⚛️ quantum physics

Broadband and long-duration optical memory in Yb:YSO

This paper demonstrates an optimized atomic frequency comb optical memory in Yb:YSO that achieves a 250 MHz bandwidth and up to 125 $\mu$ s storage time with high efficiency, utilizing a novel pumping scheme and laser setup to pave the way for future spin-wave storage and large multimode quantum networks.

Original authors: T. Sanchez Mejia, L. Nicolas, A. Gelmini Rodriguez, P. Goldner, M. Afzelius

Published 2026-02-02

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: T. Sanchez Mejia, L. Nicolas, A. Gelmini Rodriguez, P. Goldner, M. Afzelius

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 you are trying to build a super-fast, super-efficient library for light. In the world of quantum computing, this library is called an optical quantum memory. Its job is to catch a flash of light (a photon), hold onto it for a moment, and then release it perfectly intact so it can be used later.

The paper you provided describes a major upgrade to this library, built inside a special crystal doped with a rare element called Ytterbium. Here is how they did it, explained simply:

1. The Problem: The "Traffic Jam" of Light

Usually, when you try to store a lot of information in light, you face a trade-off. You can either store a lot of information quickly (high bandwidth), or you can store it for a long time. Doing both at the same time is like trying to park a million cars in a tiny garage while keeping them there for a week without them running out of gas.

Previous attempts with other crystals were like small, narrow garages. They could hold cars for a long time, but only a few at once. Or, they could hold many cars, but they had to let them go almost immediately.

2. The Solution: A Giant, Smart Parking Garage

The researchers used a crystal containing Ytterbium-171. Think of this crystal as a massive, multi-level parking garage with a very specific, clever design.

The "Teeth" of the Comb: To store information, they use a technique called an Atomic Frequency Comb (AFC). Imagine a comb where the "teeth" are tiny, perfectly spaced slots. Light enters these slots. The more teeth you have, the more information you can store at once.
The Challenge: To make a comb with thousands of teeth, you need to "burn" (create) these slots very precisely. If you try to do this one by one, it takes too long, and the light forgets what it was supposed to be.
The Innovation: The team invented a new way to "burn" the comb. Instead of painting one tooth at a time, they used a mathematical trick (a frequency-domain method) to paint the entire comb in a single, rapid burst. It's like using a stencil to paint a whole fence instantly, rather than painting each slat one by one. This allowed them to create a comb with tens of thousands of teeth across a huge range of frequencies.

3. The "Class Cleaning" Trick

Inside the crystal, the atoms are a bit messy. Some are in the right state to catch the light, but many are in the wrong state, blocking the way.

The researchers developed a "Class Cleaning" technique. Imagine a bouncer at a club who only lets people with a specific VIP pass in. They used a series of laser pulses to "kick out" all the atoms that didn't have the right pass and force them into a single, empty waiting room.

The Result: They managed to clear the floor so that 80% of the atoms were ready to catch the light. This made the "garage" much deeper and more efficient.

4. The Results: Big and Fast

By combining the "stencil" method for the comb and the "bouncer" method for cleaning the atoms, they achieved two impressive things simultaneously:

Huge Capacity (Bandwidth): They created a memory that can handle a bandwidth of 250 MHz. To put that in perspective, previous similar crystals were limited to about 10 MHz. They made the "garage" 25 times wider.
Long Duration: They held onto the light for up to 125 microseconds. While that sounds short (a fraction of a second), in the world of light, it is an eternity. It is the longest time anyone has managed to hold light in a crystal with this much capacity.

The Efficiency:

When they stored the light for a very short time, they got it back 20% of the time.
When they held it for the maximum time (125 microseconds), they still got 5% back.
This is a huge improvement over previous attempts, which struggled to get any signal back when trying to store so much data for so long.

5. The "Swiss Army Knife" Laser

To pull this off, they needed a laser system that could change frequencies instantly and precisely. They built a setup using just one laser and a single modulator (a device that tweaks the light), controlled by a computer.

Think of this like a single musical instrument that can instantly switch between playing a violin, a trumpet, and a drum, all controlled by a single sheet of music. This makes the whole system much simpler and more reliable than previous setups that required multiple lasers.

Summary

In short, the team built a super-efficient, high-capacity memory for light. They used a rare-earth crystal, cleaned it out to make room for the data, and used a new mathematical trick to organize the data into a massive, precise comb. They proved that you can store a lot of information in light for a long time, which is a crucial step toward building the future "quantum internet."

Technical Summary: Broadband and Long-Duration Optical Memory in 171Yb3+:Y2SiO5

Problem Statement
Optical quantum memories based on rare-earth (RE) ion-doped crystals are critical for long-distance quantum networks due to their long coherence times and multimode storage capabilities. To maximize temporal multimode capacity, a memory must simultaneously possess a large bandwidth and a long optical storage time. While Kramers ions like 171Yb3+ offer the potential for large bandwidths (due to GHz-scale hyperfine splittings) and long coherence times (via Zero First-Order Zeeman effects), previous demonstrations in 171Yb3+:Y2SiO5 were limited to 100 MHz bandwidth and 25 µs storage. A primary bottleneck is the "class cleaning" (CC) or "single class selection" (SCS) optical pumping scheme required to isolate specific transitions within the inhomogeneously broadened spectrum. Existing schemes for 171Yb3+ were limited to ~125 MHz, and achieving broadband pumping without sacrificing coherence or efficiency remains a significant challenge. Additionally, generating the complex waveforms required for broadband Atomic Frequency Comb (AFC) preparation (involving tens of thousands of teeth) is computationally intensive and experimentally difficult.

Methodology
The authors address these challenges through a combination of theoretical modeling, novel waveform synthesis, and experimental implementation:

Optimized Optical Pumping Scheme: The authors propose a new CC and spin polarisation (SP) scheme for 171Yb3+:Y2SiO5. By selecting a specific $\Lambda$ -system involving the $|4g\rangle \leftrightarrow |1e\rangle$ and $|1g\rangle \leftrightarrow |1e\rangle$ transitions (polarized along the $D_2$ and $b$ axes, respectively), they utilize strong optical dipole moments. They identify four specific transition frequencies to pump ions into the $|4g\rangle$ state, creating a flat, highly absorbing anti-hole. Numerical simulations using a simple optical pumping simulator predict a maximum bandwidth of 288 MHz, limited only by the excited state hyperfine splitting.
Efficient Waveform Synthesis: To generate the AFC preparation pulses (square-shaped combs) over a 250 MHz bandwidth, the authors introduce a "frequency-domain method" for synthesizing the pulse waveform. Unlike previous "circular permutation" methods that scale linearly with the number of teeth ( $O(N)$ ) in computation time, this new method constructs the spectral phase analytically using a stationary phase approximation. This reduces computation time to constant $O(1)$ (relative to $N$ ), allowing for rapid optimization of the AFC finesse.
Experimental Setup: The experiment utilizes a single frequency-locked laser and a single electro-optic modulator (EOM) driven by a complex microwave (MW) and radio-frequency (RF) circuit. This setup uses local oscillators (LOs) and IQ mixers to generate sidebands across a 10 GHz bandwidth, enabling precise phase and amplitude control for the CC, SP, and AFC pulses. The memory crystal is a 15.8 mm long 171Yb3+:Y2SiO5 sample with a reduced doping concentration of 2 ppm (down from 5 ppm in previous work) to extend optical coherence times. A quadruple-pass configuration is used to increase the effective optical depth.

Key Results

Spin Polarisation Efficiency: The experimental CC/SP scheme successfully polarized 80% of the total ion population into the $|4g\rangle$ state. This increased the optical depth of the relevant AFC transition by a factor of 3.3, reaching a peak optical depth of 3.2 over a 250 MHz bandwidth.
Broadband AFC Preparation: The new frequency-domain synthesis method enabled the creation of high-finesse AFCs with 250 MHz bandwidth. The resulting combs exhibited square-shaped teeth with minimal distortion.
Short-Delay Efficiency: At short storage times (1 µs), the AFC echo efficiency reached 20.5% at 150 MHz bandwidth and 18.4% at 350 MHz. These values are in excellent agreement with theoretical predictions (21.4%) based on the measured optical depth and background absorption.
Long-Duration Storage: The system demonstrated AFC echoes with storage times up to 125 µs. At this delay, the efficiency was 5%. The effective AFC coherence time ( $T_2^{AFC}$ ) was measured to be $(348 \pm 15)$ µs. This represents the longest coherence time achieved in any RE crystal to date, surpassing previous records in non-Kramers ions (e.g., Pr3+ and Eu3+), despite those systems having significantly smaller bandwidths (<10 MHz).
Stability: The efficiency remained constant across the entire 250 MHz bandwidth for storage times up to 125 µs.

Significance and Claims
The paper claims that these results represent a significant step toward realizing high-capacity quantum memories. By combining a 250 MHz bandwidth with a 125 µs storage time, the system achieves a temporal multimode capacity that is orders of magnitude higher than previous broadband demonstrations. The authors emphasize that the 288 MHz theoretical bandwidth limit is set by the hyperfine structure of 171Yb3+, suggesting that the demonstrated 250 MHz is close to the material's fundamental limit for this specific scheme.

The work establishes that 171Yb3+:Y2SiO5 is a viable candidate for future spin-wave AFC experiments, which could potentially extend storage times to the millisecond regime while maintaining the large bandwidth. The authors note that their specific optical pumping scheme and the new waveform synthesis method are designed specifically to facilitate these future spin-wave storage experiments. The technical advancements, including the single-laser, frequency-agile setup and the efficient computational method for waveform generation, are presented as practical tools for developing transportable and rack-mounted quantum memory systems.