Every pump pulse either excites the quantum dot or, in laser mode, attenuates a weak coherent pulse. The emitted light hits a 50/50 beamsplitter and is routed to Detector A or Detector B. A running cross-correlation of the two detector click-records builds the g²(τ) histogram, in units of the pulse period τ.
g²(τ) = ⟨n_A(t)·n_B(t+τ)⟩ / (⟨n_A⟩⟨n_B⟩)
- Quantum dot — each excitation stores at most one electron-hole pair, so at most one photon can ever leave per pulse. It can never split to both detectors at once, so g²(0) is pinned near 0: perfect antibunching.
- Attenuated laser — a coherent (Poissonian) source occasionally emits 2 or more photons in the same pulse purely by chance, so both detectors sometimes click together. g²(0) stays near 1, exactly like all the side peaks.
- Pump rate — how often pulses arrive; only changes animation speed, not the statistics.
- Mean photons / pulse (μ) — sets the average brightness for both sources so the comparison is fair; quantum-dot excitation probability is capped at 1 photon regardless of μ.
Real-world relevance: this Hanbury Brown–Twiss measurement is the standard lab test used to certify quantum-dot, colour-centre and molecule-based single-photon sources for quantum key distribution and photonic quantum computing — a device only counts as a genuine single-photon source if its measured g²(0) drops well below 0.5.