Entanglement of a Photon and a Collective Atomic Excitation

We describe a new experimental approach to probabilistic atom-photon (signal) entanglement. Two qubit states are encoded as orthogonal collective spin excitations of an unpolarized atomic ensemble. After a programmable delay, the atomic excitation is converted into a photon (idler). Polarization states of both the signal and the idler are recorded and are found to be in violation of the Bell inequality. Atomic coherence times exceeding several microseconds are achieved by switching off all the trapping fields—including the quadrupole magnetic field of the magneto-optical trap—and zeroing out the residual ambient magnetic field.

Figure 1

(a) Schematic of experimental setup. $P_1$ and $P_2$, polarizers; $D1$ and $D2$, detectors; $\lambda /4$, quarter-wave plate. (b) The structure of atomic transitions leading to generation of atom-photon entanglement and of the subsequent readout of atomic qubit.

Figure 2

Measured coincidence fringe for ${\theta }_i=67.5°$. The curve is a fit based on Eq. (5), augmented by a background contribution, with $\eta =0.81\times \pi /4$, with visibility and amplitude being adjustable parameters. The visibility of the fit is 90%. Uncertainties are based on the statistics of the photon counting events.

Figure 3

Normalized signal-idler intensity correlation function $g_{si}$ as a function of storage time. Uncertainties are based on the statistics of the photon counting events. The full curve is the best exponential fit with time constant $\tau =3.7\mu \mathrm{s}$.