⚛️ quantum physics

Detecting strongly non-Gaussian entanglement

This paper introduces a robust, multicopy continuous-variable entanglement detection method using passive linear optics and particle-number measurements that efficiently certifies strongly non-Gaussian entanglement, including NOON states, which are often missed by standard approaches.

Original authors: Serge Deside, Tobias Haas, Nicolas J. Cerf

Published 2026-09-09

📖 4 min read🧠 Deep dive

CC BY 4.0

Original authors: Serge Deside, Tobias Haas, Nicolas J. Cerf

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

In the quantum world, the most powerful resource is a strange connection called entanglement, where two particles remain linked so that the state of one instantly influences the other, no matter the distance between them. For decades, scientists have been able to identify this connection in simple, smooth systems known as Gaussian states, which behave somewhat like ripples on a calm lake. However, the most advanced quantum technologies, such as ultra-precise sensors and future quantum computers, rely on a much wilder, more complex type of entanglement found in "strongly non-Gaussian" states. These are like jagged, irregular waves that defy the smooth patterns of the past. Detecting this rugged form of entanglement has been a major stumbling block because the standard tools used to measure the smooth waves fail completely when faced with these complex shapes, leaving researchers unable to verify if their most promising quantum resources are actually working.

A team of researchers from the Université libre de Bruxelles has now developed a new method to solve this problem, offering a way to certify this difficult type of entanglement with surprising ease. Instead of trying to measure the complex state directly, which often requires impossible amounts of data, the scientists devised a technique that uses a few identical copies of the quantum state simultaneously. They pass these copies through simple, passive optical circuits—essentially networks of mirrors and beam splitters that do not add energy to the system—and then count the number of particles, such as photons or atoms, that emerge. By analyzing the statistical patterns of these particle counts, they can calculate specific mathematical values that reveal whether the state is entangled. This approach is a significant departure from previous methods, which often required measuring high-order correlations that are incredibly difficult to access in a laboratory setting.

The researchers demonstrated that their new criteria are exceptionally good at spotting the entanglement in "NOON states," a special class of quantum states where particles are in a superposition of being entirely in one place or entirely in another. These states are crucial for metrology, the science of measurement, because they can theoretically provide the highest possible precision. Previous methods, including those based on entropy or standard measurements of particle positions and momenta, completely failed to detect entanglement in these states, even when the entanglement was obvious. The new method, however, successfully identified the connection in NOON states with any number of particles, from just a few up to ten, a range where older techniques hit a wall. The researchers found that their approach works not only for pure, ideal states but also for mixed states that have been disturbed by noise, showing that the method is robust enough for real-world conditions.

To ensure their findings were not just theoretical, the team ran extensive computer simulations to test how their method would hold up under realistic experimental constraints. They modeled the inevitable losses that occur when particles are absorbed or scattered, the noise that arises when experimental conditions fluctuate slightly between copies, and the statistical uncertainty that comes from having a limited number of measurements. Even when they simulated a scenario where the system lost up to twenty percent of its particles, the method continued to detect entanglement. In cases involving the most difficult states, they found that with a reasonable number of experimental runs, specifically around one hundred thousand, they could still confirm the presence of entanglement even in the presence of significant noise. This suggests that the technique is not just a mathematical curiosity but a practical tool that could be implemented with current technology, such as the particle-counting detectors already used in photonic quantum computing and ultracold atom experiments.

The significance of this work lies in its ability to bridge the gap between theory and practice for the most advanced quantum states. By showing that entanglement can be certified using only a few copies of a state, simple optical components, and particle counting, the researchers have removed a major barrier to developing quantum sensors and computers. Their method does not require the complex, active manipulation of light or matter that often introduces errors; instead, it relies on the fundamental properties of the particles themselves. The study confirms that the entanglement in these highly non-Gaussian states is not only real but detectable with high confidence, even when the experimental setup is imperfect. This opens the door for scientists to build and verify the next generation of quantum devices, ensuring that the powerful resources they need for ultra-precise measurements and advanced computing are truly present and functional.

Technical Summary: Detecting Strongly Non-Gaussian Entanglement

Problem Statement
Efficiently certifying entanglement in continuous-variable (CV) quantum systems is a critical bottleneck for advancing quantum information processing, photonic computing, and metrology. While numerous criteria exist for Gaussian states (e.g., Simon's criterion based on second moments) and weakly non-Gaussian states (e.g., entropic criteria), these methods fail in the regime of strongly non-Gaussian entanglement. This regime includes states far beyond the convex hull of Gaussian states, such as NOON states and superpositions of Fock states. Existing approaches face two primary limitations in this domain:

Insensitivity: Entropic criteria and non-local quadrature-based methods (EPR-type) often fail to detect entanglement in highly non-Gaussian states or superpositions of coherent states.
Experimental Infeasibility: Criteria based on higher-order moments of mode operators (e.g., Shchukin-Vogel hierarchies) theoretically capture these correlations but require infeasible numbers of measurement settings or higher-order moments that grow exponentially with the state's complexity (e.g., order $2N$ for NOON states of photon number $N$ ).

Methodology
The authors extend recently introduced entanglement criteria based on moments of the partially transposed state (PT moments) to continuous-variable systems. The core methodology involves three components:

Theoretical Framework (PT Moments):
The paper generalizes two infinite hierarchies of criteria derived from the Stieltjes moment problem and Descartes' rule of signs. These criteria test the positivity of the partial transpose $\tilde{\rho}$ using moments $p_n = \text{Tr}\{\tilde{\rho}^n\}$ .
- The authors focus on third-order criteria involving the purity $p_2$ and the third moment $p_3$ .
- They derive a quadratic criterion ( $p_3 - p_2^2 \geq 0$ ) and a linear criterion ( $p_3 - (3p_2 - 1)/2 \geq 0$ ).
- Crucially, they identify an optimal third-order criterion that minimizes $p_3$ with respect to $p_2$ over separable states. This optimal criterion adapts to the state's purity, being linear for $p_2 \geq 1/2$ and following a more complex bound for $p_2 < 1/2$ .
- For pure states, these third-order criteria are shown to be necessary and sufficient for entanglement detection.
Experimental Readout (Multicopy Method):
To access these PT moments without requiring infeasible measurement settings, the authors employ a multicopy method.
- Instead of randomized measurements (which scale poorly with Hilbert space dimension in CV systems), the scheme utilizes $n$ replicas of the state $\rho^{\otimes n}$ .
- The $n$ -th PT moment is measured via a multicopy observable involving cyclic shift operators on the Fock basis.
- These shift operators are diagonalized using a Discrete Fourier Transform (DFT) implemented via passive linear optics (beam splitters and phase shifters).
- The final readout requires only particle-number-resolving detectors on the output modes.
- Specifically, $p_2$ requires a single 50:50 beam splitter per subsystem, while $p_3$ requires a specific sequence of three beam splitters and phase shifts per subsystem.
Robustness Analysis:
The authors model realistic experimental imperfections, including:
- Losses: Modeled as a lossy quantum channel acting on the input state.
- Noisy Copies: Variations in state preparation parameters (e.g., phase, loss) across different replicas.
- Finite Statistics: Statistical fluctuations due to a finite number of experimental runs.

Key Results
The paper benchmarks the proposed PT-moment criteria against standard methods (Shchukin-Vogel, EPR-type, entropic) across various state families:

Gaussian States: The third-order criteria are shown to be necessary but not sufficient for Gaussian separability (consistent with Simon's criterion being the tightest bound for $1\times1$ Gaussian states). However, the optimal third-order criterion provides a tight bound for Gaussian entanglement in terms of $p_2$ and $p_3$ .
Coherent Superpositions (Cat States): For dephased entangled Schrödinger cat states, the linear $p_3$ -PPT criterion outperforms all entropic criteria, detecting entanglement for smaller separations between coherent components ( $|\alpha| = |\beta| \gtrsim 1$ ) where entropic methods fail.
Strongly Non-Gaussian States (NOON States): This is the primary success case.
- For NOON states $|\psi\rangle = \alpha|N,0\rangle + \beta|0,N\rangle$ , the third-order criteria detect entanglement for all $N$ and all non-trivial superpositions ( $\alpha, \beta \neq 0$ ).
- Standard methods (EPR, entropic) fail completely for these states.
- Mode-operator-based criteria would require moments up to order $2N$ , which is experimentally prohibitive for large $N$ . The proposed method requires only third-order moments regardless of $N$ .
Robustness:
- Losses: The method detects NOON entanglement up to $\sim 20\%$ loss for high photon numbers ( $N=10$ ). Entanglement certification fails only when losses exceed 50%.
- Noisy Copies: The criteria remain robust against variations in phase and loss parameters between replicas, tolerating up to $\sim 30\%$ variation in loss parameters.
- Finite Statistics: Numerical simulations indicate that $\sim 10^3$ samples are sufficient to witness entanglement for NOON states with $N=1$ at 10% loss, and $\sim 10^5$ samples suffice for $N=10$ at 15% loss.

Significance and Claims
The paper claims to provide a strong and efficient entanglement criterion specifically tailored for the strongly non-Gaussian regime, a domain where previous methods are either insensitive or experimentally infeasible.

Efficiency: By utilizing the multicopy method with passive linear optics and particle-number measurements, the scheme avoids the exponential scaling of measurement settings associated with high-order mode-operator moments.
Generality: The criteria are dimension-independent and applicable to both pure and mixed states, covering the entire class of NOON states and other non-Gaussian families (e.g., photon-added/subtracted states, Hermite-Gauss states).
Practicality: The method is demonstrated to be robust against realistic experimental constraints (losses, noise, finite statistics), making it a viable tool for certifying entanglement in current photonic and ultracold atom platforms.
Limitations: The authors acknowledge that the primary experimental burden is the need for interferometric stability across multiple state copies. They also note that generating multiple copies of probabilistic non-Gaussian states remains a challenge, though they argue that probabilistic preparation is effectively equivalent to reduced statistics, which their method can accommodate.

In conclusion, the work establishes PT-moment-based criteria as a superior alternative for detecting strong non-Gaussian entanglement, bridging the gap between theoretical necessity and experimental feasibility in continuous-variable quantum systems.

More like this