A high-performance quantum memory for quantum interconnects

Abstract

Single photons are the flying qubits of choice for distributing entanglement in a quantum internet. Quantum memories embedded in quantum repeaters are crucial to overcome transmission loss and enhance the rate of quantum communication. A multimode memory can further boost the channel capacity. However, benchmarking and building a practical quantum memory that simultaneously optimizes multiple performance metrics poses two key challenges. Here, we introduce quantum interconnect rate to comprehensively quantify quantum memories, and further demonstrate a high-performance quantum memory that simultaneously integrates three essential criteria at once: large multimode capacity, high efficiency, and high fidelity. Operating on 11-dimensional spatial modes, our memory achieves a uniform efficiency exceeding 80% and qubit storage fidelities above 99%, enabling the efficient storage of high-dimensional qudits. Based on these capabilities, we estimate a distribution of 3.56 bits of quantum information over a 1000-km repeater link in one minute, highlighting a practical pathway toward scalable quantum interconnects and quantum networks.

Figures (9)

• Figure 1: Perfomance of multimode quantum memory (a) Scheme of a quantum interconnect based on quantum repeaters. Entanglement swapping within a repeater node (dashed blue box) extends the range of entanglement between end-nodes (blue cubes) and repeater nodes (yellow spheres). The multimode quantum memory enhances the performance of the quantum interconnect in two ways: i. increasing the entanglement swapping rate through multiplexing, and ii. increasing the channel capacity via high-dimensional storage. (b) Comprehensive performance of a high-dimensional quantum memory. The performance is quantified for a 1000-km, 2-layer quantum repeater channel. The QIR, $\mathcal{R}_{qm}$, is plotted as a function of the number of modes $M$ and the storage efficiency $\eta_s$, under different levels of depolarizing noise ($p_n = 1\%, 10\%, 50\%$). The other parameters are set as follows: $\eta_d=1$, $p=1$, and $L_{\text{att}}=22~\mathrm{km}$.
• Figure 2: Efficient quantum memory for single photons in spatial modes. (a) Experimental setup. Two $^{85}$Rb atomic ensembles are employed. After transmission through 200 m of fiber, the anti-Stokes photons are encoded into OAM or POV modes, using SLM 1. The encoded photons are stored in MOT 2 , retrieved by programmable control pulses, decoded with SLM 2, and guided to SPCMs 2 and 3 via a fiber beam splitter. The inset shows three examples of phase patterns on SLM 1 and their corresponding beam profiles at the atomic ensemble, including an OAM mode with $\ell=+1$, a POV mode with $\ell=+5$, and a superposed POV mode $\ell_1+\ell_2$ where $\ell_1=+5$ and $\ell_2=-5$. The energy levels involved in the EIT process are shown in the dashed box: the ground state $\vert g \rangle=\vert 5S_{1/2},~F=2 \rangle$, the storage state $\vert s \rangle=\vert 5S_{1/2},~F=3 \rangle$, and the excited state $\vert e \rangle=\vert 5P_{1/2},~F=3 \rangle$. (b) Storage efficiency. Single photons encoded with OAM (blue) and POV (yellow) modes are stored and retrieved to quantify the storage efficiency. (c) Single-photon autocorrelation. The second-order correlation function $g^{(2)}(0)$ is plotted as a function of POV mode order. Error bars indicate one standard deviation.
• Figure 3: Storage fidelity of photonic qubits. (a) Density matrix of a photonic qubit before and after storage. Two POV modes, $\ell_1=+5$ and $\ell_2=-5$, form a qubit. The density matrix of the state $\vert \varphi \rangle=(\vert \ell_1 \rangle + i\vert \ell_2 \rangle)/\sqrt{2}$ is reconstructed as $\rho_o$ (before storage) and $\rho_s$ (after storage), respectively. (b) Storage fidelity within a qubit subspace. Six states are tested in the two-dimensional subspace defined by $\vert \ell_1 \rangle$ and $\vert \ell_2 \rangle$: $D=(\vert \ell_1 \rangle+\vert \ell_2 \rangle)/\sqrt{2}$, $M=(\vert \ell_1 \rangle-\vert \ell_2 \rangle)/\sqrt{2}$, $L=(\vert \ell_1 \rangle+i\vert \ell_2 \rangle)/\sqrt{2}$, $R=(\vert \ell_1 \rangle-i\vert \ell_2 \rangle)/\sqrt{2}$, $\vert \ell_1 \rangle$, and $\vert \ell_2 \rangle$. These states are represented on the Bloch sphere. The storage fidelity for each state is shown below, with the average fidelity $\bar{F}=99.3\%$ indicated by the dashed line. (c) Average storage fidelity of qubits. From 55 possible qubit pairs, 11 are randomly selected for fidelity measurement. The average fidelity for the six states of each selected qubit is displayed in the corresponding cell of the table (in percentage).
• Figure 4: Characterization of photonic qudits. Qudit states are characterized before storage (a) and after storage (b). For both the input and retrieved states, we measure the heralding rate $h_o^{(i)}$ (before) and $h_s^{(i)}$ (after), as well as the population distribution $q_o^{(i)}$ (before) and $q_s^{(i)}$ (after) for each POV mode.
• Figure S1: Experiment setup. a, Schematic of the experiment. Two $^{85}$Rb atomic ensembles serve as the biphoton source and the quantum memory, respectively. The intensities of the pump and coupling beams are monitored by photodetectors (PDs) and actively stabilized using half-wave plates (HWPs), quarter-wave plates (QWPs), and polarizing beam splitters (PBSs) for polarization control. The anti-Stokes photons are projected onto spatial light modulators (SLMs) for encoding and decoding qudit states. A flip mirror (FM) directs the beam toward a Fourier lens (FL) and an auxiliary camera to monitor the spatial mode. b, Relevant energy levels of $^{85}$Rb and the corresponding optical transitions for the pump, coupling, and photon fields. c, Experimental timing sequence. During the measurement window, the trapping beams are turned off. Upon detection of a Stokes photon, the control beam is switched off and on to store and retrieve the correlated anti-Stokes photon. The control beam amplitude is linearly ramped to compensate for the decreasing optical depth and to maximize the storage efficiency.