Memory-assisted squeezed light velocimetry under realistic loss and incoherent noise

This paper proposes a memory-assisted velocimetry scheme using a two-memory Mach-Zehnder interferometer with squeezed vacuum that, despite realistic loss and incoherent noise, achieves a few-percent sensitivity improvement over coherent homodyne detection for near-term parameters.

The Gist

Imagine you are trying to measure how fast a car is driving, but you can't see the car directly. Instead, you have to listen to the sound of its engine. If the engine is a standard, noisy one (like a regular laser), it's hard to hear the tiny changes in pitch that tell you the speed. But what if you could tune the engine to be "quieter" in a specific way, so those tiny speed changes stand out clearly? That is the basic idea behind this paper, but instead of a car engine, they are using light.

Here is a simple breakdown of what the scientists are doing:

The Setup: A Quantum Race Track

The researchers built a "race track" for light called a Mach-Zehnder interferometer. Think of it as a fork in the road where a beam of light splits into two paths:

1. The Reference Path: One path stays still. It acts like a stationary stopwatch.
2. The Moving Path: The other path goes into a "memory box" (a quantum memory) that is physically moving.

When light travels through a memory box that is moving, the motion changes the light's "phase" (imagine it as the timing or rhythm of the light wave). The faster the box moves, the bigger the change in rhythm. By comparing the rhythm of the moving light to the stationary light when they meet back up, the scientists can calculate the speed.

The Problem: Noise and Loss

In the real world, this is tricky for two main reasons:

• The "Static" (Noise): The memory boxes aren't perfect. They add their own static noise, like a radio picking up static between stations.
• The "Dimming" (Loss): The longer you keep the light in the memory box to get a better speed reading, the more the light fades away (gets dimmer). If it gets too dim, you can't measure it accurately.

Usually, scientists use a standard, bright laser beam for this. But lasers have a natural "fuzziness" (called shot noise) that limits how precise the measurement can be.

The Solution: "Squeezed" Light

To beat the fuzziness, the researchers tried using squeezed light.

• The Analogy: Imagine a balloon. A normal balloon is round and bouncy in all directions. Squeezed light is like taking that balloon and squeezing it tight on one side. It becomes very thin and flat in one direction (making it very quiet and precise in that specific measurement) but bulges out on the other side.
• By "squeezing" the light, they reduce the noise in the specific direction they need to measure speed, making the signal much clearer than a standard laser.

The Big Question

The paper asks: Does this "squeezed" trick still work when you have to store the light in a memory box that adds noise and makes the light dim?

In a perfect, theoretical world, squeezed light is always better. But in the messy real world, the memory box might ruin the advantage.

What They Found

The scientists created a detailed mathematical model to test this. Here are their main conclusions:

1. It Still Works (But Not by Much): Even with the noise and dimming from the memory boxes, the squeezed light still provides a better speed measurement than a standard laser. However, the improvement is modest—roughly 5% to 10% better under realistic conditions.
2. The "Noise Floor" Isn't the Enemy: You might think the static noise from the memory box is the biggest problem. Surprisingly, the paper says that even if the memory box is a bit noisy (up to a certain level), it doesn't kill the advantage. The squeezed light is robust enough to handle it.
3. The Real Bottlenecks: The things that actually stop the improvement are loss (the light getting too dim) and instability (the timing of the experiment drifting). If the light fades too much or the setup wobbles, the squeezed light can't help as much.
4. The Sweet Spot: There is a "Goldilocks" time for how long to store the light.
   • If you store it too briefly, the speed signal is too weak to detect.
   • If you store it too long, the light fades away too much.
   • The scientists found the perfect middle ground where the speed signal is strong enough, but the light hasn't faded too much.

The Takeaway

This paper proves that using "squeezed" quantum light to measure speed is a viable idea, even when using imperfect, noisy memory boxes. It won't give you a super-powerful speedometer overnight (the gain is small), but it proves that the quantum advantage survives the messy reality of the lab.

The main lesson for future experiments is: Don't just worry about the noise in the memory box. To get the best results, you need to focus on keeping the light bright (reducing loss) and keeping the setup steady (reducing vibration and timing errors). If you can do that, the "squeezed" trick will give you a measurable edge over standard lasers.

Technical Summary

Technical Summary: Memory-assisted squeezed light velocimetry under realistic loss and incoherent noise

Problem Statement
Interferometric phase sensing is a mature setting for quantum enhancement, where injecting squeezed vacuum into a Mach–Zehnder interferometer can reduce phase noise below the coherent-state limit. However, this advantage is fragile, surviving only when attenuation and technical fluctuations are sufficiently small. While optical memories offer a mechanism to map external perturbations (such as velocity) into an accumulated phase of retrieved light, they introduce storage-time-dependent loss and incoherent noise. The central question addressed in this work is whether memory-assisted velocimetry can retain a metrological advantage over coherent-state benchmarks when the full sensing cycle—including propagation, storage/retrieval, and readout—is subjected to realistic loss, detector inefficiency, phase noise, and unconditional memory noise floors.

Methodology
The authors propose and model a velocity sensor based on a two-memory Mach–Zehnder interferometer. The setup involves:

• Input: A bright coherent probe state $|\alpha\rangle$ and a single-mode squeezed vacuum state $S(r)|0\rangle$ injected into the unused port.
• Architecture: One arm contains a stationary reference memory ($Q_{ML}$), while the second arm contains a moving memory ($Q_{MU}$) with average velocity $v$ during the storage interval $\tau$.
• Signal Encoding: The motion of the upper memory induces a differential phase shift $\phi_v = -k_p v \tau$, where $k_p$ is the probe wave number.
• Noise Modeling: The system is described by a Gaussian model incorporating:
  • Gaussian memory lifetime with efficiency $\eta_{mem}(\tau) = \eta_0 \exp[-(\tau/\tau_{mem})^2]$.
  • Unconditional memory noise floors ($p_n$) for both memories.
  • External optical loss, detector inefficiency, squeezing-angle jitter, electronic noise, and residual phase noise.
• Readout: Balanced homodyne detection measures the dark output port.
• Analysis: The authors derive the classical Fisher information and the Cramér-Rao bound for velocity estimation. They compare the performance of the squeezed-input scheme against a coherent-state homodyne benchmark under equal optical resources (same photon number $N$ and number of runs $M$).

Key Contributions

1. Derivation of Sensitivity Bounds: The paper derives the shot-normalized velocity sensitivity $\Delta v \sqrt{M}$, explicitly accounting for the trade-off between signal phase accumulation (which increases with $\tau$) and retrieval efficiency/noise (which degrade with $\tau$).
2. Optimization of Storage Time: An analytic expression for the optimum storage time is derived using the Lambert-W function, showing how the optimal $\tau$ shifts based on the detected squeezing contrast and the noise floor.
3. Threshold Analysis: The authors determine the minimum total transmission ($\eta_{min}$) required to achieve a target fractional quantum gain over the coherent benchmark, given specific noise parameters.
4. Noise Budget Evaluation: The study quantifies the impact of unconditional memory noise ($p_n$) on the viability of the squeezed scheme, distinguishing it from transmission losses and phase instability.

Results

• Quantum Gain Window: The squeezed scheme improves sensitivity over coherent homodyne only within an operating window defined primarily by total transmission and phase stability.
• Impact of Memory Noise: For representative near-term parameters, unconditional memory noise floors up to approximately $10^{-1}$ photons per trial do not, by themselves, eliminate the quantum advantage. The improvement remains at the few-percent level even with realistic noise floors ($p_n \sim 10^{-3}$).
• Limiting Factors: In the parameter regime studied, total transmission and phase stability are found to be more restrictive constraints on the quantum gain than the unconditional memory noise floor itself.
• Optimal Performance: For a reference working point (10 dB injected squeezing, zero-storage transmission $\eta(0) \approx 0.316$), the optimized improvement over coherent homodyne is modest, approximately 5%. The optimal storage time for the squeezed scheme is slightly shorter than that for the coherent scheme ($\approx 0.95\tau_{mem}$ vs. $1.01\tau_{mem}$).
• Saturation: At fixed transmission, the gain saturates once loss and technical noise dominate the detected variance, regardless of further increases in injected squeezing.

Significance and Claims
The paper claims that memory-assisted velocimetry with squeezed light is a viable, albeit modest, step toward memory-enhanced quantum metrology. The authors emphasize that while the improvement over coherent-state detection is small (few to ten percent) under current realistic conditions, the advantage persists despite significant optical loss, detector inefficiency, and memory noise.

The primary significance lies in identifying the practical bottlenecks: the unconditional memory noise levels reported in current experiments are not the primary barrier. Instead, the preservation of the squeezed quadrature through the full write–store–read cycle and the maintenance of interferometric phase stability are the critical requirements for realizing measurable advantages. The work provides a concrete framework for evaluating when quantum memories act as active sensing elements rather than passive storage devices, applicable to platforms ranging from warm vapors to cold atoms and rare-earth-ion-doped crystals.