quantum information · entangled photon pairs from one dot

Two photons,
born entangled.

The last layer made one exciton emit one photon. Pump the dot harder and you load two electron–hole pairs — a biexciton. It can't release both at once; it decays in a two-step cascade, emitting two photons in sequence. And when the dot is symmetric enough, those two photons come out polarization-entangled — a single semiconductor crystal acting as an on-demand source of quantum-correlated light.

Real physics: the XX→X→0 radiative cascade, biexciton binding shift, and the fine-structure-splitting condition for entanglement (Benson et al. 2000; demonstrated by Akopian 2006, Stevenson 2006). Idealized model; limits stated at the end.

// instrument

The two-photon cascade

XX → X → 0 · pump, cascade, entangle
InAs/GaAs dot
▸ pump the dot ↻ auto
FSS = 0 → the two decay paths are indistinguishable → photons entangled.
photon 1 (XX→X)
photon 2 (X→0)
energy splitting
two-photon state

Two decay routes connect XX to ground — one through each intermediate exciton spin state. If those two states have the same energy (FSS = 0), you can't tell which route the dot took, the routes interfere, and the pair emerges entangled. Open a splitting and the routes become distinguishable — the entanglement washes out into ordinary correlated light.

// the cascade

It can't let go of both at once

Loading two excitons gives the biexciton state |XX⟩, sitting near twice the single-exciton energy but shifted by the biexciton binding energy (the two pairs interact). A photon can only carry away one exciton's worth of energy, so the dot decays in steps: |XX⟩ → |X⟩ emits the first photon, then |X⟩ → |0⟩ emits the second. Because the biexciton is shifted, the two photons have slightly different energies — which is exactly how experimentalists tell the cascade photons apart.

|XX⟩ ── photon 1 ──▶ |X⟩ ── photon 2 ──▶ |0⟩ E(photon 1) = E_XX − E_X E(photon 2) = E_X the two differ by the biexciton binding: ΔE = E_b(XX) entanglement requires the intermediate |X⟩ states degenerate: FSS → 0 ⇒ |Ψ⟩ = (|H₁H₂⟩ + |V₁V₂⟩)/√2 ← maximally entangled FSS ≠ 0 ⇒ which-path info leaks ⇒ entanglement degrades

The instrument tracks both photon energies live from your binding setting, and judges the two-photon state from the fine-structure splitting — degenerate paths give the entangled Bell-type state, a nonzero splitting collapses it toward classical correlation.

// the entanglement condition

Two roads, indistinguishable

From the biexciton there are two ways down to the ground state: through the horizontally-polarized exciton or the vertically-polarized one. Each path emits its pair with matching polarization — |H,H⟩ on one road, |V,V⟩ on the other. Quantum mechanics says: if nothing in the universe records which road was taken, the dot takes both, and the emitted pair is the superposition (|HH⟩ + |VV⟩)/√2 — a maximally entangled Bell state. The thing that can betray the road is the fine-structure splitting: a tiny energy difference between the two intermediate states, caused by the dot being slightly asymmetric. Make the dot symmetric (FSS → 0) and the roads become identical — entanglement restored.

biexciton XX

Two electron–hole pairs. The starting line of the cascade. Bound a few meV from twice the single exciton.

two roads down

Via the H exciton or the V exciton. Same energy ⇒ indistinguishable ⇒ interference ⇒ entanglement.

FSS = which-path

Asymmetry splits the two roads in energy, tagging the path and erasing entanglement. The engineering goal is FSS → 0.

// why anyone cares

A crystal that makes quantum light

// honesty about the model

Where this simplifies

Background: O. Benson et al., "Regulated and Entangled Photons from a Single Quantum Dot" (Phys. Rev. Lett., 2000); N. Akopian et al. and R. M. Stevenson et al., experimental entangled-pair emission (2006). Biexciton cascade is standard quantum-dot quantum-optics.