---
title: Dissipative Quantum Limit (DQL) Overview
url: https://www.emergentmind.com/topics/dissipative-quantum-limit-dql
type: topic
---

# Dissipative Quantum Limit (DQL) Overview

Searching arXiv for recent papers on “dissipative quantum limit” and closely related formulations.
The **Dissipative Quantum Limit (DQL)** denotes a class of fundamental constraints imposed by dissipation and decoherence on quantum evolution, sensing, metrology, and nonequilibrium state transformation. In the literature summarized here, the term appears in several technically distinct but structurally related settings: as a lower bound on effective force-noise spectral density in stationary linear sensors, \(S_{\rm DQL}(\Omega)=\hbar|\Im\chi^{-1}(\Omega)|\), arising from dissipative probe dynamics [2011.14716]; as a set of operator-valued uncertainty relations for non-stationary force sensing governed by the antisymmetric part of the inverse susceptibility \(\chi_a^{-1}(t,t')\) [2605.03559]; and, in open-system quantum dynamics, as fundamental inequalities constraining evolution speed and information acquisition under Lindblad dynamics through the interplay of unitary variance, dissipative deformation, and dissipative fluctuations [2507.02501]. Related work also realizes DQL-type behavior as universal bounds on nonadiabatic entropy production near dissipative quantum phase transitions [2512.05074]. Across these formulations, the common feature is that dissipation is not merely a source of degradation: it both enables and constrains distinguishability, sensitivity, irreversibility, and control.

## 1. Historical definitions and domain-specific formulations

The term **Dissipative Quantum Limit** was explicitly proposed for stationary force sensing in "Quantum limits for stationary force sensing" [2011.14716]. In that setting, the DQL is a lower bound on the meter’s effective force-noise spectral density caused by the dissipative dynamics of the probe, even when correlations are used to evade the Standard Quantum Limit (SQL). The defining expression is
\[
S_{\rm DQL}(\Omega)=\hbar\,|\Im\chi^{-1}(\Omega)|,
\]
where \(\chi(\Omega)\) is the probe susceptibility and \(\Im\chi^{-1}(\Omega)\) encodes dissipation [2011.14716].

A later extension to non-stationary linear sensors reformulated the concept in time domain. For stationary systems, the same irreducible noise floor appears after optimization over the meter noise subject to a generalized uncertainty relation, but for non-stationary sensors the strict scalar DQL disappears for a single known waveform and is replaced by multi-parameter uncertainty relations involving the antisymmetric kernel
\[
\chi_a^{-1}(t,t')=\frac12\big[\chi^{-1}(t,t')-\chi^{-1}(t',t)\big].
\]
These are called “DQL-like” because they still depend purely on dissipation [2605.03559].

A distinct but closely related usage appears in open-system quantum speed-limit theory. "Quantum speed limit under decoherence: unitary, dissipative, and fluctuation contributions" [2507.02501] does not introduce DQL as a pre-existing formal term, but its synthesis explicitly organizes the results around that notion: dissipation sets fundamental limits on the speed of evolution and on information acquisition in Markovian Lindblad dynamics. In that formulation, the DQL is encoded in a quantum speed limit and in a short-time quantum Fisher information bound governed by the quantities \(\Delta H_0\), \(\mathcal{G}\), and \(\mathcal{E}\) [2507.02501].

A further thermodynamic realization appears in finite-time driving across second-order dissipative quantum phase transitions. There, a DQL is not named as such in the original formalism, but the nonadiabatic entropy production \(\Sigma_{\rm na}\) plays the role of a universal lower bound on irreversibility under finite-time driving, with scaling fixed by critical exponents \(\alpha\) and \(\gamma\) [2512.05074]. This suggests a broader unifying interpretation: the DQL is any universal lower bound set specifically by dissipative response.

## 2. Stationary linear sensing: spectral DQL and force-noise floor

In stationary linear force sensing, the probe obeys
\[
\chi^{-1}(\Omega)\hat{x}(\Omega)=F_{\rm sig}(\Omega)+\hat{F}_{\rm ba}(\Omega)+\hat{F}_T(\Omega)+\dots,
\]
while the meter output and backaction are
\[
\tilde{x}(\Omega)=\hat{x}_{\rm fl}(\Omega)+\hat{x}(\Omega),\qquad
\hat{F}_{\rm ba}(\Omega)=\hat{F}_{\rm fl}(\Omega)-K(\Omega)\hat{x}(\Omega).
\]
Referencing the output back to force yields
\[
\hat{F}_{\rm sum}(\Omega)=\chi_K^{-1}(\Omega)\hat{x}_{\rm fl}(\Omega)+\hat{F}_{\rm fl}(\Omega),
\]
with \(\chi_K^{-1}=\chi^{-1}+K\), and sum-noise spectral density
\[
S_{\rm sum}(\Omega)=|\chi_K^{-1}|^2S_{xx}+2\,\Re[\chi_K^{-1}S_{xF}]+S_{FF}.
\]
The meter noises satisfy a generalized uncertainty relation,
\[
S_{xx}S_{FF}-|S_{xF}|^2 \ge \hbar|\sigma|+\frac{\hbar^2}{4},
\]
with \(\sigma(\Omega)=\Im[K(\Omega)S_{xx}(\Omega)+S_{xF}^*(\Omega)]\) [2011.14716].

Optimizing \(S_{\rm sum}\) under this constraint yields two regimes. In the weak-backaction regime the minimum is QCRB-like, while for sufficiently large backaction it saturates at the DQL,
\[
S_{\rm sum}(\Omega)=S_{\rm DQL}(\Omega)=\hbar|\Im\chi^{-1}(\Omega)|.
\]
The paper identifies a threshold \(S_{\rm thr}(\Omega)\) separating the QCRB-dominated and DQL-dominated regimes, and reports a phase-transition-like nonanalyticity at the boundary: \(S_{\rm sum}\) remains continuous, but the second derivative of \(S_{\rm sum}\) with respect to \(S_{FF}\) and the first derivative of the optimal cross-correlation \(S_{xF}\) are discontinuous at \(S_{FF}=S_{\rm thr}\) [2011.14716].

This stationary DQL is more fundamental than the SQL in the sense stated in the source material: SQL depends on the full \(|\chi^{-1}|\), whereas the DQL depends only on the dissipative part \(|\Im\chi^{-1}|\) [2011.14716]. It is also distinct from thermal noise. The thermal-force spectral density from fluctuation-dissipation theory is
\[
S_{\rm FDT}(\Omega)=\hbar|\chi^{-1}(\Omega)|\coth\!\left(\frac{\hbar\Omega}{2k_BT}\right)\ge \hbar|\chi^{-1}(\Omega)|,
\]
so the DQL constrains the additional meter quantum noise rather than the total probe-plus-meter noise [2011.14716].

A central physical interpretation follows from commutators. The internal thermal force satisfies
\[
[\hat{F}_T(t),\hat{F}_T(t')]=i\hbar[\chi^{-1}(t,t')-\chi^{-1}(t',t)],
\]
and because the total output \(\tilde F(t)\) must commute at different times, the meter sum noise must carry the opposite non-autocommutativity. This forces
\[
S_{\rm sum}(\Omega)\ge \hbar|\Im\chi^{-1}(\Omega)|,
\]
so the DQL originates from the non-autocommutativity of internal thermal noise [2011.14716].

## 3. Non-stationary sensing: disappearance of scalar DQL and reappearance as operator uncertainty

For non-stationary linear sensors, the response is described by kernels rather than spectral densities. The probe obeys
\[
\chi_K^{-1}(t,t')\,\hat x(t')\,dt' = F_{\rm sig}(t)+\hat F_{\rm fl}(t)+\hat F_T(t),
\]
and the meter output after linear processing is
\[
\tilde F(t)=F_{\rm sig}(t)+\hat F_{\rm sum}(t)+\hat F_T(t),
\]
with
\[
\hat F_{\rm sum}(t)=\chi_K^{-1}(t,t')\hat x_{\rm fl}(t')\,dt'+\hat F_{\rm fl}(t).
\]
The crucial commutator is
\[
[\hat F_{\rm sum}(t),\hat F_{\rm sum}(t')]=-2i\hbar\,\chi_a^{-1}(t,t'),
\]
where
\[
\chi_a^{-1}(t,t')=\frac12[\chi^{-1}(t,t')-\chi^{-1}(t',t)].
\]
Because the thermal noise satisfies
\[
[\hat F_T(t),\hat F_T(t')]=2i\hbar\,\chi_a^{-1}(t,t'),
\]
the total output commutes,
\[
[\tilde F(t),\tilde F(t')]=0
\]
[2605.03559].

The conceptual shift is that for a **single known force waveform** there is no irreducible DQL in the strict stationary sense. With a chosen filter \(\Phi(t)\), one estimates an amplitude using
\[
\tilde{\mathcal F}=\int \Phi(t)\tilde F(t)\,dt.
\]
Within the linear model, the paper shows that the meter state can be chosen so that the variance of the integrated meter noise \(\langle \hat{\mathcal F}_{\rm sum}^2\rangle\) becomes arbitrarily small, even zero in the idealized limit. In a memoryless example with delta-correlated noises,
\[
B_{xx}(t,t')=S_{xx}(t)\delta(t-t'),\quad
B_{FF}(t,t')=S_{FF}(t)\delta(t-t'),\quad
B_{xF}(t,t')=S_{xF}(t)\delta(t-t'),
\]
the minimal integrated noise becomes
\[
\langle \hat{\mathcal F}_{\rm sum}^2\rangle_{\min}
=\int \frac{\hbar^2}{4}\,\frac{\Psi^2(t)}{S_{FF}(t)}\,dt,
\]
which can be pushed arbitrarily low as \(S_{FF}(t)\) increases [2605.03559].

However, for **multiple force components** or multiple waveform shapes, DQL-like uncertainty relations re-emerge. For filters \(\Phi_j(t)\) and \(\Phi_k(t)\),
\[
[\hat{\mathcal F}_{\rm sum}^{(j)},\hat{\mathcal F}_{\rm sum}^{(k)}]
=-2i\hbar\iint \Phi_j(t)\chi_a^{-1}(t,t')\Phi_k(t')\,dt\,dt',
\]
which implies
\[
\langle(\hat{\mathcal F}_{\rm sum}^{(j)})^2\rangle\,
\langle(\hat{\mathcal F}_{\rm sum}^{(k)})^2\rangle
\ge
\hbar^2\left(
\iint \Phi_j(t)\chi_a^{-1}(t,t')\Phi_k(t')\,dt\,dt'
\right)^2.
\]
These relations are the non-stationary DQL-like bounds [2605.03559].

For a narrow-band force
\[
F_{\rm sig}(t)=F_c(t)\cos\Omega_0 t + F_s(t)\sin\Omega_0 t,
\]
with stationary dissipation approximated near \(\Omega_0\) by \(\chi_a^{-1}(\Omega)\approx -i\Omega H\), the two quadratures obey
\[
\langle(\hat{\mathcal F}_{\rm sum}^c)^2\rangle\,
\langle(\hat{\mathcal F}_{\rm sum}^s)^2\rangle
\ge
\frac{\hbar^2|\chi_a^{-1}(\Omega_0)|^2}{4}
\left(\int \Phi_{c0}(t)\Phi_{s0}(t)\,dt\right)^2.
\]
This establishes a multi-parameter DQL-like trade-off that depends only on dissipation [2605.03559].

## 4. Markovian open-system dynamics: DQL as speed and information bound

For open quantum systems governed by time-independent Markovian Lindblad dynamics,
\[
\frac{d\rho_t}{dt} = -i[H,\rho_t] + \mathcal{D}[L]\rho_t,
\]
with
\[
\mathcal{D}[L]\rho = L\rho L^\dagger - \frac12\{L^\dagger L,\rho\},
\]
and pure initial state
\[
\rho_0 = |\psi_0\rangle\langle\psi_0|,
\]
the Bures angle from the initial state is
\[
\Theta_t = \arccos\!\left(\sqrt{\mathrm{Tr}(\rho_0\rho_t)}\right),\qquad \Theta_t\in[0,\pi/2].
\]
The time derivative satisfies
\[
\frac{d\Theta_t}{dt}
=
\frac{1}{2\sin\Theta_t\cos\Theta_t}
\left(
\mathrm{Tr}\{i[\rho_0,H]\rho_t\}
-
\mathrm{Tr}\{\rho_t\mathcal{D}^\dagger[L]\rho_0\}
\right),
\]
and is bounded by
\[
\frac{d\Theta_t}{dt}
\le
\frac{1}{\sin 2\Theta_t}
\left(
2\Delta H_0\sin\Theta_t
+
\sqrt{2}\,\mathcal{G}\sin\Theta_t
+
\mathcal{E}
\right),
\tag{1}
\]
where
\[
\Delta H_0=\sqrt{\mathrm{Tr}(H^2\rho_0)-\mathrm{Tr}(H\rho_0)^2},
\]
\[
\mathcal{G}=\|\mathcal{D}^\dagger[L]\rho_0\|_{\rm F},
\qquad
\mathcal{E}=\|L|\psi_0\rangle\|^2 - |\langle\psi_0|L|\psi_0\rangle|^2.
\]
The three contributions are, respectively, a unitary term, a dissipative deformation term, and a fluctuation term [2507.02501].

This decomposition is the basis for a DQL in the dynamical sense. Integrating the inequality yields a closed-form quantum speed limit \(T_{\rm QSL}\) depending on the target Bures angle \(\Theta_T\), the effective speed coefficient
\[
\mathcal{V}=2\Delta H_0+\sqrt{2}\,\mathcal{G},
\]
and the fluctuation strength \(\mathcal{E}\) [2507.02501]. The synthesis states the following limiting behaviors.

In the no-decoherence limit \(L\to0\), \(\mathcal{G}=\mathcal{E}=0\), so
\[
T_{\rm QSL}\to \frac{\sin\Theta_T}{\Delta H_0},
\]
recovering a Mandelstam–Tamm-type bound [2507.02501].

In the strong-decoherence regime, for large \(\mathcal{E}\),
\[
T_{\rm QSL}\approx \frac{\sin^2\Theta_T}{\mathcal{E}}.
\]
This means the fluctuation term dominates and the minimal time scales as \(1/\mathcal{E}\). The source explicitly states that as \(\mathcal{E}\to\infty\), distinguishability can be achieved in arbitrarily short time, even without coherent driving [2507.02501].

If both \(\mathcal{V}\) and \(\mathcal{E}\) scale with a dissipation strength \(\gamma\) as
\[
\mathcal{V}\sim \gamma,\qquad \mathcal{E}\sim \gamma,
\]
and their ratio remains constant, then
\[
T_{\rm QSL}\propto \frac{1}{\gamma},
\]
so the overall speed limit is set by dissipation [2507.02501]. This is a direct operational DQL: the environment determines the shortest transformation time.

The same framework gives a short-time quantum Fisher information bound. Using
\[
F(\rho_0,\rho_T)=1-\frac{F_Q}{4}T^2+o(T^3),
\]
one has
\[
\Theta_T \approx \frac{\sqrt{F_Q}}{2}T,\qquad \sin\Theta_T\approx \frac{\sqrt{F_Q}}{2}T.
\]
From a simplified QSL,
\[
T_{\rm QSL}\ge \frac{\sin^2\Theta_T}{\mathcal{E}+\mathcal{V}\sin\Theta_T},
\]
the paper derives
\[
F_Q \le \left(\mathcal{V}+\sqrt{\mathcal{V}^2+\frac{4\mathcal{E}}{T}}\right)^2.
\]
This expresses a speed–precision trade-off: the same dissipative quantities that accelerate state-space motion also constrain attainable short-time estimation precision [2507.02501].

## 5. Representative dynamical examples of the dissipative limit

The single-qubit dephasing example in [2507.02501] uses
\[
|\psi_0\rangle = \cos\frac{\theta}{2}|0\rangle + \sin\frac{\theta}{2}|1\rangle,\qquad
H=\frac{\omega}{2}\sigma_x,\qquad
L=\sqrt{\gamma}\,\sigma_z.
\]
For this model,
\[
\Delta H_0=\frac{\omega}{2}\cos\theta,\qquad
\mathcal{G}=2\gamma\sin\theta,\qquad
\mathcal{E}=\gamma\sin^2\theta,
\]
and
\[
\mathcal{V}=\omega\cos\theta + 2\sqrt{2}\gamma\sin\theta.
\]
The source states that for small \(\gamma\) the bound behaves like the unitary Mandelstam–Tamm limit, while for large \(\gamma\), \(T_{\rm QSL}\propto 1/\gamma\), so dephasing accelerates evolution and sets the speed limit [2507.02501].

For amplitude damping with
\[
L=\sqrt{\gamma}\,\sigma_-=\sqrt{\gamma}|1\rangle\langle 0|,\qquad H=0,\qquad |\psi_0\rangle=|0\rangle,
\]
the exact state is
\[
\rho_t=
\begin{pmatrix}
e^{-2\gamma t} & 0\\
0 & 1-e^{-2\gamma t}
\end{pmatrix},
\qquad
F_t=e^{-2\gamma t},
\qquad
\Theta_t=\arccos(e^{-\gamma t}).
\]
The exact time to reach angle \(\Theta_T\) is
\[
T_{\rm exa}= -\frac{1}{\gamma}\ln(\cos^2\Theta_T),
\]
while the QSL uses \(\mathcal{V}=\sqrt{2}\gamma\) and \(\mathcal{E}=\gamma\). The synthesis states that \(T_{\rm QSL}\le T_{\rm exa}\) for all \(\Theta_T\), that the bound is asymptotically tight for small \(\Theta_T\), and that both scale as \(1/\gamma\). This is a paradigmatic purely dissipative DQL [2507.02501].

A many-body example uses \(N\) qubits with
\[
H=\sum_{i=1}^N \frac{\omega}{2}\sigma_z^{(i)},
\qquad
L=\sqrt{\gamma}\,\sigma_x^{(i)}
\]
acting locally on qubit \(i\), and product initial state with Bloch angle \(\theta\). The scaling is
\[
2\Delta H_0=\mathcal{O}(\sqrt{N}),
\qquad
\mathcal{G}=2N\gamma\sin\theta=\mathcal{O}(N\gamma),
\qquad
\mathcal{E}=N\gamma\sin^2\theta=\mathcal{O}(N\gamma).
\]
In the dissipation-dominated regime \(\gamma\gg\omega\),
\[
\mathcal{V}\simeq \sqrt{2}\,\mathcal{G}\sim N\gamma,\qquad
T_{\rm QSL}\propto \frac{1}{N\gamma}.
\]
The source interprets this as collective dissipative speedup due to many independent channels of dissipation, rather than entanglement [2507.02501].

## 6. Thermodynamic DQL near dissipative quantum phase transitions

A thermodynamic form of DQL emerges in finite-time driving across second-order dissipative quantum phase transitions described by Lindblad dynamics,
\[
\frac{d\hat\rho}{dt}=\mathcal{L}_g[\hat\rho]
=
-i[\hat H(g),\hat\rho]+\sum_j\mathcal{D}[\hat L_j(g)]\hat\rho,
\]
with control parameter ramp
\[
g(t)=g_0+(g_c-g_0)\frac{t}{\tau_q}.
\]
For each \(g\), the nonequilibrium steady state \(\hat\pi_g\) satisfies
\[
\mathcal{L}_g[\hat\pi_g]=0.
\]
The Liouvillian gap closes near criticality as
\[
\Delta_g\sim |g-g_c|^\gamma,
\]
and the KMB quantum Fisher information diverges as
\[
\mathcal{I}_g^{\rm KMB}\sim |g-g_c|^{-\alpha}.
\]
The nonadiabatic entropy production rate is approximated by
\[
\dot\Sigma_{\rm na}(t)\approx \dot g(t)\,\zeta_{g(t)}\,\dot g(t),
\]
with
\[
\zeta_g=\tau_g\,\mathcal{I}_g^{\rm KMB},
\qquad
\tau_g\sim 1/\Delta_g.
\]
Hence
\[
\zeta_g\sim |g-g_c|^{-(\alpha+\gamma)}.
\]
The main scaling law for total nonadiabatic entropy production is
\[
\Sigma_{\rm na}(0,\tau_q)\sim \tau_q^{\frac{\alpha-2}{1+\gamma}}.
\]
This means that excess dissipation cannot decay faster than the indicated universal power law when the ramp is slowed [2512.05074].

The source highlights a particularly rigid case: for single-mode bosonic Gaussian dissipative quantum phase transitions,
\[
\alpha=2.
\]
Then
\[
\Sigma_{\rm na}(0,\tau_q)\sim \tau_q^0,
\]
so the nonadiabatic entropy production becomes independent of driving speed to leading order. The paper states that any finite-time traversal of the transition incurs a finite, universal amount of nonadiabatic entropy production regardless of how slowly the system is driven [2512.05074]. This is a hard DQL in the thermodynamic sense: no adiabatic limit exists with respect to \(\Sigma_{\rm na}\).

The driven-dissipative Dicke model in the thermodynamic limit realizes this scenario, with \(\gamma\simeq1\), \(\alpha=2\), \(\zeta_g\sim |g-g_c|^{-3}\), and total \(\Sigma_{\rm na}(0,\tau_q)\) saturating to a constant as \(\tau_q\to\infty\) [2512.05074]. Finite-size scaling in the open Kerr model approaches the same \(|g-g_c|^{-2}\) KMB-QFI divergence, supporting universality within the same class [2512.05074].

## 7. Conceptual scope, misconceptions, and relation to other limits

One recurring misconception is to treat the DQL as a single universal formula. The literature surveyed here does not support that. Instead, the same phrase refers to several precise but context-dependent objects: a spectral force-noise floor \(S_{\rm DQL}(\Omega)\) in stationary sensing [2011.14716], time-domain multi-parameter uncertainty relations in non-stationary sensing [2605.03559], QSL- and QFI-based bounds for Lindblad state evolution [2507.02501], and universal entropy-production scaling at dissipative critical points [2512.05074]. These are not mutually reducible by any formula given in the source material.

A second misconception is that dissipation is always purely detrimental. The cited works show a more structured picture. In stationary sensing, dissipation sets an irreducible residual noise floor even after backaction-evasion is optimized [2011.14716]. In non-stationary sensing, it can be circumvented for a single known waveform but reappears as an incompatibility constraint for multiple waveform components [2605.03559]. In Lindblad speed-limit theory, dissipation can accelerate distinguishability growth through \(\mathcal{G}\) and \(\mathcal{E}\), yet the same quantities limit information acquisition [2507.02501]. Near dissipative critical points, critical slowing down and susceptibility divergences conspire to impose unavoidable irreversibility [2512.05074].

The DQL also differs from the SQL and from energy-based quantum speed or metrological limits. In stationary sensing, the SQL scales as
\[
S_{\rm SQL}(\Omega)=\hbar|\chi^{-1}(\Omega)|,
\]
while the DQL is
\[
S_{\rm DQL}(\Omega)=\hbar|\Im\chi^{-1}(\Omega)|.
\]
The QCRB or Energetic Quantum Limit instead constrains sensitivity for finite probing strength \(S_{FF}\) [2011.14716]. In the open-system speed-limit setting, the unitary Mandelstam–Tamm term \(2\Delta H_0\sin\Theta_t\) is only one part of the full bound; dissipative deformation and fluctuations contribute independently [2507.02501].

A plausible implication is that “Dissipative Quantum Limit” functions best as a family resemblance term: it names irreducible constraints that arise specifically because a quantum system is open, dissipative, and coupled to an environment. In every formulation summarized here, the decisive mathematical object is the dissipative part of the generator—\(\Im\chi^{-1}\), \(\chi_a^{-1}(t,t')\), \(\mathcal{D}[L]\), or the Liouvillian gap and associated nonequilibrium metric—and the limit concerns what cannot be overcome once that dissipative structure is fixed.

Source: https://www.emergentmind.com/topics/dissipative-quantum-limit-dql