First-hand information for everyone
This paper proposes and analyzes a way to connect separate quantum processors using a single trapped atom as a short-term memory. Instead of requiring two photons to arrive at the same time and interfere, the scheme lets one processor send a photon that becomes entangled with the atom and is stored there while the other processor keeps trying. When the second photon arrives and is detected, a final readout of the atom completes the entanglement between the two processors.
The authors describe a concrete implementation based on a single rubidium atom held in a high-finesse optical cavity and a near-deterministic photon–atom controlled-Z (CZ) gate. A CZ gate is a simple two-qubit operation that imprints a phase only when both qubits are in a certain state; here it is implemented by reflecting one rail of a dual-rail photonic qubit from the cavity so that the photon picks up an atom-state-dependent phase. Detecting the reflected photon heralds that the processor’s communication qubit is now entangled with the atomic memory. The process repeats with the second processor, and a final optical readout of the atom projects the three-party state onto a Bell pair between the two processors.
This approach matters because it removes two major bottlenecks of common linear-optics methods. Standard Bell-state measurements with linear optics (the type-II fusion gate) need both photons to be indistinguishable and to arrive in the same attempt, so the success rate falls like the square of the already-low probability that a processor delivers a photon. By storing the first link in the atom, the new protocol does not require simultaneous arrivals and does not need the photons from the two processors to be identical. That change lets the entanglement rate scale roughly linearly with the per-attempt photon-delivery probability over a wide range. The authors also note that the cavity gate and atomic readout operate on nanosecond time scales, which can give orders-of-magnitude higher entanglement rates compared with the linear-optics approach.
The paper does more than sketch the idea. It derives the expected entanglement rates under realistic device parameters and includes two key error sources: the decay of the stored atom–processor link while waiting for the second photon, and decoherence caused when photons interact with the memory but are not detected (unheralded interactions). The analysis is based on analytical formulas and Monte Carlo simulations, using projected parameters for an integrated atom–cavity system called QS-LINK. The authors also discuss practical details such as operation at 795 nm (the rubidium transition) and the option to convert telecom photons to 795 nm at the interconnect, with conversion efficiency folded into the delivery probability.
Important caveats remain. The results are predictions based on models and projected hardware parameters, not a completed experimental demonstration. The scheme’s benefit depends on having a memory qubit with long enough coherence to wait for the second photon, and on keeping unwanted photon interactions with the atom low enough not to corrupt the stored link. Real-world losses, coupling efficiencies, and atomic coherence times will determine the actual speed-up. Nevertheless, the study shows that a single-atom, cavity-based memory could remove a central scaling bottleneck in modular quantum computing if the projected device performance can be realized.