Although the double-slit experiment is now often referred to in the context of quantum mechanics, it was originally performed by the English scientist Thomas Young some time around 1805 in an attempt to resolve the question of whether light was composed of particles (the "corpuscular" theory); or rather consisted of waves travelling through some aether, just as sound waves travel in air.
The interference patterns observed in the experiment seemed to discredit the corpuscular theory, and the wave theory of light remained well accepted until the early 20th century, when evidence began to accumlate which seemed instead to confirm the particle theory of light.
The double-slit experiment then became a classic "gedanken" or thought experiment for its clarity in expressing the central puzzles of quantum mechanics; although in this form the experiment was not actually performed until 1961 (using electons), and not until 1989 in the form of "one electron at a time".
In September 2002, the double-slit experiment was voted "the most beautiful experiment" by readers of Physics World.
Explanation of experiment
In Young's original experiment, sunlight passes first through a single slit, and then through two thin vertical slits in otherwise solid barriers, and is then viewed on a rear screen.
When either slit is covered, a single peak is observed on the screen from the light passing through other slit.
But when both slits are open, instead of the sum of these two singular peaks that would be expected if light were made of particles, a pattern of light and dark fringes is observed.
This pattern of fringes was best explained as the interference of the light waves as they recombined after passing through the slits, much as waves in water recombine to create peaks and swells. In the brighter spots, there is "constructive interference", where two "peaks" in the light wave coincide as they reach the screen. In the darker spots, "destructive interference" occurs where a peak and a trough occur together.
The Thought Experiment
If sunlight is replaced with a light source that is capable of producing just one photon at a time, and the screen is sensitive enough to detect a single photon, Young's experiment can, in theory, be performed one photon at a time - with identical results.
If either slit is covered, the individual photons hitting the screen, over time, create a pattern with a single peak - much as if gunshot were being poorly aimed at a target.
But if boths slits are left open, the pattern of photons hitting the screen, over time, again becomes a series of light and dark fringes.
This result seems to both confirm and contradict the wave theory. On the one hand, the interference pattern confirms that light still behaves much like a wave, even though we send it one particle at a time.
On the other hand, each time a photon with a certain energy is emitted, the screen detects a photon with the same energy. Since the photons are emitted one at a time, the photons are not interfering with each other; so what is the "interference" with?
Modern quantum theory resolves these questions by postulating probability waves which describe the likelihood of finding the particle at a given location; these waves interfere with each other just like ordinary waves do.
A refinement of this experiment consists in putting a detector at each of the two slits, to determine which slit the photon passes through on its way to the screen. But when the experiment is arranged in this way, the fringes disappear - for reasons related to the collapse of the wave function.
Conditions for interference
The waves interfering must be coherent, ie the light has the same frequency and is in the same phase. In Young's experiment, this was achieved by passing the light through the first slit, and thereby diffracting it, producing a coherent wave; this more typically achieved now by using a laser and removing the first slit.
The source must be polarised or have significantly resolved parts in the same plane.
The slits must be close (about 1000 times the wavelength of the source), otherwise the interference pattern would be too close to be seen.
The width of the slits is usually slightly smaller than the wavelength (λ) of the light; this allows the slits to be treated as point-sources of spherical waves, and reduced the effects of single slit diffraction on the results.
The bright bands observed on the screen happen when the light has interfered constructively - where a crest of a wave meets a crest. The dark regions show destructive interference - a crest meets a trough.
A formula linking the slit separation s, wavelength of light λ, distance from the slits to the screen D, and fringe width (the distance between the centres of the observed bands of light - x) exists:
It is possible to work out the wavelength of light using this equation and the above apparatus. If s and D are known and x is observed then λ can be easily calculated.