Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

# One-hour coherent optical storage in an atomic frequency comb memory

## Abstract

Photon loss in optical fibers prevents long-distance distribution of quantum information on the ground. Quantum repeater is proposed to overcome this problem, but the communication distance is still limited so far because of the system complexity of the quantum repeater scheme. Alternative solutions include transportable quantum memory and quantum-memory-equipped satellites, where long-lived optical quantum memories are the key components to realize global quantum communication. However, the longest storage time of the optical memories demonstrated so far is approximately 1 minute. Here, by employing a zero-first-order-Zeeman magnetic field and dynamical decoupling to protect the spin coherence in a solid, we demonstrate coherent storage of light in an atomic frequency comb memory over 1 hour, leading to a promising future for large-scale quantum communication based on long-lived solid-state quantum memories.

## Introduction

Quantum repeater1 is proposed to establish the long-distance entanglement in the presence of channel loss. However, the construction of a practical quantum repeater is of great challenge due to its demanding requirements and system complexity2,3,4,5,6. Space-borne quantum communication7,8 and transportable quantum memory9 can avoid using optical fibers, but optical quantum memories with a lifetime on the order of hours would be essential for extending the communication distance to global scale.

With the help of a zero-first-order-Zeeman (ZEFOZ) magnetic field and dynamical decoupling (DD), Zhong et al. have reported a 6 h spin coherence lifetime9 in europium-doped yttrium orthosilicate (Eu3+:Y2SiO5). This long-lived spin coherence is a key resource for constructing quantum memories for global quantum communication. Although hours of spin coherence time have been demonstrated, long-lived optical storage remains a challenge because of the complicated and unknown energy structures in ZEFOZ fields9,10,11 and a reduced effective absorption in magnetic fields12 because only one subsite can be used for the long-term storage11. To date, the longest optical storage time is ~1 min realized in 87Rb atoms13 and a Pr3+:Y2SiO5 crystal using the electromagnetically induced transparency protocol10. For single-photon-level storage, the longest storage time is ~1 s realized in 87Rb atoms14.

To take advantages of the long-lived spin coherence, a spin-wave based optical storage protocol should be employed. So far, atomic frequency comb (AFC)15,16 has been the only successful protocol for spin-wave storage of photonic qubits17,18,19 in the rare-earth-ion doped crystals. In Eu3+:Y2SiO5, spin-wave storage of polarization qubit20 and the generation and multimode storage of single photons as spin waves21 have been demonstrated based on AFC protocol. A recent work extends the AFC memory time to 0.5 s in Eu3+:Y2SiO522.

Here we demonstrate a coherent optical memory with a storage time of 1 h using a 151Eu3+:Y2SiO5 crystal, based on the spin-wave AFC protocol16 in a ZEFOZ field, namely ZEFOZ-AFC method. DD is used to protect the spin coherence and extend the storage time. The coherent nature of this device is verified by implementing a time-bin-like interference experiment after 1 h storage with a fidelity of 96.4%, which shows the feasibility of qubit storage. Our work improves the AFC memory time by ~6000 times22.

## Results

### Characterization of the energy level structure

In order to implement optical storage in a ZEFOZ field, the knowledge of the energy level structures in both the ground and excited state is a prerequisite. Previous works11,23,24 have determined the spin Hamiltonians for the ground-state 7F0 and excited-state 5D0, which can be used to predict the level structures in any given magnetic field. However, the theoretical predictions may have an error comparable to the storage bandwidth and lead to a wrong choice of pumping strategies. In order to precisely determine the structures, we first find the ZEFOZ transition $${\left|3\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|4\right\rangle }_{{\rm{g}}}$$ by adjusting the strength of the field and orientation of the sample. We then use the continuous-wave Raman heterodyne detection (RHD)25 to probe the ground-state resonances in the ZEFOZ field. However, continuous-wave RHD fails to detect the excited-state resonances due to the short-lived population in the optically excited states and the weaker interaction strength with the radio-frequency (RF) field. Here we employ the pulsed RHD to probe the excited-state resonances (see Supplementary Note 1). The experimentally determined energy level structure is presented in Fig. 1a.

### Experimental setup

In order to increase the effective absorption, we use an isotopically enriched sample of 151Eu3+:Y2SiO5 crystal (Scientific Materials), with a doping level of 1000 ppm and an isotopic enrichment of 99.3%. The sample has an optical depth αL of 2.6 for Eu3+ at site 1. In order to quickly find the correct direction of the sample in the ZEFOZ field, the optical surface of the sample is cut perpendicularly to the direction of the known ZEFOZ magnetic field with a cylindrical shape, such that the field of the superconductor magnet along the vertical direction (Fig. 1b) is approximately aligned with the ZEFOZ field (a photo is provided in Supplementary Fig. 2b). The field is 1.280 T in the direction of [−0.535, −0.634, 0.558] in the [D1, D2, b] frame of the crystal as indicated in previous works9,11). The sample has a diameter of 4.5 mm and a thickness of 6 mm, and is mounted on two goniometers. The goniometers allow tilting in two dimensions, with a tilting range within 6.6° and a positioning resolution of 0.002°. A pair of 4-turn Helmholtz coils with a diameter of 6 mm is placed at the two sides of the sample. The coils are driven by an arbitrary waveform generator with the output amplified by an RF amplifier with an output power of 300 W, to implement the DD sequences. The experiments are conducted after the sample is cooled to 1.7 K.

The laser is locked to an ultra-stable plano-concave cavity at 580.039 nm, with a linewidth well below 10 kHz. The pump and probe beam are independently generated by acoustic-optic modulators in the double-pass configuration. Then the pump and probe beam are overlapped with each other inside the crystal with a diameter of 200 and 50 μm respectively, and an angle of ~30 mrad with each other. The beams double pass the crystal, and the transmitted probe beam is collected by a single-mode fiber. The probe beam is then combined with a reference laser at a locked frequency offset of 43.66 MHz for optical heterodyne detection. The beat signal is captured by a photodetector and demodulated to give the amplitude of the signals.

### Spin-wave AFC and dynamical decoupling

The experimental sequence is shown in Fig. 2a. The experiment begins with a preparation procedure based on spectral hole burning26 (see Supplementary Note 2). The first step is the so-called “class-cleaning”, which can isolate a single class of ions in the inhomogeneously broadened sample. Six chirped pulses which are resonant with $${\left|1\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|1\right\rangle }_{{\rm{e}}}$$, $${\left|2\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|2\right\rangle }_{{\rm{e}}}$$, $${\left|3\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|3\right\rangle }_{{\rm{e}}}$$, $${\left|4\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|3\right\rangle }_{{\rm{e}}}$$, $${\left|5\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|3\right\rangle }_{{\rm{e}}}$$, and $${\left|6\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|5\right\rangle }_{{\rm{e}}}$$ transitions are applied to choose the target ions as shown in Fig. 1a. Other classes of ions will relax to other nonresonant ground-state levels and no longer interact with the six pulses. The chirp bandwidth of these pulses is 1 MHz. The second step is the so-called “spin polarization”, which initializes all the population of the chosen ions into the state $${\left|3\right\rangle }_{{\rm{g}}}$$ by using the five pulses mentioned above except for $${\left|3\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|3\right\rangle }_{{\rm{e}}}$$. After these two steps, a Λ system has been well isolated, which comprises the level $${\left|3\right\rangle }_{{\rm{g}}}$$, $${\left|4\right\rangle }_{{\rm{g}}}$$, and $${\left|3\right\rangle }_{{\rm{e}}}$$. The third step is the “AFC preparation”. An AFC is prepared in the $${\left|3\right\rangle }_{{\rm{g}}}$$ level by frequency-selected hole burning while keeping the $${\left|4\right\rangle }_{{\rm{g}}}$$ level empty. The AFC has a bandwidth of 1.0 MHz and a periodicity of Δ = 100 kHz, which leads to an echo emission at t = 1/Δ = 10 μs after input. After the AFC preparation, a 2 μs (full width at half maximum) probe pulse of Gaussian profile with a power of 150 μW resonant with $${\left|3\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|3\right\rangle }_{{\rm{e}}}$$ is absorbed by the sample and creates a collective excitation. A control pulse with a width of 4 μs and a power of 360 mW resonant with $${\left|4\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|3\right\rangle }_{{\rm{e}}}$$ immediately transfers the optical coherence into a spin-wave excitation. The control pulse has a complex hyperbolic secant profile27 to achieve efficient control over the bandwidth of the AFC. Then a DD sequence is applied to protect the spin coherence. After that, a second control pulse transfers the spin coherence back to the optical excitation. The collective state will continue rephasing and emits an AFC echo when the excited-state storage time reaches 1/Δ = 10 μs.

Two kinds of DD sequences, i.e., CPMG and KDDx, are used to protect the spin coherence in our experiments. As shown in Fig. 2a, each of them consists of sequences of π pulses with different phase shifts, where τ is the spacing between the center of the π pulses. As a comparison with the optical storage times, the spin coherence times with DD are also measured using RHD (see Supplementary Note 3). CPMG and KDDx can reach a 1/e spin coherence lifetime of 2.68 ± 0.06 and 0.95 ± 0.03 h when τ = 20 ms, respectively. For storage of light, we measure the decay of the echo with various intervals τ shown in Fig. 3a, c. The longest storage times are achieved at τ = 100 ms, where CPMG and KDDx have a lifetime of 52.9 ± 1.2 and 33.3 ± 1.1 min, respectively. These results are in good agreement with the spin coherence times measured with the same τ = 100 ms, which are 50.6 ± 2.0 and 38.2 ± 2.0 min (see Supplementary Note 3).

In Fig. 2b, we present oscilloscope traces of the spin-wave AFC storage with a CPMG sequence with τ = 100 ms, which consists of four π pulses. Traces recorded after longer storage times can be found in Fig. 3b. When τ is shorter than 100 ms, unlike the spin coherence times, the optical storage lifetimes are shortened as shown in Fig. 3a, c. This is because DD sequences with smaller intervals result in a greater heating effect of the RF coils, which broadens the optical homogeneous linewidth28, leading to a reduction of the AFC efficiency.

The storage efficiency is analyzed as follows:

$${\eta }_{{\rm{total}}}\,=\,{\eta }_{{\rm{AFC}}}\,\cdot\, {({\eta }_{{\rm{control}}})}^{2}\,\cdot\, {\eta }_{{\rm{spin}}}.$$
( 1)

The two-level AFC efficiency ηAFC without spin-wave storage and DD is 4.5%. This efficiency can be enhanced close to unity by using cavity enhancement29 technique to achieve an efficient photon absorption. We also note that this efficiency will be reduced if a RF sequence is applied to the sample due to the heating. For example, ηAFC is reduced to 2.5% after a DD sequence with τ = 100 ms. The transfer efficiency of the control pulse ηcontrol is determined to be 38.5% (see Supplementary Note 2). This transfer efficiency is not very high because the polarization of the incident light cannot be parallel to D1 axis (maximum absorption) of the crystal due to the special sample orientation described above. The total storage efficiencies for 5 min storage with CPMG and KDDx when τ = 100 ms are 0.035% and 0.052%, respectively. ηspin is the efficiency of storage in the spin states, which can be estimated from Eq. (1) to be 9.5% and 14.1% for 5 min storage with CPMG and KDDx when τ = 100 ms, respectively. ηspin is primarily limited by the insufficient bandwidth of the π pulses and the inhomogeneity of the RF coils in our case. The width of the π pulse is determined at 65.1 μs (see Supplementary Note 3). However, the inhomogeneous broadening Γinh of the spin transition $${\left|3\right\rangle }_{{\rm{g}}}\,\leftrightarrow\, {\left|4\right\rangle }_{{\rm{g}}}$$ of the Eu3+ ions in the crystal is 30 kHz, which cannot be entirely covered by the bandwidth of the π pulse Γπ ≈ 1/(65.1 μs) ≈ 15 kHz. The decoherence of those ions, with transition frequencies outside the bandwidth of the π pulse, cannot be recovered by the DD sequences, thus leading to an efficiency loss. ηspin is also related to the inhomogeneity of the RF field generated by the coils, because the thickness of the sample is comparable to the diameter of the coils. A more efficient spin manipulation can be achieved by RF coils with an improved homogeneity, thus increasing the storage efficiency. The total storage efficiencies can be greatly enhanced by the combination of those techniques mentioned above.

### Coherence verification

Quantum information can be encoded into various degrees of freedom of photons, such as frequency, time and polarization. The time-bin encoding has the particular advantage of good robustness to environmental fluctuations. Therefore, here we implement the coherent light storage with time-bin-like encoding17. The experimental sequence is shown in Fig. 4a. Here the AFC preparation and DD sequences are the same as that described above. In order to verify its long-term coherence protection, CPMG with an interval τ = 100 ms is used to extend the spin coherence time. Two input pulses with an interval t are stored in the memory and partially read out with two π/2 pulses separated by the same t, respectively, after the CPMG sequence. Each input pulse is read out as two AFC echoes, and therefore the later echo of the earlier input and the earlier of the later are overlapped. The relative phase Δφ between the two input pulses is varied to generate the interference pattern presented in Fig. 4b while fixing the relative phase Δθ between the two π/2 pulses. The visibility of the interference after a storage time of 5, 30, and 60 min is V = 93.0 ± 5.5%, 95.3 ± 4.2%, and 92.9 ± 4.9%, corresponding to a fidelity of F = (1 + V)/2 = 96.5 ± 2.8%, 97.6 ± 2.1%, and 96.4 ± 2.5%, respectively, which indicates a promising application as a faithful quantum memory for time-bin qubits in the future. Compared with the longest optical storage time demonstrated so far, which is 1 min realized in Pr3+:Y2SiO510, this coherent storage time is approximately enhanced by 60 times.

## Discussion

In summary, spin-wave AFC protocol is combined with ZEFOZ technique for the purpose of long-lived quantum storage of photonic qubits. An optical memory with a coherent optical storage time of 1 h and a time-bandwidth product of 3.6 × 109 is achieved using a 151Eu3+:Y2SiO5 crystal. In order to extend this work to single-photon regime, higher efficiencies and good signal-to-noise ratios (SNR) are required. There are two sources of noise when it comes to single-photon regime. One is the coherent noise caused by the intense control pulse, which can be filtered out with spectral holes17, Fabry–Pérot cavities12, etc. The other is the photon noise caused by extra population in the ground state introduced by the DD sequences30, which can be suppressed by better optical pumping and high-fidelity π pulses. The storage time in this work can be further extended by using a stronger ZEFOZ field23, where a higher SNR can be expected since less DD pulses are required.

## Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

## References

1. Briegel, H.-J., Dür, W., Cirac, J. I. & Zoller, P. Quantum repeaters: the role of imperfect local operations in quantum communication. Phys. Rev. Lett. 81, 5932–5935 (1998).

2. Sangouard, N., Simon, C., Riedmatten, H. D. & Gisin, N. Quantum repeaters based on atomic ensembles and linear optics. Rev. Mod. Phys. 83, 33–80 (2011).

3. Duan, L.-M. & Monroe, C. Colloquium: Quantum networks with trapped ions. Rev. Mod. Phys. 82, 1209–1224 (2010).

4. Specht, H. P. et al. A single-atom quantum memory. Nature 473, 190–193 (2011).

5. Kalb, N. et al. Entanglement distillation between solid-state quantum network nodes. Science 356, 928–932 (2017).

6. Yu, Y. et al. Entanglement of two quantum memories via fibres over dozens of kilometres. Nature 578, 240–245 (2020).

7. Gündoğan, M. et al. Space-borne quantum memories for global quantum communication. Preprint at http://arxiv.org/abs/2006.10636 (2020).

8. Yin, J. et al. Satellite-based entanglement distribution over 1200 kilometers. Science 356, 1140–1144 (2017).

9. Zhong, M. et al. Optically addressable nuclear spins in a solid with a six-hour coherence time. Nature 517, 177–180 (2015).

10. Heinze, G., Hubrich, C. & Halfmann, T. Stopped light and image storage by electromagnetically induced transparency up to the regime of one minute. Phys. Rev. Lett. 111, 033601 (2013).

11. Ma, Y. et al. A Raman heterodyne study of the hyperfine interaction of the optically-excited state 5D0 of Eu3+:Y2SiO5. J. Lumin. 202, 32–37 (2018).

12. Jobez, P. et al. Coherent spin control at the quantum level in an ensemble-based optical memory. Phys. Rev. Lett. 114, 230502 (2015).

13. Dudin, Y. O., Li, L. & Kuzmich, A. Light storage on the time scale of a minute. Phys. Rev. A 87, 031801 (2013).

14. Yang, S.-J., Wang, X.-J., Bao, X.-H. & Pan, J.-W. An efficient quantum light-matter interface with sub-second lifetime. Nat. Photon. 10, 381–384 (2016).

15. de Riedmatten, H., Afzelius, M., Staudt, M. U., Simon, C. & Gisin, N. A solid-state light-matter interface at the single-photon level. Nature 456, 773–777 (2008).

16. Afzelius, M., Simon, C., de Riedmatten, H. & Gisin, N. Multimode quantum memory based on atomic frequency combs. Phys. Rev. A 79, 052329 (2009).

17. Gundogan, M., Ledingham, P. M., Kutluer, K., Mazzera, M. & de Riedmatten, H. Solid state spin-wave quantum memory for time-bin qubits. Phys. Rev. Lett. 114, 230501 (2015).

18. Kutluer, K. et al. Time entanglement between a photon and a spin wave in a multimode solid-state quantum memory. Phys. Rev. Lett. 123, 030501 (2019).

19. Yang, T.-S. et al. Multiplexed storage and real-time manipulation based on a multiple degree-of-freedom quantum memory. Nat. Commun. 9, 3407 (2018).

20. Laplane, C. et al. Multiplexed on-demand storage of polarization qubits in a crystal. N. J. Phys. 18, 013006 (2015).

21. Laplane, C., Jobez, P., Etesse, J., Gisin, N. & Afzelius, M. Multimode and long-lived quantum correlations between photons and spins in a crystal. Phys. Rev. Lett. 118, 210501 (2017).

22. Holzäpfel, A. et al. Optical storage for 0.53 s in a solid-state atomic frequency comb memory using dynamical decoupling. N. J. Phys. 22, 063009 (2020).

23. Longdell, J. J., Alexander, A. L. & Sellars, M. J. Characterization of the hyperfine interaction in europium-doped yttrium orthosilicate and europium chloride hexahydrate. Phys. Rev. B 74, 195101 (2006).

24. Cruzeiro, E. Z. et al. Characterization of the hyperfine interaction of the excited 5D0 state of Eu3+:Y2SiO5. Phys. Rev. B 97, 094416 (2018).

25. Mlynek, J., Wong, N. C., DeVoe, R. G., Kintzer, E. S. & Brewer, R. G. Raman heterodyne detection of nuclear magnetic resonance. Phys. Rev. Lett. 50, 993–996 (1983).

26. Lauritzen, B. et al. Spectroscopic investigations of Eu3+:Y2SiO5 for quantum memory applications. Phys. Rev. B 85, 115111 (2012).

27. Rippe, L., Nilsson, M., Kröll, S., Klieber, R. & Suter, D. Experimental demonstration of efficient and selective population transfer and qubit distillation in a rare-earth-metal-ion-doped crystal. Phys. Rev. A 71, 062328 (2005).

28. Könz, F. et al. Temperature and concentration dependence of optical dephasing, spectral-hole lifetime, and anisotropic absorption in Eu3+:Y2SiO5. Phys. Rev. B 68, 085109 (2003).

29. Sabooni, M., Li, Q., Kroll, S. & Rippe, L. Efficient quantum memory using a weakly absorbing sample. Phys. Rev. Lett. 110, 133604 (2013).

30. Cruzeiro, E. Z., Fröwis, F., Timoney, N. & Afzelius, M. Noise in optical quantum memories based on dynamical decoupling of spin states. J. Mod. Opt. 63, 2101–2113 (2016).

## Acknowledgements

This work is supported by the National Key R&D Program of China (No. 2017YFA0304100), the National Natural Science Foundation of China (Nos. 11774331, 11774335, 11504362, 11821404, and 11654002), Anhui Initiative in Quantum Information Technologies (No. AHY020100), the Key Research Program of Frontier Sciences, CAS (No. QYZDY-SSW-SLH003), the Science Foundation of the CAS (No. ZDRW-XH-2019-1), and the Fundamental Research Funds for the Central Universities (No. WK2470000026, No. WK2470000029). Z.-Q.Z. acknowledges the support from the Youth Innovation Promotion Association CAS.

## Author information

Authors

### Contributions

Z.-Q.Z. and C.-F.L. conceived the experiment. Y.M. and Y.-Z.M. performed the experiment. Y.M. and Z.-Q.Z. wrote the paper. Z.-Q.Z., C.-F.L., and G.-C.G. supervised the project. All authors discussed the experimental procedures and results.

### Corresponding authors

Correspondence to Zong-Quan Zhou or Chuan-Feng Li.

## Ethics declarations

### Competing interests

The authors declare no competing interests.

Peer review information Nature Communications thanks the anonymous reviewer(s) for their contribution to the peer review of this work. Peer reviewer reports are available.

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Reprints and Permissions

Ma, Y., Ma, YZ., Zhou, ZQ. et al. One-hour coherent optical storage in an atomic frequency comb memory. Nat Commun 12, 2381 (2021). https://doi.org/10.1038/s41467-021-22706-y

• Accepted:

• Published:

• DOI: https://doi.org/10.1038/s41467-021-22706-y

• ### High-performance cavity-enhanced quantum memory with warm atomic cell

• Lixia Ma
• Xing Lei
• Kunchi Peng

Nature Communications (2022)

• ### Elimination of noise in optically rephased photon echoes

• You-Zhi Ma
• Ming Jin
• Guang-Can Guo

Nature Communications (2021)

• ### Proposal for space-borne quantum memories for global quantum networking

• Mustafa Gündoğan
• Jasminder S. Sidhu
• Markus Krutzik

npj Quantum Information (2021)