Light Smaller Than Light: Rayleigh's Diffraction Barrier Broken

Regular posting resumes.

People who work with really really small things--in our own humble example, neurons--are generally acquainted with Rayleigh's diffraction barrier. It goes by a number of names: Rayleigh's law of resolution maxima, the Rayleigh criterion, and so forth. But the gist of is that there is a hard limit to the maximum resolving power one can achieve via optical microscopy, partially due to the properties of visible light.

If you haven't encountered this idea before, click on the link above, as wiki can explain it better than I can. But the basic version: for our purposes, I think we can get away with describing the resolving power of a microscope as one's basic ability to distinguish between objects in a field being viewed at a certain magnification. This is dependent on a function whose variables include the wavelength of the light involved, the refractive index of the medium being viewed through, and the sine of the angular aperture. Examining our three variables: the sine function maxes out at 1 (although for practical reasons, the limit is generally 0.94). For reasons I cannot claim any knowledge of, the best refractive index achievable seems to be 1.56, with an oil immersion. And the lowest wavelength of visible light available is thus 187 nm. Therefore, the lowest wavelength that can be visualized into visible light is 187 nm. Thus, the minimum distance that can be between any two objects we would like to visualize is about 200 nm.

Now, I know what you're thinking: WOW, that's REALLY REALLY REALLY SMALL. And you're right.

But: let's compare it to a few numbers:
Size of an average neuron's soma (the cell body): 10-25 micrometers. That's around 50-125 times the size of our barrier. Not bad.
Size of an average dendrite (the processes that receive information from other cells) or an axon (the processes that send information to other cells): ~1 micrometer. Now, that's still five times the resolution barrier.

But at the same time, realize that this limit is the minimum space objects need to be separated by to be distinguishable from each other. This means that you're pretty much screwed when it comes to trying to visualize things smaller than the level of an individual dendrite or axon--which is where all the interesting stuff is happening. To be sure, there are a number of way to get around this. You can tag specific proteins with fluorescent dyes, so that only they will appear on the scope. But if it's a highly localized protein, then you're still in trouble. You can also visualize via electron microscopy, but the preparation required prevents you from working with living biological systems.

But this morning, thanks to Alex Palazzo, I discovered that a new technique--called stimulated emission depletion (STED) microscopy--has been developed to work around this limitation. Apparently Stefan Hell and his colleagues take his name literally*, as this technique can be used on fluorescently labelled preparations to achieve a resolution of approximately 50 nm. And, in theory there is no fixed limit to the resolution achievable by this technique. Last week, they published their first paper from practical use of this technique, and Nature has published a brief overview of the technique.

So how'd they do it?

I'm glad you asked.

From Nature's overview, by Garth Simpson:

The first beam excites molecules just as in traditional fluorescence microscopy — as the molecule absorbs the energy from the light, it is promoted up to a higher energetic state, and as it relaxes back to the ground state, it releases the energy in the form of light. The second beam, at a different wavelength, suppresses this fluorescence by 'stimulated emission', in which molecules are actively pumped down out of the excited state by light (in essence, absorption driven backwards). In sufficiently intense optical fields, stimulated emission becomes more efficient than fluorescence and is the dominant effect, drastically reducing the fluorescence.

In STED microscopy, the normal fluorescence is suppressed by a depletion beam that is shaped like a doughnut and contains a node of zero intensity — a hole — at the centre of the focus. So, any fluorescent light that is detected arises just from this hole of about 50 nm across where the depletion beam is absent (Fig. 1). Scanning this narrowed viewing point across the sample allows a higher-resolution image to be built up, much as a computer screen made up of smaller pixels will provide a sharper image...

But a picture is worth a thousand words, so here you go:



In the upper left corner, you can see a blue light produced by excitation of a fluorophore. In the upper right corner, you can see the orange STED beam. This is basically blocking the flourescence of the excited molecules in its range by knocking them down a peg (in terms of energy states) before they can release any light. Then in the lower left corner, you see a green light representing what's left after the two lights interact. And the graph in the bottom corner represents the flourescence profile after their interaction.

So what have they used this technique on to get me so enthused? Well, if you really really care you've probably already clicked on the link and found out. If not, you can either click on it now (if you're the impatient sort), or wait until later tonight for the exciting conclusion of our two-part post. Because, um, right now I've gotta get running to class.
--
*In German, "Hell" means "brightly."

References:
Simpson GJ (2006) The diffraction barrier broken. Nature 440:879-880.
Willig KI, Rizzoli SO, Westphal V, Jahn R, Hell SW (2006) STED microscopy reveals that synaptotagmin remains clustered after synaptic vesicle exocytosis. Nature 440:935-939.

EDITED at 7:30 PM because I accidentally linked to the U-of-M-access Nature pages.

EDITED on Wednesday, 26 April 2006 after a fundamental physics error was called to my attention.