Take a picture, if you can.

July 24, 2006 5:58 PM

Does an actual photos of small things like atoms, molecules or DNA?

Google didn't show anything for photo, just lots of models. I am looking for real bona fide made from reality image captures. Is it even possible to do photograph these things? If not, are we moving to a point where we weill be able to take such photos or even movies?

posted by anonpeon to science & nature (15 comments total)

It is impossible to take a photo (in the sense of a recording of visible light) of these things because they are smaller than the wavelength of visible light.

The closest that can be done is to do things like bomarding the object with electrons and measuring how they bounce off and so forth. The images produced by an electron microscope are not all that interesting to someone not trained to read them. Hence, the data from such source is combined and processed into the models you have been finding (which themselves are at best an aproximation of "reality").
posted by Riemann at 6:05 PM on July 24, 2006 [1 favorite has favorites]

High resolution image of a crystal.
What you may want to look into is a Scanning Tunnelling Microscope.

More pics here.
The IBM logo in atoms.

A photo of some DNA.
posted by Jimbob at 6:05 PM on July 24, 2006

There are two books here and here that seem like they contain the sort of photos you want. Atoms are specifically mentioned in the write-up for Heaven and Earth. Maybe your local library has copies of them?
posted by Joh at 6:30 PM on July 24, 2006

Scanning Tunneling Microscopes are not actually optical. They use an electrical probe to create an image at an atomic scale. Scanning Electron Microscopes and Transmission Electron Microscopes use a beam of electrons, and produce an image more like what you could see, if you could see at that scale. (example)
posted by ijoshua at 6:30 PM on July 24, 2006

Riemann, that's not really correct... perhaps you're thinking of X-ray crystallography, which produces a pattern of dots like these. Electron microscopy, as in the pictures linked above, produces a rather viewable picture. The snake-like photos of DNA are a classic example.
posted by rxrfrx at 6:49 PM on July 24, 2006

For X-ray crystallography, the diffraction pattern (of the sort rxrfrx linked) can be transformed into a map of where atoms in a molecule are in relation to one another, resulting in pictures like this one of hemoglobin. The diffraction pattern and the atomic map (electron density map, technically) are related by a Fourier transform. So crystallography can locate atoms very precisely, but it doesn't give a direct image in the way that electron microscopy or scanning tunneling microscopy does.

X-ray crystallography is being extended to time-resolved applications, with the idea of creating a molecular movie. This is cutting edge stuff and not many researchers have the know-how and/or access to facilities to do it. As one example, here's a movie of carbon monoxide being released from myoglobin [mpg, 784 kB]. Article describing the movie.
posted by Joe Invisible at 7:22 PM on July 24, 2006

Visible light waves are in the range of 380 to 780 nanometers (10^-9). The diameter of a carbon atom is 70 picometers (10^-12). You can't use visible light to look at an atom and thus cannot take a photograph in the traditional sense.

Instead, we have to use indirect mechanisms (like x-ray crystallography or electron microscopy) whose results we then convert to images like those linked above.
posted by event at 7:28 PM on July 24, 2006

Even when you're using electrons in any of several ways to try to image molecules, the result won't be anything like a photo. It cannot be, because atoms aren't actually little spheres. The outer shell of an atom is defined by the quantum wave function of the electrons in that outer shell, and what any kind of electron microscope detects is the boundaries of the outer electron orbitals.
posted by Steven C. Den Beste at 7:33 PM on July 24, 2006

It cannot be, because atoms aren't actually little spheres.

How do you know? You can't see 'em. That's not really snark. It's something important to remember with quantum mechanics.

Throw your intuition out the window. All you have is the math.
posted by teece at 8:30 PM on July 24, 2006

Seconding what event said above. And electron microscopic images are probably the closest to what you're looking for. Like this or this.
These are pictures of DNA and proteins - BIG molecules. Even here you can only see the fuzzy outlines, not the individual atoms in the chains, so you can imagine that you can't get much smaller.

Another example: this is how close you can see into a cell's nucelus using electron microscopy, and this is pretty much the range using light microscopy (scroll down to the picture of the cell - that's pretty close up for light microscopy!)
posted by easternblot at 8:59 PM on July 24, 2006

With a good Scanning Electron Microscope (SEM), you can see large molecules, such as DNA. To distinguish individual atoms, you can use either Scanning Tunneling Microscope (STM) or Atomic Force Microscope (AFM).

An STM works by moving a tiny metallic tip across the surface you want to study, while feeding a current from the tip to the surface. You then measure the deflection of the tip required to maintain this constant current. It depends on the topology of the sample as well as the electron distribution.

The AFM comes in many varieties but generally consists of a fine tip at the edge of a cantilever, which will be deflected by atom-to-atom forces between the tip and the sample. This deflection too depends on topology and charge distribution, and is typically measured by bouncing a laser off the back of the cantilever and seeing where the spot ends up.

How do you know? You can't see 'em. That's not really snark. It's something important to remember with quantum mechanics. Throw your intuition out the window. All you have is the math.

With these techniques, you can actually see the electron distributions (orbitals) of individual atoms. Depends of course on what you mean by "see", but I don't think an STM is different in this regard from eg a radio telescope. They're all tools enabling us to detect what is going on when direct eyesight is insufficient.
posted by springload at 11:59 PM on July 24, 2006

The 3D looking pictures that you will see of this magnitude are likely created using a Scanning Electron Microscope. These require the surface to be metallic so often the object is first coated in metal.
Such as this bug eye. So you aren't seeing a photo of the actual thing.
posted by metaname at 12:14 AM on July 25, 2006

metaname:

SEM does not require a metallic sample, and neither does AFM. STM does, however, need a reasonably well conducting sample, since there is a current flowing between the probe and the measurement spot.
posted by springload at 12:46 AM on July 25, 2006

springload: "
SEM does not require a metallic sample, and neither does AFM. STM does, however, need a reasonably well conducting sample, since there is a current flowing between the probe and the measurement spot."

But SEM does also require a reasonably conducting sample, to avoid a buildup of charge, which distorts the image. So usually, the samples are coated in a thin gold or carbon film to allow the charge to dissipate.

I would have thought, having done both SEM and STM, that SEM requires a more conducting sample than STM. After all, the currents in STM are typically picoamps, so anything other than a really good insulator will allow the current to flow.
posted by Boobus Tuber at 4:36 PM on July 25, 2006

Boobus Tuber:

Charge buildup can be a problem, but the current in an SEM is also typically in the picoamps, and you can get a picture before the distortion gets too bad, as the electron beam strikes a comparativley large surface. I often do SEM on oxidized silicon and polymers (photoresist and e-beam resist) and it works. The biggest problems in my experience are heating of the sample and accumulation of soot. While the current in an STM is low, the current density is high, with all electrons preferraby tunneling through the atom closest to the tip. That is not really possible with an insulating sample.
posted by springload at 12:45 PM on July 26, 2006