What's the deal with Fluorescence Microscopy?

d.w.rowlands [at] gmail.com

In 2014, the Nobel Prize in Chemistry was won by a physicist and two physical chemists who were involved in the development of fluorescence microscopy techniques that allow optical imaging on a scale smaller than the "diffraction limit". While I was glad to see that the prize was for physical chemistry---albeit physical chemistry that's mostly of interest to biologists---instead of the all-too-frequent chemistry prizes that are really a second prize in biology, I was a bit embarrassed to admit that I didn't really understand how fluorescence spectroscopy works. I'd heard it mentioned in freshman biology at Caltech, but the professor didn't go into any detail about how it gets around the diffraction limit and the TAs didn't seem to know. And, since then, I hadn't really thought about it at all, since it's a technique that's really only of use in biology labs. However, the Nobel announcement made me decide that I needed to try to actually get a bit of an understanding. And, since I've been reading about it, I thought I'd write an essay to try to give a simple lay explanation of it.

What is the diffraction limit?

Light consists of waves, with the color of the light solely a function of the wavelength. Visible light has wavelengths between 390 nm (nanometers) and 700 nm and ultra-violet light has even shorter wavelengths. Because light consists of waves, it experiences diffraction: the patterns of interference you see when two waves with different sources (for example, the ripples from two stones dropped into a pond) intersect. As in that example, the size of the diffraction patterns produced by two intersecting light waves will be of a scale related to the wavelength of the light.

The reason this matters is that, to produce an image, you need to focus light on a spot, causing a large number of light waves from different directions to all intersect in one location. When you do this, you produce diffraction, and the scale of this diffraction effect controls the minimum size of the spot you can focus the light.

It turns out that the size of this spot is the wavelength of the light divided by 2nsin(θ). θ here is the angular size of the lens that is focusing the light and n is the "index of refraction" of the medium you're trying to focus the light in, essentially how much slower light travels in it than in a vacuum. The index of refraction matters because what is actually fixed for light of a given color is its frequency: the wavelength is the speed divided by the frequency. Modern optical systems can reduce the denominator here to 2.8 or so, so in practice the diffraction limit is that you can't focus light (or image structures) at a scale smaller than the wavelength of the light you're using divided by 2.8.

The obvious solution here is to use higher-energy, and thus shorter-wavelength, light, and you can do this to some degree. However, many interesting biological systems don't show as much contrast with ultraviolet light, and many of them can be damaged by the higher-energy photons. (This shouldn't be surprising, given that ultraviolet light is what causes sunburn and sometimes even skin cancer!)

One other way to get around this problem is to use a very small light source very close to the thing you're trying to image: if the light source is smaller than the wavelength of the light, and closer to the sample than it, you can get around diffraction effects by essentially forcing the light to behave as particles. This can be done with a scanning tip and a very small fiber optic light source. However, you need to be able to get within nanometers of your sample, and it will take a substantial time to scan your whole target. Furthermore, the amount of light you can use is very limited.

What is fluorescence?

To change the subject for a bit, we also need to discuss fluorescence. Normally, when we shine light on something and see the light return to our eyes, the light is reflecting off the surface. While the quantum mechanics of this is complicated, the essence is that light waves hit the surface and immediately bounce back into space. However, light that its a material doesn't always bounce back immediately. Light with an energy equal to the energy difference between two states of the electrons in an atom can be absorbed, elevating an electron from a lower to a higher energy state. However, this higher state is generally not permanently stable, and the electron will eventually drop to a lower-energy state, emitting light in the process. The emitted light is "fluorescence".

How does this help?

This is the hard part, and I admit I don't completely understand all of the details. But there are several techniques that you can use to image things smaller than the usual diffraction limit by taking advantage of fluorescence.

One technique, called 4Pi microscopy, uses fluorescence to maximize the angular size of the light source. While reflection happens over a half-sphere in the direction of the light source, once a fluorescent molecule is excited, it can emit light in every direction. This means that you can use two in-phase lasers on opposite sides of the sample to illuminate the sample, making a very large effective angular size for the light source. This reduces the volume of the maximum resolution allowed by the diffusion limit by a factor of roughly five to seven.

Another technique, STED microscopy, uses two laser pulses at different frequencies in quick succession. The second pulse knocks excited electrons into an intermediate energy state, so that when they return to the low-energy ground state, they'll release light with a different energy (and so color). Tricks with the phase of the second pulse (which I don't understand) cause it to be shaped like a donut with a hole in the middle that is significantly smaller than the diffraction limit. Then, if you observe the sample using light at the original fluorescence wavelength, you'll only see light from this smaller region that was struck by the first but not the second laser.

One limit on these techniques, however, is that you can only see fluorescent molecules with them. Biologists tend to solve this problem by tagging the genomes of whatever cell part they want to see with the genes for naturally-occurring fluorescent proteins, particularly "green fluorescent protein". The 2008 Nobel Prize in Chemistry, in fact, was given to the discoverers of green fluorescent protein.