State-of-the-art microscope is a hybrid of Michael Phelps and an arcade game.
Most of the world of biology, and, indeed, a significant part of physics, is focused on trying to generate a clear image of really tiny stuff. For many reasons, it simply isn't possible to get better images. Still we keep trying, and a recent success reminded me that sometimes all you can do is stare in awe at utter genius.

I present to you a microscope that is a hybrid of Michael Phelps and an arcade game. It's a swimming lens that you steer over a surface, bringing back fascinating views of the terrain below. It goes one better than Phelps, in that it gets its energy locally, so there's no need to drag it out of the pool for food and recreational drugs. And, finally, it's really, really simple. So simple that if you have a decent microscope already, you can probably use one of these swimming lenses with a few minor modifications.

Enhance my image

As always, to reveal the full beauty of the idea, we need to see why an ordinary version of the technology fails. The basic ideas behind a microscope's failure go back to two things: conservation of energy and conservation of momentum. Light consists of waves that have a certain energy and momentum, depending on the frequency and wavelength of the light. When light scatters off a surface, the frequency and wavelength don't change, which means that the total momentum and energy of the light is unchanged.

As a consequence, all the information about the surface is carried in the angles at which the light scatters—how the individual directional components of momentum are changed.

The angular distribution of photons is determined by the surface features. Any surface can be described by a set of waves, where the long wavelength waves describe the big features and the short wavelength waves describe tiny ones. Each of these waves has its own momentum. So, when the light wave scatters from the surface, the surface waves give the light additional momentum along the direction of the surface. To conserve momentum, this is removed from the light's momentum vertical to the surface.

So, small features impart a large sideways momentum and scatter at angles that are close to parallel to the surface. Large features scatter at angles close to vertical. Since a lens has a limited size, it cannot capture light that is scattering at angles that are close to parallel to the surface. That means the finer features of the surface cannot be seen.

However, even if the lens could be wrapped around the surface and capture all the light, it still would not resolve all features. This still comes down to conservation of momentum. Imagine that there's a small feature; to conserve momentum, the light must scatter at an angle that is close to parallel to the surface. For very, very small features, however, even scattering parallel to the surface cannot conserve momentum.

These very fine features do scatter light, but not in waves that travel like ordinary light. Instead, these waves decay away over a very short distance (just a wavelength or two). But while they live, these waves hold information about these fine spatial features. If they can be captured, the finer details of the image can be resolved.

The process of capturing the dying waves is called near-field imaging and is very difficult. The lens has to be very close to the surface. Indeed, often you have to use a narrow probe that kind of scans along the surface, which means the image you capture at any time is very, very tiny. These difficulties have basically made near-field imaging a speciality for a few labs and inaccessible to most of us.

Cheap and cheerful

This is what makes this latest research so exciting. It uses a form of rough near-field imaging that provides startlingly good results with minimal modifications to preexisting microscopes.

The researchers from the University of California-San Diego created tiny spherical lenses from polystyrene. These lenses are floated on top of the sample to be imaged. Now, they are so close—practically touching the sample—that they can capture some of these non-propagating waves. The lens then produces a magnified, high-resolution image in the air just slightly above it. That image can then be captured by an ordinary microscope.

In this setup, you end up with a large area in the microscope field of view, with a small part that is highly detailed. This by itself is pretty useless, because you can't control the placement of the tiny lens, so you have no way of scanning the surface you're imaging.

To get around this, the researchers put a small patch of titanium, nickel, and platinum on the side of the small lens; they then dumped some hydrogen peroxide into the solution it was floating in. The titanium is just there to make sure the rest of the metals stick to the lens. The platinum supplies propellant by decomposing hydrogen peroxide into water and oxygen. When the reaction products fly off the surface, they kick the lens along in the opposite direction, moving the lens across the sample. The speed of motion is simply controlled by changing the concentration of hydrogen peroxide. That is just genius.

It gets better though—the nickel is there for a reason, too. Nickel is a magnetic material, so it will try to align to a magnetic field. So the lens can be rotated by rotating a magnet. Combined with the little hydrogen peroxide rocket, you have a lens that can be steered across a surface, imaging as it goes. The technique has the feel of those old-fashioned arcade games, where you steer your ship along a surface, but you can't really see what is coming, so you have to react very quickly to stay on the path you want to track.

The speed is also very respectable. The researchers calculate that with a single lens, they already outperform most near-field and high-resolution imaging techniques. And, even better, you don't have to stick with one lens. Multiple lenses in parallel can be used, making it even faster.

Too good to be true?

I'd imagine that the platinum catalyst gets clogged up with contaminants (probably sulphur and carbon) pretty quickly, so you probably have to replace the lenses fairly regularly. There is also a tradeoff to be made: imaging speed versus how much hydrogen peroxide you add; the chemical can potentially react with anything you're imaging. The researchers showed that they can image single frames very fast, but the hydrogen peroxide concentration was up to 7.5 percent by then. I suspect that is already high enough to influence live cell cultures, as the disinfectant solutions you can find in a pharmacy are only three percent.

So, as with all new technologies, it can't be used for everything. But I expect it will certainly be very useful.