It’s all too tempting to imagine that our macroscopic intuitions can be transferred to the nanoscale world, but these analogies can be dangerous and misleading. For an example, take the case of the buckyball bearings. It seems obvious that the almost perfectly spherical C60 molecule, Buckminster fullerene, would be an ideal ball bearing on the nanoscale. This intuition underlies, for example, the design of the “nanocar”, from James Tour’s group in Rice, that recently made headlines. But a recent experimental study of nanoscale friction by Jackie Krim, from North Caroline State University, shows that this intuition may be flawed.
The study, reported in last week’s Physical Review Letters (abstract here, subscription required for full article), directly measured the friction experienced by a thin layer sliding on a surface coated with a layer of buckminster fullerene molecules. Krim was able to directly compare the friction observed when the balls were allowed to rotate, with the situation when the balls were fixed. Surprisingly, the friction was higher for the rotating layers – here the ball-bearing analogy is seductive, but wrong.
Actually friction at the macro scale often works in this manner as well – it’s the same thing which makes hydroplaneing in your car so dangerous. When the wheels are turning the fiction with the road is *higher* because the tire molecules and the road molecules can temporairly bond. When the wheels stop rotating and start sliding, friction is reduced and your car slides all over the place and can take 2x+ the distance to stop. This is why anti-lock brakes are so valuable! You can get more information by looking up ‘static’ versus ‘kinetic’ friction coefficients – static is pretty much always higher.
Obviously it’s not your mistake, but something is extremely fishy about the article…
An interesting point, which illustrates that I should perhaps been more careful in how I expressed this. You’re quite right to say that in rolling friction, the frictional force at the contact is often greater than it is in sliding friction, but the frictional energy dissipated, is still much less for the rolling case, because in the absence of sliding no work is done by the force. I should have specified that it was the energy dissipated that was being compared in the two cases, not the contact force.
The latest version of the rice nanocar replaced the C60 wheels with spherical molecules of carbon, hydrogen and boron called p-carborane. It seems that the old ones drained too much power away from the light powered motor that was mounted recently.
Buckyballs are crappy bearing because their sp2 bonds are too high energy. Some fullerenes even react with air. Surely there are better molecules for diminishing friction (Noble gas molecules?).
Phillip, I don’t think that can be the reason, as the SP2 bond is pretty stable and graphitic carbon doesn’t start oxidising till rather high temperatures. In any case, the Krim experiments were done in ultra-high vacuum. I suspect the reason is more to do with the fact that the C60, being hollow, is not that rigid, and energy is dissipated during cyclic deformation of the molecule as it rolls. A more compact molecule like the p-carborane that Nanoenthusiast cites would do better, as would “bucky-diamond” (that is to say, small diamond clusters with a sp2 surface reconstruction).
Is energy loss via cyclic deformation the same principle as macroscale “ringing”? I was assuming the C60s would be squished enough to deform plastically. My bad.
It doesn’t really make sense to talk about plastic deformation unless you get wholesale displacement of atoms (i.e. formation of new defects like vacancies and dislocations) which I don’t think you’d see here. I’m not quite sure what you mean by “ringing”. The way I think about it, (which of course isn’t guaranteed to be right), is that in the cyclic deformation you are exciting low frequency normal modes of the system, and energy subsequently leaks from these modes into higher frequency modes to end up as thermal phonons. The condition for this kind of mode mixing to occur is simply that bonds are being deformed enough to take them out of the harmonic regime (i.e. to a point at which the potential can no longer be approximated as a simple quadratic) – this needs a much less severe deformation than one that actually leads to bonds being broken.
Sweet. Free physics lesson. For every different rigid molecule is the bond deformation force needed to induce phonons, a known value?