The Most Important Experiment in Physics
Richard Feynman called the double-slit experiment the phenomenon that “has in it the heart of quantum mechanics” and that is “impossible, absolutely impossible, to explain in any classical way.” It demonstrates, in the simplest possible setup, that the rules governing the microscopic world are fundamentally different from anything in everyday experience.
Historical Context
In 1801, Thomas Young passed sunlight through two narrow slits and observed an interference pattern on a screen behind them: alternating bright and dark bands. This settled the centuries-long debate between Newton’s corpuscular theory of light and Huygens’ wave theory. Waves from the two slits interfere constructively (bright bands) where they arrive in phase and destructively (dark bands) where they arrive out of phase.
The experiment seemed definitive. Light is a wave. The story should have ended there.
The Quantum Version
In the twentieth century, the experiment was repeated with particles: electrons, neutrons, atoms, and even large molecules (C60 buckyballs). The results were identical. An interference pattern appeared on the detector screen.
Setup
A source emits particles one at a time toward a barrier with two slits. On the far side, a detector screen records where each particle lands. The source intensity is low enough that only one particle is in flight at any given time. There is no possibility of particles interacting with each other.
Results
Each particle lands at a single point on the screen, producing a single dot. There is no smearing, no wave-like spread for any individual detection event. But as thousands of particles accumulate, the dots form an interference pattern: bright bands where many particles land and dark bands where almost none land.
The pattern is identical to what would be produced by a wave passing through both slits simultaneously. Yet each particle is detected as a single, localized event. The particle appears to “know” about both slits, even though it is a single, indivisible entity.
Which-Path Detection
The mystery deepens. Place a detector at one of the slits to determine which path each particle takes. The interference pattern vanishes. The distribution on the screen becomes the simple sum of two single-slit patterns, exactly as classical particles would produce.
The act of acquiring which-path information destroys the interference. It does not matter how gently the observation is performed. Any interaction that could in principle reveal which slit the particle passed through eliminates the quantum interference.
Delayed-Choice Experiments
John Archibald Wheeler proposed an even more disturbing variant. What if the decision to observe which-path information is made after the particle has already passed through the slits?
Experiments confirm the prediction: the result depends on the measurement choice, even when that choice is made after the particle has traversed the barrier. If which-path information is available at the time of detection, no interference. If it is erased before detection, interference returns.
This rules out any simple story in which the particle “decides” to be a wave or a particle at the slits. The outcome depends on the entire experimental configuration, including elements that are spatially and temporally separated from the slits.
Mathematical Description
The quantum formalism accounts for all of these results with a single rule: add probability amplitudes, not probabilities.
Let $A_1$ be the amplitude for the particle to arrive at a point on the screen via slit 1, and $A_2$ the amplitude via slit 2. The probability of detection at that point is:
$$P = |A_1 + A_2|^2 = |A_1|^2 + |A_2|^2 + 2,\text{Re}(A_1^* A_2)$$
The last term, $2,\text{Re}(A_1^* A_2)$, is the interference term. It can be positive (constructive interference, bright bands) or negative (destructive interference, dark bands).
In the classical case, probabilities add directly:
$$P_{\text{classical}} = |A_1|^2 + |A_2|^2$$
No cross term. No interference. The entire difference between quantum and classical physics, in this experiment, is the cross term.
When which-path information is obtained, the two amplitudes acquire distinct, uncontrollable phase relationships (decoherence), and the cross term averages to zero. The quantum result reduces to the classical one.
Interpretations
Copenhagen Interpretation
The particle has no definite path. The wave function passes through both slits and interferes with itself. Upon measurement, the wave function collapses to a single outcome. The question “which slit did it go through?” has no answer when no measurement is performed.
Many-Worlds Interpretation
There is no collapse. The wave function always evolves unitarily. When which-path information is recorded, the observer becomes entangled with the particle. The “observer who saw slit 1” and the “observer who saw slit 2” both exist, in separate branches of a universal wave function. Interference is not destroyed; it becomes unobservable within a single branch.
Pilot Wave Theory
The particle always goes through one slit. But it is guided by a pilot wave (the wave function) that passes through both slits and creates an interference pattern in the guiding field. The particle follows a deterministic trajectory, but that trajectory depends on the global configuration of the experiment, including the other slit. Detecting which slit disturbs the pilot wave and destroys the interference in the guiding field.
Modern Variations
Quantum Eraser
A which-path detector is placed at the slits, destroying interference. But a second measurement downstream erases the which-path information before it can be read. Interference returns, but only in the subset of events correlated with the erasure. The full pattern, without post-selection, shows no interference.
Elitzur-Vaidman Bomb Tester
A Mach-Zehnder interferometer (a refined double-slit setup) can detect whether an object blocks one path without any particle actually interacting with the object. This “interaction-free measurement” exploits the fact that even the possibility of which-path information affects the interference pattern. It is one of the clearest demonstrations that quantum mechanics is not about what happens, but about what could happen.