Reader Request: Graphene

Last week’s Reader Request Thread produced a bunch of good suggestions, some of which I’ll be responding to this week as I put the last touches on the book draft and send it off. We’ll start with a good physics question from Moshe:

So, what do you think about graphene? the next big thing, or just the latest fad?

Doug might be a better person to answer this, as it’s a little closer to his field. It’s unquestionably the latest fad, the question is whether it’s a fad with legs or not.

Graphene, for those not up on the subject, is basically just carbon in one-atom-thick sheets. It’s two-dimensional graphite, more or less– lots and lots of carbon atoms bound together in a regular structure that looks a little like chicken wire.

Graphene is a relatively recent development– the Wikipedia link above says it was first studied in 2004, which sounds about right– but interest in it has really exploded. Not a week goes by without a story in Physics World about some amazing new development involving graphene– they account for a fair fraction of the materials stories I tag on del.icio.us. It has all sorts of interesting electrical and thermal properties, and there’s a lot of hype about the revolutionary applications that are just around the corner.

In terms of the “fad” question, this really reminds me of the explosion of hype about carbon nanotubes a few years back– you couldn’t flip open a science magazine without seeing three or four stories about the nifty things people would be doing with nanotubes any day now. Progress there has slowed substantially– not to a halt, certainly, but the promised wonders haven’t quite emerged.

It’s not clear whether graphene will follow the same trajectory or not. My half-assed guess would be that graphene stands a better chance of being put to use quickly, just because it seems like it should be easier to produce and manipulate– my impression is that it’s kind of hard to grow nanotubes in the arrangements you would like for interesting electronic applications, but you get graphene in sheets by peeling it off graphite, and can make whatever connections you like. I could be way off base, though.

I guess I’d say I’m cautiously optimistic that it might be a real Big Thing and not just a passing fad. That’s mostly a guess, though, not a terribly well-informed opinion.

Comments

  1. #1 Uncle Al
    July 28, 2008

    Graphene is stuff (chemists). Its support arises from projected things (engineers). Engineers feigning chemistry ooze putrescence. Add PERT-charted management for improved means to deteriorated ends, a triumph of overwhelming ignorance impelled by overweening arrogance.

    What assembles graphene into graphite? Aromatic stacking. Graphite intercalcates like a fiend, levitating single graphene sheets. Surface energy and wetting are not mysteries. Sonicate graphite in a suitable solvent (NMP has nice numbers, maybe hot DEET). A pi-acid solvent should disassemble graphite at will.

    That is monthly progress reports dripping milestones (bullbleep). Stop whining; make the stuff to order. Add revealable thiol terminations. That stuff will parallel self-assemble into things on a nano-patterned gold grid. Was that so difficult?

    Somebody whisper “UNKNOWN HAZARDS!”. $millions for R&D, $billions for peremptory apologies.

  2. #2 djlactin
    July 28, 2008

    what have you been snorting, uncle al?

  3. #3 Max
    July 28, 2008

    Uncle Al sounds an awful lot like timecube.

  4. #4 John Novak
    July 28, 2008

    As a nitpick, I don’t think the carbon nanotube results have slowed, per se. They just haven’t accelerated, and while they *were* over-hyped at the time, the hype has gone in search of new targets so the overall effect looks like they’ve died. I read IEEE Trans. Nano on a regular basis, and there are always a slew of CNT papers there.

    (Don’t ask me for details, I’m looking for other things in that journal and I only skim the CNT entries at best.)

    Graphene strikes me as being about the same.

  5. #5 Chad Orzel
    July 28, 2008

    what have you been snorting, uncle al?

    Graphene, of course.

    As a nitpick, I don’t think the carbon nanotube results have slowed, per se. They just haven’t accelerated, and while they *were* over-hyped at the time, the hype has gone in search of new targets so the overall effect looks like they’ve died. I read IEEE Trans. Nano on a regular basis, and there are always a slew of CNT papers there.

    I don’t follow nano that closely, so I’m mostly going by “news” type stories– Physics World and that sort of thing. In other words, all I see is the hype.

  6. #6 Moshe
    July 28, 2008

    Thanks Chad, this is consistent with everything I hear. From where I am standing the dynamics of this fad is slightly different from other fads in that there is a lot of excitement from theorists as well, so it is not driven just by (often overhyped) technological applications, as you mention this stuff has some really interesting properties. It would be amusing if the stuff in our pencils is the next big thing, so I am hoping it will not go away anytime soon.

  7. #7 Art Vandelay
    July 28, 2008

    As a grad student working on a graphene project, I think it will be incredibly difficult for graphene to make any significant inroads in terms of mainstream applications in the semiconductor industry, which is really where the primary interest in the material lies.

    Take a look at how the stuff is made. Many, if not most, of the papers in the present literature make use of “exfoliated” graphene, which means that a grad student peeled some layers off of a hunk of graphite using Scotch tape and stuck it down on a wafer with a layer of SiO2. We’re trying to talk about using tape off someone’s desk to manipulate a layer of material that is exactly 1 atom thick, and then using that material to replace what is probably one of the most highly engineered devices in the history man, the silicon CMOS transistor. As Shankar Das Sarma has pointed out (PNAS 104, 18392 or PRL 98, 186806), graphene obtained in this way has so many charged impurities that the transport properties are completely dominated by defects, and the intrinsic carrier mobility cannot be measured.

    OK, this scotch tape business was a cute lab experiment and spurred some initial interest in graphene, but it sure would be nice to get a more pure form of the stuff in some other way. A number of groups have succeeded in forming epitaxial graphene on silicon carbide by heating the substrate to 1400C or so, such that a few monolayers of Si are desorbed, leaving behind carbon that happens to recrystallize as graphene. This approach is not without its problems. First, temperatures of 1400C are incompatible with existing CMOS processes due to diffusion issues, so a completely new process flow would have to be engineered. Second, people have been trying to make electronics on SiC for many years – it is an attractive wide-bandgap substrate. But SiC is chock full of defects; until recently, it was impossible to get a wafer without micropipes, which are literally holes in the substrate due to agglomerations of dislocations. Compared to the perfection with which a 300mm diameter Si wafer can be produced, SiC is complete crap.

    Next comes gate dielectrics. Silicon forms a nice dielectric – throw in a little steam at 800C or so (a process that has been refined in more detail than you might imagine over the course of 50 years) and you form a native oxide with phenomenally low defect density. Makes a nice capacitor. In the last few years, device scaling has run the SiO2 thickness down to the point that gate dielectric thicknesses were getting into the 10-20 angstrom ballpark, at which point tunneling of electrons through the barrier became a significant issue. So, quite a few years and many millions (probably hundreds of millions) of dollars were spent on figuring out how to throw in a little bit of high-k material, like hafnium silicate, so that dielectric can be thicker to cut down on the tunneling. Now – what material can we use to make such an interface with graphene, and how can we deposit it, and do so without disturbing the electronic properties of this one-atom-thick layer? Many things are being tried, but nobody really knows.

    So… until someone can lay down a few-layer graphene film on an atomically flat 300mm substrate, then put a 10-30 angstrom thick dielectric layer on it with less than 10^10 charged impurities per cm^2 (perhaps much less), and subject to a great many other constraints that really just begin to boggle the mind… I say I’ll believe it when I see it.

    Cliffs notes version: Incredible new technology comes along, greatest thing since sliced bread, people go nuts with ideas, then slowly begin to realize that it’s pretty damn tough to replace a technology with 50 years of development in an industry worth hundreds of billions of dollars a year.

  8. #8 GF
    July 28, 2008

    Art is right, however, … that doesn’t mean there can’t be some interesting science experiments. In principle, at least, electrons in graphene obey Dirac’s equation; meaning they have a linear dispersion relation near the Fermi level and hence act relativistically (at least that’s my understanding – someone correct me if I’m wrong). You have to admit that’s really cool. I have yet to see that effect eploited to significant inpact, but I suspect it will be soon (or has and I’ve just missed it.)

    Anyway, not everything has to beat silicon to be interesting physics. Moreover, not all technology has to compete with silicon.

  9. #9 Frederick Ross
    July 29, 2008

    It’s even better than obeying a Dirac equation. The electrons act like massless relativistic particles in the sheets! It’s our own little 2D, relativistic, quantum mechanical sandbox. And if you dope the graphene you can add effective mass terms. Who cares about the applications? It’s the best toy to come along in years, since the 2D electron gas.

    Which is why it’s a fad. Up until now only the particle folks got to play with the relativistic equations. Think of it as condensed matter spring break.

  10. #10 Uncle Al
    July 29, 2008

    Tsunamis of studies and no applications. It is not a thing problem (engineering), it is a stuff problem (chemistry). No stuff, no things. Optimization won’t change that.

    Synthesize graphene to spec in Pyrex (linked above) for a monodisperse product with selectively reactive opposite edges for spontaneous assembly with gold, silver, or copper traces. 400 carbon atoms is MW = 4800 (call it 5000 with decoration). One microgram of synthetic reactive graphene is 0.2 nanomoles or 10^14 graphene sheets. Given 1 billion gates/chip that builds 10^5 chips. Stop whining.

    Chemically disassociate well-crystallized bulk graphite into graphene with a (pi-acid) wetting solvent that disrupts pi-stacking. NMP and DEET (high temp) plus sonication have the correct surface energy. Aluminum-based ionic solvents with solvocation-solvoanion tunable Lewis acidity are obvious. IT’S BEEN DONE,

    J. Chem. Ed. 81 819 (2004)
    http://www.jce.divched.org/Journal/Issues/2004/Jun/abs819.html

    Chemistry has upped the standards; up yours.