Electron Microscopes

We're not too far from the end of the Physics 202 class I'm helping teach, and as we finish things out we're learning about the particle nature of light and the wave nature of matter. It's really the very basics of quantum mechanics. One of the applications of this kind of knowledge is the electron microscope.

Light microscopes have a problem. As a rule, you can't resolve features smaller than the wavelength of the light you're using. Since this might be in the neighborhood of 600 nanometers for visible light, you have no real hope of seeing smaller things, or even of seeing objects a few times bigger than 600 nm very clearly. Bacteria can easily be this small, and viruses are smaller still. To see them clearly you'd need a smaller wavelength light. Ultraviolet light or x-rays might fit the bill, but they also tend to be very difficult to use in microscopes and their photons are too energetic to see many things without destroying them in the process.

Fortunately the wave nature of matter gives another option: use electrons instead. The de Broglie wavelength of a particle of matter is given by Planck's constant divided by the momentum of the particle. Doing a little bit of math, we find out that an electron accelerated by just 2 volts has a wavelength equal to about 600 nm. Cranking the voltage up to an easily achievable 15,000 volts means the electron wavelength is less than a tenth of a nanometer. Viruses and bacteria aren't so difficult to see clearly at this resolution:

i-faadfab1cfff05d4cedb4cf84baa2618-bacteria.png

While in theory the wavelength can be made as small as you like by cranking up the power, there's not much point. The most advanced electron microscopes can make out individual atoms as blurry fuzzballs (which they are), and there's simply no well-defined structure to "see" at lower scales. There's no shortage of things to learn at smaller scales, which is why we keep building bigger and bigger particle accelerators to probe those scales, but spatial structure isn't really one of those things.

More like this

Are there any microscope pictures (any kind of fancy high-tech microscope will do) that DO show small things... bigger, though? I mean, the bacteria on the image are really incredibly tiny, and here they're about 150 pixels long... but that's a *far* cry from seeing what they're made up of (i.e. looking inside cells, etc).

(Don't ask me how we should cut 'em up, though! I leave that to the people who might have the knowledge to answer my question(s) :)

@1:

The image shown is an SEM (scanning electron microscope) image, which produces its results by, well, scanning the electron beam over the surface of the sample.

What you are thinking about is a TEM (transmission electron microscope) image, which does show the insides of things like cells. To get to the insides, you have to mount the beasties in an epoxy, then shave a thin slice for the electron beam to pass through.

Of course, Matt is leaving out the truly fun stuff: The electrons producing characteristic X-rays from the material under the beam. This allows for qualitative and quantitative chemical analysis of very small areas. Hence, the electron microprobe.

@serenity

We published images from a field emission-scanning electron microscope which showed several individual pili on their surface. Each pilus measured ~5 nm wide. For the curious we ran the scope at 1 kev. (Mol. Micro. Varga et al. 2006)

You can also look at thin sections using TEM and see larger structures inside a bacterial cell. In a paper from my old lab (J. Bacteriol. Harry et al. 2009) there are TEMs which show condensed chromosomes, crystals of enterotoxin, masses of some carbohydrate, and all sorts of division septa inside the bacterial cells.

I don't have any links handy, but in our images of bacteria inside of macrophages you could make out host ribosomes (they are little black dots, nothing exciting :P), vesicle membranes and the like.

scanning tunneling microscopes (STM) can achieve atomic resolution. for instance, here:

http://groups.physics.umn.edu/stmlab/gallery/gallery_fr.html

fig 1 shows a nice view of the reconstructed silicon surface. the surface is made of rows of atoms. each row is 2 Si atoms wide. dark spots are missing atoms or pairs of atoms (dimers).

fig 2 is a close up of the Si(001) surface. you can make out individual atoms in the pairs that make up the rows. some rows are buckled, which makes it even easier to see the individual atoms. (like the row 2nd from the bottom)

you can't resolve features smaller than the wavelength of the light you're using.

That should be "0.6 wavelength." That is not true, either. Ash and Nichols in 1972 reduced to practice with lambda/60 resolution re near-field scanning optical microscopy.

Physics showed IR spectroscopy cannot possibly work for its massive line broadenings, difference-sum lines, band splitting by Fermi resonance... Chemists listened, turned their backs, and did it anyway.

I recently watched a fellow blow apart gallium nano-wires just by 'zooming in' on them with his SEM! Apparently, it's possibly to make your sample really, really hot if you're reckless.

Ah, I meant "possible" there. Preview Post is my new best friend.

[delurks]

"...particle nature of light and the wave nature of matter."

Oooh! Dead cool!

Too bad I didn't get exposed to this in my undergrad days. I may have changed my major (and faculty) over just this point.
Matt, could we have some more posts on this topic, please?

Apparently, it's possibly to make your sample really, really hot if you're reckless.

More insidiously, you can alter the elemental distribution of what you are looking at. For some materials, the beam heating can actually cause some atoms to diffuse away from the focal spot. If you are attempting an analysis, you can literally watch the count rate for those elements tail off as they move away.

It can be depressing, as to get usable data you have to alter your analytical conditions, then restandardize, then find a new spot to zap.

While in theory the wavelength can be made as small as you like by cranking up the power, there's not much point. The most advanced electron microscopes can make out individual atoms as blurry fuzzballs (which they are), and there's simply no well-defined structure to "see" at lower scales. There's no shortage of things to learn at smaller scales, which is why we keep building bigger and bigger particle accelerators to probe those scales, but spatial structure isn't really one of those things.

This is really REALLY incorrect.

There is well defined structure to see at lower scales, whether it is the detailed radial distribution of protons and neutrons in a nucleus, the football-like shape of deformed nuclei, or the actual radial distribution of quarks in the proton and neutron.

Visible substructure in the charge distribution of the neutron was the first evidence that there had to be constituents (first called "partons") making up "elementary" particles such as the neutron and proton. Heck, the mere fact that the neutron has a charge distribution tells you there is something like a quark inside. Further, the magnetic distribution in protons and neutrons is how we know that quarks have a magnetic moment.

You can read some of what you need to know in the Nobel bio and background information for the award that was given to Robert Hofstadter for his work in determining the spatial structure of nuclei and nucleons with high energy electrons. He use SLAC as a giant electron microscope.

Those looking for constituents of quarks and leptons are continuing that work, but it is hard to match the clarity of experiments done with electrons.

UMM.. this is all VERY confusing... I don't really understand any of it!! Of course I am still in middle school. but i am looking for a website where I could find out the effects of these electron microscopes on modern day medicine and healthcare. It would be really great if someone could help me!!

this is 2 advanced