Kramers-Kronig Relations and Causality in Non-Markovian Open Quantum Dynamics: Kernel, State, and Effective Kernel
Kejun Liu · 2026
Read Kramers-Kronig Relations and Causality in Non-Markovian Open Quantum Dynamics: Kernel, State, and Effective Kernel by Kejun Liu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-17058.txt.
Introduction
Liu links three traits of a memory kernel for open quantum systems: delayed support, upper-half-plane holomorphy, and Hardy boundary control. He gives conditions linking them for the Nakajima-Zwanzig kernel when the system space is finite-dimensional. The paper also derives Kramers-Kronig formulas, treats a perturbative force-fit equation, and proves them in Lean 4, closing with a Jaynes-Cummings check.
Catalog record
| CREATOR | Kejun Liu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 105472 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.17058
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Kramers-Kronig Relations and Causality in Non-Markovian Open Quantum Dynamics: Kernel, State, and Effective Kernel Kejun Liu Abstract Kramers-Kronig (KK) relations have not, to the best of our knowledge, been systematically formulated for the central object of non-Markovian open quantum dynamics: the Nakajima-Zwanzig (NZ) memory kernel. We organize the analysis around three distinct objects – the memory kernel 𝒦 ~ ( z ) \tilde{\mathcal{K}}(z) , the reduced-state Laplace transform σ ~ ( z ) \tilde{\sigma}(z) , and the force-fit effective kernel 𝒦 ~ eff ( z ) \tilde{\mathcal{K}}_{\text{eff}}(z) – each with different sensitivity to initial system-bath correlations. Under an explicit real-axis spectral-representation hypothesis for the projected generator, with a coupling-weighted spectral density in L 1 ∩ L p L^{1}\cap L^{p} , 𝒦 ~ ( z ) \tilde{\mathcal{K}}(z) belongs to the vector-valued Hardy space H + p ( ℬ ) H^{p}_{+}(\mathcal{B}) and obeys the corresponding KK or subtracted KK relations. Three consequences follow: a CPTP-Hardy consistency criterion (uncancelled upper-half-plane kernel poles imply dynamics incompatible with uniformly bounded CPTP reduction), a passivity-analyticity compatibility statement via the Herglotz-Nevanlinna class for passive bosonic baths, and a finite-truncation Carleman diagnostic for Padé reconstructions. The reduced-state transform σ ~ ( z ) \tilde{\sigma}(z) is analytic in the upper half-plane for any initial state, factorized or correlated; microscopic unitarity forces ‖ σ ( t ) ‖ op ≤ 1 \|\sigma(t)\|_{\mathrm{op}}\leq 1 . In scalar channels, the force-fit effective kernel 𝒦 ~ eff ( z ) = 𝒦 ~ ( z ) + ℐ ~ ( z ) σ ~ ( z ) − 1 \tilde{\mathcal{K}}_{\text{eff}}(z)=\tilde{\mathcal{K}}(z)+\tilde{\mathcal{I}}(z)\tilde{\sigma}(z)^{-1} can develop upper-half-plane poles at simple zeros of σ ~ ( z ) \tilde{\sigma}(z) . Passive bosonic baths occupy a robust regime where these zeros stay on the real axis, while near-resonant systems can enter a fragile regime in which coherence-channel zeros migrate into the upper half-plane – an intrinsic property of the channel symmetry, existing already for factorized initial states, with initial correlations controlling only whether the inhomogeneous numerator fails to cancel the pole. We verify the bath-correlator KK at Born order (a proxy check), the robust-regime self-consistency for the independent boson model, and – in the Jaynes-Cummings model, whose projected generator has real spectrum – the full operator-valued KK relation for the extracted 4 × 4 4\times 4 NZ memory kernel. The measured integrated relative L 2 L_{2} residual of 3.8 % 3.8\% sits at or below the circular FFT-Hilbert pipeline’s calibrated noise floor ( ≈ 5 % {\approx}5\% at N t = 4000 N_{t}{=}4000 on a known-analytic test function), so the data are consistent with exact matrix-valued KK within the protocol’s accuracy and do not certify a physical violation. In the Jaynes-Cummings model, a scalar force-fit extraction of the coherence-channel effective kernel gives a slight increase of the KK residual under initial correlation (ratio ≈ 1.17 \approx 1.17 at Laplace shift ε = 0.3 \varepsilon=0.3 , rising where finite-truncation leakage dominates) – a qualitative finite-truncation diagnostic rather than a quantitative prediction. I Introduction Apparent acausality in reduced open-quantum-system dynamics does not, in general, signal a breakdown of physical causality. When a modeller discards the inhomogeneous initial-correlation term and force-fits a homogeneous Volterra equation to correlated initial data, the resulting effective memory kernel can develop upper-half-plane poles. In scalar channels, the pole mechanism is simple: an uncancelled simple zero of the reduced-state Laplace transform \tilde{\sigma}(z) becomes a pole of \tilde{\mathcal{K}}_{\mathrm{eff}}(z) through the quotient term \tilde{\mathcal{I}}(z)\tilde{\sigma}(z)^{-1} . The reduced-state transform itself is never acausal, a consequence of microscopic unitarity: \|\sigma(t)\|_{\mathrm{op}}\leq 1 forces \tilde{\sigma}(z) to be analytic in \mathbb{H}_{+} for any initial state, factorized or correlated. The exact memory kernel is also well-behaved when the projected dynamics admits a real-axis spectral representation with an L^{p} coupling-weighted density; in that case it belongs to the vector-valued Hardy space H^{p}_{+}(\mathcal{B}) and obeys a rigorous Kramers-Kronig (KK) relation. Only the force-fit kernel \tilde{\mathcal{K}}_{\mathrm{eff}}(z)=\tilde{\mathcal{K}}(z)+\tilde{\mathcal{I}}(z)\tilde{\sigma}(z)^{-1} can fail KK, and in scalar channels it does so at frequencies determined by uncancelled UHP zeros of \tilde{\sigma}(z) . Recognising these three distinct analytic objects — the kernel, the state, and the force-fit effective kernel — localises the geometric origin of “acausality” in the complex plane and suggests a possible frequency-domain residual in spectroscopic reconstructions. This mechanism is the open-quantum-system analogue of the coarse-graining constructions of Gavassino and collaborators for relativistic hydrodynamics and classical field theory 16 ; 15 ; 17 ; 19 , realised here diagnostically rather than constructively. The impact of initial correlations on the complete positivity of reduced maps has been studied extensively from a set-theoretic and master-equation viewpoint 32 ; 35 ; 36 ; 10 ; 33 ; 12 ; 28 : those works ask whether a reduced map or generator can be assigned for a given family of \rho_{0} . By contrast, we ask where in the complex plane the force-fit effective object develops non-analyticity, and relate that geometry to KK and to diagnostic breakdowns of CPTP. Mapping the problem onto the operator-valued Hardy space H^{p}_{+}(\mathcal{B}) converts a static binary check into a geometric diagnostic: a robust/fragile boundary controlled by Laplace-domain zeros, with candidate predictions at the level of specific frequencies and residuals rather than set membership alone. The central objects — memory kernel \tilde{\mathcal{K}}(z) 31 ; 41 ; 29 ; 6 ; 39 , reduced-state Laplace transform \tilde{\sigma}(z) , and force-fit effective kernel \tilde{\mathcal{K}}_{\mathrm{eff}}(z) — are treated here within one analytic-structure framework, addressing a gap left by prior work that either assumes KK without proof or focuses on numerical kernel extraction 21 ; 26 ; 27 . The paper establishes two analytic-structure theorems. A causality-preservation theorem (Sec. III ) shows that, under an explicit real-axis spectral hypothesis on the projected generator, the NZ memory kernel lies in H^{p}_{+}(\mathcal{B}) in an admissible range of p . A state-robustness theorem (Sec. V ) shows that the reduced-state Laplace transform \tilde{\sigma}(z) is analytic in \mathbb{H}_{+} for any initial state, factorized or correlated; this is the precise statement that acausality cannot be blamed on the state alone. The Hardy framework then supplies four diagnostic tools (Sec. IV ): KK reconstruction formulas with a subtracted variant for algebraic tails; a CPTP-Hardy consistency criterion, by which generic uncancelled upper-half-plane poles in approximate kernels imply dynamics incompatible with CPTP; a passivity-analyticity compatibility statement, by which the Hardy class for passive bosonic baths is reduced to the Nevanlinna subclass; and a finite-truncation Carleman diagnostic for MKCT Padé reconstructions. At the force-fit level, a perturbative modified KK relation for scalar-channel effective kernels (Sec. VI ) is accompanied by a parametric robust-versus-fragile criterion controlled by uncancelled zeros of \tilde{\sigma}(z) . Numerically, the full operator-valued KK relation is checked on the extracted 4\times 4 NZ memory kernel of the Jaynes-Cummings model: the measured integrated relative L_{2} residual is 3.8\% , at or below the calibrated circular FFT-Hilbert noise floor ( {\approx}5\% ), providing a concrete consistency check of the matrix-valued Hardy-space claim within the protocol’s accuracy (Sec. VII ). II Preliminaries II.1 Nakajima-Zwanzig projection We consider a composite Hilbert space \mathcal{H}=\mathcal{H}_{s}\otimes\mathcal{H}_{b} with Hamiltonian \hat{H}=\hat{H}_{s}+\hat{H}_{b}+\hat{H}_{sb} and total Liouvillian \mathcal{L}\,\cdot\equiv[\hat{H},\,\cdot\,] . We note that \mathcal{L} is self-adjoint in the Hilbert-Schmidt inner product \langle\!\langle A,B\rangle\!\rangle=\mathrm{Tr}[A^{\dagger}B] : using cyclicity of trace, \langle\!\langle\mathcal{L}A,B\rangle\!\rangle=\mathrm{Tr}[(HA{-}AH)^{\dagger}B]=\mathrm{Tr}[A^{\dagger}HB]-\mathrm{Tr}[A^{\dagger}BH]=\langle\!\langle A,\mathcal{L}B\rangle\!\rangle . This statement concerns \mathcal{L}