Quantum Storage of Frequency-Multiplexed Photons Exhibiting Nonclassical Correlations with Telecom C-Band Photons

Abstract

Multiplexing is essential for improving entanglement distribution rates in quantum communication. Frequency multiplexing provides a promising and scalable path toward large-capacity quantum networks. Further progress requires increasing the number of frequency modes and developing broadband photon-pair sources and quantum memories that are spectrally compatible. Here, we report the integration of a cavity-enhanced spontaneous parametric down-conversion source in the telecom C-band with a frequency-multiplexed atomic frequency comb memory. The bow-tie cavity source was simultaneously resonant at 606 nm and 1550 nm, generating non-degenerate photon pairs exhibiting a clustered frequency-comb spectrum. The atomic frequency comb memory, implemented in Praseodymium-doped Yttrium Orthosilicate crystals, provided up to 83 frequency modes with 123 MHz spacing and enabled broadband storage of 606 nm signal photons. By filtering the main cluster, we obtained $32.7 \pm 4.8$ effective modes, as confirmed from coincidence measurements. Importantly, we observed strong nonclassical correlations after storage, with cross-correlation values of $g_{s,i}^{(2)} = 8.1\pm0.7$. Our experimental results demonstrate the feasibility of integrating cavity-enhanced photon-pair sources with rare-earth-ion-doped solid-state memories. The integration reveals a high frequency multiplicity that is essential for scalable quantum networks.

Figures (4)

• Figure 1: (a) Cluster structure of our PPS. Since the SPDC output bandwidth is restricted to the frequencies where the two wavelengths are resonant simultaneously (double resonance), the output spectrum of each photon forms a cluster structure. The central cluster (main cluster) has a width of approximately 10 GHz. The orange dashed line indicates the broadband spectral envelope of SPDC without the cavity, illustrating the full phase-matching bandwidth before spectral narrowing by the cavity. (b) Time correlation of photon pairs generated by the PPS without cavity-length locking. The histogram shows the time difference between the detection of an idler photon and that of a signal photon, and the black line represents a fit to the envelope. The pump power was 0.1 mW, and the measurement time was 10 min.
• Figure 2: Absorption spectrum of the Pr:YSO crystal showing an AFC with approximately 83 frequency modes. Insets show 20-MHz-wide magnified views of representative modes $0$, $\pm20$, and $\pm41$.
• Figure 3: Schematic of the experimental setup. The PPS pump laser (436 nm) is incident on a PPKTP crystal inside the cavity. The signal and idler photons are separated by a DM and directed to their respective paths. Among them, the 1550 nm idler photons are reflected by a VBG so that only the main cluster is selectively filtered, and they are detected by an SSPD. The 606 nm signal photons are stored in the quantum memory, and the retrieved photons are subsequently detected by an SPCM. An etalon is placed in front of the SPCM to reduce noise by filtering out photons outside the inhomogeneous broadening of Pr:YSO. Both lasers used in this experiment are frequency-stabilized using an optical frequency comb. The inset shows the experimental sequence. This sequence is cyclically repeated and indicates the timing when the AOM shutters in the setup are opened.
• Figure 4: (a) Time correlation between idler photons and signal photons retrieved from the AFC. The inset shows the comb structure. The SPDC pump power was 1 mW, and the measurement time was 2 h. (b) Number of effective modes as a function of the number of AFC modes created in the Pr:YSO crystal. The SPDC pump power was 2.2 mW. (c) Second-order cross-correlation function $g_{s,i}^{(2)}(400~\mathrm{ns} )$ as a function of the pump power in cSPDC. The black dashed line indicates the classical limit ($g_{s,i}^{(2)} = 1.03$) for 33 modes.