Quantum physics · Landmark · Young 1801 → understood
The Double-Slit Experiment
One thing at a time, and still it interferes.
Make a beam of light so faint that it comes out one tiny speck at a time. Think of each speck as a little bullet of light. You fire them one by one at a wall with two narrow slits cut in it, and you watch where each one lands on a screen behind.
Every speck lands at a single point and makes one bright dot, just like a thrown pebble would. Now watch what happens as the dots pile up. After hundreds, then thousands of them, they don't gather into two clumps behind the two slits. Instead they slowly build a striped pattern of bright and dark bands, as if each single speck had gone through both slits at once and overlapped with itself.
Then comes the strangest part. The moment you peek to check which slit each speck actually goes through, the stripes vanish and you are left with two plain bands. Looking at it changes what happens. Try the button under the picture.
When particle waves pass through two slits, they overlap. Bright stripes (fringes) form on a back screen wherever the waves arrive perfectly in step, meaning the distance travelled from one slit is exactly a whole number of wavelengths longer than the distance from the other. This creates a predictable pattern where the spacing between the stripes depends directly on the wavelength of the particle and the distance between the slits.
Quantum mechanics assigns each particle a complex probability amplitude, and the probability of finding it somewhere is the squared modulus. That is the Born rule, \(P = |\psi|^2\). With both slits open the amplitude is a sum over paths, \(\psi = \psi_1 + \psi_2\), so
\[ P = |\psi_1 + \psi_2|^2 = |\psi_1|^2 + |\psi_2|^2 + \underbrace{2\,|\psi_1||\psi_2|\cos\Delta\varphi}_{\text{interference}} . \]
The first two terms are just the dull single-slit blobs. The cross term is the fringes, oscillating with the relative phase \(\Delta\varphi = \tfrac{2\pi}{\lambda}\,(r_2 - r_1)\). Picture each amplitude as a little rotating arrow, a phasor. Where the two phasors point the same way they add and you get bright; where they oppose they cancel and you get dark.
Now introduce a which-path measurement. It entangles the particle with a detector, \(\psi_1|D_1\rangle + \psi_2|D_2\rangle\), and ignoring the detector multiplies the cross term by the overlap \(\langle D_1|D_2\rangle\). As the path becomes knowable that overlap shrinks, the relative phase smears, and the interference term averages to zero. This is decoherence, and it recovers the classical sum \(|\psi_1|^2 + |\psi_2|^2\). The fringe visibility \(V\) and the path distinguishability \(\mathcal{D}\) trade off as \(V^2 + \mathcal{D}^2 \le 1\). Dial the which-path slider and watch \(V\) fall.
Specks fired one at a time land as single dots, yet dot by dot a striped pattern appears. Press “watch which slit” to make the stripes collapse. Circular wavelets from the two slits overlap; dark channels are where they cancel, and the screen reads out the fringes. Slide \(\lambda\) and \(d\) to change the spacing. Each screen point sums two phasors: aligned ⇒ bright, opposed ⇒ dark. Add which-path info and watch the cross-term fade.
No one is watching. Dot by dot, the stripes build up.
Reduced-motion is on: the figure shows a resolved still frame and the controls are paused.