Entanglement distribution: To herald or not to herald

High-rate, high-fidelity entanglement distribution is essential for the creation of a quantum internet, and spontaneous parametric downconverters (SPDCs) are, at present, the preferred sources of entangled signal-idler photon pairs for transmission to Alice and Bob’s quantum nodes. SPDCs using phase-matched spectral islands are especially attractive, in this regard, because they provide wavelength-division multiplexed signal-idler pairs with single-mode temporal behavior. This paper compares the entanglement distribution rates of three islands-based systems. Two use idler detections for heralding: islands-based zero-added-loss multiplexing (ZALM), and an islands-based Sagnac SPDC source with signal-path erasure. The third employs an unheralded Sagnac SPDC source. For 90% or lower heralding efficiencies, ZALM’s per-pump-pulse entanglement distribution rate exceeds that of the signal-path erasure source, and both rates are inferior to unheralded operation’s when all three systems employ $N_I$ spectral islands and allocate $N_M = N_I$ quantum memories to each pump pulse. These behaviors, however, must be weighed against the three systems’ differing equipment requirements, e.g., ZALM requires a pair of perfectly-matched Sagnac sources, which is a significant burden not incurred by the signal-path erasure approach, and both heralded systems will suffer, in comparison with unheralded operation, if they cannot realize high heralding efficiencies.

💡 Research Summary

This paper presents a comprehensive comparative analysis of three distinct entanglement distribution systems designed for building a high-rate, high-fidelity quantum internet. All systems are based on Spontaneous Parametric Down-Converters (SPDCs) that utilize “phase-matched spectral islands,” a technology that provides wavelength-division multiplexed signal-idler photon pairs with desirable single-mode temporal behavior.

The three architectures under scrutiny are: 1) Islands-based Zero-Added-Loss Multiplexing (ZALM), which employs a pair of perfectly-matched Sagnac SPDC sources and heralds entanglement via a partial Bell-state measurement (BSM) on the idler photons. 2) An islands-based Sagnac SPDC source with signal-path erasure (following Chahine et al.’s design), a simpler heralded source using a single Sagnac loop. 3) An unheralded Sagnac SPDC source, which transmits signal photon pairs directly without any idler-based heralding mechanism.

The core of the paper rigorously evaluates and compares the per-pump-pulse entanglement distribution rates of these systems. A key finding is that for heralding efficiencies at or below 90%, the ZALM scheme outperforms the signal-path erasure source in terms of distribution rate. However, a more significant result emerges when considering scenarios where multiple entangled pairs can be generated per pump pulse, utilizing multiple spectral islands and quantum memories. In this case, when all systems employ the same number of islands (N_I) and allocate an equal number of quantum memories (N_M = N_I), the unheralded source’s distribution rate surpasses that of both heralded systems, even when assuming an optimistic 90% heralding efficiency. This highlights the substantial rate penalty imposed by losses in the heralding detection chain.

The paper strongly emphasizes that raw distribution rate is not the sole criterion for choosing a system. The analysis weighs these performance behaviors against the practical technological burdens associated with each approach. For instance, ZALM requires a complex setup with two matched Sagnac sources and a multi-channel BSM with many single-photon detectors, a significant challenge not faced by the simpler single-Sagnac heralded source. Furthermore, the performance of both heralded systems is highly sensitive to their achievable heralding efficiency; if high efficiency (e.g., >90%) cannot be realized, their disadvantage compared to unheralded operation becomes more pronounced.

Additional practical considerations are also addressed, including the impact of detector dark counts on the heralded systems and the effect of background light during free-space transmission to the quantum receivers (Alice and Bob). The paper concludes that the choice between heralded and unheralded entanglement distribution is a non-trivial trade-off. It depends on a balance between the target application’s required fidelity and rate, the achievable heralding efficiency with available components, and the acceptable level of system complexity and cost. The work provides crucial insights for engineers designing practical quantum network hardware, moving beyond theoretical performance limits to grapple with real-world implementation constraints.