⚛️ quantum physics

Quantum Fisher information and quadrature squeezing in Janus superpositions of squeezed vacua

This paper demonstrates that while Janus superpositions of squeezed vacua do not surpass single squeezed vacua in principal second-moment squeezing at fixed mean photon number, they can offer significant metrological advantages in specific laboratory quadratures and quadratic-generator sensing when compared within a fixed two-state span or at fixed observed squeezing levels.

Original authors: Arash Azizi

Published 2026-03-18

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Arash Azizi

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 Picture: The "Janus" State

Imagine you are trying to measure something incredibly tiny, like the distance between two atoms or a slight shift in gravity. To do this, you need a "probe"—a beam of light or a particle—that is as quiet and stable as possible.

In the world of quantum physics, the standard "quiet" tool is a Squeezed Vacuum. Think of this like a balloon that you have squeezed in one direction. It gets very thin and quiet in that direction (low noise), but it puffs out and gets noisy in the other direction. This is the gold standard for precision measurements.

Now, imagine a new tool called a Janus State. Named after the two-faced Roman god, this is a "superposition" of two different squeezed balloons mixed together. It's like taking two different types of noise-canceling headphones and blending their signals. The author, Arash Azizi, asks: Does mixing two squeezed states together create a "super-probe" that is even better than the best single squeezed state?

The Core Discovery: It Depends on How You Compare Them

The paper's main conclusion is a bit like a sports debate: "Who is the better athlete?" The answer depends entirely on the rules of the game. The author found that Janus states can win, lose, or tie depending on which "benchmark" (rulebook) you use.

Here are the three main ways the paper compares them:

1. The "Fair Energy" Contest (The No-Go)

The Rule: We compare the probes based on how much "fuel" (energy/photons) they use. If both probes use the exact same amount of energy, who is quieter?
The Result: The Single Squeezed Vacuum wins.
The Analogy: Imagine two runners. One is a professional sprinter (the single squeezed vacuum). The other is a runner trying to combine the styles of two different sprinters (the Janus state). If you give them both the exact same amount of food (energy), the professional sprinter is still faster. The "mixing" of the Janus state doesn't magically create more speed; in fact, at the fundamental level of "principal squeezing," the single squeezed vacuum is mathematically the best you can get for a given amount of energy.
Takeaway: You cannot beat the standard squeezed vacuum if you are strictly limited by energy.

2. The "Fixed Lab Bench" Contest (The Local Win)

The Rule: Instead of comparing against the theoretical best, we compare the Janus state against the specific two ingredients used to make it. We ask: "Can mixing these two specific squeezed states create a result that is quieter in a specific direction than either of the original states?"
The Result: Yes! The Janus state wins.
The Analogy: Imagine you have two muddy puddles. One is muddy in the North, the other in the East. If you mix them carefully (using a specific "interference" recipe), you might create a spot in the middle that is surprisingly clean, even though both original puddles were dirty.
Takeaway: By mixing two specific states, you can cancel out noise in a specific direction better than either of the original states could do alone. This is a "local" victory, not a global one.

3. The "Fixed Squeezing Level" Contest (The Operational Win)

The Rule: In real-world labs, scientists often measure how much they have "squeezed" the noise (e.g., "We reduced noise by 10 decibels"). The question is: If two probes show the same amount of noise reduction on a meter, which one is actually better at sensing?
The Result: The Janus state wins big time.
The Analogy: Imagine two cars that both claim to get 50 miles per gallon on the dashboard display.

Car A (Squeezed Vacuum): Actually gets 50 mpg.
Car B (Janus State): Also shows 50 mpg, but it's doing something sneaky. It's using a different engine trick (higher-order fluctuations) that makes it incredibly efficient at climbing hills (sensing quadratic changes), even though the dashboard looks the same.
Takeaway: If you look only at the "noise meter," the Janus state looks the same as the standard one. But if you look under the hood at how it handles complex changes, the Janus state is secretly much more powerful. It can be 10 times more sensitive to certain types of measurements while showing the same noise level.

The "Secret Sauce": Higher-Order Fluctuations

Why does the Janus state win in the third contest?
Think of noise not just as a simple "hiss" (like static on a radio), but as a complex pattern of waves.

The Standard Squeezed Vacuum is great at smoothing out the basic "hiss" (second-order noise).
The Janus State is a "non-Gaussian" state. This means it has a more complex, jagged, or "spiky" internal structure.

When you mix the two squeezed states, you create a pattern where the "spikes" cancel out in a way that the standard vacuum can't. This allows the Janus state to hide its true power. It looks quiet on the surface (same squeezing level), but inside, it has a secret superpower that makes it incredibly sensitive to specific, complex changes in the environment.

Summary for the General Audience

The Goal: Build the ultimate sensor for measuring tiny things.
The Old Tool: A "Squeezed Vacuum" (a balloon squeezed in one direction). It's the best you can do if you are limited by energy.
The New Tool: A "Janus State" (a mix of two squeezed balloons).
The Verdict:
- If you compare them by energy, the Old Tool wins. The new tool can't break the laws of physics to be more efficient.
- If you compare them by mixing specific ingredients, the New Tool can create a "quiet spot" that is better than its parts.
- If you compare them by what a lab meter actually reads, the New Tool is a hidden champion. It can be vastly more sensitive to complex changes while looking just as "quiet" as the Old Tool.

The Bottom Line: The paper teaches us that "better" in quantum physics isn't a single number. It depends entirely on how you ask the question. The Janus state proves that by mixing quantum states, we can create "hidden" advantages that standard tools miss, provided we know how to look for them.

1. Problem Statement

The paper addresses a fundamental question in quantum metrology: Can non-Gaussian states, specifically coherent superpositions of squeezed vacua (Janus states), offer a genuine metrological advantage over the standard single-mode squeezed vacuum (SMSV)?

While SMSVs are the canonical resource for precision sensing, they are Gaussian states. Non-Gaussian states are often hypothesized to offer superior performance, but the conditions under which they outperform Gaussian references are not always clear. The core challenge lies in defining a "fair" comparison. Previous studies on Janus states focused on photon statistics (e.g., $g^{(2)}$ ), but this paper investigates whether Janus interference can enhance the Quantum Fisher Information (QFI)—the ultimate bound on estimation precision—under different resource constraints (benchmarks).

2. Methodology

The author employs an exact analytic treatment of the Janus state, defined as a coherent superposition of two single-mode squeezed vacua:
$|\psi\rangle = \chi |\xi\rangle + \eta |\zeta\rangle$
where $|\xi\rangle$ and $|\zeta\rangle$ are squeezed vacua with potentially different squeezing strengths ( $r, s$ ) and phases ( $\theta, \phi$ ).

The methodology involves three main analytical steps:

Second-Moment Analysis: Deriving closed-form expressions for laboratory-axis quadrature variances ( $\Delta Q, \Delta P$ ), the covariance matrix, and the principal squeezing (minimum variance achievable via homodyne rotation).
Span-Constrained Optimization: Treating the Janus state as a vector in a two-dimensional subspace spanned by $|\xi\rangle$ and $|\zeta\rangle$ . The author formulates the minimization of a fixed-axis variance as a generalized eigenvalue problem ( $M_Q \mathbf{c} = \lambda S \mathbf{c}$ ), where $M_Q$ is the operator matrix and $S$ is the Gram (overlap) matrix.
QFI Evaluation: Calculating the QFI for two distinct sensing scenarios:
- Number-generated phase estimation: $G = \hat{n} = a^\dagger a$ .
- Quadratic-generator sensing: $G \propto a^2$ or $a^{\dagger 2}$ (sensitivity to squeezing strength or axis rotation).
  The analysis compares the QFI of Janus states against the SMSV under two specific benchmarks:
- Fixed Mean Photon Number ( $\bar{n}$ ): A global resource constraint.
- Fixed Measured Squeezing ( $u$ ): An operational constraint based on the observed minimum quadrature noise (in dB).

3. Key Contributions and Results

A. Principal Squeezing vs. Fixed-Axis Squeezing

Principal Squeezing (Global Benchmark): The paper proves a "no-go" theorem for principal second-moment squeezing. Under a fixed mean photon number ( $\bar{n}$ ), the single-mode squeezed vacuum remains the optimal state for minimizing the principal quadrature variance. No non-Gaussian Janus superposition can beat the SMSV in this specific metric. The SMSV saturates the Robertson-Schrödinger uncertainty bound at fixed energy.
Fixed-Axis Squeezing (Span-Constrained): Within a prescribed two-state span (fixed constituents, varying coefficients), Janus interference can reduce the variance of a specific laboratory quadrature (e.g., $Q$ ) below that of either constituent squeezed vacuum. This reduction is achieved through destructive interference (subtractive superposition), though it typically increases noise in the conjugate quadrature or the mean photon number.

B. Quantum Fisher Information (QFI) Enhancements

The results regarding QFI are highly dependent on the chosen benchmark:

Fixed Mean Photon Number ( $\bar{n}$ ):
- Phase Estimation: A Janus state can outperform the SMSV in number-generated phase estimation ( $F_Q(\phi)$ ) at fixed $\bar{n}$ . This advantage arises not from improved second-moment squeezing, but from higher-order fluctuation structure. Specifically, Janus interference can suppress the mean photon number contribution while enhancing the factorial moment $N_2$ , leading to a larger $g^{(2)}(0)$ and thus higher QFI.
- Mechanism: The advantage is driven by the ability to cancel lower-order terms (e.g., the $|2\rangle$ Fock component) while retaining higher-order components, effectively reshaping the photon statistics to maximize variance in the number operator.
Fixed Measured Squeezing (Operational Benchmark):
- Quadratic-Generator Sensing: When comparing probes with the same observed minimum quadrature noise (fixed $u$ ), Janus states can substantially outperform the SMSV (by more than an order of magnitude in the examples provided).
- Reasoning: Fixing the measured squeezing fixes the second moments ( $N_1, |M_2|$ ) but leaves the fourth-order connected moments ( $M_4 - M_2^2$ ) and higher-order factorial moments ( $N_2$ ) free to vary. Janus interference can amplify these higher-order terms without changing the observed squeezing level, thereby boosting the QFI for quadratic generators.

C. Analytical Framework

The paper provides a complete set of closed-form analytic expressions for:

Overlap integrals and cross-moments ( $\langle \zeta | a^{2k} | \xi \rangle$ ).
Generalized eigenvalues for minimizing fixed-axis variance.
QFI expressions for both linear (phase) and quadratic generators in terms of the superposition coefficients and squeezing parameters.

4. Significance and Conclusion

Conceptual Clarity: The paper resolves a confusion in the literature by distinguishing between different notions of "advantage." It demonstrates that metrological advantage is benchmark-dependent.

If the resource is energy (fixed $\bar{n}$ ), the SMSV is optimal for principal squeezing, but Janus states can win in phase estimation via higher-order statistics.
If the resource is observed noise (fixed $u$ ), Janus states can vastly outperform SMSVs in quadratic sensing.

Practical Implications:

Experimental Design: The results suggest that in experiments where squeezing is reported in dB (fixed $u$ ), non-Gaussian superpositions (Janus states) offer a significant pathway to enhanced sensitivity for parameters encoded in quadratic generators, without requiring higher energy.
Non-Gaussianity as a Resource: The work establishes Janus states as a tractable, analytically solvable platform for studying how non-Gaussian interference reshapes quantum fluctuations beyond the Gaussian limit. It highlights that while non-Gaussian states may not beat the Gaussian limit in all metrics simultaneously, they offer unique advantages in specific operational regimes.

Summary: The paper concludes that Janus superpositions do not universally "win" over squeezed vacua. Instead, their utility is controlled by the comparison metric: they cannot beat the SMSV in principal second-moment squeezing at fixed energy, but they can simultaneously beat their constituents in fixed-axis noise and beat the SMSV in phase QFI (at fixed energy) or quadratic QFI (at fixed measured squeezing) by leveraging higher-order interference effects.