Quantum-memory-assisted on-demand microwave-optical transduction

Microwave-optical transducers and quantum memories are fundamental components of quantum repeaters, essential for developing a quantum internet in which solid-state quantum computers serve as nodes interconnected by optical fibers for data transmission. Although both technologies have made significant advancements, the integration of microwave-optical conversion and quantum memory functionalities remains a challenge. Here, we theoretically propose and experimentally demonstrate a memoryenhanced quantum microwave-optical transduction using a Rydberg ensemble. By utilizing a cascaded electromagnetically induced transparency process, we store microwave photons in a highly excited collective state and subsequently convert them into optical photons during the retrieval process. Taking advantage of the optical depth with order of millions for microwave photons in Rydberg ensemble, combined with a minimal storage dephasing rate at the single-photon level, the transducer achieves an areanormalized storage efficiency greater than 90%, a bandwidth of 2.1 MHz, and a noise equivalent temperature as low as 26 K, even in cavity-free conditions. Our findings pave the way for the practical implementation of quantum repeaters based on atomic ensembles in quantum information processing.

💡 Research Summary

The paper presents a novel on‑demand microwave‑optical quantum transducer (OMQT) that integrates quantum memory and frequency conversion within a single device, using a cold ensemble of rubidium atoms excited to high‑lying Rydberg states. The authors first motivate the need for such a device in a quantum internet architecture where solid‑state qubits (e.g., superconducting circuits) operate at microwave frequencies, while long‑distance communication is carried out over optical fibers. Conventional direct conversion schemes suffer from continuous optical pumping noise and lack of temporal synchronization, which limits entanglement distribution rates.

The proposed OMQT relies on two cascaded electromagnetically induced transparency (EIT) processes. In the first stage, a microwave photon resonant with the Rydberg transition |3⟩→|4⟩ is slowed and stored as a collective excitation (spin‑wave) in the Rydberg state |5⟩ by turning off a “write” control field. After a programmable storage interval, a “read” control field drives the reverse transition |6⟩→|1⟩, converting the stored excitation into an optical photon at 780 nm. Theoretical modeling uses first‑order Maxwell‑Bloch equations for the microwave and optical field envelopes, incorporating a filling factor F that accounts for imperfect mode overlap between the waveguide TEM mode and the atomic cloud. The key parameter is the microwave optical depth d_M, which can reach ~10⁶ because the Rydberg transition possesses a gigantic electric dipole moment and a very narrow decay rate, yielding an absorption cross‑section orders of magnitude larger than that of typical optical transitions.

Efficiency analysis shows that the overall conversion efficiency η factorizes into three contributions: transmission loss η_t, spin‑wave survival η_s, and pulse‑compression factor η_c. In the regime of short storage times and large optical depths, η_s and η_c approach unity, leaving η≈η_t. An analytical expression η_t≈η₀ exp(−2γ₅₁ t_d d) is derived, where γ₅₁ is the decoherence rate of the |1⟩↔|5⟩ coherence, t_d the total delay, and d a dimensionless optical‑depth ratio. Using realistic parameters (F = 0.01, γ₅₁/2π = 10 kHz, t_d = 0.5 µs, d_M = 10⁶, d_L = 100, α_M = 50, α_L = 0.5) the model predicts η≈0.9, i.e., >90 % conversion.

Experimentally, the authors implement a proof‑of‑concept OMQT in a free‑space configuration. A two‑dimensional magneto‑optical trap (MOT) provides a cigar‑shaped cloud of ⁸⁷Rb atoms. Six optical fields (probe, write, read, and two auxiliary fields) are arranged to create the required ladder of Rydberg states: |3⟩ = |47S₁/₂⟩, |4⟩ = |47P₃/₂⟩, |5⟩ = |46D₅/₂⟩, and |6⟩ = |5P₃/₂⟩. Microwave photons at 37.5 GHz are launched via a helical antenna at a small angle to the cloud. By adjusting the probe power, the population in |3⟩ is controlled, thereby tuning the microwave optical depth. The optical readout photons are collected with a single‑mode fiber, filtered, and detected by a single‑photon counting module (SPCM). Coincidence measurements (Hanbury‑Brown‑Twiss) confirm the single‑photon nature of the retrieved light.

Key results include: (i) an area‑normalized internal conversion efficiency η_I exceeding 90 % for an average input microwave photon number ⟨N⟩ = 0.1, compared with only ~32 % for a six‑wave‑mixing direct conversion; (ii) a bandwidth of 2.1 MHz, limited by the EIT transparency window; (iii) a noise‑equivalent temperature of 26 K, derived from measured noise counts (thermal photons ≈0.109 per pulse, Rydberg fluorescence, stray light, and detector dark counts); (iv) observation of storage‑time‑dependent dephasing of the Rydberg coherence, scaling with the stored photon number. The authors also calculate entanglement distribution rates for two distant solid‑state qubits using OMQT versus direct conversion. The OMQT scheme yields a rate of ~65 kHz at η = 0.5, roughly ten times higher than the direct scheme, and the advantage grows at lower efficiencies because the memory suppresses noise photons.

Importantly, the demonstrated transducer operates without an optical cavity, making it compatible with cryogenic environments required for superconducting qubits. The authors argue that scaling up the atomic optical depth (e.g., by increasing atom number or using tighter confinement) and optimizing control pulse shapes could push the single‑photon conversion efficiency close to the theoretical limit. Future work may involve integrating the atomic ensemble with on‑chip microwave waveguides, multiplexing multiple channels, and interfacing directly with solid‑state qubits.

In summary, this work provides both a theoretical framework and an experimental validation for a memory‑enhanced, on‑demand microwave‑to‑optical quantum transducer. By leveraging the enormous microwave optical depth of Rydberg ensembles and the well‑established techniques of EIT‑based quantum memory, the authors achieve high efficiency, low noise, and on‑demand operation—all essential ingredients for scalable quantum repeaters and ultimately a functional quantum internet.