Entanglement Enhanced Sensing with Qubits affected by non-Markovian Dephasing

arXiv:2604.22368 · quant-ph · Submitted 2026-04-24 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: We're looking at the paper: "Entanglement Enhanced Sensing with Qubits affected by non-Markovian Dephasing".

Mira: Here's the gist of the paper. The gist: Entanglement can provide a substantial improvement for Ramsey spectroscopy subject to correlated noise, demonstrating that entanglement can enhance the sensitivity of sensors with entangled probes compared to those with separable probes under non-Markovian temporal noise Fundamental Limits and Noise Characterization The analysis establishes a fundamental limit for quantum sensing stating that the signal-to-noise ratio of a sensing experiment cannot exceed the signal-to-noise ratio experienced by the system sensing the signal This limit is derived from considering stationary, translationally invariant, classical, pure dephasing noise The estimation uncertainty for a time-independent frequency estimate best depends on the number of qubits N, the total sensing time T, and the noise affecting the system For uncorrelated spatial and temporal noise, this bound is given by a scaling of 1/√N in the long time limit.

Kai: Entanglement Enhancement in Ramsey Spectroscopy The work demonstrates that for non-Markovian temporal noise, entanglement can indeed substantially enhance the sensitivity of sensors with entangled probes compared to those with separable probes Specifically, for spin-squeezed states, an uncertainty scaling of 1/N3/4 is achievable provided the environment has a suitable noise spectrum.

Mira: The metrological gain r is defined as the ratio of the uncertainty of standard Ramsey spectroscopy (separable), ∆bestsep, with that of the entangled protocol, ∆bestent Performance Under Different Noise Spectra The estimation uncertainty is evaluated for various noise spectra including white, Gaussian, linear, and Ohmic spectra For spatially uncorrelated noise with a Gaussian spectrum SG(ω) = Ae− ω2/2σ2, the uncertainty approaches the fundamental limit of Eq.

Kai: (6), which is ∆best = pA/NT.

Paper summary: Kai: So, this paper is about entanglement improving sensing when the noise isn't just random between shots, but actually correlated across those shots.

Mira: Right, it tackles how decoherence messes with quantum sensors and whether entanglement can fight that kind of noise. The authors call this "Entanglement Enhanced Sensing with Qubits affected by non-Markovian Dephasing".

Lev: From an engineering standpoint, the core question is whether we can actually build something that sees a better signal than what we get from separate probes when the noise is correlated.

Kai: Exactly. The paper sets up a fundamental limit based on the idea that your sensitivity can't possibly beat what you see in terms of signal-to-noise ratio by the probe itself.

Mira: And they show that when you focus on Ramsey spectroscopy with probes suffering from pure classical dephasing, entanglement gives a better scaling of sensitivity with the number of probes than using separable states.

Lev: That's a big claim because we usually expect some kind of noise advantage to disappear once you introduce temporal correlations between shots.

Kai: Well, this specific work looks at scenarios where those temporal correlations exist and they find that for suitable spatial and temporal noise correlations, entangled probes really do scale better than separable states.

Mira: It seems like the main point is that entanglement provides a substantial improvement specifically when dealing with correlated noise in Ramsey spectroscopy.

Lev: What does this mean for the hardware we're building? Does it mean we can actually use these states to get a better measurement bound than standard quantum limits?

Kai: The results they show, focusing on spin-squeezed states, can achieve an uncertainty scaling of one over N to the power of three-quarters, which is one/N to the power of three-fourths <ref:2604.22368#pg1>.

Mira: That's a big jump from what we usually see in simpler cases, like the standard quantum limit scaling with one over square root of N <ref:2604.22368#pg1>.

Lev: But what about those other noise spectra they looked at, like white or Gaussian noise? Does that advantage still hold there?

Kai: For white and Gaussian noise, the metrological gain stays constant as N increases for those specific noise types. The scaling advantage is really more pronounced when you look at linear or Ohmic noise.

Mira: That's interesting because Ohmic and linear noises don't have that same hard limit on the uncertainty that some other noise spectra do.

Lev: If the gain stays N dependent for those noises, it suggests this isn't just a fluke for white noise; it points to a deeper way entanglement interacts with those specific environmental correlations.

Kai: They extend this to general spatial and temporal noise correlations, which they define by that term involving Q tau l minus l prime times P n minus n prime.

Mira: And they then derive an uncertainty expression based on that correlation structure, which incorporates those terms accounting for the shots being correlated.

Lev: So if I'm trying to run this on real hardware with many qubits, this equation is what tells me how much overhead we're really looking at due to these shot correlations.

Kai: The paper concludes that in the large N limit, the smallest improvement happens when there is a zero-frequency component of the noise spectrum that is non-zero.

Mira: But they find the largest improvement actually occurs for noise where that zero-frequency component is zero, which means gamma(tau) scales with tau squared.

Lev: So what does that imply about how long we can effectively use these entangled states before the correlation structure hurts us?

Kai: For a temporal noise scaling like tau to the power of one, they get an uncertainty scaling of one over N to the power of two-thirds <ref:2604.22368#pg1>.

Mira: And for a different temporal noise scaling, specifically tau squared, they find an uncertainty scaling that is one over N to the power of three-fourths <ref:2604.22368#pg1>.

Lev: That latter scaling is interesting because it matches the results you see from GHZ states in some other literature when you look at those specific noise conditions.

Kai: The overall message of this work on "Entanglement Enhanced Sensing with Qubits affected by non-Markovian Dephasing" is that entanglement improves measurement uncertainty in all the cases they considered.

Mira: It shows that for certain types of noise correlations, like the ones they explored here, we can get a real scaling advantage over separable states in Ramsey spectroscopy.

Lev: So what's the practical implication? Are we looking at a path to achieving better quantum sensors in noisy environments than just using standard single-qubit setups?

Kai: It suggests that if you can engineer noise correlation patterns, entanglement is a tool you can use to lower your measurement uncertainty below what's achievable with separable states.

Conclusion: Kai: So, we've been looking at how entanglement helps quantum sensors when the noise isn't just random between shots, but actually correlated across those shots for this paper "Entanglement Enhanced Sensing with Qubits affected by non-Markovian Dephasing."

Mira: Yeah, it’s about showing that for Ramsey spectroscopy under specific types of noise, using entangled probes gives you a better sensitivity scaling than using separate probes.

Lev: From what I see on the page, they're focusing on pure classical dephasing noise and looking at how different noise spectra—like white or Ohmic—affect the gain.

Kai: They show that for some correlations, like when the zero-frequency part of the noise spectrum is zero, you can actually get a scaling beyond what we normally expect from just using separable states.

Mira: That scaling improvement depends heavily on the temporal correlation function they define in equation nineteen and how it relates to those noise properties.

Lev: If we were to try this on real hardware, the required state preparation for spin-squeezed states is where things get tricky with fidelity issues and control overhead.

Kai: The authors specifically find that for certain noise behaviors, like tau squared dependence in the dephasing rate gamma of tau, you can achieve an uncertainty scaling of one over N to the power of three-fourths.

Mira: That specific scaling is what links up nicely with results seen in GHZ states under different noise conditions, which is pretty neat.

Lev: But we have to be careful; for linear or Ohmic noise, the improvement doesn't keep growing with N as much as it does for white or Gaussian noise.

Kai: It means the utility of entanglement isn't universal across every single type of decoherence we face in a lab setting.

Mira: So, this paper really tells us that while entanglement is powerful, you need to match it carefully to the exact nature of the noise you are dealing with.

Lev: It sets up a clear path for future work where we can tailor our error correction or state preparation specifically to the noise environment we're trying to measure.

Noah Kaufmann, Kasper H. Nielsen, Eva M. González-Ruiz, Anders S. Sørensen

Center for Hybrid Quantum Networks (Hy-Q) · Niels Bohr Institute

quant-ph

Submitted: 2026-04-24

Updated: 2026-10-19

Comments: 17 pages, 8 figures

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 78/100

The gist: The gist: Entanglement can provide a substantial improvement for Ramsey spectroscopy subject to correlated noise, demonstrating that entanglement can enhance the sensitivity of sensors with entangled

Key concepts

Fundamental Limit: This is the theoretical maximum signal-to-noise ratio achievable in any sensing experiment. It sets a baseline; no measurement technique, including quantum ones, can surpass this limit for a given system and noise conditions.
Spin-Squeezed States: These are specific types of entangled quantum states used as probes. They are prepared to reduce the inherent 'shot noise' (randomness) in a measurement, allowing the sensor to be more sensitive than classical or separable states under certain noisy conditions.
Non-Markovian Dephasing: This refers to noise where the system's future state depends on its entire past history, rather than just the immediate present. It describes complex environmental interactions that are not simple or memoryless, leading to noise correlations across multiple experimental shots.

Terminology

Entanglement — 3× in this paper · explained in 4 episodes

Summary

The gist: Entanglement can provide a substantial improvement for Ramsey spectroscopy subject to correlated noise, demonstrating that entanglement can enhance the sensitivity of sensors with entangled probes compared to those with separable probes under non-Markovian temporal noise

Fundamental Limits and Noise Characterization

The analysis establishes a fundamental limit for quantum sensing stating that the signal-to-noise ratio of a sensing experiment cannot exceed the signal-to-noise ratio experienced by the system sensing the signal This limit is derived from considering stationary, translationally invariant, classical, pure dephasing noise The estimation uncertainty for a time-independent frequency estimate best depends on the number of qubits N, the total sensing time T, and the noise affecting the system For uncorrelated spatial and temporal noise, this bound is given by a scaling of 1/√N in the long time limit.

Entanglement Enhancement in Ramsey Spectroscopy

The work demonstrates that for non-Markovian temporal noise, entanglement can indeed substantially enhance the sensitivity of sensors with entangled probes compared to those with separable probes Specifically, for spin-squeezed states, an uncertainty scaling of 1/N3/4 is achievable provided the environment has a suitable noise spectrum. The metrological gain r is defined as the ratio of the uncertainty of standard Ramsey spectroscopy (separable), ∆bestsep, with that of the entangled protocol, ∆bestent

Performance Under Different Noise Spectra

The estimation uncertainty is evaluated for various noise spectra including white, Gaussian, linear, and Ohmic spectra For spatially uncorrelated noise with a Gaussian spectrum SG(ω) = Ae− ω2/2σ2, the uncertainty approaches the fundamental limit of Eq. (6), which is ∆best = pA/NT. In contrast, for linear and Ohmic noise, no such limit exists, allowing for a substantial improvement of the estimation uncertainty with spin-squeezed states The metrological gain r saturates with increasing N for white and Gaussian noise, while it remains N dependent for Ohmic and linear noise.

Temporal and Spatial Correlations

The analysis extends to general spatial and temporal noise correlations, where the correlation is defined as ⟨ϕnlτϕn′l′τ⟩ = Qτ (l − l′)P(n − n′) The estimation uncertainty in this scenario is described by Eq. (19), which incorporates terms accounting for correlations among experimental shots The last term in Eq. (19) accounts for correlations among experimental shots.

Conclusion on Entanglement Utility

In the large-N limit, the smallest improvement occurs for correlated noise with S(0) ≠ 0, while the largest improvement occurs for noise with S(0) = 0, with γ(τ) ∝ γ2 The optimal scaling using spin-squeezed states depends on the short-time behavior of the dephasing γ(τ). For γ(τ) ∝ τ1, an uncertainty scaling as ∆best ∝ N−2/3 is achieved For γ(τ) ∝ τ2, the optimal scaling is ∆best ∝ N−3/4, which corresponds to the scaling reported for GHZ states in Refs. [18, 19]. The work concludes that entanglement improves the measurement uncertainty in all the cases considered

How it works

The enhancement is achieved by preparing qubits in a spin-squeezed state, which lowers the shot noise and can exceed the shot-noise limit The protocol involves preparation of a spin-squeezed state, followed by Ramsey pulses, free evolution for time τ, and measurement The uncertainty expression is given by Eq. (19) which includes terms related to the squeezing parameter κ and the noise correlations.

Key Findings

Entanglement provides a considerable gain in the sensitivity of quantum sensing compared to Ramsey spectroscopy with separable states
An uncertainty scaling beyond the standard quantum limit of 1/√N can be achieved for classical pure dephasing noise if the zero-frequency component of the noise spectrum is zero
For Gaussian and Ohmic temporal noise, entanglement improves the uncertainty for higher N compared to the separable case
The scaling advantage in Ohmic noise is r ∝ N1/4, whereas for linear noise it is r ∝ N1/6

Comparison with GHZ States

GHZ states improve sensitivity as the phase of a collective N-qubit state evolves N times faster compared to a single qubit state However, for white noise, the metrological gain r is 1 in Fig. 8b For linear noise, GHZ states perform worse than separable input states Spin-squeezed states have the advantage that the amount of entanglement is tunable.

Appendix A: Spin-squeezed States

The appendix extends the discussion to other classes of spin-squeezed states, including one-axis twisted (OAT) and two-axis twisted (TAT) states OATs have been realized experimentally in several hardware platforms through collective nonlinear interactions

Appendix B: Noise Properties

The appendix provides exact expressions for the dephasing behavior of the temporal noise spectra used in Sec. IV The autocorrelation functions of these spectra are given by Eq. (B3)

Appendix C: Time Dependence

The analysis formalizes the notion of taking the long-time limit and discusses short-time behavior for a single qubit across all noise spectra The long-time limit requires that T is much longer than the correlation time of the noise

Appendix D: GHZ States

GHZ states improve sensitivity as the phase of a collective N-qubit state evolves N times faster compared to a single qubit state The measurement operator for the l-th experiment is defined in Eq. (D1) The response of the measurement to signal changes is given by Eq. (D4)

Appendix D: GHZ States Derivation

The estimation uncertainty in Ramsey spectroscopy with GHZ states is given by Eq. (D7) In the absence of noise, this recovers the Heisenberg scaling ∆best = 1/N1/2T τ [8] The optimal interrogation time scales as τopt ∝ N−1/3 for ν = 1

Appendix D: GHZ States Protocol

The protocol for Ramsey spectroscopy with GHZ states consists of the following steps: preparation of an N-qubit GHZ state, free evolution of the system for a time τ, projection measurement onto the initial GHZ state, and repetition The total measurement operator is defined in Eq.

Improvements for AI systems

The AI system can perform high-precision parameter estimation in quantum sensing protocols like Ramsey spectroscopy under correlated noise by utilizing spin-squeezed states to achieve an uncertainty scaling of 1/N3/4, provided the environment has a suitable noise spectrum.
The AI system can distinguish between different types of noise correlations by analyzing the zero-frequency component of the noise spectrum, noting that for noise spectra with D(0, 0) ≠ 0, no sensing strategy exceeds a scaling of 1/√N in the long time limit, which allows for an asymptotic improvement over separable strategies.
The AI system can optimize experimental parameters (squeezing parameter κ and interrogation time τ) to maximize metrological gain by solving the trade-off between shot noise reduction and dephasing, as shown by the scaling results in Table I for different noise spectra.
For specific non-Markovian temporal noise spectra, such as Ohmic noise where γ(τ) ∝ τ2, the AI system can determine that an optimal interrogation time scales as τopt ∝ N −3/4, leading to a scaling of (∆best)2 ≈ N −3/4 for the estimation uncertainty.
The AI system can assess the metrological gain between different entangled states, finding that for noise with S(0) = 0 or G(0) = 0, squeezing improves the uncertainty only up to a constant factor compared to the separable case.