During the formative years of quantum theory at the beginning of the 20th Century, people found out that electrons can behave quite like light - a phenomenon described as wave-particle duality1. When electrons are accelerated to high speed2 the wavelength of the electron wave becomes thousands of times smaller than the wavelength of visible light3. This was great news for people interested in improving microscopes, because the resolution of a normal optical microscope cannot be better than half the wavelength of the light used. Using electron-beams instead of light would seriously enhance the obtainable resolution.
Electron beams, being composed of fast-moving charged particles, can be bent by magnetic fields, and can be focussed in a similar manner to light waves (using magnetic lenses rather than the glass lenses used for light). This was also great news for microscopy people, because they could design electron microscopes using the same blue-prints they used to design common microscopes, only replacing the lenses by magnets and using a electron-beam as a 'light' source. The first electron microscope (a TEM; see below) was built in 1931 by German engineers Max Knott and Ernst Ruska4.
General Working Mechanism
A common electron microscope looks takes the form of a tube, between 50cm and 2m tall, with some measuring gadgets at the bottom (and possibly the middle as well). Older microscopes also have large, insanely complex consoles with a variety of knobs, buttons, gauges, displays and suchlike. Modern electron microscopes are generally controlled by computer, and thus look much less impressive. Inside the tube is vacuum, as an electron beam cannot penetrate more than a few mm of air. At one end of the tube (normally the top) is the electron gun, which is composed of a pointed cathode and one or more anode rings. (The cathode is connected to the negative terminal of the high-voltage power supply, and the anode(s) to the positive terminal.) The electron beam is generated at the cathode and accelerated by the anode(s). The voltage difference between the cathode and the final anode determines the speed of the electrons, and generally lies in the range 30 - 300 kV. The higher the voltage, the faster the electrons go and the more they penetrate into a sample (and the faster they can fry a sample as well).
In the section of the tube after the electron gun are the magnets which will steer and focus the beam onto the object under investigation. Electrons may pass through the object and be detected on the other side, or they can scatter off electrons from the surface which are detected from above the sample. In the first case the electron microscope is called a 'transmission electron microscope', or TEM. In the second case the electron microscope is called a 'scanning electron microscope', or SEM.
Scanning Electron Microscopes (SEM)
In the SEM, the electron beam is focussed down to a small spot and scanned over the surface of the object that is to be investigated. (This is done with magnetic coils placed at right angles to the beam. A similar arrangement is used to scan the electron beam in a cathode ray tube based TV set.) At the focal spot, the beam knocks out electrons from the object's surface. These scattered electrons are then detected by suitable detectors which are located inside the vacuum tube slightly above the sample. The signals from these detectors are displayed by a TV tube, or by a computer with modern instruments. The detectors can sometimes be moved to different angles, to generate different impressions of depth, which looks like shadows on the bumpy parts of the object.
As mentioned above the electron beam knocks electrons off the object's surface. This spot of the object then normally becomes ionised. If the object is made of a conducting material the charge can flow away, regenerating the surface to a neutral charge. Otherwise, if the object is made out of a non-conducting material then the charge remains at the same spot (it cannot flow away). This is a problem because the electron beam next to this spot will interact with the charge messing up the whole focussing and thus the whole image. For this reason only conducting materials or probes coated with metal films can be observed. Another problem is that the resolution is not as improved as expected for the wavelength of the electrons. This is because the interaction volume (the area where the beam knocks out the electrons from the sample) tends to spread out a bit in the bulk of the sample.
Transmission Electron Microscopes (TEM)
In the transmission electron microscope (TEM) most of the electrons are not scattered from the object but pierce through, provided the object is thin enough (generally this means less than 100 nm thick, for 100 - 200 kV microscopes). The interactions of the electrons with the object allow a great deal of information to be extracted form the sample. The physical micro-structure of the material can be imaged in great detail, and spectroscopic information allows the chemical make-up of the sample to be determined. The sample preparation is complicated, however - basically, you have to slice your sample to about one-thousandth the width of a human hair, preferably without damaging the structure of the sample too much - and modern TEMs are extremely expensive.
The most serious limitation for electron microscopy is the necessity of vacuum and metallic surfaces, meaning no living stuff can be inspected, although it is actually possible to examine live samples such as bacteria in highly specialised electron microscopes. These are normally very large, rare and incredibly expensive. Another big problem, which is often swept under the rug, is the sample preparation procedure. Cells, for instance, are gigantic bodies for a good electron microscope. For that reason the cells must be cut in slices (by a device called a microtome), to enable the full use of the electron microscope's high resolution. The most interesting stuff one will want to observe (eg, cell slices) is commonly made out of insulating material, and so it must be coated with conducting material first (usually by deposition from silver or gold vapour) - as you can imagine, this is an art form itself. Taking a micrograph is commonly the last and relatively most easy step; good images are only obtained from well-prepared samples. The sample preparation procedure can take days of hard work.