---
title: Quantum Stopwatches and QFI Metrology
url: https://www.emergentmind.com/topics/quantum-stopwatches
type: topic
---

# Quantum Stopwatches and QFI Metrology

“Quantum stopwatches” (*Editor’s term*) denotes quantum metrological constructions in which elapsed evolution, a phase, or a more general encoded parameter is inferred from the response of a quantum state. In the cited literature, the formal vocabulary is not “stopwatch” but quantum Fisher information (QFI), quantum Fisher information matrix (QFIM), measurement-induced Fisher information, and sub-QFI. Across these works, the common technical theme is that precision is characterized by the QFI or QFIM, while the physically accessible information depends on the measurement architecture, the subsystem retained, the reduced description employed, and whether the state is in equilibrium or driven out of equilibrium [1208.0104][2311.12596][2408.12918].

## 1. Parameter sensitivity as the operational core

The central object is the QFI of a parameterized state family. For a one-parameter family \(\rho_\theta\), the symmetric logarithmic derivative \(L_\theta\) is defined by
\[
\partial_\theta \rho_\theta = \frac12(L_\theta \rho_\theta + \rho_\theta L_\theta),
\]
and the QFI is
\[
F(\rho_\theta) := \mathrm{tr}(\rho_\theta L_\theta^2).
\]
For a POVM \(M=\{M_i\}\), the induced probability distribution \(p_\theta(i)=\mathrm{tr}(\rho_\theta M_i)\) has classical Fisher information
\[
F(\rho_\theta\mid M) := \sum_i p_\theta(i)\left(\partial_\theta \ln p_\theta(i)\right)^2,
\]
and the Braunstein–Caves relation states
\[
F(\rho_\theta)=\sup_M F(\rho_\theta\mid M).
\]
This gives the standard operational meaning: QFI is the maximal measurement-induced Fisher information over all POVMs [1208.0104].

In the many-body formulation of [2311.12596], the parameterization is multiparameter and unitary:
\[
\rho(\bm\phi)=\hat U(\bm\phi)\rho \hat U^\dagger(\bm\phi),\qquad
\hat U(\bm\phi)=\exp\!\Big(i\sum_\alpha\sum_{ij}\phi_\alpha^{ij}\hat J_\alpha^{ij}\Big),
\]
with one-body generators
\[
\hat J^{ij}_\alpha = \tfrac12 (\hat b_i^\dagger, \hat b_j^\dagger)\,\sigma_\alpha\,(\hat b_i,\hat b_j)^T.
\]
For pure states, the QFIM reduces to a covariance matrix of the generators:
\[
M_{\alpha\beta}^{ijkl}[\psi] =
2\langle\psi|\{\hat J_\alpha^{ij}, \hat J_\beta^{kl}\}|\psi\rangle
-4\langle\psi|\hat J_\alpha^{ij}|\psi\rangle
\langle\psi|\hat J_\beta^{kl}|\psi\rangle.
\]
This is the precise many-parameter sensitivity tensor used in the reduced-functional approach [2311.12596].

A useful clarification is that the cited works do not formulate autonomous quantum clocks or time operators. They analyze parametric sensitivity of quantum states under unitary encoding. This suggests that “quantum stopwatch” is best understood as an umbrella label for QFI-based timing or phase-estimation schemes rather than a distinct formalism.

## 2. Composite systems and subsystem-accessible timing information

For bipartite systems, [1208.0104] distinguishes between intrinsic reduced-state QFI and the Fisher information actually accessible under restricted measurement classes. For a bipartite family \(\rho_\theta^{ab}\), the reduced states are
\[
\rho_\theta^a = \mathrm{tr}_b \rho_\theta^{ab},\qquad
\rho_\theta^b = \mathrm{tr}_a \rho_\theta^{ab}.
\]
Local, product, adaptive, and global measurement classes are introduced:
\[
\mathcal M^a,\quad \mathcal M^b,\quad \mathcal M^{a,b},\quad \mathcal M^{a\rightarrow b},\quad \mathcal M^{a\leftarrow b},\quad \mathcal M^{ab}.
\]
They satisfy the hierarchy
\[
F(\rho_\theta^{ab}\mid \mathcal M^a)\le
F(\rho_\theta^{ab}\mid \mathcal M^{a,b})\le
F(\rho_\theta^{ab}\mid \mathcal M^{a\rightarrow b})\le
F(\rho_\theta^{ab}),
\]
and similarly with \(a\) and \(b\) exchanged. The endpoints are especially important:
\[
F(\rho_\theta^{ab}\mid \mathcal M^a)=F(\rho_\theta^a),\qquad
F(\rho_\theta^{ab}\mid \mathcal M^b)=F(\rho_\theta^b),\qquad
F(\rho_\theta^{ab}\mid \mathcal M^{ab})=F(\rho_\theta^{ab}).
\]
Thus, the reduced-state QFI equals the maximum Fisher information accessible by measurements on that subsystem alone [1208.0104].

The same paper classifies three extremal distribution types. For product states, QFI is locally owned and additive:
\[
F(\rho_\theta^{ab})=F(\rho_\theta^a)+F(\rho_\theta^b).
\]
For the entangled state
\[
|\Psi_\theta\rangle=\frac{1}{\sqrt{2}}\bigl(|0^a0^b\rangle+e^{i\theta}|1^a1^b\rangle\bigr),
\]
one has
\[
F(\rho_\theta^{ab})=\frac12,\qquad F(\rho_\theta^a)=F(\rho_\theta^b)=0,
\]
so the information is locally inaccessible: it resides entirely in correlations. By contrast, for
\[
|\Phi_\theta\rangle=\cos\frac{\theta}{2}|0^a0^b\rangle+\sin\frac{\theta}{2}|1^a1^b\rangle,
\]
the paper states
\[
F(\rho_\theta^{ab})=F(\rho_\theta^a)=F(\rho_\theta^b)=1,
\]
which is the fully shared case [1208.0104].

This hierarchy matters directly for any stopwatch-like interpretation. A global state can contain parameter sensitivity that is absent from every reduced state, or duplicated across them, or movable between these configurations by global unitaries. The CNOT examples in [1208.0104] show both directions: a unitary can convert locally owned Fisher information into locally inaccessible correlation-based Fisher information, and the inverse unitary can concentrate that information back into one subsystem without changing the total QFI.

## 3. Subnormalized, truncated, and auxiliary-assisted architectures

A second line of work treats cases where only a restricted sector, postselected branch, or small auxiliary system is experimentally available. In [2010.02904], the generalized fidelity for subnormalized states is
\[
F_*(\tau,\sigma)=\Big\|\sqrt{\tau}\sqrt{\sigma}\Big\|_1
+\sqrt{(1-\operatorname{Tr}[\tau])(1-\operatorname{Tr}[\sigma])},
\]
and the truncated QFI is defined by
\[
\mathcal{I}_*(\theta;\rho^{(m)}_\theta)
=
8\lim_{\delta\to 0}
\frac{1-F_*(\rho^{(m)}_\theta,\rho^{(m)}_{\theta+\delta})}{\delta^2},
\]
where \(\rho_\theta^{(m)}\) is obtained by projecting onto the \(m\) largest eigenvalues of \(\rho_\theta\). The construction yields a lower bound
\[
\mathcal{I}_*(\theta;\rho^{(m)}_\theta)\le I(\theta;\rho_\theta),
\]
with equality at full rank \(m=r\), and it is monotone under completely positive trace-non-increasing maps:
\[
\mathcal{I}_*(\theta;\Phi(\rho^{(m)}_\theta))
\le
\mathcal{I}_*(\theta;\rho^{(m)}_\theta).
\]
This does not explicitly constitute a subsystem-QFI theory, but it is mathematically aligned with reduced or postselected states, since partial trace and heralded operations are CPTNI maps [2010.02904].

A complementary strategy is developed in [2408.12918]. A high-dimensional parameterized state \(\rho^a(x)\) is coupled to a parameter-independent auxiliary state \(\sigma^b\), and a controlled unitary
\[
U=\sum_{i=1}^{N}\Pi_i^a\otimes O_i^b
\]
is applied. After tracing out the original system,
\[
\rho^b(x)=\mathrm{Tr}_a\!\left[U(\rho^a(x)\otimes \sigma^b)U^\dagger\right]
=
\sum_{i=1}^N p_i(x)\,O_i^b\sigma^b O_i^{b\dagger},
\]
with
\[
p_i(x):=\mathrm{Tr}\big(\Pi_i^a\rho^a(x)\Pi_i^a\big).
\]
The key inequality chain is
\[
F(\rho^a(x))
=
F\big(U\rho^{ab}(x)U^\dagger\big)
\ge
F(\rho^b(x))
\ge
F^{(\mathrm{sub})}(\rho^b(x)),
\]
so the measurable sub-QFI of the auxiliary subsystem lower-bounds the QFI of the original large system [2408.12918].

The same paper shows that maximal transfer is achieved when the auxiliary state is pure and the branch states are mutually orthogonal. In that case, the maximal auxiliary QFI is the classical Fisher information of the distribution \(\{p_i(x)\}\):
\[
F_{\mathrm{max}}(\rho^b)=\sum_{i=1}^{N}\frac{(\partial_x p_i)^2}{p_i}.
\]
It also gives a criterion for dimensional reduction by coarse-graining projectors. If projectors are grouped into blocks and, within each group,
\[
p_j^{(i)}\partial_x p_l^{(i)}=p_l^{(i)}\partial_x p_j^{(i)},
\]
then the coarse-grained measurement preserves the classical Fisher information and therefore the maximal QFI transferable to the auxiliary system [2408.12918].

The Bell-state example is exact. For
\[
|\psi_x\rangle=\frac{1}{\sqrt{2}}(e^{-ix}|00\rangle+e^{ix}|11\rangle),
\]
the QFI is
\[
F[|\psi_x\rangle,J_z]=4.
\]
After mapping to a single auxiliary qubit, the reduced auxiliary state becomes
\[
\rho^b(x)=\frac{1}{2}(1+\cos 2x)|0\rangle\langle 0|
+\frac{1}{2}(1-\cos 2x)|1\rangle\langle 1|,
\]
and, because sub-QFI equals QFI for a single qubit, one obtains
\[
F^{(\mathrm{sub})}(\rho^b)=F(\rho^b)=4.
\]
This is an explicit small-system stopwatch architecture in the sense of parameter readout from a reduced subsystem [2408.12918].

## 4. Many-body compression through one-body reduced data

The most direct many-body compression result is the 1-RDM functional framework of [2311.12596]. For a ground state of
\[
\hat H=\hat h+\hat W,
\]
with interaction
\[
\hat W=
\sum_{\alpha,\beta}\sum_{ijkl}
u_{\alpha\beta}^{ijkl}\{\hat J_\alpha^{ij},\hat J_\beta^{kl}\},
\]
the one-body reduced density matrix is encoded as
\[
\gamma_\alpha^{ij}\equiv \mathrm{Tr}[\rho\,\hat J_\alpha^{ij}].
\]
The universal constrained-search functional is
\[
\mathcal F[\bm\gamma;\bm u]
\equiv
\min_{\psi\rightarrow \bm\gamma}
\sum_{\alpha,\beta}\sum_{ijkl}
u_{\alpha\beta}^{ijkl}
\langle\psi|\{\hat J_\alpha^{ij},\hat J_\beta^{kl}\}|\psi\rangle.
\]
Because 1-RDMFT provides a minimizer \(|\psi[\bm\gamma]\rangle\), the QFIM becomes a functional of the 1-RDM:
\[
\mathcal M_{\alpha\beta}^{ijkl}[\bm\gamma]
\equiv
M_{\alpha\beta}^{ijkl}[\psi[\bm\gamma]].
\]

The central derivative relation is
\[
\mathcal{M}_{\alpha\beta}^{ijkl}[\bm{\gamma};\bm{u}]
=
4\Big[
\Big(\frac{\partial \mathcal{E}[\bm{\gamma};\bm{u}]}{\partial u_{\alpha\beta}^{ijkl}}\Big)_{\!\bm{\gamma}}
-\gamma_\alpha^{ij}\gamma_\beta^{kl}
\Big],
\]
with \(\mathcal E[\bm\gamma;\bm u]=\mathrm{Tr}[\hat h\bm\gamma]+\mathcal F[\bm\gamma;\bm u]\). Equivalently,
\[
\frac{\partial \mathcal{F}[\bm\gamma;\bm u]}{\partial u_{\alpha\beta}^{ijkl}}
=
\langle\psi[\bm\gamma]|\{\hat J_\alpha^{ij},\hat J_\beta^{kl}\}|\psi[\bm\gamma]\rangle.
\]
Hence the universal functional \(\mathcal F\) is a generating functional for QFI: all QFIM entries with respect to the one-body generator basis are obtained from derivatives with respect to the couplings \(u_{\alpha\beta}^{ijkl}\) [2311.12596].

This route does not treat QFI of a spatial subregion’s reduced density operator. The reduced object is instead the one-body sector. That distinction is explicit in [2311.12596]: the framework computes the global many-body QFIM for one-body generators from one-body reduced data, not the QFI of a proper subsystem in the spatial-partition sense.

The two-site Bose–Hubbard model provides the concrete bosonic example. With
\[
\hat H=-t\sum_{\langle ij\rangle}\hat b_i^\dagger \hat b_j
+
u\sum_j \hat n_j(\hat n_j-1),
\]
one has
\[
\mathcal{M}_{zz}[\bm\gamma]
=
4\Big(\frac{\mathcal F[\bm\gamma;u]}{u}-\gamma_z^2\Big)-N^2+2N.
\]
For \(N=2\), exact universal QFIM functionals are given analytically over the Bloch disk in terms of \(\gamma_x,\gamma_z\), including the repulsive and attractive cases [2311.12596]. This suggests a compressed stopwatch-like description in which many-body sensitivity to one-body generators is reconstructed without explicitly handling exponentially large wave functions.

## 5. Criticality, squeezing, and reduced-state stopwatches

In the Dicke-model study [1312.1426], the total ground state is pure but entangled, and the field and atomic subsystems are mixed:
\[
\hat\rho_B=\mathrm{Tr}_A(|g\rangle\langle g|),\qquad
\hat\rho_A=\mathrm{Tr}_B(|g\rangle\langle g|).
\]
The chosen generators are
\[
\hat G_B=\hat b^\dagger\hat b,\qquad \hat G_A=\hat J_x.
\]
The mixed-state QFI is written as
\[
F=
4\sum_n p_n(\Delta \hat G)_n^2
-
\sum_{m\neq n}
\frac{8p_mp_n}{p_m+p_n}
\left|\langle\psi_m|\hat G|\psi_n\rangle\right|^2.
\]
The paper compares subsystem QFI to classical limits \(F_B^{\mathrm{cl}}=4\bar n\) for a coherent field state and \(F_A^{\mathrm{cl}}=N\) for a coherent spin state [1312.1426].

Near the superradiant critical coupling, both subsystems can surpass their classical limits. For finite and large enough atom number, the paper reports that near the critical atom-field coupling the QFIs of the atomic and field subsystems can surpass the classical limits because of nonclassical squeezed states. Far beyond the critical point, the two reduced states become highly mixed, which degrades the QFI and therefore the ultimate phase sensitivity [1312.1426].

In the thermodynamic limit, the atomic subsystem obeys
\[
F_A = N\,\frac{\mu\omega_0}{\Omega}
\frac{e^{\beta\Omega}-1}{e^{\beta\Omega}+1},
\qquad
\xi^2 =
\frac{\mu\Omega}{\omega_0}
\frac{e^{\beta\Omega}+1}{e^{\beta\Omega}-1},
\]
which implies the exact relation
\[
F_A\,\xi^2=N\mu^2.
\]
At resonance and at the critical point, the paper gives
\[
\xi^2(\lambda_{\mathrm{cr}})=\frac{1}{\sqrt{2}},\qquad
\frac{F_A}{N}(\lambda_{\mathrm{cr}})=\sqrt{2}.
\]
For the field subsystem, the phase-quadrature variance satisfies
\[
(\Delta \hat X_{\pi/2})^2
=
\frac{\Omega}{4\omega}\frac{e^{\beta\Omega}+1}{e^{\beta\Omega}-1},
\]
and in the superradiant phase
\[
F_B(\Delta \hat X_{\pi/2})^2 \approx \bar n.
\]
At the critical point on resonance,
\[
4(\Delta \hat X_{\pi/2})^2(\lambda_{\mathrm{cr}})=\frac{1}{\sqrt{2}},
\qquad
\frac{F_B}{4\bar n}(\lambda_{\mathrm{cr}})\approx \sqrt{2}.
\]
These formulas make the link between squeezing and reduced-state QFI explicit [1312.1426].

The singular behavior is equally important. In the thermodynamic limit, the derivatives of the scaled subsystem QFIs diverge at the critical point:
\[
\partial_\lambda(F_A/N)\sim |\lambda-\lambda_{\mathrm{cr}}|^{-1/2},
\qquad
\partial_\lambda(F_B/4\bar n)\sim |\lambda-\lambda_{\mathrm{cr}}|^{-1/2}.
\]
Accordingly, subsystem QFI is not merely a metrological witness; it is also a sharp diagnostic of quantum criticality [1312.1426].

## 6. Nonequilibrium amplification and limits

The out-of-equilibrium analysis of [2503.21905] is formulated entirely in terms of subsystems. For an infinite spin chain and a contiguous block \(A=\llbracket l,r\rrbracket\), the observable is the restricted extensive order parameter
\[
O_A=\sum_{\ell\in A} O_\ell,
\]
with the transverse-field Ising example
\[
X_A=\sum_{\ell\in A}\sigma_\ell^x.
\]
For pure states, the QFI reduces to variance,
\[
\frac{1}{4}F(\rho,O)=\Delta_\rho O^2,
\]
but for mixed reduced states the paper emphasizes the normalized quantity
\[
\chi(\rho;O)=\frac{F(\rho,O)}{4\|O\|^2}.
\]
If \(\chi\) remains finite at large \(|A|\), then \(F(\rho_A,O_A)\propto |A|^2\), which the paper interprets as macroscopic quantum behavior within the subsystem [2503.21905].

The equilibrium statement is restrictive: in equilibrium local quantum many-body systems, subsystem QFI with respect to an extensive observable is typically proportional to subsystem volume, and the QFI per unit volume squared becomes negligible. The nonequilibrium result is the contrast case. At zero or sufficiently low temperatures in a spin chain with an ordered phase, a transient localized perturbation enhances the subsystem QFI so that it scales quadratically with subsystem length. More general localized kicking protocols can control this enhancement [2503.21905].

The same work revisits global quenches. After a global quench in the thermodynamic limit, the density of localized multipartite entanglement approaches zero at late times, but there is an optimal time frame proportional to the subsystem length in which subsystems fall into macroscopic quantum states. This means the stopwatch-like sensitivity of compact subsystems can become superextensive, but only within a specific nonequilibrium window [2503.21905].

A central technical contribution is the identity
\[
\frac{1}{4}F(\rho,O)=
\int_{-\infty}^{\infty}\frac{d\beta}{\cosh(\pi\beta)}\,
I_{\frac12+i\beta}(\rho,O),
\]
which expresses QFI in terms of rotated Wigner–Yanase–Dyson skew information. Together with Gaussian free-fermion methods, this permits direct computation of subsystem QFI in noninteracting spin chains [2503.21905].

Taken together, these results delimit the scope of “quantum stopwatches” in the present corpus. They are not theories of time observables or autonomous clockwork. They are theories of parameter sensitivity: how timing or phase information is quantified by QFI, how much of it survives in reduced states, how it can be compressed into auxiliary systems or one-body reduced descriptions, and how equilibrium, criticality, and nonequilibrium dynamics respectively suppress, sharpen, or transiently amplify that sensitivity [1208.0104][2010.02904][2311.12596][2408.12918][1312.1426][2503.21905].

Source: https://www.emergentmind.com/topics/quantum-stopwatches