Unity Efficiency Achieved in Raman Quantum Memory: 100 % Retrieval and Storage of Single Photons

↑
↑

  //science-technology.news-articles.net/content/2 .. 100-retrieval-and-storage-of-single-photons.html
  Published in Science and Technology on Tuesday, December 9th 2025 at 11:38 GMT by Phys.org
      Locale: UNITED KINGDOM

Unity‑Efficiency Raman Quantum Memory: A Major Step Toward Scalable Quantum Networks

A team of physicists has announced a breakthrough in the field of quantum memories, reporting that a Raman‑type memory can now operate with an overall efficiency of 100 %. The study, published in Nature Communications (and highlighted by Phys.org on November 5 , 2025), demonstrates that it is possible to store and retrieve single‑photon quantum states with perfect fidelity—an achievement that could dramatically accelerate the development of quantum repeaters, secure quantum communications, and distributed quantum computing.

Why Quantum Memories Matter

Quantum memories are the “RAM” of a quantum computer or network. They temporarily hold quantum information encoded in photons, enabling tasks such as entanglement swapping, quantum error correction, and long‑distance quantum key distribution. Without efficient, high‑fidelity memories, a photon that is lost en route between nodes would break the chain of quantum correlations, limiting the distance over which quantum communication can be achieved.

Two broad families of quantum memories have emerged over the past decade: electromagnetically induced transparency (EIT)–based memories and Raman‑type memories. Raman memories are prized for their broadband operation, large acceptance bandwidth, and flexibility in choosing storage wavelengths. However, achieving high storage and retrieval efficiencies—especially while maintaining low noise—has been challenging. The new work resolves these challenges, achieving unity efficiency for the first time.

The Experimental Architecture

The researchers used a cold ensemble of ^87Rb atoms held in a magneto‑optical trap (MOT) inside a high‑finesse optical cavity. Key to the success of the experiment was the creation of an optical depth (OD) exceeding 200, a figure that boosts the probability that an incoming photon interacts with the atomic medium. The cavity mirrors were tuned to the D1 transition of rubidium (795 nm), and the cavity linewidth (∼1 MHz) matched the natural linewidth of the atomic transition, ensuring that the stored spin wave could be efficiently coupled back into a photon mode.

The Raman protocol proceeds as follows:

1. Write Pulse: A weak “signal” photon—often the result of a heralded single‑photon source—is incident on the ensemble while a strong, far‑detuned “control” laser (detuned by +2.5 GHz from the excited state) illuminates the atoms. The control field couples the ground‑state hyperfine levels via a virtual excited state, inducing a Stokes Raman scattering event that maps the quantum state of the photon onto a collective spin wave.

2. Storage: The spin wave is stored in the ground‑state coherence of the atoms for a programmable time, typically up to a few microseconds. The long storage time is maintained by magnetic field shielding and a buffer gas (neon) that suppresses collisional dephasing.

3. Read Pulse: A second, counter‑propagating control pulse retrieves the stored excitation by stimulating the anti‑Stokes transition, re‑emitting the photon with the same temporal mode and polarization as the input.

To suppress noise from spontaneous emission and four‑wave mixing (FWM), the team employed several strategies:

• Temporal Gating: The control pulses were shaped with 10 ns leading edges to avoid early leakage.
• Spectral Filtering: Two cascaded Fabry‑Perot etalons suppressed out‑of‑band photons by > 70 dB.
• Polarization Isolation: The Raman process preserves the input polarization; a polarizing beam splitter (PBS) further rejected control‑field leakage.

By optimizing these parameters, the researchers achieved a measured storage‑and‑retrieval efficiency of 1.00 ± 0.02, as confirmed by both coincidence counting with a heralded single‑photon source and by homodyne tomography that quantified the fidelity of the retrieved quantum state.

Comparisons to Previous Work

Prior demonstrations of Raman memories have typically reported efficiencies in the 70–90 % range. A 2023 breakthrough by the same group achieved 95 % efficiency with a longer storage time, but the product of storage and retrieval efficiencies fell short of unity. Other teams, using EIT or rare‑earth‑doped crystals, have reached near‑perfect retrieval but at the cost of narrow bandwidth or limited operating wavelengths.

The current achievement represents a quantum leap: the product of storage and retrieval efficiencies—i.e., the overall quantum transmittance—has reached 100 %. This indicates that, for a single‑photon input, the probability that it survives the complete write‑store‑read sequence is essentially one. The fidelity, measured through Wigner‑function tomography, remained above 99 %, confirming that the memory preserves the quantum coherence of the input state.

The Significance for Quantum Networks

With unity efficiency, the overhead required to repeat quantum signals over long distances decreases dramatically. Quantum repeaters rely on swapping entanglement between nodes via photon‑mediated Bell‑state measurements. Each swap introduces a probabilistic loss; if the memory efficiency is low, the success probability scales unfavorably with the number of segments. A 100 % efficient memory essentially removes this bottleneck, allowing for the construction of repeaters that can span thousands of kilometers with only modest resources.

Beyond repeaters, the new memory also opens the door to photonic cluster‑state generation for measurement‑based quantum computing. In such schemes, entangled photon states are prepared and stored until all necessary gates have been performed. The high efficiency ensures that the cluster remains intact, reducing the need for extensive redundancy.

Finally, the memory’s broadband nature—thanks to the Raman protocol—makes it compatible with existing photonic sources, including those that emit at telecom wavelengths (1.55 µm). By simply adding a frequency conversion stage, the same architecture could serve as a building block for a global quantum internet.

Future Directions

The researchers plan to extend the storage time beyond the current microsecond regime by employing magnetic field–insensitive “clock” states and dynamical decoupling techniques. They also intend to integrate the system onto a chip‑scale platform, which would make the memory more amenable to field deployment.

Another avenue is the incorporation of this memory into a continuous‑wave quantum repeater scheme, where the storage time is matched to the coherence time of entangled photon pairs produced by spontaneous parametric down‑conversion. This would allow real‑time synchronization of entanglement across a network.

The study also references a parallel effort by a group at the University of Stuttgart, who achieved a similar efficiency using a solid‑state ensemble of praseodymium‑doped yttrium orthosilicate at cryogenic temperatures. While the solid‑state approach offers long storage times (milliseconds), it suffers from lower bandwidth; the Raman approach offers a complementary route that could be combined in hybrid architectures.

Conclusion

The report of a Raman quantum memory with unity efficiency marks a watershed moment in quantum information science. By overcoming the longstanding trade‑off between bandwidth, noise, and efficiency, the new system delivers a near‑perfect quantum interface between light and matter. The implications are far‑reaching: from enabling scalable quantum repeaters and robust quantum key distribution, to providing a key component for future quantum processors and networks. As the field moves from laboratory demonstrations to practical implementations, such breakthroughs will pave the way toward a truly global quantum internet.