scanning tunneling microscope

Seeing things with light is relatively easy - it just requires good lenses. However, the resolution is ultimately limited by the wavelength of light; it is difficult to resolve subcellular structures. One solution is to use electrons instead, but this too fails at atomic level (and could only (until relatively recently) be used with dead tissue).

Enter the STM, a device very different from other microscopes in that it 'feels' rather than 'sees'. That is to say, a probe is run across the surface of a sample and the shape is recorded. This is achieved by quantum tunneling between the atoms of the probe tip and the atoms of the sample. The signal is converted into a map of the sample, producing an image of its surface. Similar in principle to an atomic force microscope it also has to be operated at unimaginably low (few kelvin) temperatures. However, it allows accurate visualisation of Ångstrom detail.

The scanning tunneling microscope - STM - was invented by Gerd Binnig and Heinrich Rohrer, who shared the Nobel Prize in physics in 1986 with Ernst Ruska, creator of the electron microscope.

The STM uses - as the previous writeup says - the quantum tunneling effect to make an image of the specimen. The STM probe is usually made out of tungsten, and is incredibly sharp; the tip is usually only made up of one or two metallic atoms. The specimen has to be metallic too for electric currents to run, and therefore it is common to cover it with a thin layer of gold.

The operation of the STM is controlled by piezoelectric elements that keep the tunneling current constant by holding the probe tip on a constant distance to the surface. The tunneling current is exponentially proportional to the distance between the needle and the metallic surface. As the probe sweeps over the surface of the sample, a topographical picture of it is created.

As mentioned in Scanning Tunneling Microscopy, the scanning tunneling microscope (STM) can have a wide variety of working conditions. The tip is typically tungsten, but platinum-iridium alloy is used if the surface (or adsorbates) are too reactive. Some STMs can operate at room temperature and atmospheric pressure, but under those conditions most surfaces will get dirty very quickly. (That is, atoms and molecules will stick to the surface and get in the way.) So, for most work, STMs are operated at ultra-high vaccuum (about 10^-10 torr) and cryogenic temperatures (less than 100 kelvins).

General sequence for operating an STM:

1. Seal STM assembly and chamber, containing sample. The chamber is usually made of thick, rolled stainless steel to prevent air from leaking through the chamber walls.
2. Pump out the chamber until about 10^-8 torr is reached. A variety of vacuum pumps are used, including the ion pump and titanium pump.
3. At 10^-8 torr, further pumping is futile, since there is usually a thin layer of water continually degassing from the walls of the chamber. So, the entire assembly is heated to about 150 Celsius, vaporizing the water and allowing it to be pumped out. After this, a vaccuum of about 10^-10 torr is achieved.
4. Clean the sample. We want a surface that is pretty much atomically flat and clean. This is usually done with sputter and anneal cycles. High-energy argon or neon atoms bombard the surface, removing the first few layers of atoms. Heating the sample causes the surface to smooth from thermal diffusion.
5. Measure surface with STM.

The surface can be probed in either constant-height or constant-current mode. In constant-height mode, the tip is held at one height and the current is measured as it passes over the surface. This is fast, but the tip could crash into an unexpected rise in the surface. In constant-current mode, the tip's height is controlled by a feedback signal so that the tunneling current is constant. This is slower, but you avoid tip-crashing. You can also do things like use the reactions between the tip atoms and surface (substrate or adsorbate) to move things on the surface. If there aren't enough things on the surface, you can drive a tungsten tip into the surface, digging up a lot of substrate atoms. It's quite popular to make things quantum corrals, logos, or quantum guns and to manipulate individual chemical reactons under situations where they normally would not occur.

Correction to Bigmouth_Strikes:

The sample does not have to be metallic, it simply has to be a reasonably good conductor, like a doped semiconductor.

Addition:

Scanning tunnelling spectroscopy is a technique where the bias voltage of the STM is altered while the position of the tip is held constant. At different potentials, the electron density of states (and therefore the tunnelling current) may be different due to the allowed energy states of the sample material. Given this information, electronic properties of the material can be deduced, and some idea of the sample's composition can be formulated.