals in a sheet of graphite or buckytube all blend together, and depending on the tube's chirality (how the hexagons are oriented relative to the cylindrical axis), this can either allow electrons to move up and down the tube very easily, or it can give semiconductor-like behavior. So a trick to building these kinds of circuits is to find joints that will allow you to join tubes of differing chirality.

Fullerenes are generally quite stable molecules, so it's not too surprising that they describe difficulty in getting current into and out of the buckytube. It turns out that it's not too hard to stick little molecular pieces onto the sides of buckytubes. Al Globus at NASA [nasa.gov] has done a lot of thinking and simulations relating to applications of nanotubes, including adding teeth to make them function as gears. [nasa.gov]

Possibly the best source of information on fullerenes is Richard Smalley's Center for Nanoscale Science and Technology [rice.edu] at Rice University. Smalley received a Nobel prize for the discovery of fullerenes. The CNST has an interesting-looking PDF document [rice.edu] describing the Carbon Nanotechnology Laboratory, and discussing the science of fullerenes and some of the potential applications. Fun stuff.

After the TubeFET

[Fly]

The TubeFET is the precursor to the Bipolar, Optical Bose-Einstein Amplifing Field Effect TransisTor, or BOBAFETT, which actually developed a long time ago, in a galaxy far, far away.

Tiny researchers

[shadowdark]

there was an article in Science magazine back about a year ago explaining that some researchers had discovered that changing a nanotube's shape (I believe by changing a very few bonds) that it would go from conducting _very_ well to insulating quite well.
as far as schools doing research in fullerenes, we are a little tiny school (3500, including grad students in the seminary) and a few of my friends spend the summer producing buckyballs.
(which just goes to show that the really cool stuff doesn't have to be done by the big boys, a nice garage can be the spawning pool for new technologies :)

Maybe just optimistic?

[Anonymous Shepherd]

Yeah, I understand your caution, but the article really is hype. As I understand it, science has always been hyped to the public as the next best thing since sliced bread =)

There are definite possibilites and strange characteristics, which we may be able to harness even if not in a traditional computing sense. Even the fact that fullerenes can display metallic, semi-metallitc, insulative, and superconductive properties, depending on impurities and construction, gives us a lot of hope that we can do something unique and wonderful with them. Perhaps the electrical contact issue can be resolved by using the tubes themselves as wires, rather than metals... I'm sure many options are being considered, discarded, re-examined, and discovered.

Though I applaud your voicing caution amid hype =)

Couldn't help but notice your email... A physicist, perhaps? Materials science? Chemist? I think cory is the EE/CS network, because my friend is also at Berkeley right now. What real options/opinions and ideas are being done right now for fullerenes?

AS
AS

Nanotubes for Sale!!!

[slothbait]

I was browsing the link someone posted to the "Center for Nanoscale Science and Technology" at Rice University, and they have a link where you can buy nanotubules created at Rice for research purposes. Of course, it would probably take an electron microscope to SEE the things, but somehow this strikes me as cool. Tubes for sale:


http://cnst.rice.edu/tubes/


--Lenny