In Barcelona, researchers have quietly extended one of the most delicate boundaries in modern physics: how long a single photon's quantum identity can be held in stillness before the world erases it. By coaxing individual particles of light into a rare-earth crystal and applying carefully choreographed radiofrequency pulses to keep the quantum ensemble in step with itself, the team at ICFO preserved photon coherence for 180 microseconds — long enough, in principle, for light to cross 36 kilometers of fiber. The achievement does not yet bridge continents, but it demonstrates that the patient in
Researchers achieve record quantum memory storage, advancing path to practical quantum internet
Quantum information can survive storage while the network catches up
Mark
Why does 180 microseconds matter? That sounds impossibly short.
Mimi
It's short in human terms, but in quantum systems it's substantial. Light travels about 36 kilometers of fiber in that time. For a quantum repeater—a node that needs to hold quantum information while other parts of the network synchronize—that's the difference between a connection that works and one that collapses.
Luke
But they also showed classical light stored for 3 milliseconds. Why the huge gap?
Mimi
That's the real question. The spin rephasing technique works, but quantum operation introduces lower efficiency and more noise. Some of that noise comes from magnetic field imperfections.
Mark
So this isn't a finished product.
Mimi
Not at all. It's a proof that the architecture can work. They're showing the path forward—better magnetic field control could push single-photon storage much longer.
Luke
How much longer? What's the actual target for a practical quantum network?
Mimi
The source doesn't specify. We know large multisegment networks will need longer storage than 180 microseconds, but the exact requirement depends on network design.
Mark
Is this the only approach to quantum memory?
Mimi
No. There are other architectures—different materials, different techniques. But this one uses telecom-wavelength photons, which is a real advantage for fiber networks.
Luke
And they published in Physical Review Letters, so the work has been peer-reviewed.
Mimi
Yes. The team is at ICFO in Barcelona, a well-established quantum research center.
Mark
What happens next?
Mimi
They'll work on magnetic field control, try to push storage times longer, and see if this architecture can scale to multiple nodes. But this is foundational work—it shows the concept is viable.
The Pulse
- Quantum networks face a fundamental constraint classical networks do not: quantum information cannot be copied, so it must be held in memory while distant nodes laboriously synchronize — and until now, solid-state memories lost coherence far too quickly to be useful.
- The Barcelona team's crystal-based memory was undermined by its own atomic diversity — each praseodymium ion drifting out of phase with its neighbors like clocks set slightly wrong, causing the stored quantum signal to dissolve before it could be retrieved.
- Researchers fought back with XY4 radiofrequency pulse sequences that periodically flip the spin evolution, forcing drifting ions back into alignment — a technique analogous to resetting a room full of wandering clocks in coordinated waves.
- The result: nonclassical photon correlations survived 180 microseconds of storage, with a cross-correlation of 4.6 — well above the threshold that separates quantum behavior from anything classical physics can explain.
- A gap remains between the 180-microsecond quantum result and the 3-millisecond classical benchmark, pointing toward residual magnetic noise as the next frontier — one the team believes better field control could substantially close.
In Barcelona, researchers have quietly extended one of the most delicate boundaries in modern physics: how long a single photon's quantum identity can be held in stillness before the world erases it. By coaxing individual particles of light into a rare-earth crystal and applying carefully choreographed radiofrequency pulses to keep the quantum ensemble in step with itself, the team at ICFO preserved photon coherence for 180 microseconds — long enough, in principle, for light to cross 36 kilometers of fiber. The achievement does not yet bridge continents, but it demonstrates that the patient infrastructure of a quantum internet is not merely theoretical; it is, slowly, becoming material.
A research team at ICFO in Barcelona has set a new record for solid-state quantum memory, storing individual photons inside a rare-earth crystal for 180 microseconds while preserving the quantum correlations that make such storage meaningful. Published in Physical Review Letters, the result is equivalent to the time light takes to travel 36 kilometers through optical fiber — a modest distance, but a significant milestone for the architecture of a future quantum internet.
The challenge the team was solving is structural to quantum networking itself. Unlike classical networks, which freely copy and amplify signals, quantum networks must work with states that cannot be duplicated. Instead, they will operate by entangling short network segments and stitching them together — a process that requires each segment to hold its quantum information patiently while neighboring links complete their operations. Memory coherence time is not a performance metric; it is the load-bearing wall.
The memory was built from a praseodymium-doped crystal cooled to near absolute zero. Photon pairs were generated through spontaneous parametric down-conversion: one photon at 606 nanometers, matched to the crystal's absorption, and a telecom-wavelength partner at 1,436 nanometers that traveled freely through fiber. Detection of the telecom photon heralded the storage of its partner inside the crystal, where an atomic frequency comb — a precisely tuned pattern of absorption across many ions — absorbed the photon collectively. Control pulses then transferred this excitation into a long-lived spin state, effectively freezing the photon's information until a later pulse converted it back to light.
The central obstacle was dephasing. Each ion in the crystal sits in a slightly different microscopic environment, causing their quantum phases to drift apart over time until the ensemble can no longer reconstruct the original photon. The team countered this with XY4 radiofrequency pulse sequences — a form of dynamical decoupling that periodically reverses the spin evolution, nudging the drifting ions back into coherence. Classical tests showed the technique extended measurable echoes to beyond 3 milliseconds. The quantum test, more demanding, confirmed nonclassical correlations surviving to 180 microseconds, with a peak cross-correlation of 4.6 — comfortably beyond the classical bound.
The gap between the quantum and classical results points to remaining sources of noise, particularly residual magnetic fields interacting with the pulse sequences. The researchers believe that improved magnetic shielding and field control could push single-photon storage toward the millisecond range. For now, the experiment's most important contribution may be conceptual: it demonstrates that quantum information encoded in individual photons can survive spin rephasing in a solid-state system while a telecom-compatible partner travels through a real network — giving quantum repeaters, and the long-distance connections they would enable, a little more room to breathe.
A team of researchers in Barcelona has pushed a fundamental boundary in quantum memory, storing individual photons inside a crystal for 180 microseconds while keeping their quantum properties intact. The achievement, published in Physical Review Letters, marks the longest storage time yet achieved using this particular type of solid-state memory architecture—long enough for light to travel roughly 36 kilometers through optical fiber.
The work matters because a quantum internet, unlike today's networks, cannot simply copy and regenerate signals the way classical networks do. Quantum states—the qubits that would carry information across a quantum network—cannot be duplicated without destroying them. Instead, quantum networks will rely on a different strategy: establishing entanglement across short segments of the network, then connecting those segments together. For this to work, quantum information must wait in memory while other parts of the network catch up. A memory that loses its coherence before neighboring links finish their operations cannot synchronize the whole system. Storage time, in other words, is not a luxury. It is essential infrastructure.
The Barcelona team, working at ICFO, built their memory from a praseodymium-doped yttrium orthosilicate crystal cooled to about 3 kelvin. They generated pairs of photons using a technique called spontaneous parametric down-conversion. One photon had a wavelength of 606 nanometers, suited to the memory's optical transition. Its partner had a wavelength of 1,436 nanometers—the telecom range used in fiber-optic cables. When the telecom photon was detected, it acted as a herald, confirming that its partner was ready to be stored.
Inside the crystal, the researchers used what is called an atomic frequency comb. Rather than storing a photon in a single atom, they prepared groups of praseodymium ions with a carefully spaced pattern of absorption frequencies. The incoming photon was collectively absorbed across many ions at once. Control pulses then transferred that collective excitation into a long-lived spin state, effectively pausing the normal process by which the photon would be re-emitted. A later pulse converted the excitation back into light. This method also allows the retrieval time to be chosen while the photon is stored—a useful property for synchronizing different parts of a quantum network.
But holding the excitation in a spin state created a new problem. Each praseodymium ion sits in a slightly different microscopic environment inside the crystal. Those tiny differences cause the ions' quantum phases to evolve at different rates. Over time, the ensemble falls out of step. The collective excitation dephases, and the ions can no longer reconstruct the original photon efficiently. The researchers countered this using radiofrequency pulses applied to the crystal in a carefully phased sequence called XY4. The pulses flip the evolution of the spins, making individual phases converge again after another period of evolution. This technique, known as spin rephasing or dynamical decoupling, is similar in principle to resynchronizing a group of clocks that have begun drifting apart.
Tests with classical light showed how dramatically the rephasing pulses extended storage. Without effective rephasing, spin dephasing rapidly suppressed useful photon echoes. With the more advanced XY4 and XY16 sequences, measurable classical-light echoes persisted beyond 3 milliseconds—nearly two orders of magnitude longer than earlier spin-wave storage in the same material without spin rephasing. The harder test involved genuine single-photon-level quantum light. The researchers stored the 606-nanometer photons while retaining their relationship with the 1,436-nanometer heralding photons, then measured correlations between detections after retrieval. Nonclassical correlations remained detectable for storage times reaching 180 microseconds, with the highest measured cross-correlation reaching 4.6 with an uncertainty of 0.4, comfortably above the classical bound.
The 180-microsecond result is still far shorter than the storage times that large multisegment quantum networks may eventually require. The gap between the 180-microsecond quantum result and the roughly 3-millisecond classical result also reveals that spin coherence itself was not the only limitation. Quantum operation suffered from lower storage efficiency and increasing noise as the delay grew. Part of that noise appears connected to small residual magnetic fields interacting with the radiofrequency sequence. Better control of the magnetic environment could extend useful single-photon storage significantly. The researchers estimate that moderate controlled fields could push individual-ion spin coherence toward milliseconds, while specially engineered magnetic conditions could eventually support far longer storage.
For a quantum internet, the important advance is not simply keeping light trapped for another fraction of a millisecond. It is showing that quantum information encoded in individual photons can survive spin rephasing inside a solid-state memory while a telecom-compatible partner travels through a network. That gives quantum repeaters a little more time to wait, and long-distance quantum connections a little more room to grow.
Notable Quotes
Our scheme establishes praseodymium-doped quantum memories as a major candidate for the scalable implementation of quantum networks— Hugues de Riedmatten, senior researcher at ICFO
Want the full story?
Read the original at The Brighter Side of News ↗