- The paper introduces squeezing-enhanced Hamiltonian engineering that achieves higher-than-quadratic QFI scaling at short times despite noise.
- The paper shows that non-Markovian reservoirs induce information backflow, temporarily restoring coherence and boosting estimation accuracy.
- The paper reveals that while Gaussian measurements perform near-optimally under moderate conditions, strong squeezing necessitates non-Gaussian strategies to fully capitalize on QFI.
Beating Noise in Frequency Estimation with Squeezing and Memory in Continuous-Variable Systems
Introduction
This work provides a systematic analysis of frequency estimation in noisy continuous-variable (CV) quantum systems, focusing on two central metrology-enhancement strategies: Hamiltonian engineering via squeezing and the exploitation of non-Markovian reservoir dynamics. Quantum metrology in CV systems frequently aims to surpass the shot-noise limit (SNL), leveraging resources such as squeezing and entanglement. However, noise and decoherence, especially in realistic open systems, degrade such quantum advantages. The research addresses this challenge by comprehensively investigating how internal system manipulation and external reservoir engineering can jointly overcome noise-induced losses in metrological precision.
Frequency estimation is carried out on a single-mode CV probe, with the performance quantified by the quantum Fisher information (QFI). The optimal precision, as per the quantum Cramér-Rao bound, is determined by the maximum attainable QFI over all possible POVMs. For Gaussian states, the QFI admits analytically tractable expressions in terms of first and second moments, providing an efficient route to performance benchmarking. Measurement strategies are considered within the general-dyne detection framework, which encompasses homodyne, heterodyne, and general Gaussian measurements.
Squeezing-Enhanced Hamiltonian Engineering
The study first considers the enhancement achievable through direct incorporation of a squeezing term into the system Hamiltonian. For a single-mode harmonic oscillator, the time-scaling of the QFI in the presence of a squeezing term, both under unitary and dissipative dynamics, is analytically explored. In the unitary regime, squeezing introduces a tunable nonlinearity into the parameter encoding, enabling higher-than-quadratic short-time QFI scaling. Under Markovian dissipative noise, the analysis demonstrates that squeezing robustly mitigates the decay in QFI, consistently yielding higher estimation precision relative to unsqueezed systems across all times until eventual exponential decay. The optimal operational regime is the short-interrogation-time window where decoherence is not yet dominant and the nonlinear enhancement is maximal.
A key technical result is that the phase of the squeezing parameter can be selected to ensure constructive cubic scaling in QFI, constituting a strictly stronger claim than traditional quadratic scaling from phase rotations alone. Numerical analysis confirms that squeezing counteracts environmental damping, maintaining a significant QFI gap in favor of the engineered Hamiltonian.
Non-Markovian Memory Effects
The work advances to scenarios where the probe system is coupled to structured, finite-memory reservoirs, modeled via quantum Brownian motion (QBM) with Lorentz-Drude spectral density. In the non-Markovian regime, characterized by a bath correlation time commensurate with the system timescales, there are substantial deviations from monotonic information loss. Transient negativity in the decoherence rates signals the breakdown of CP-divisibility and the presence of information backflow from the environment to the probe.
This backflow effect can temporarily restore quantum coherence and, consequently, estimation precision beyond that allowed under Markovian or even unitary evolution. The QFI exhibits pronounced non-monotonicity: for suitable parameter choices, it can exceed the unitary QFI during intermediate times. The relative size of this enhancement is contingent on probe amplitude and temperature: at small probe amplitudes and moderate temperatures, the non-Markovian memory effect is most pronounced.
Comparison and Competition of Mechanisms
A targeted comparison of squeezing-induced (Hamiltonian-engineered) and memory-induced (reservoir-engineered) enhancements reveals nontrivial competition. In regimes of moderate squeezing and small coherent amplitudes, non-Markovianity can exceed the precision gain from squeezing alone under Markovian noise. However, when squeezing or probe amplitude is increased, the Hamiltonian engineering rapidly dominates. When both mechanisms are present, non-Markovianity further protects the off-diagonal covariance correlations generated by squeezing, leading to amplified metrological advantages in specific regimes.
Achievability of QFI: Limitations of Gaussian Measurements
The attainability of the QFI is interrogated for Gaussian measurement strategies. In the absence of Hamiltonian squeezing, optimal general-dyne measurements can saturate the QFI at specific times, with homodyne or heterodyne detection being sufficient. When squeezing is present, especially at larger strengths, a persistent gap emerges between the dyne-optimal classical Fisher information (CFI) and the full QFI. This gap widens as squeezing increases, conclusively showing that Gaussian measurements become progressively suboptimal. In such regimes, non-Gaussian strategies (e.g., photon counting) are implicated as necessary for attaining quantum-limited precision.
Notably, even when non-Markovian noise is included, the general-dyne measurement retains near-optimality in a broad parameter window, but the temporal variability of optimal measurement settings (due to information backflow altering phase-space orientation) creates intervals where adaptive or non-Gaussian measurement bases would be advantageous.
Practical and Theoretical Implications
The analysis establishes two synergistic axes for quantum probing in noisy CV systems: (i) Hamiltonian engineering through accessible internal controls (e.g., optical squeezing), and (ii) structured environmental engineering to induce beneficial non-Markovian behavior. The findings prescribe that optimal quantum metrology protocols in open systems should jointly exploit these resources, tuning both system parameters and environmental structure. Experimentally, the results are directly relevant for quantum optics and optomechanics platforms, where both squeezing and reservoir spectral structure are tunable.
From a theoretical standpoint, the result that non-Markovian QFI can transiently exceed the unitary bound sharpens the understanding of quantum noise—not merely as an adversary but as a potential metrological resource. Furthermore, the identification of non-Gaussian measurement requirements for strong squeezing underscores the necessity of measurement innovation alongside state and reservoir engineering.
Conclusion
This study provides a detailed quantitative framework for analyzing and optimizing parameter estimation in noisy CV quantum systems. By unifying Hamiltonian engineering and environmental memory exploitation, it delineates the fundamental and practical limits of metrological precision in open systems. The results point toward hybrid protocols combining squeezing, non-Markovian bath engineering, and non-Gaussian measurement as the future architecture for quantum-enhanced sensing, with implications for a wide range of quantum technologies.
Essay based on "Beating noise in frequency estimation with squeezing and memory in continuous-variable systems" (2605.06263).