Spin rephasing pushes solid-state quantum memory to record single-photon storage times

The vision of a quantum internet, a global network in which information is carried not by classical bits but by qubits that can exist in superpositions of zero and one, has moved one step closer to reality. Researchers at ICFO-The Institute of Photonic Sciences in Barcelona have demonstrated that a solid-state quantum memory can store single photons for the longest time yet reported for this class of devices, using a technique known as spin rephasing to protect the fragile quantum information from degradation. The work, published in Physical Review Letters and carried out within the Quantum Internet Alliance, addresses one of the most persistent bottlenecks on the road to distributing entanglement across continental distances: the limited time a quantum memory can hold a photon without destroying the correlations that make it useful.

To understand why storage time matters so much, it helps to consider how a quantum internet would differ from the network we use today. Classical communication deals in bits, which are unambiguously either zero or one, and any signal that degrades in transit can simply be amplified and copied. Qubits obey different rules. A qubit can be zero, one, or any superposition of the two, and it can become entangled with other qubits, exhibiting correlations that no classical system can reproduce. These properties enable forms of information processing and secure communication that are impossible classically, but they come at a price: a quantum state cannot be copied or amplified without being destroyed. Every stage of a quantum network must therefore preserve the state exactly as it is, which makes losses and timing mismatches far more damaging than in conventional telecommunications.

The architecture most widely proposed for overcoming these obstacles is the quantum repeater. Rather than sending entanglement directly over a long lossy channel, a repeater scheme divides the route into shorter segments and establishes entanglement independently across each one. The segments must then be connected, and this is where quantum memories become indispensable. To stitch together entanglement from separate segments, the outcomes of measurements on one segment must be synchronized with those on another, which requires that quantum states be held in storage long enough for classical signals to travel between the nodes and for the appropriate operations to be selected. If the memory loses coherence before this synchronization is complete, the entanglement is lost with it. The storage time of the memory thus directly determines the maximum length of the individual links, and therefore the practical reach of the whole network.

Solid-state quantum memories, in which light is stored in the collective excitation of atoms embedded in a crystal, are among the most promising candidates for this role. They have already demonstrated high storage efficiency, the ability to store entanglement, and multiplexing capabilities that allow several modes to be stored simultaneously. In the experiment reported by the ICFO team, the memory consisted of a crystal doped with praseodymium ions and cooled to a temperature of 3 kelvin inside a cryostat. The researchers began by generating a pair of entangled photons: one at a telecom wavelength, chosen for compatibility with the optical fibers used to distribute light over long distances, and the other matched to the absorption profile of the memory crystal. The telecom photon served as a herald, its detection certifying that a single photon had been delivered to the memory.

Inside the crystal, the team employed the Atomic Frequency Comb protocol, a well-established technique for photon storage. In this scheme, the atoms are prepared so that their absorption spectrum takes the shape of a comb of narrow peaks. An incoming photon is collectively absorbed by the ensemble, creating a delocalized excitation spread across all the ions rather than localized in any single one. Left alone, the atoms would naturally re-emit the photon after a fixed delay determined by the comb spacing. To extend the storage beyond this natural limit, an optical control pulse transfers the collective excitation into the spin state of the ions, an energy level that does not emit light. This effectively pauses the re-emission, holding the photon in the material. A second control pulse later transfers the excitation back to the optical excited state, triggering the release of a photon that carries the same quantum information as the original.

The difficulty lies in keeping the collective spin excitation intact during this pause. Although the spin state does not radiate light, it is not immune to its environment. Each ion in the crystal experiences a slightly different local surroundings, and as a result each accumulates phase at a slightly different rate. The initial superposition, in which all ions contribute coherently to a single collective excitation, gradually dephases into a jumble of unrelated phases. Once this coherence is lost, the photon that is eventually retrieved no longer preserves the quantum properties of the original, and the information is effectively destroyed. Extending storage time is therefore not simply a matter of waiting longer; the dephasing itself must be actively counteracted.

This is precisely what the spin rephasing protocol achieves. The technique involves applying a train of radiofrequency pulses to the memory at carefully chosen intervals. Each pulse flips the phase that the ions have accumulated up to that moment. After each flip, the researchers wait for the same interval, allowing the newly accumulated phase to compensate the one that was flipped. When the sequence is complete, the accumulated phases cancel out and the collective spin excitation is restored, as if no time had passed at all. By controlling the number and timing of these pulses, the team could dictate exactly when the rephasing would occur, and at that moment they applied the second optical control pulse to read the photon out of the memory.

Using this approach, the researchers stored single photons for up to 180 microseconds, the longest storage time reported to date for this kind of solid-state quantum memory. That duration corresponds to an equivalent fiber-link distance of more than 30 kilometers, a scale that begins to be meaningful for network architectures in which memories must bridge real telecommunications distances. Crucially, the experiment was performed not with classical light but with quantum light, and the team confirmed that quantum correlations survived between the spin-rephased memory and the telecom photon that had traveled the fiber-compatible channel. Demonstrating that the correlations endure through the full storage-and-retrieval cycle is what elevates the result from an exercise in coherence control to a genuine enabling step for quantum networking.

The distinction between classical and quantum demonstrations is one the researchers themselves emphasize. Spin rephasing had been shown before with classical input states, but extending it to single photons, the currency of any future quantum internet, is what establishes its practical relevance. Alberto Rodríguez Moldes, first author of the study, notes that longer storage times should become available in the future by applying small magnetic fields to the quantum memory, a route that could push the achievable storage durations even further. ICREA Prof. Hugues de Riedmatten, who led the research, explains that the results show the technique can be extended to quantum light and that the scheme establishes praseodymium-doped quantum memories as a major candidate for the scalable implementation of quantum networks. As repeater designs mature, the combination of telecom-compatible heralding, long spin-state storage, and demonstrated preservation of quantum correlations positions this platform among the leading contenders for the memories at the heart of a future quantum internet.

Subject of Research: Long-lived storage of single photons in a spin-rephased solid-state quantum memory

Article Title: A spin rephased quantum memory for single photons

Article References: A spin rephased quantum memory for single photons. (n.d.). Original publication

Image Credits: AI Generated

DOI: Not provided

Keywords: quantum memory, quantum internet, spin rephasing, single photons, Atomic Frequency Comb, praseodymium-doped crystal, quantum repeaters, entanglement, solid-state quantum memory, quantum networks, ICFO, Physical Review Letters