Interfacing a two-photon NOON state with an atomic quantum memory

Multiphoton entangled states play a crucial role in quantum information applications such as secure quantum communication, scalable computation, and high-precision quantum metrology. Quantum memory for entangled states is a key component of quantum repeaters, which are indispensable in realizing quantum communications. Storing a single photon or an entangled photon has been realized through different protocols. However, there has been no report demonstrating whether or not a multiphoton state can be stored in any physical system. Here, we report on the experimental storage of a two-photon NOON state in a cold-atomic ensemble. Quantum interference measured before and after storage clearly shows that the properties of the two-photon NOON state are preserved during storage. Our experiment completes a step towards storing a multiphoton entangled state.

Figure 1

Generation and storage of a two-photon NOON state. (a) Energy diagram and timing sequence. Both pumps 1 and 2 are pulses with durations of $\mathrm{\Delta }{t}_1=\mathrm{\Delta }{t}_2=70\mathrm{ns}$. The time delay of pump 2 is $\mathrm{\Delta }T$, and the storage time in MOT $B$ is set by $\mathrm{\Delta }{t}_3$; here, $\mathrm{\Delta }{t}_3=150\mathrm{ns}$. ${\mathrm{\Delta }}_1$ and ${\mathrm{\Delta }}_2$ represent single-photon detuning of the pump-1 and pump-2/coupling light. Here, ${\mathrm{\Delta }}_2=70\mathrm{MHz}$, and ${\mathrm{\Delta }}_1–{\mathrm{\Delta }}_2$ equals the energy difference between $|1\rangle$ and $|2\rangle$. (b) Simplified experimental setup. PBS: polarizing beam splitter; BS: beam splitter; λ/2: half-wave plate; λ/4: quarter-wave plate; $D0$: photodetector; $D1(D2$/$D3$/$D4$): single-photon detector (avalanche diode, PerkinElmer SPCM-AQR-16-FC); PZT: piezoelectric ceramics; L/R: represents the memory path. Pump 1 counterpropagates with pump 2 through the atomic ensemble in MOT $A$. Experimentally, there is another PBS [not shown in (b)] used for filtering before detecting by $D1(D2$/$D3$/$D4$). Photons of signal 1 and signal 2 are collected with 1.5° to the direction of the pumping light. The coupling light is obliquely incident on the atomic ensemble in MOT $B$ with the same angle of 1.5° with respect to memory paths $L$ and $R$. The power for the light beams (pump 1, pump 2, and coupling) is about 100, 20, and 24 mW, respectively.

Figure 2

Observation of interference dip and interference peak in signals 1 and 2. Coincidence counts rates of $D1$ and $D2$ taken over 200 s (a) without and (b) with BS2 with increasing $\mathrm{\Delta }T$ (time step: 10 ns). The black squares are experimental data; the red curve is a theoretical fit using a Gaussian function. For data points in (b), the coincidence window is 450 ns (see Appendix pp2). Error bars represent ±standard deviation (s.d.) estimated from Poisson statistics.

Figure 3

Interference curves at BS2. Interference curves for (a), (b) single-photon states and (c), (d) two-photon NOON states, (a), (c) before and (b), (d) after 150 ns storage at BS2. Error bars represent ±s.d. estimated from Poisson statistics. The red curve is a fit using the Cosine function.

Figure 4

Interference curves at BS3. (a), (b) Interference curves for the two-photon NOON state before and after 150 ns storage. (c) Theoretical simulation of interference pattern for coherent light at BS3. Error bars represent ±s.d. estimated from Poisson statistics. The red curve in (a), (b) is a fit using the Cosine function.

Figure 5

The wave package of a single photon (a) before and (b) after storage. In (b), the first pulse is the leaked signal, and the second pulse is the stored signal.