Quantum Cramér–Rao Bound in Quantum Metrology

• Quantum Cramér–Rao Bound is a theoretical limit on the variance of unbiased estimators in quantum systems, defined via the quantum Fisher information.
• It generalizes the classical Cramér–Rao bound by optimizing over quantum measurements, proving essential for applications in metrology, sensing, and spectroscopy.
• The bound informs optimal experimental protocols and measurement strategies, including projective, entangling, and LOCC-based techniques.

The quantum Cramér–Rao bound (QCRB) establishes the lowest achievable variance in unbiased parameter estimation on quantum systems, providing the mathematically rigorous precision limit for quantum metrology. Grounded in the quantum Fisher information (QFI), the QCRB is both foundational in quantum statistics and central to modern quantum technologies ranging from sensing and spectroscopy to fundamental studies of quantum geometry and information theory. The QCRB generalizes the classical Cramér–Rao bound by optimizing over all quantum measurements and explicitly incorporates the quantum structure of the estimation problem.

1. Fundamental Definition and Formulation

Given a finite- or infinite-dimensional Hilbert space, consider a family of quantum states $\rho(\theta)$, smooth in a real parameter $\theta$. For $M$ independent repetitions with any locally unbiased estimator $\hat\theta$, the quantum Cramér–Rao bound is

$$\operatorname{Var}(\hat\theta) \geq \frac{1}{M F_Q[\rho(\theta)]}$$

where $F_Q[\rho(\theta)]$ is the quantum Fisher information. The QFI is defined via the symmetric logarithmic derivative (SLD) $L(\theta)$, the Hermitian solution of

$$\frac{\partial \rho(\theta)}{\partial\theta} = \frac{1}{2}\left(L(\theta) \rho(\theta) + \rho(\theta) L(\theta)\right),$$

with

$$F_Q[\rho(\theta)] = \operatorname{Tr}\!\left[\rho(\theta) L(\theta)^2\right].$$

For pure states $$\rho(\theta) = |\psi(\theta)\rangle\langle\psi(\theta)|$$, this reduces to

$$F_Q = 4 \left(\langle \partial_\theta \psi | \partial_\theta \psi \rangle - |\langle \psi | \partial_\theta \psi \rangle|^2\right)$$

and, for unitary encoding $$|\psi(\theta)\rangle = e^{-i \theta H} |\psi_0\rangle$$, $F_Q=4\,\operatorname{Var}(H)_{|\psi_0\rangle}$ (Volkoff et al., 2022, Yu et al., 2020).

For a parameter vector $\boldsymbol\theta\in\mathbb{R}^d$, the QFI generalizes to the quantum Fisher information matrix (QFIM),

$$[F_Q(\boldsymbol\theta)]_{jk} = \frac{1}{2} \operatorname{Tr}\left[ \rho(\boldsymbol\theta)\, \{L_j, L_k\} \right],$$

with SLDs $L_j$ for each $\theta_j$. The multiparameter QCRB is

$$\mathrm{Cov}(\hat{\boldsymbol\theta}) \succeq [F_Q(\boldsymbol\theta)]^{-1}$$

[in the Loewner sense], so for any positive weight matrix $W\succ 0$: $$\operatorname{Tr}[W\,\mathrm{Cov}(\hat{\boldsymbol\theta})] \ge \operatorname{Tr}[W\,F_Q^{-1}]$$ (Nurdin, 18 Feb 2024, Goldberg et al., 2021).

2. Attainability and Measurement Strategies

Saturability Conditions: In single-parameter quantum estimation, the QCRB can always be saturated in the asymptotic limit by optimal POVMs—frequently available as projective measurements onto SLD eigenbases (Zhou et al., 2018). In the multiparameter scenario, attainability is subject to informational incompatibility: commutativity of the SLDs is necessary and sufficient for simultaneous attainability of the matrix QCRB for all parameters. When $[L_\mu, L_\nu]=0$ for all $\mu,\nu$, there exists a common eigenbasis for joint measurements and the SLD-CRB is attainable (Li et al., 2022).

When SLDs are non-commuting, as in generic multi-parameter estimation, the bound is unattainable via single measurements. Attainability can sometimes be restored on an extended Hilbert space or by collective measurements over many copies, or in specific subspaces determined by the rank structure of $\rho(\boldsymbol\theta)$ (Conlon et al., 1 Apr 2024). In the single-copy scenario for general mixed states, recent work gives necessary and sufficient conditions in terms of projected SLD commutativity and unitary solutions to nonlinear PDEs. Explicit algebraic and block structure criteria have been established for optimal projective measurements, illustrating when the multiparameter QCRB can or cannot be saturated (Nurdin, 2 May 2024, Nurdin, 18 Feb 2024).

LOCC and Entangling Measurements: For pure states, local operations and classical communication (LOCC) protocols exist that achieve the QCRB, even though local measurements alone can be sub-optimal except in special cases such as GHZ states (Zhou et al., 2018). For multiparameter estimation, LOEM strategies—preparing orthogonal pure states and applying global entangling measurements—can achieve saturability with Heisenberg scaling for all parameters simultaneously (Mi et al., 12 Sep 2025).

3. Quantum Fisher Information: Properties and Calculation

The QFI encapsulates the ultimate statistical distinguishability of neighboring quantum states under infinitesimal parameter shifts and serves as a Riemannian metric in quantum state space. For Gaussian states, explicit QFI formulas can be given in terms of first and second moments, and the corresponding QCRB can be saturated via homodyne measurement in the optimal quadrature mode (Pinel et al., 2010, Woodworth et al., 2022).

For statistical models where the quantum state rank is parameter-dependent, the standard QFI formula can become discontinuous and SLDs unbounded at rank-changing points, but the QCRB remains valid when using the continuous (Bures-metric) QFI and imposing local unbiasedness (Ye et al., 2021). Caution must be exercised in implementing the QFI and QCRB at such singularities.

4. Multiparameter Scenario: Geometry and Tradeoffs

In the multiparameter setting, the QFIM serves as a quantum geometric tensor on the parameter manifold. Geometrically, the real part of the quantum-geometric tensor (Fubini–Study/Bures metric) characterizes the attainable information, while the imaginary part (Berry curvature) encodes incompatibility via the commutators of the SLDs (Li et al., 2022). The quantumness metric $\gamma = \|i 2 J_Q^{-1} F\|_\infty$ quantifies the degree of incompatibility, interpolating between classical, fully compatible models ($\gamma=0$) and maximal incompatibility ($\gamma=1$), directly determining the gap between the matrix-inverse QCRB and the true attainable bound (e.g., the Holevo bound).

For pure-state two-parameter estimation, the most informative achievable QCRB has been characterized by a closed-form minimization over a single variable parameterizing the geometry, revealing when the Helstrom (SLD) bound is tight and when informational trade-offs fundamentally limit precision (Yung et al., 18 Nov 2025). This resolves the quantum multiparameter estimation problem for arbitrary pure-state probes.

5. Attainability in Open and Noisy Systems

In open quantum systems, the QCRB is generalized via vectorization in the Liouville–superoperator formalism. For Markovian quantum channels, the quantum Fisher kernel can be described in terms of the Lindblad generator, enabling analytical lower bounds on estimation variance even in the presence of dissipation and decoherence (Alipour et al., 2013). Techniques such as quantum noise cancellation and optimal quantum smoothing can saturate the QCRB in force estimation with continuous monitoring (Tsang et al., 2010).

6. Experimental Verifications and Applications

Experimental Verification: Recent work has demonstrated direct measurement and saturation of the QCRB in solid-state qubits (e.g., nitrogen-vacancy centers), confirming that experimentally achieved phase estimation matches the theoretical QCRB, with QFI extracted independently via coherent spectroscopy without full state tomography (Yu et al., 2020). Macroscopic quantum light—bright two-mode squeezed states—achieves QCRB-limited transmission estimation, showing significant quantum enhancement over classical strategies in practical protocols (Woodworth et al., 2022).

Protocols: The twist–untwist protocol using one-axis twisted spin coherent states achieves the Heisenberg limit and saturates the QCRB in collective spin interferometry, even for finite-range interactions, demonstrating robustness and practical simplicity (Volkoff et al., 2022).

Algorithmic and Diagnostic Value: Generalized Cramér–Rao bounds based on higher statistical moments, such as third-order absolute moment bounds, provide sensitive diagnostics for detecting unnoticed estimator biases in quantum sensors that would be invisible to the variance-based QCRB (Cimini et al., 2020).

7. Extensions, Limitations, and Special Cases

Energy-Constrained Unitary Estimation: In noiseless unitary parameter estimation, the QCRB can be unattainable due to geometric reasons and energetic constraints. Specialized risk functions and optimal probe preparation under energy constraints achieve a true Heisenberg scaling bound, while the naive QCRB can become vacuously loose (Hayashi et al., 2016).

Approximate QCRBs: Practical calculation of Helstrom QFI can become intractable for high-dimensional mixed states; the Wigner–Yanase skew information provides a computable approximation to the QCRB and, in certain limits, matches Helstrom information exactly (Luati, 2011).

Singular Cases: At parameter values where the density matrix rank changes, apparent violations of the QCRB arise from the discontinuity of the algebraic QFI formula and the unboundedness of the SLD. The continuous Bures–QFI restores the correctness of the bound even in these singular models (Ye et al., 2021).

Optimal Scalarization: In the multiparameter setting, the “intrinsic” scalarization of the QCRB by the Lie-algebra metric yields a bound that is independent of parametrization and captures the minimal achievable error across all generators (Goldberg et al., 2021).

---

References

• "Saturating the one-axis twisting quantum Cramér–Rao bound with a total spin readout" (Volkoff et al., 2022)
• "Saturating the quantum Cramér-Rao bound using LOCC" (Zhou et al., 2018)
• "A geometric perspective: experimental evaluation of the quantum Cramer-Rao bound" (Li et al., 2022)
• "Quantum Metrology in Open Systems: Dissipative Cramér-Rao Bound" (Alipour et al., 2013)
• "General Cramér-Rao bound for parameter estimation using Gaussian multimode quantum resources" (Pinel et al., 2010)
• "Quantum Fisher information measurement and verification of the quantum Cramér-Rao bound in a solid-state qubit" (Yu et al., 2020)
• "Saturability of the Quantum Cramér-Rao Bound in Multiparameter Quantum Estimation at the Single-Copy Level" (Nurdin, 18 Feb 2024)
• "Approaching the Multiparameter Quantum Cramér-Rao Bound via Classical Correlation and Entangling Measurements" (Mi et al., 12 Sep 2025)
• "Diagnosing Imperfections in Quantum Sensors via Generalized Cramér-Rao Bounds" (Cimini et al., 2020)
• "Quantum Limits on Parameter Estimation" (Goldstein et al., 2010)
• "Intrinsic Sensitivity Limits for Multiparameter Quantum Metrology" (Goldberg et al., 2021)
• "Quantum Cramér-Rao bound for quantum statistical models with parameter-dependent rank" (Ye et al., 2021)
• "Quantum-Limited Estimation of Phase Gradient" (Larson et al., 2019)
• "Fundamental Quantum Limit to Waveform Estimation" (Tsang et al., 2010)
• "Saturation of the Multiparameter Quantum Cramér-Rao Bound at the Single-Copy Level with Projective Measurements" (Nurdin, 2 May 2024)
• "Unattainable & attainable bounds for quantum sensors" (Hayashi et al., 2016)
• "An approximate quantum Cramér--Rao bound based on skew information" (Luati, 2011)
• "Transmission Estimation at the Fundamental Quantum Cramér-Rao Bound with Macroscopic Quantum Light" (Woodworth et al., 2022)
• "Role of the extended Hilbert space in the attainability of the Quantum Cramér-Rao bound for multiparameter estimation" (Conlon et al., 1 Apr 2024)
• "The Most Informative Cramér--Rao Bound for Quantum Two-Parameter Estimation with Pure State Probes" (Yung et al., 18 Nov 2025)