🔬 optics

Erbium-Doped Fibre Quantum Memory for Chip-Integrated Quantum-Dot Single Photons at 980 nm

This paper presents the first experimental demonstration of a coherent hybrid light-matter interface that successfully stores and retrieves deterministic single photons from a chip-integrated InAsP/InP nanowire quantum dot in an erbium-doped fiber quantum memory at 980 nm without requiring spectral tuning.

Original authors: Nasser Gohari Kamel, Arsalan Mansourzadeh, Ujjwal Gautam, Vinaya Kumar Kavatamane, Ashutosh Singh, Edith Yeung, David B. Northeast, Paul Barclay, Philip J. Poole, Dan Dalacu, Daniel Oblak

Published 2026-07-16

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Nasser Gohari Kamel, Arsalan Mansourzadeh, Ujjwal Gautam, Vinaya Kumar Kavatamane, Ashutosh Singh, Edith Yeung, David B. Northeast, Paul Barclay, Philip J. Poole, Dan Dalacu, Daniel Oblak

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the internet of the future, but instead of sending emails and cat videos, it's sending the most delicate secrets in the universe: quantum information. To build this "quantum internet," scientists need to do something tricky. They need to catch a flying particle of light (a photon) that carries a message, pause it in mid-air to store that message safely, and then let it fly again without losing any of its special quantum magic. The problem is that light is fast and forgetful; it zips through fiber optic cables at the speed of light and doesn't like to stop. To make it stop, we need a "quantum memory"—a special box that can catch the light, hold it, and release it on command.

But here's the catch: the light coming from our best quantum light sources (tiny artificial atoms called quantum dots) speaks a different "color" language than the boxes we usually use to store it. It's like trying to plug a USB-C cable into a USB-A port; they just don't fit. For a long time, scientists had to try to stretch or squeeze the light or the memory to make them match, which is messy and unstable. This paper is about a clever new way to make two very different things fit together perfectly, creating a smooth bridge between a light source and a memory bank, all while keeping the light's precious quantum secrets safe.

The Great Quantum Matchmaking

In this study, a team of scientists from Canada pulled off a rare feat of quantum matchmaking. They successfully connected two very different systems: a tiny, chip-based light source that spits out single photons, and a long glass fiber filled with special atoms that can act as a memory bank. The goal was to show that these two systems could talk to each other without needing any awkward adjustments or "spectral tuning" (which is like trying to force a square peg into a round hole).

The Light Source: The Tiny Flashlight
On one side of the experiment, they used a "quantum dot." Think of this as a microscopic, chip-integrated flashlight made from a semiconductor material called InAsP. When you zap it with a laser, it doesn't just flash a beam; it fires off single particles of light, one by one, on demand. This is crucial because for a quantum internet to work, you can't rely on guessing when a photon will arrive; you need to be able to say, "Send one now!" This specific quantum dot is designed to emit light at a wavelength of 980 nanometers (a specific shade of near-infrared light).

The Memory: The Glass Sponge
On the other side, they used a 10-meter-long piece of glass fiber (about the length of a small car) that had been "doped" with Erbium ions. Imagine this fiber as a giant, super-cooled sponge made of glass. Inside this sponge are trillions of tiny atoms (Erbium ions) that can absorb light and hold onto its energy for a while before spitting it back out. Usually, scientists use these fibers for a different color of light (telecom wavelengths), but this team decided to try using them for the 980 nm light from their quantum dot.

The Big Discovery: They Fit Without Forcing
The team's main finding is that these two systems naturally speak the same language. The 980 nm light from the quantum dot perfectly matches the "absorption bandwidth" of the Erbium-doped fiber. In the past, scientists often had to tweak the light source or the memory to make them align, which is hard to keep stable. Here, the researchers found that the quantum dot's emission and the fiber's memory capabilities just happened to line up perfectly at cryogenic temperatures (super cold, around 10 millikelvin, which is just a hair above absolute zero).

How They Stored the Light
To prove the memory worked, they used a technique called the "Atomic Frequency Comb" (AFC). You can imagine this as tuning the glass sponge so that its atoms are arranged in a specific, comb-like pattern. When a photon hits this comb, it gets absorbed by the "teeth" of the comb. Because of the way the atoms are arranged, the photon is forced to wait for a specific amount of time before it is automatically re-emitted.

The team demonstrated this in two ways:

Storing "Weak" Light: They sent in 59 very faint pulses of light (weak coherent pulses) and successfully stored and retrieved them. They managed to hold these pulses for up to 100 nanoseconds. The system was so good at handling multiple pulses at once that it could theoretically store up to 144 different time slots, though they tested 59.
Storing Single Photons: This was the big test. They took the actual single photons from the quantum dot and stored them in the fiber. Without changing the color of the light or the fiber, they successfully caught the single photon, held it, and let it go again.

The Results and Limits
The experiment showed that the memory worked, but it wasn't perfect yet.

Efficiency: The system was able to store and recall the light, but only about 1% of the photons made it through the whole process. The paper notes this is mostly because of losses in connecting the fiber to the rest of the equipment (splicing and connectors), not because the memory itself failed.
Quality: When they checked the "fidelity" (how well the memory kept the quantum information), they found it was high enough to prove it was a true quantum process, beating what a classical computer could do.
Speed: The memory could hold the light for different durations, ranging from 5 to 100 nanoseconds.

What They Didn't Do (and Why It Matters)
The paper explicitly rules out the idea that they needed to "tune" the quantum dot. Often, to make a quantum dot match a memory, scientists have to apply magnetic fields or stretch the material to shift its color. In this case, the researchers showed that the quantum dot's natural emission was already a perfect match for the fiber's 980 nm transition. This is a huge deal because it means the system is more stable and easier to build for real-world use.

They also clarified that while the 980 nm transition is commonly used in lasers to pump energy into systems, this is the first time its properties as a quantum memory have been explored and proven to work with a quantum dot.

The Bottom Line
This paper suggests that Erbium-doped fiber is a very promising candidate for building the "hard drives" of a future quantum internet. By showing that a chip-based quantum dot can talk directly to a fiber-optic memory without needing complex adjustments, the researchers have taken a significant step toward making quantum networks that are practical, scalable, and ready to be integrated into the existing fiber-optic cables that already crisscross our world. While the efficiency needs improvement (currently at 1%), the fundamental bridge between the light source and the memory is solid, proving that the two technologies can coexist and cooperate.

Technical Summary: Erbium-Doped Fibre Quantum Memory for Chip-Integrated Quantum-Dot Single Photons at 980 nm

Problem Statement
The realization of long-distance quantum communication and a scalable quantum internet requires coherent hybrid light–matter interfaces that connect deterministic quantum light emitters with quantum memory (QM) systems. While significant progress has been made in developing both quantum emitters (such as quantum dots) and photonic quantum memories (based on rare-earth ion-doped solids, REIDS), experimental demonstrations of hybrid interfaces between these specific systems remain scarce. A primary obstacle is achieving spectral compatibility between the absorption wavelength of the QM and the emission spectrum of the quantum emitter. Although spectral tuning of emitters is possible, it often lacks long-term stability and offers limited tuning ranges, hindering the practical implementation of scalable networks. Furthermore, while REIDS-based storage has been demonstrated using probabilistic sources (e.g., spontaneous parametric down-conversion), the storage and retrieval of photons from deterministic semiconductor quantum dots (QDs) remains underexplored.

Methodology
The authors present an experimental realization of a hybrid interface between a chip-integrated InAsP/InP nanowire quantum dot (QD) and a solid-state QM based on Erbium-doped fiber (EDF). The experimental setup operates at cryogenic temperatures (~10 mK) and utilizes a 10-meter-long EDF doped with 200 ppm natural abundance Er³⁺ ions.

Spectroscopic Characterization: The team first characterized the 980 nm optical transition ( $4I_{15/2} \leftrightarrow 4I_{11/2}$ ) in the EDF. They measured the inhomogeneous broadening, spin population lifetime ( $T_1^S$ ), and optical coherence time ( $T_2^O$ ) using persistent spectral hole burning and Hahn echo techniques under varying magnetic fields.
Quantum Memory Implementation: An Atomic Frequency Comb (AFC) protocol was implemented on the 980 nm transition. By modulating the amplitude and phase of optical pump pulses, the authors created an 8 GHz bandwidth AFC with squarish-shaped absorption teeth. This protocol allows for the storage of photons for a time $t_s = 1/\Delta$ , where $\Delta$ is the tooth spacing.
Quantum Light Source: The study utilized InAsP nanowire QDs integrated onto Si₃N₄ waveguides. These QDs were excited above-band using 532 nm picosecond pulses or 633 nm continuous-wave lasers. The emission of the negatively charged exciton ( $X^{1-}$ ) was spectrally filtered and coupled to the EDF.
Hybrid Interface Demonstration: The researchers aligned the fixed emission wavelength of the $X^{1-}$ state with the 8 GHz bandwidth of the AFC. They demonstrated the deterministic storage and retrieval of both weak coherent pulses and single photons from the QD without requiring spectral tuning of the QD emission.

Key Contributions and Results

First Spectroscopic Investigation of 980 nm EDF Transition: The paper reports the first detailed spectroscopic study of the 980 nm transition in EDF at cryogenic temperatures. Measurements revealed a substantial inhomogeneous broadening of 10 nm (FWHM) and a long spin population lifetime, with components of $T_1^S \approx 6.8$ s and $T_1^S \approx 6.4$ min. The optical coherence time ( $T_2^O$ ) was measured to be approximately 2.6 $\mu$ s at 0.06 T and 0.09 T magnetic fields.
Broadband Multimode QM: The authors successfully implemented an 8 GHz bandwidth AFC, enabling the storage and retrieval of 59 weak coherent temporal modes (laser probe pulses) with a storage time of 90 ns. The storage fidelity for time-bin qubits was assessed at 87.6%, exceeding the classical limit of 66.7%.
Deterministic Single-Photon Storage: The study demonstrates the first storage and retrieval of single photons from a chip-integrated InAsP/InP nanowire QD in an EDF-based QM. The $X^{1-}$ emission at ~980 nm was directly compatible with the EDF absorption bandwidth without spectral tuning.
Performance Metrics: For the single-photon storage, the system achieved a signal-to-noise ratio (SNR) of 1.92. The retrieved photons maintained their temporal profile, and the estimated output second-order correlation function $g^{(2)}_{out}(0)$ was 0.5547, remaining below the classical limit of 1, confirming the quantum nature of the storage process. The total memory efficiency was measured at 1%, limited primarily by splicing and connector losses (83% loss) rather than the intrinsic memory protocol.

Significance and Claims
The paper claims to present the first experimental realization of a coherent hybrid light-matter interface between a chip-integrated semiconductor QD and a solid-state QM based on EDF. The authors emphasize that this achievement is notable because it was accomplished without spectral tuning of the QD emission, demonstrating direct compatibility between the two systems.

The work highlights the potential of the 980 nm transition in EDF for quantum storage applications, a regime previously unexplored for coherence properties despite its common use in optical pumping for lasers. The large inhomogeneous broadening of the EDF is identified as a critical resource for bridging the interface between broadband quantum light sources and quantum memories. The authors conclude that while the current platform shows viability, future improvements in scalability and performance depend on the fabrication of isotopically purified rare-earth doped fibers to minimize decoherence and the continued development of tunable semiconductor QD growth. This research represents a significant step toward robust quantum memories compatible with hybrid quantum architectures for long-distance communication.

The Great Quantum Matchmaking

More like this