Electron N-Slit Interference

Single electrons build up fringes one by one; slit number, width, spacing, voltage and a which-path detector reshape the pattern.

Switch to 3D

Presets

Playback

Electron source

The accelerating voltage sets the wavelength.

Slit plate

Number, width and spacing of the slits.

Detector

Distance to the screen and the which-path detector.

Layers

Model checks

The science

What is shown

Electrons accelerated through a voltage V have a de Broglie wavelength lambda = h / sqrt(2 m e V (1 + eV / 2mc^2)) (about 5.4 pm at 50 kV). They pass a plate with N slits of width a and spacing d and are detected one at a time on a screen a distance L away. Each electron lands at a random position; the distribution of many landings follows the intensity |psi|^2.

Equations (Fraunhofer limit)

With beta = pi a sin(theta) / lambda and gamma = pi d sin(theta) / lambda the coherent pattern is I = sinc^2(beta) (sin N gamma / sin gamma)^2 / N^2: a single-slit envelope times the N-slit grating function. Principal maxima sit at d sin(theta) = m lambda, i.e. a fringe spacing lambda L / d on the screen; the envelope first vanishes at lambda L / a. If d / a is an integer r, every r-th principal maximum is missing. Between principal maxima there are N - 2 weaker secondary maxima.

Which-path information

If a detector can tell which slit an electron used, the fringes fade. For a pure detector state the distinguishability D and visibility V obey V^2 + D^2 = 1 (Greenberger-Yasin, Englert). The page uses mu = sqrt(1 - D^2) as the weight of the cross terms: I = sinc^2(beta) [N + mu (S - N)] / N^2. At D = 1 the pattern is the incoherent sum of the single-slit patterns. For N > 2 this uniform scaling of all cross terms is a model choice, not a derived result.

Checks on the page

1) An explicit sum of 200 Huygens point sources per slit reproduces the closed form, and the same sum with exact path lengths shows how far the Fraunhofer formula is from the exact result (set by the Fresnel number). 2) The integral of the pattern over all angles is 1/N for any coherence (conservation of probability), N - 2 secondary maxima are found numerically, and the principal maximum sits at gamma = pi. 3) 80,000 electrons sampled from the pattern give chi^2/ndf near 1 against the expected counts.

Simplifications (honest)

Scalar waves, one dimension, perfectly monochromatic and coherent beam, infinitely thin opaque plate, no Coulomb or magnetic phase, no thick-slit or vector effects. Real electron interferometers (Jonsson 1961, Tonomura 1989, Bach 2013) use much larger d / lambda and the screen distances here are idealised. The 2D wave field is drawn with the angular scale magnified so that fringes are visible; only the histogram and the numbers use the true scale.

What to try
  1. Press Single slit: only the envelope remains, with its first zero at lambda L / a.
  2. Switch to the five-slit preset: principal maxima keep their positions but get narrower, with three weak secondary maxima between them.
  3. Raise D to 1: the fringes vanish while the total number of electrons stays the same (energy sum rule).
  4. Choose d = 3a and look for the missing third order inside the central lobe.
  5. Lower the voltage to 5 kV: slower electrons have a longer wavelength and the whole pattern widens.
Predicted intensity and detected electrons on the screen
Fringe visibility against which-path distinguishability
Electron wavelength against accelerating voltage
Pattern for 1 to 5 slits with the same width and spacing