Entanglement swapping across a five-node relay in a multiplexed quantum-classical network

AI summary

Overview

Research area: Experimental quantum networking — specifically "quantum-classical coexistence," where entangled photons and conventional high-power optical data traffic share the same standard telecommunications fiber.

Technical level: Intermediate. The core results are accessible with a basic understanding of entanglement and fiber optics, but the modeling section (Gaussian continuous-variable simulations) and the Bell-state-measurement details require graduate-level familiarity with quantum optics.

Scope: The paper reports the first experimental demonstration of entanglement swapping (a five-node quantum relay) operating simultaneously with 10-Gbps classical data traffic over the same four standard fiber segments, with total span up to 40 km.

What This Paper Is About

Real fiber infrastructure is already full of bright classical laser light carrying Internet traffic, and that light generates noise photons through spontaneous Raman scattering (SpRS) that can drown out fragile quantum signals. Entanglement swapping — the protocol that entangles two photons that never interacted, by measuring two other photons at a middle node — had previously been shown only in fibers dedicated exclusively to quantum signals, and quantum-classical coexistence experiments had involved at most three connected nodes. This paper asks whether swapping still works across a longer, five-node relay when every one of its four fiber links simultaneously carries ordinary 10-Gbps data.

Key Contributions

1. First entanglement swapping alongside live classical traffic. The authors demonstrate time-bin entanglement swapping over a five-node relay in which all four long-distance fibers simultaneously carry 10-Gbps C-band classical data, a configuration not previously reported (prior swapping demonstrations used quantum-dedicated fiber, and prior coexistence work spanned at most three nodes).

2. A quantified fidelity-versus-classical-power trade-off. They measure the spontaneous Raman scattering rate in each fiber and show how swapped-state visibility degrades monotonically as classical launch power rises, backed by a theoretical model of the Raman noise physics.

3. Operation at a realistic link length. Swapping is demonstrated across 40 km (four 10-km SMF-28 spools) and 20 km (four 5-km spools), with all four photons near 1536.6 nm, chosen for compatibility with erbium-ion quantum memory transitions.

4. A design tool for future networks. A detailed simulation — extended from Gaussian continuous-variable formalism to include spectral/temporal photon correlations and realistic filtering, using 3,000 to 10,000 frequency/time modes per arm — is used to project performance for alternative wavelength allocations (C/C-band, C/O-band, O/C-band) and fiber lengths up to roughly 350 km.

Main Findings

• Dark-fiber baseline (20 km relay): Without classical light, swapped entanglement visibilities are V_X = (73.4 ± 1.8)% and V_Z = (85.9 ± 1.4)%. The lower X-basis value is attributed to imperfect indistinguishability; both are reduced by multiphoton emission.

• Noise scaling: Measured SpRS rates per fiber are [γ₁, γ₂, γ₃, γ₄] = [1.20, 1.49, 2.61, 1.12] × 10⁶ counts/s/mW, scaling linearly with launch power. Fiber lengths and launch powers varied by 3.6% and 17.5% respectively across the four links.

• Operational threshold: For the 10-Gbps transceivers used, error-free communication requires more than −18.5 dBm at the receiver; roughly 1.2 dB attenuation over the 5-km fibers dictated that fiber input power P₀ exceed −17.3 dBm. Visibility degradation becomes significant for average classical launch power above −15 dBm.

• Full coexistence achieved (20 km): With all classical receivers operating above the error-free threshold, the average visibility across measurement bases is approximately 75% — above the 70.7% limit required for a Bell test and far above the 33.3% entanglement bound. The Z basis is more noise-sensitive, attributed to an additional 3 dB loss in the measurement interferometers.

• Robustness headroom: The authors estimate that an initially ideal swapping setup could sustain V > 70.7% up to approximately −7 dBm of classical power, indicating the underlying entanglement correlations are highly robust to substantial noise.

• Source-limited, not noise-limited, fidelity in dark fiber: Individual source visibilities are V_X,S1 = (95.3 ± 0.1)% and V_X,S2 = (96.6 ± 0.1)%; four-fold Hong–Ou–Mandel visibility is V_HOM = (85.9 ± 1.1)% in the main text (the appendix reports (86.1 ± 1.0)% at a higher mean photon number near μ ≈ 0.02). Correcting for multiphoton degradation gives an estimated indistinguishability of I ≈ 0.92 at the Bell-state measurement.

• 40-km relay results: In dark fiber, V_X+ = (71.4 ± 5.1)% and V_X− = (74.1 ± 4.8)%. With classical signals at an average launch power of −18.7 dBm per spool through all four 10-km fibers, V_X+ = (74.5 ± 3.8)%, V_X− = (75.0 ± 4.7)%, and the Z-basis average is (85.2 ± 1.5)%.

• Fidelity above the Bell-test threshold: Using F = (1 + 3V_avg)/4 with an averaged visibility of (78.2 ± 2.1)%, the fidelity to |Ψ⁻⟩ is F = (83.7 ± 1.6)% — exceeding the Bell-test requirement by 3.6 standard deviations.

• Projected reach: Simulations suggest the current C-band system could reach roughly 100 km of fiber, beyond which noise and loss cause rapid visibility decline. O-band quantum signals with C-band classical traffic could in principle reach roughly 350 km because O-band SpRS can be nearly five orders of magnitude weaker, but at the cost of an additional 0.13 dB/km loss relative to C-band.

• Source-limited scaling path: The authors state that improved spectral purity (via optimized filter and pump spectral profiles or intrinsically pure sources) could improve tolerance to classical power by nearly an order of magnitude.

Methodology in Plain English

Two identical sources each produce entangled photon pairs by pumping a nonlinear crystal with a pulsed laser (a 1536.6-nm mode-locked laser, 6-ps pulses at 396.83 MHz, frequency-doubled to 768.3 nm for type-II spontaneous parametric down-conversion). A Michelson interferometer turns each pump pulse into a pair of pulses separated by 346 ps, so each photon pair is emitted in a superposition of "early" and "late" time bins — the time-bin Bell state |Ψ⁺⟩.

One photon from each source travels L/4 to a central node (Charlie), where the two photons interfere on a 50:50 beam splitter and are detected. A successful detection pattern there projects them onto |Ψ⁻⟩ and heralds entanglement between the two photons that went to Alice and Bob, which never interacted. Alice and Bob each use Michelson interferometers to scan measurement phase and read out the resulting fringes.

The classical light is 10-Gbps data from standard pluggable transceivers at 1547.7 nm (ITU channel 37), on the red side of the quantum channel at 1536.6 nm (ITU channel 51). Standard 100-GHz dense wavelength-division multiplexers combine and separate the channels at each fiber; a second, narrower 50-GHz filter stage (roughly 41-GHz passband) sits in each quantum path to suppress Raman noise and to improve spectral purity for interference.

Because temperature and stress drift the polarization and arrival time of photons over hours, automated fiber polarization controllers update every few seconds, and a variable optical delay is recalibrated every 10 minutes by briefly (for 10 seconds) offsetting one source by 173 ps — half the time-bin spacing — achieving roughly 500 fs RMS alignment precision. Detection uses superconducting nanowire single-photon detectors (~90% efficiency, 40-ps FWHM jitter) with ~2-ps time-tagging and 300-ps coincidence windows. Classical link health was monitored in real time via field-programmable gate arrays running the open-source Quantum Instrumentation Control Kit (QICK) with integrated bit error ratio tests.

Why This Matters

Research impact: This moves quantum-classical coexistence from simple single-photon transmission and short chains to a genuine multi-node swapping protocol — the primitive underlying quantum repeaters and relays. It also supplies a calibrated model that separates noise-induced degradation from source-intrinsic imperfections, giving other groups a template for predicting performance in lit fiber.

Real-world applications:

• Quantum key distribution and post-quantum-secure communications over existing carrier fiber, without dedicated dark fiber.
• Quantum repeater and relay links for extending entanglement across metropolitan and inter-city distances.
• Distributed quantum sensing and clock networks that require entanglement between separated nodes.
• Distributed quantum computing, where swapping distributes entanglement between processor modules.

Industry relevance: Telecom operators own the fiber, and their C-band channels routinely carry more than 20 dBm of power. This work quantifies how much classical power a C-band quantum channel can tolerate, what wavelength plans help (O-band quantum signals, or hollow-core fiber with ultra-low nonlinearity for high-power C-band coexistence), and what filtering and detector performance are needed — concrete inputs for equipment and network architects.

Future Directions

• O-band swapping experiments. The authors state that near-term experimental work will explore O-band swapping, building on recent high-power O-band teleportation results, to test the predicted ~350 km reach.
• Improving source purity and multiphoton performance through optimized filter and pump spectral profiles or intrinsically high-purity sources, which the authors estimate could raise classical-power tolerance by nearly an order of magnitude and enable terabits-per-second classical rates.
• Adding memory and purification. Extending the relay to a full quantum repeater requires quantum memories and entanglement purification; the paper flags that how these behave over classical-populated infrastructure "has yet to be evaluated."
• Mapping the broader design space. A systematic study is needed across number of segments, segment lengths, node placement, co- versus counter-propagation, and wavelength assignments, since the authors note that direct comparison with end-to-end quantum transmission is ill-posed in configurations like their four-fiber design.

Target Audience

Experimental quantum-networking and quantum-optics researchers; engineers designing quantum-safe or hybrid quantum-classical telecom infrastructure; program managers and network architects at national labs, carriers, and quantum hardware companies; and graduate students entering the field who want a worked example of how photonic entanglement swapping is built, stabilized, and benchmarked under realistic noise.

Authors’ abstract

Quantum networks are resources for scaling quantum computers and distributed sensing technologies while offering post-quantum security benefits. Teleporting non-classical resources like entanglement, via so called entanglement swapping, is essential for networks in particular overcoming rate-loss limits via quantum repeaters. Deploying these systems on real infrastructure will likely require multiplexing photonic qubits into fibers carrying 'classical' light encoding standard Internet communications and control plane signals for multi-node quantum protocols. Here, we report the first demonstration of entanglement swapping and conventional communications operating over the same fibers. Entanglement is swapped across a five-node quantum relay topology connected by four long-distance fibers, each populated with classical data signals. Time-bin entangled photons in the C-band are multiplexed alongside C-band classical signals using dense-wavelength division multiplexing, introducing noise photons generated by high-power classical light. We experimentally and theoretically characterize the trade-off between quantum fidelity and Raman noise photons. Entanglement swapping is demonstrated over a maximum fiber length of 40 km (four 10-km fibers) while simultaneously transmitting 10-Gbps classical data through all fibers. These results represent a significant advancement in the demonstrated complexity of coexisting quantum and classical networks and provide a roadmap for achieving the widespread deployment of advanced quantum technologies.