arXiv:2609.25977
Dynamical, thermal, and ground-state multiparameter quantum metrology: Fundamental limits and their attainability in magnetometry
Riccardo Pedroni, Víctor Izquierdo, Advay Burte, Paolo Abiuso, John Calsamiglia, Martí Perarnau-Llobet
Abstract
We consider the simultaneous estimation of multiple parameters in Hamiltonians of the form $H_{\vecθ}= H_{\vecθ}^P + H^C$, where $H_{\vecθ}^P$ encodes the unknown parameters and $H^C$ is a control term. We consider three basic paradigmatic frameworks for metrology: (i) dynamical, where the parameters are encoded unitarily, (ii) thermal, and (iii) ground state. We generalize previous bounds for single-parameter estimation to simultaneous multiparameter estimation for all three metrological settings and establish the conditions required for their saturability. We then focus on vector magnetometry, deriving tight bounds for the simultaneous estimation of magnetic-field components and identifying optimal many-body control Hamiltonians $H^C$ that saturate them. These results establish the fundamental limits of vector magnetometry for unitary dynamics, thermal equilibrium states, and ground states.
AI Summary
The paper asks how parameter-independent Hamiltonian controls limit simultaneous estimation of parameters in $H_{\boldsymbol\theta}=H^P_{\boldsymbol\theta}+H^C$, comparing unitary evolution, Gibbs states, and nondegenerate gapped ground states. It studies the local, asymptotic weighted error bound $\operatorname{Tr}(WQ^{-1})$, separating optimization of the quantum Fisher information matrix from measurement compatibility. General bounds follow from $\operatorname{Tr}(WQ^{-1})\geq\sum_i w_i/Q_{ii}$, variance bounds for Hamiltonian derivatives, and spectral-width bounds. Saturation requires compatible variance trade-offs and a diagonal QFIM; equilibrium and ground-state settings impose additional commutation or excitation-coupling conditions. Weak commutativity is checked for the proposed optimal probes to establish asymptotic attainability. For $N$ spin-$1/2$ particles sensing $d$ magnetic-field components, with $\sum_i\sqrt{w_i}=d$, the bounds are $4d^2/[\beta^2N(N+2)]$ for thermal encoding, $d^2\Delta^2/[N(N+2)]$ for ground states, and $d^2\hbar^2/[t^2N(N+2)]$ for dynamics. For equal weights, constructions saturate the three-component bounds and the two-component ground-state and dynamical bounds. These use a symmetric-subspace thermal mixture generated by $-MJ^2$, Dicke probes, and an XYZ-GHZ ground state with engineered excitation energies. The two-component thermal bound is not saturable. General bounds need not be tight, and arbitrary-weight saturation is not established. Results assume unrestricted controls and local estimation; some constructions require particular particle numbers, strong interactions, or nonlocal projector Hamiltonians. Explicit optimal thermal measurements, locality constraints, and finite-resource Bayesian extensions remain open.
Error found: The claimed necessity of [H,∂ᵢH] = 0 for saturation of the unitary seminorm bound is false because noncommuting dynamics confined away from the extremal sensing subspace can leave the bound exactly attainable. The two-component ground-state construction Hᶜ = gJ_z² − J² saturates the stated gap-dependent bound only when g ≤ N, contrary to the main text's unrestricted claim for g > 0, although the appendix supplies a sufficient restriction.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.