So much for the diffraction limit.
Way back in the 20th century, it was thought that microscopes could never see anything smaller than half a wavelength of light. That's because an ordinary lens can't focus any better than that.
The first science essay we published, one year ago, described several ways that the diffraction limit could be beaten, or at least stretched. Several of them, in theory, could be used for imaging.
Now there's another to add to the list. It's very useful--and it's been demonstrated. Here's the article.
Plasmons can be created when light interacts with electrons in metal-coated glass. These plasmons are packets of energy derived from photons--but far more compact. So the plasmons can interact with a sample, then bounce back out again, convert themselves back to photons, and be picked up by an ordinary optical microscope. Features 150 nm apart have been imaged. That's about 1/3 the wavelength of the light they used. And they think they can get down to 10 nm.
The sample doesn't require special preparation--just the metallized surface to sit on, and a carefully placed drop of glycerin to help guide the plasmons. (They're already talking about how to eliminate the glycerin.) And the system doesn't need vacuum. That is a significant advantage over electron microscopy. Another advantage is the cost. One of the team members said, "This would move everybody into the world of electron microscopy for the cost of a regular microscope."
The imaging would be fast (unlike near-field scanning optical microscopy) because they can see all the plasmon-derived photons at once, rather than using a scanning technology. If they can use a normal microscope, they can pick up the photons with an ordinary CCD camera. The article didn't say how far off the surface the sample could be. My guess is that with this technology, a single layer closest to the metal could be imaged, with a thickness on the order of the pixel size. But I wouldn't rule out some variant of this being able to get actual 3D images, using other ways of delivering plasmons or photons to the sample. I also wonder whether they'll be able to get spectroscopic data, not just reflectance.
Once this is developed, it will make nanoscale imaging (and sample access/manipulation) vastly easier than it has been. This looks like a major enabling technology for R&D of nanoscale mechanical systems.