I had to miss a few of the talks today, but even so I heard quite a lot of interesting stuff. Cool trivia: A DVD has 7.5 miles of track on it. And a possible correction: On Day 1, I described a report of a mechanical vibrating beam that was so sensitive its vibrational energy was quantized. While talking about nano-mechanics, James Hone today said "there are about ten or twenty reasons why this is complete hogwash."
This post will cover lots of cool technical stuff on nanoscale optics and molecule-scale mechanics, including a different kind of quantized mechanical property: strength in the face of defects--with some speculations from me on why this may make nanomachine design easier. Again, here's a link to the schedule and a link to the abstracts.
"Nanoscale Optical Devices" by Rajeev Ram
Semiconductor lasers are more or less the same as ordinary light emitting diodes, except that they're surrounded by mirrors. ... but wait, there are now new technologies to make them even more efficient, such as building quantum dots into them. The nano-optics talk went on to discuss a lot of details about how stuff works, and I learned even more from talking with the speaker afterward.
Fluorescent semiconductor quantum dots have a broad absorption range but a narrow emission range. This is useful for labeling cells under a microscope. It means you can light up several colors of dots with one light source, which is very convenient compared with molecular dyes where you need a different light source for each different color you want to use. I also learned the reason: that it takes time for a highly excited electron to emit a photon, and much less time for it to lose energy back to the level of the basic fluorescent frequency. This is the kind of science detail I love to collect, and this conference is full of such snippets--and people who can explain them well enough to weave them into insights. But I can't convey the insights in a blog post, so I'm leaving out most of the more arcane snippets.
Organic LED's are being made into displays and appearing in cell phones and cameras. But they had a picture of a TV screen two millimeters thick, and another display on a piece of plastic that was curved into a half-circle. And OLED's can be printed "reel to reel" just like putting a print pattern on a piece of fabric, so it should soon be possible to make very cheap displays.
Hot things usually emit heat (light, if they're hot enough) in a broad spectrum. That's why light bulbs make white light. But there's a thing called a photonic crystal that can allow only certain frequencies of light to pass through. They've built a light bulb filament out of a photonic crystal, and it produces a much narrower range of frequencies. Photonic crystals are also useful for new kinds of fiber optics. Most fiber optics use solid glass or plastic to trap and transport light along the length of the fiber. But if you make a long photonic crystal with a hole in the middle, then light will be confined to the hole--the same function as traditional fiber optics, but a new way of achieving it. If the hole is made of air, then you can transport frequencies of light (ten micron) that are absorbed by most fibers. If the hole is filled with plastic, then you can shoot in a single color of light at very high intensity, and it will broaden into "white" light--an effect that's a nuisance in communication fibers, but if it's strong enough, it's useful for making brief intense white light pulses--which are useful for white-light interferometry--which is a very precise way of measuring things.
"Mechanical properties of materials: Mechanics of Nanostructures" prepared by Rod Ruoff and presented by Hugh Bruck
They can strength-test individual carbon nanotube (buckytube) molecules by sticking them to an atomic force microscope probe. They found that the tubes do in fact exhibit "extraordinarily high" tensile strengths--up to 60 GPa (ordinary steel is about 1-2 GPa). They warn that if used as additives in plastic, the mix will reach a strength "only" about as strong as steel. But I'd point out that when it becomes possible to fasten buckytubes directly together, most of the extraordinary strength will be available.
They also described some very interesting theoretical work on fractures and defects. Normally, fractures are simulated as though the material was continuous, but of course at the nanoscale you have to think about discrete atoms. So Ruoff's group developed a "quantized fracture mechanics" (QFM) model. They compared it with detailed molecular mechanics simulations, and got answers that agreed within a fraction of a percent. QFM is much faster to run--"minutes" instead of "hours and hours" to do a calculation.
With the QFM theoretical tool, they were able to run a lot of simulations of carbon nanotubes with missing atoms. And they found something very interesting, and then reviewed experimental results and found the same thing. The strength of a tube with just a few defects is quantized--the tube can only have a few different levels of strength. It makes sense--a tube can't be missing half an atom, so it can't have a strength that's halfway between perfect and one-atom-missing.
To me, this is an intriguing and useful way to think about the mechanical properties of covalent solids. A given structure will have certain properties, and copies of it will have identical properties, and different structures will have properties that differ by a noticeable amount. The bad news is that if you need exactly a certain value for a property, you may not be able to get it. But the good news is that if you're trying to characterize a structure--to understand how it works and how to use it--you don't need to find as many numbers. It seems to make things more predictable, easier to understand and to simulate, and should make it easier to use nanodevices in multi-level designs.
Nanomechanical Devices (a partial overview) by James Hone
This talk focused on MEMS (Micro ElectroMechanical Systems), which are bigger than nanotech. When an audience member challenged him on that (the audience is quite interactive, sometimes more like a classroom than a conference), Hone provided a demonstration of how the term "nanotechnology" has been broadened: his response was, "There's always a little bit of flexibility when you say nano. So you choose your smallest dimension and you make that nano." But some of Hone's MEMS are being used to play with buckytubes--and that is definitely nanotech.
Hone started by showing a few slides of MEMS and a five-minute summary of how they're made, and mentioned Zettl's rotating electrostatic motor made out of a single multi-walled buckytube. Then he discussed how floating micron-scale beams can be used as very high-quality resonators; for you electrical engineers out there, they're getting Q's of 10,000 and sometimes up almost to 100,000. And some of the resonator designs are tunable. And as the beams become nanoscale (including buckytubes) they're already developing ways to detect the motion. By the way, this may result in less power-hungry cell phones and detectors that can weigh just a few atoms and thus detect and characterize single molecules--useful for detecting dangerous bio-substances.
Now the nanotubes... they can make a silicon slab with holes in it, deposit catalyst particles, and grow nanotubes suspended across the holes. Then they can deposit gold on top of the tubes--they have a very cool picture of a single nanotube (which, remember, is a single molecule) with stripes of gold on it. They can make these suspended tubes resonate--and they actually move far enough for the motion to be detected optically.
But the thing I found most exciting is that they can now characterize buckytubes and then place them where they want them. Buckytubes come in a variety of flavors, some semiconducting and some metallic--like the components of a microchip. But it's hard to sort them, and it's hard to do research on nano-electronic devices if (as has been the case) you have to drop them randomly on a surface and poke around till you find two that are crossed, and if you don't even know what kinds of tubes you're playing with. So these people grow nanotubes across a gap, then use Rayleigh spectroscopy to tell what kind of tube they have at each position. Then, they transfer the nanotube to an electronic device they want to build by simply turning the chip carrying the tube upside-down over the target chip and dropping liquid through the hole to flush it onto the target position. So they can build exactly the kind of electronic device they want to study. This isn't usful for building whole computer chips, but it's a big step forward for basic research to help design the chips.