Long-Lived Quantum Memory Enabling Atom-Photon Entanglement over 101 km of Telecom Fiber

Long-distance entanglement distribution is the key task for quantum networks, enabling applications such as secure communication and distributed quantum computing. In this work, we take a crucial step toward this task by sharing entanglement over long optical fibers between a single ${}^{87}$$\mathrm{Rb}$ atom and a single photon. High fidelity of the atomic state could be maintained during long flight times through such fibers by prolonging the coherence time of the single atom to 10 ms based on encoding in long-lived states. In addition, the attenuation in the fibers is minimized by converting the wavelength of the photon to the telecom S band via polarization-preserving quantum frequency conversion. These improvements enable us to observe entanglement between the atomic quantum memory and the emitted photons transmitted through standard spooled telecom fibers with a length of 101 km with a fidelity of $70.8\pm 2.4$%. This fidelity is comparable to recent demonstrations over 20 km, despite the channel loss now significantly exceeding 20 dB. In fact, now the reduction in fidelity is due to detector dark counts rather than loss of coherence of the atom or photon, proving the suitability of our platform to realize city-to-city-scale quantum network links.

Popular Summary

Future quantum networks will connect quantum computers efficiently over large distances and provide a platform for applications of multiparty quantum communication such as distributed quantum computing, quantum sensing, and secure communication. Here, we present a major milestone toward this grand goal, that is, the distribution of entanglement between a single-atom-based quantum network node and a photon transmitted over up to 101 km of telecom fiber.

Our quantum network node consists of a single rubidium atom where the entanglement is generated in a spontaneous emission between the polarization of the photon with the final state of the atom. To achieve very long transmission distances, we convert the wavelength of the photon into the telecom window without disturbing its quantum state. While the photon is traveling through the long fiber link, about half a millisecond for 100 km, the atomic quantum memory has to be protected from decoherence. Transferring the atomic state to a memory basis, a coherence time of 7 ms was obtained, which effectively would allow the entanglement to be distributed over fiber links of almost 1000 km.

The observed fidelity of 71% is largely limited by dark counts of the current detectors, showing the suitability of our platform to realize high-fidelity city-to-city-scale quantum network links. The next steps to increase its efficiency and to improve the rate is using atom arrays, enabling parallelization of the entanglement generation as well as implementation of quantum gates, thereby paving the way to efficient quantum repeaters for weaving future quantum networks.

Figure 1

The architecture of a neutral atom-based quantum network node. (a) A single ${}^{87}$$\mathrm{Rb}$ atom is stored in a strongly focused ODT (wavelength 849.5 nm, waist 2.05 µm, and trapping potential 2.3 mK for an intensity of 4.97 mW µm⁻²). Emitted photons are collected by a high-numerical-aperture (NA) objective (NA = 0.5) and coupled into a 5 m single-mode fiber. The photons first pass a microelectromechanical systems (MEMS) fiber switch, which selectively guides the photons either to a silicon avalanche photodiode (APD) during atom loading and atomic state readout or to a quantum frequency-conversion (QFC) device during entanglement-generation attempts. The QFC device converts the emitted photons at 780 nm to 1517 nm. Next, the telecom photons propagate through up to 101 km of spooled fiber, after which their polarization is analyzed with a polarizing beam splitter (PBS) and superconducting nanowire single-photon detectors (SNSPDs). (b) The experimental sequence to generate and analyze atom-photon entanglement, which is divided into three stages: cooling (green background), entanglement generation (light-gray background), and atomic readout (dark-gray background). The entanglement-generation attempts are executed in bursts of 11, after which the atom is recooled. Successful detection of a single photon interrupts a burst and starts the atomic state readout. (c) The scheme of the atom-photon entanglement generation. A $\pi$-polarized laser pulse excites the atom, followed by a spontaneous decay to the ${|\downarrow ⟩}_z$ or ${|\uparrow ⟩}_z$ state. Here, $\{{|\uparrow ⟩}_z,{|\downarrow ⟩}_z\}$ is defined as the initial basis for the atomic qubit. (d) The scheme of the state-selective readout. Depending on the polarization of the readout laser pulse, i.e., ${\sigma }^{-}$, ${\sigma }^{+}$ or a superposition of them, either the state ${|\downarrow ⟩}_z$, ${|\uparrow ⟩}_z$, or a superposition of them, is initially excited to the $5^2{P}_{1/2}$ state and subsequently ionized. An additional cycling laser pulse reexcites the atoms, which decay to the $5^2{S}_{1/2}|F=2⟩$ hyperfine state during the readout process, to the $5^2{P}_{3/2}|F^{\prime }=3⟩$ state, from which they are also ionized.

Figure 2

Characterization of the quantum memory coherence time. The measured atom-photon state visibility as a function of the storage time (see also Appendix pp2). The red dots are measured visibilities in the initial basis $\{{|\uparrow ⟩}_z,{|\downarrow ⟩}_z\}$, achieved with ODT ramping and a magnetic bias field. To obtain a longer coherence time, the atomic qubit is mapped into the memory basis via a Raman transfer (blue dots). The data are fitted with exponential decay functions showing an increase in the coherence time from $279.1\pm 31.5\text{μ}\mathrm{s}$ in the initial basis to $10.51\pm 1.04\mathrm{ms}$ in the memory basis. The top axis indicates the corresponding fiber length by approximating the speed of the photon in the optical fiber with $\frac{2}{3}c$, where $c$ is the speed of light. The inset shows the atom-photon correlation for a storage time of about 267.92 µs.

Figure 3

Characterization of the Zeeman-state-selective Raman transfer. (a) The scheme of the Raman transfer for the atomic qubit: a Raman $\pi$ pulse, which couples either ${|\uparrow ⟩}_z$ to $5^2{S}_{1/2}|F=2,m_F=+1⟩$ or ${|\downarrow ⟩}_z$ to $5^2{S}_{1/2}|F=2,m_F=-1⟩$ via a two-photon Raman process—by choosing the former, we change from the initial basis to the memory basis $\{{|\downarrow ⟩}_z,|F=2,m_F=+1⟩\}$. (b),(c) The Raman transfer for varying two-photon detuning $\delta$ (b) at zero magnetic field and (c) with a bias field of $B_z=244.5$ mG. An atom is prepared in the $|F=1⟩$ manifold in the state ${|\downarrow ⟩}_z$ (blue) or ${|\uparrow ⟩}_z$ (orange) state, after which the Raman-transfer pulse is applied. A reduced $|F=1⟩$ population signals a successful state transfer to the $|F=2⟩$ manifold. A bias field is applied to achieve a Zeeman-state-selective transfer with a contrast of $97.8\pm 2.2\mathrm{\%}$. This contrast is limited by the pulse power drift and the polarization purity of the Raman-transfer beam.

Figure 4

The quantum network node performance for entanglement distribution over 50 and 101 km of optical fiber. (a) The detection-time histogram of the converted photons. Photon detections are accepted within a window of 50 ns, indicated with dashed black lines, accepting approximately 62 % of the single-photon events. (b) The signal-to-noise ratio (SNR) characterization as a function of the fiber length. The lines represent a model of the SNR as a function of the fiber lengths for $R_{\mathsf{d}\mathsf{c}}=6$ as well as for 1 count/s. For completeness, we have also measured the SNR for a fiber length of 5 km. (c) The atom-photon state correlations over a fiber length of 50 km. In the 50-ns photon window, $6548$ photon-detection events were recorded in 25.2 h. The resulting state fidelity is $\ge 84.8\pm 1.1\mathrm{\%}$. (d) The atom-photon state correlations over a fiber length of 101 km. We have recorded $656$ photon-detection events in 47.7 h. The atom-photon correlations have been measured in three bases, resulting in a state fidelity of $\ge 70.8\pm 2.4\mathrm{\%}$. All error bars indicate one standard deviation.

Figure 5

The energy-level diagram of the Raman-transfer scheme. We utilize a pair of ${\sigma }^{+}$-polarized Raman beams on the ${}^{87}$$\mathrm{Rb}$ $D_1$ line at $795$-nm wavelength to employ the Raman transfer; thus two transfers are possible. For a three-level Raman transfer, we define $|m⟩=|F=1,m_F=+1⟩$, $|n⟩=|F=2,m_F=+1⟩$, and $|k⟩=|F^{\prime }=2,m_{F^{\prime }}=+2⟩$. For a four-level Raman transfer, we define $|a⟩=|F=1,m_F=-1⟩$, $|b⟩=|F=2,m_F=-1⟩$, $|3⟩=|F^{\prime }=1,m_{F^{\prime }}=0⟩$, and $|4⟩=|F^{\prime }=2,m_{F^{\prime }}=0⟩$. We also define $|1⟩=|F=1,m_F=0⟩$ and $|2⟩=|F=2,m_F=0⟩$. The energy difference between the two levels $|1⟩$ and $|2⟩$ ($|3⟩$ and $|4⟩$) is defined as ${\mathrm{\Delta }}_{\mathsf{h}\mathsf{f}\mathsf{s}}$ (${\mathrm{\Delta }}_{\mathsf{h}\mathsf{f}{\mathsf{s}}^{\prime }}$). Assuming a weak bias field $B_z$, the ground (excited) levels will be split with energy difference $g_F{m}_F{\mu }_B{B}_z$ ($g_{F^{\prime }}{m}_{F^{\prime }}{\mu }_B{B}_z$), where $|g_F|=1/2$ ($|g_{F^{\prime }}|=1/6$). The frequencies of the Raman pair are defined with ${\omega }_{\text{L1}}$ and ${\omega }_{\text{L2}}$, respectively.

Figure 6

An analysis of the quantum memory coherence time in the memory basis by measuring atom-photon correlations at different storage times.

Figure 7

The optimization of the quantum memory coherence time. During the storage time, two sets of dipole trap depths are used to determine the coherence time under different conditions. (a) The dipole trap depth is 2.3 mK (referred to as “high-ODT”). For these measurements, four different conditions have been tested: high ODT (light gray), a bias field along the $y$ axis of $B_y$=64 mG (light green), a bias field along the $z$ axis of $B_z$=244.5 mG (light red), and a Raman transfer to the memory basis together with a bias field of $B_z$=244.5 mG (light blue), respectively. The corresponding coherence times for these four conditions are $194.3\pm 32.6\text{μ}\mathrm{s}$, $560.2\pm 67.3\text{μ}\mathrm{s}$, $71.4\pm 1.6\text{μ}\mathrm{s}$, and $2.32\pm 0.37\mathrm{ms}$, respectively. (b) The dipole trap depth is adiabatically ramped down to 0.19 mK (referred to as “low-ODT”) and the same conditions yield coherence times of $184.9\pm 20.7\text{μ}\mathrm{s}$, $1.24\pm 0.18\mathrm{ms}$, $279.1\pm 31.5\text{μ}\mathrm{s}$, and $10.51\pm 1.04\mathrm{ms}$, respectively.

Figure 8

An analysis of the data recorded during the 101-km atom-photon experiment: the effects of time filtering on the SNR, the fidelity, and the number of events. The length of the acceptance window is varied by fixing the start timing of the acceptance window and changing the end timing.

Figure 9

The detector-click probabilities for varying fiber length. Three processes can be distinguished that generate single-photon detector clicks: the signal photons (blue), noise introduced by the QFC device (gray solid line), and noise introduced by the detectors, i.e., detector dark counts (gray dashed line). In our setup, the detector dark counts become the prominent noise source for fiber lengths above 50 km.