arXiv:2611.04526 · Quantum Frontiers
Long-lived quantum memories with rare-earth ions doped in crystals
Yu-Yu Sa, Pei-Xi Liu, Zong-Quan Zhou, Chuan-Feng Li
Abstract
Quantum memory is a core enabling device for large-scale quantum networks, serving as the fundamental building block for quantum repeaters and transportable quantum memories. Among various physical platforms, rare-earth-ion-doped crystals stand out as one of the most promising solid-state candidates for long-lived photonic quantum memories, owing to their exceptionally long spin coherence time and their compatibility with integration. In this review, we first introduce the physical mechanisms underlying long-lived spin coherence in rare-earth systems, with particular emphasis on the zero-first-order-Zeeman transition and dynamical decoupling techniques. We then review the experimental progress in this field, highlighting three distinct lifetimes--the spin coherence time, the light storage time, and the quantum storage time--and covering experiments that have achieved hour-scale spin coherence, minute- to hour-scale light storage, and minute-scale single-photon-level quantum storage. Finally, we discuss the remaining challenges that hinder the realization of practical long-lived quantum memories, explore potential strategies to overcome these obstacles, and outline the prospects for their future applications in quantum networks.
AI Summary
This review examines how rare-earth-ion-doped crystals support long-lived photonic memories and what prevents long spin coherence from translating into efficient quantum storage. It surveys published experiments rather than presenting new data, focusing on zero-first-order-Zeeman (ZEFOZ) transitions, which suppress linear magnetic-field sensitivity, and dynamical decoupling (DD), which refocuses spin dephasing. The discussion distinguishes ensemble spin coherence, coherent optical storage, and storage with demonstrated quantum fidelity. Reviewed milestones include approximately six-hour spin coherence in Eu$^{3+}$:Y$_2$SiO$_5$ under ZEFOZ conditions and DD; atomic-frequency-comb optical storage with a $52.9\pm1.2$ minute decay time and $92.9\pm4.9\%$ interference visibility after one hour; and noiseless-photon-echo storage of weak coherent inputs with a $27.6\pm0.5$ second decay time. In the latter experiment, time-bin fidelity was $88.0\%$ after 5.6 seconds, exceeding the corresponding $82.1\%$ classical benchmark. Retrieval after 42 seconds had a single-photon-equivalent signal-to-noise ratio of $1.34\pm0.33$, inferred from inputs averaging 4.14 photons; quantum fidelity was not directly demonstrated at that duration. The review identifies competing requirements for absorption and coherence: higher dopant concentration improves optical depth but increases interactions and broadening, while comb preparation further reduces effective absorption. Long DD sequences introduce pulse-error noise and heating, and ZEFOZ operation remains sensitive to higher-order magnetic fluctuations and field alignment. Proposed directions include improved crystals and control sequences, alternative ZEFOZ transitions, cavity-enhanced absorption, and compact cryogenic and magnetic hardware for integrated or transportable memories.
Central claim may not hold: The review claims minute-scale quantum storage, but its explicitly reported above-classical qubit-fidelity demonstration is at 5.6 s, with no equivalent certification reported at tens-of-seconds delays. At 42 s, the reported single-photon-equivalent SNR is normalized from coherent pulses with mean photon number 4.14, and its interpretation as beyond-classical storage is conditional on an ideal single-photon input rather than established for the measured input ensemble. The conclusion identifies the 27.6 s signal-decay constant as a quantum storage time without establishing that the applicable quantum-fidelity benchmark remains exceeded for that duration.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Journal reference
- Quantum Frontiers 5, 14 (2026)
- DOI
- 10.1007/s44214-026-00110-1
- Comments
- The incorrect value 4.14 on page 7 of the published version has been corrected to 1.16