🔬 optics

Robustness and optimization of N00N-state interferometry

This paper establishes a comprehensive theoretical framework for folded Franson interferometry using partially entangled N00N states, demonstrating that while perfect fringe visibility can be recovered by compensating loss asymmetry with input imbalance, the Fisher information peaks at a distinct operating point, thereby defining the critical loss and entanglement thresholds required to maintain a genuine quantum advantage over optimized single-photon strategies.

Original authors: Romain Dalidet, Anthony Martin, Louis Bellando, Mathieu Bellec, Nicolas Fabre, Sébastien Tanzilli, Laurent Labonté

Published 2026-03-16

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Romain Dalidet, Anthony Martin, Louis Bellando, Mathieu Bellec, Nicolas Fabre, Sébastien Tanzilli, Laurent Labonté

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

The Big Idea: Measuring the Unmeasurable with Light

Imagine you are trying to measure the thickness of a single hair, or the distance to a star, with extreme precision. Scientists use interferometers for this. Think of an interferometer like a race track for light. You split a beam of light into two paths (like two runners on two lanes), send them around a loop, and then bring them back together.

If one runner gets slightly delayed (by hitting a hair or a gravitational wave), the two beams interfere when they meet, creating a pattern of light and dark stripes called fringes. By looking at how much the stripes shift, you can calculate the delay with incredible accuracy.

The Problem:
In the real world, things aren't perfect. Light gets absorbed by dust, mirrors aren't 100% reflective, and fibers have tiny cracks. This is called loss. When light is lost, the signal gets weaker, and your measurement gets fuzzy.

The "Magic" Solution (N00N States):
To beat the limits of normal light, scientists use special "entangled" bundles of photons called N00N states.

Analogy: Imagine normal light as a group of 100 people walking randomly down a hallway. If 10 get lost, you still have 90.
N00N State: Imagine a group of 100 people who are magically linked. They are either all in the left hallway or all in the right hallway, but never mixed. If even one person gets lost, the whole group's "quantum magic" breaks, and the measurement fails.

This paper asks: How do we keep using these fragile, super-sensitive "magic groups" when the hallway is full of obstacles (loss)?

The Two Different Goals: The "Look" vs. The "Score"

The authors discovered that there are two different ways to judge how well the experiment is working, and they don't always agree.

1. Fringe Visibility (The "Look")

This is how clear the light-and-dark stripes look.

The Analogy: Imagine you are trying to see a pattern on a foggy window. If the fog is uneven (asymmetric loss), the pattern looks washed out.
The Fix: The paper shows that if you adjust the "balance" of your magic light group before you start, you can perfectly cancel out the fog. You can make the stripes look perfectly crisp and clear (100% visibility), even if half the light is getting lost.
The Catch: Just because the stripes look perfect doesn't mean you have all the information you need.

2. Fisher Information (The "Score")

This is the actual amount of useful data you get to make a measurement.

The Analogy: Imagine you are trying to guess the weight of a package by looking at a scale.
- Visibility is how clearly you can read the numbers on the scale.
- Fisher Information is how many times you get to look at the scale before the battery dies.
The Conflict: Even if you fix the "fog" to make the numbers look crystal clear (high visibility), you might have lost so much light that the "battery" (the total number of photons) is running low. You have a clear picture, but you don't have enough data points to be sure.
The Result: The setting that makes the stripes look the best is not the same setting that gives you the most accurate measurement.

The Key Findings

1. You Can "Tune" Your Way Out of Trouble

The authors found a "knob" you can turn. If your experiment has uneven losses (one path is dirtier than the other), you don't have to throw it away. You can change the initial mix of your entangled light (the "imbalance") to compensate.

Simple term: If the left path is leaky, you send more light down the left path initially to balance it out. This restores the perfect "look" of the interference.

2. The "Sweet Spot" is Different for Everyone

For the "Look" (Visibility): You can always find a perfect balance to make the stripes clear, no matter how much loss there is.
For the "Score" (Sensitivity): There is a limit. If the loss is too high, no amount of tuning can save you. The "sweet spot" for getting the best measurement is usually at a lower loss level than the "sweet spot" for just making the stripes look good.

3. When is it Worth It? (The Quantum Advantage)

The paper calculates exactly when using these fancy "magic groups" (N00N states) is better than just using normal single photons.

The Verdict: For small groups of photons (like 2 or 3), you can handle a surprising amount of loss (up to about 64% loss for pairs of photons) and still beat the best possible normal measurement.
The Warning: As you try to use more photons to get even better precision, the system becomes incredibly fragile. The "loss budget" shrinks. You need almost perfect conditions to use big groups of entangled photons.

The Takeaway for Everyday Life

Think of this like taking a photo in a dark room:

Visibility is like adjusting the focus and contrast so the image looks sharp on your screen. You can do this even if the room is dark.
Fisher Information is like the actual detail in the photo. If the room is too dark, even if the image is sharp, it will be grainy and full of noise because you didn't capture enough light.

The paper's lesson: Don't just chase a "sharp image" (high visibility). If you want the most accurate measurement (high sensitivity), you have to accept that you might need to change your strategy, and sometimes, no matter how much you tune the settings, the darkness (loss) will eventually win. However, for small, manageable amounts of darkness, we now know exactly how to tune our "magic light" to get the best possible result.

1. Problem Statement

Quantum-enhanced interferometry using N00N states is theoretically capable of surpassing the Standard Quantum Limit (SQL) and achieving Heisenberg-limited scaling. However, in realistic experimental settings, performance is severely degraded by a complex interplay of imperfections, primarily:

Asymmetric losses: Different attenuation rates in the two interferometer arms.
State imbalance: Non-maximal entanglement in the input state (often due to unbalanced beam splitters).
Photon number ( $N$ ): The fragility of N00N states increases with $N$ .

A critical, often overlooked issue is the distinction between fringe visibility (coherence) and metrological sensitivity (Fisher Information). High visibility does not guarantee high sensitivity. While visibility can often be restored by compensating for losses with input imbalance, this compensation does not necessarily maximize the information available for phase estimation. The paper addresses the need to understand these trade-offs to determine the true operating regimes where a genuine quantum advantage exists over optimized single-photon strategies.

2. Methodology

The authors develop a comprehensive analytical framework based on a folded Franson interferometer (Mach-Zehnder-like configuration) fed with partially entangled N00N states.

Model Setup:
- Input State: A path-encoded $N$ -photon state $|\psi_{in}\rangle = [\sqrt{\alpha}(\hat{a}^\dagger)^N + \sqrt{1-\alpha}(\hat{b}^\dagger)^N]|00\rangle$ , where $\alpha$ controls the degree of entanglement ( $\alpha=0.5$ is maximal).
- Imperfections: One arm accumulates a phase shift $\phi$ and suffers relative loss $p$ (modeled by a fictitious beam splitter coupling to a vacuum mode).
- Output: The state is recombined on a balanced beam splitter, and $N$ -photon coincidence counts are measured.
Analytical Derivation:
- The authors derive exact closed-form expressions for the joint detection probabilities $P_{ij}$ .
- From these probabilities, they calculate two key figures of merit:
  1. Fringe Visibility ( $V$ ): A measure of interference contrast.
  2. Normalized Fisher Information ( $F_N$ ): A measure of phase estimation sensitivity, normalized by photon number to allow comparison with the SQL ( $F_N=1$ ).
Optimization Strategy:
- The study analytically optimizes $\alpha$ (entanglement) and $p$ (loss) to maximize $V$ and $F_N$ independently.
- It benchmarks multi-photon strategies against single-photon strategies under identical loss conditions to define "Quantum Advantage."

3. Key Contributions and Results

A. Divergence Between Visibility and Sensitivity

The paper demonstrates a fundamental decoupling between restoring coherence and maximizing information:

Visibility Restoration: Perfect fringe visibility ( $V=1$ ) can always be recovered in the presence of asymmetric losses by tuning the input entanglement parameter $\alpha$ . The optimal condition is $\alpha_V = \frac{(1-p)^N}{1+(1-p)^N}$ . This converts relative losses into a global attenuation factor, leaving the interference contrast intact.
Sensitivity Degradation: Unlike visibility, the Fisher Information ( $F_N$ ) cannot be fully restored by tuning $\alpha$ . The Fisher Information depends on both the slope of the fringe and the absolute detection probability. Since losses reduce the total photon count (signal attenuation), the information is irreversibly lost.
Different Optima: The value of $\alpha$ that maximizes visibility ( $\alpha_V$ ) is distinct from the value that maximizes Fisher Information ( $\alpha_{opt}$ ). Specifically, $\alpha_V(p, N) = \alpha_{opt}(p, 2N)$ . The operating point for maximum contrast is not the operating point for maximum sensitivity.

B. Optimization of Fisher Information

Loss Dependence: Fisher Information decreases monotonically with increasing loss $p$ . There is no non-trivial optimum for loss; the best performance is always in the lossless limit ( $p=0$ ).
Entanglement Dependence: For a fixed loss $p$ , there is an optimal entanglement $\alpha_{opt} = \frac{(1-p)^{N/2}}{1+(1-p)^{N/2}}$ that maximizes sensitivity. As loss increases, the optimal state becomes increasingly unbalanced (less entangled) to compensate for the attenuation.
Scaling: In the ideal lossless case, $F_N = N$ (Heisenberg scaling). However, for any $p > 0$ , the sensitivity degrades rapidly as $N$ increases, eventually falling below the SQL for large $N$ .

C. Quantum Superiority vs. Quantum Advantage

The authors define and map two distinct regimes:

Quantum Superiority: $F_N > 1$ $F_{N} > 1$ (Beating the SQL).
- Requires $N > 1$ .
- The maximum tolerable loss decreases as $N$ increases. While larger $N$ relaxes the requirement for perfect entanglement, it tightens the loss budget.
Quantum Advantage: $F_N > F_1$ $F_{N} > F_{1}$ (Outperforming the best single-photon strategy under the same loss conditions).
- This is a more practical metric. Even if the SQL is not beaten, entangled probes can still be better than single photons.
- Key Result for $N=2$ : The paper derives a concrete threshold for two-photon interference. A genuine quantum advantage survives up to a relative loss of $p \approx 0.64$ (64%). Beyond this point, single-photon strategies become superior even with optimized entanglement.

D. Operational Maps

The study provides analytical boundaries in the $(\alpha, p)$ parameter space:

Fig. 5 & 6: These figures map the regions of sub-SQL sensitivity and practical quantum advantage. They illustrate that as $N$ increases, the "safe" operating window for losses shrinks significantly, highlighting the fragility of high- $N$ N00N states.

4. Significance

Theoretical Framework: The paper establishes a unified theoretical description that bridges the gap between idealized quantum metrology bounds and realistic experimental constraints. It clarifies that "restoring visibility" is not synonymous with "restoring sensitivity."
Experimental Guidance: By identifying the specific trade-offs between loss, entanglement, and photon number, the work provides a roadmap for experimentalists. It suggests that for high-loss environments, one should not necessarily aim for maximal entanglement ( $\alpha=0.5$ ) but rather tune the input imbalance to the specific loss regime to maximize information extraction.
Benchmarking: The derivation of the $p \approx 0.64$ threshold for $N=2$ provides a concrete, experimentally testable benchmark for demonstrating quantum advantage in current photonic setups, validating the experimental results reported in the authors' companion work [1].
General Principle: The findings underscore a broader principle in quantum sensing: optimizing interference contrast (coherence) is insufficient for optimizing metrological gain, as the latter is fundamentally limited by signal attenuation (photon loss).

In summary, this work moves beyond the asymptotic scaling arguments of N00N states to provide a practical, exact analysis of how to operate quantum interferometers in lossy environments, defining the precise limits where quantum advantages persist and how to tune experimental parameters to achieve them.