Failure of the mean-field Hartree approximation for a bosonic many-body system with non-Hermitian Hamiltonian

Introduction

Mean‑field Hartree theory is a cornerstone of many‑body physics, allowing the reduction of a complex interacting system to an effective single‑particle evolution. While its reliability is well‑established for Hermitian Hamiltonians, recent interest in non‑Hermitian models—used to describe particle gain/loss, open‑system dynamics between quantum jumps, and even quantum algorithms for nonlinear differential equations—has prompted the assumption that Hartree’s approach extends unchanged. In a recent study, Ginzburg, Rademacher, and De Palma rigorously test this assumption and uncover fundamental breakdowns of the Hartree approximation in the non‑Hermitian setting.

Why the Hartree Approximation May Fail

The authors first consider the most general bosonic mean‑field Hamiltonian with two‑body interactions. By analysing the exact many‑body Schrödinger dynamics and comparing it with the Hartree equation, they prove that, for non‑Hermitian generators, the second time derivative of the one‑particle marginal at $t\!=\!0$ does not coincide with the Hartree prediction in the thermodynamic limit ($N\to\infty$). This discrepancy already signals that the mean‑field closure—replacing higher‑order correlations by products of one‑particle states—cannot be justified by a simple large‑$N$ argument when the Hamiltonian lacks Hermiticity.

Explicit Counterexample: An Anti‑Hermitian Model

To make the failure concrete, the paper solves analytically a model of $N$ bosonic qubits governed solely by an anti‑Hermitian Hamiltonian. In the $N\to\infty$ limit the exact one‑particle marginal can be computed explicitly. The resulting dynamics differ markedly from the solution of the corresponding non‑Hermitian Hartree equation. Moreover, the authors identify an initial pure state that evolves into a mixed state after a finite time—a transition that never occurs in Hermitian mean‑field dynamics, where purity is preserved. This finite‑time mixing illustrates a genuinely non‑Hermitian effect that the Hartree framework cannot capture.

Beyond Purely Anti‑Hermitian Cases

The investigation does not stop at pathological anti‑Hermitian examples. The authors also simulate a non‑Hermitian Hamiltonian with a real spectrum—an ostensibly benign situation often assumed to be safe for mean‑field methods. Even here, they find families of initial conditions for which the Hartree equation fails to reproduce the correct one‑particle marginal in the large‑$N$ limit. These results suggest that the presence of a real spectrum alone is insufficient to guarantee Hartree validity.

Implications and Outlook

The findings challenge the prevailing belief that the Hartree approximation is universally applicable to non‑Hermitian many‑body systems. They imply that additional criteria—beyond particle number scaling and spectral reality—must be imposed to ensure a faithful mean‑field description. Potential directions include:

• Deriving rigorous error bounds that explicitly depend on the non‑Hermitian part of the Hamiltonian.
• Identifying subclasses of non‑Hermitian interactions (e.g., PT‑symmetric or dissipative terms with specific structure) where Hartree remains accurate.
• Developing alternative reduced‑dynamics frameworks, such as stochastic mean‑field or cluster expansions, tailored to open quantum systems.

Practical Takeaway

Researchers employing Hartree‑type mean‑field models for bosonic systems with gain, loss, or other non‑Hermitian effects should verify the approximation’s validity on a case‑by‑case basis. Simple checks—such as comparing early‑time derivatives of observables or testing purity preservation—can reveal hidden inconsistencies before large‑scale simulations are undertaken.

--- This summary distills the core arguments and results of the paper without introducing new data or quotations, offering a concise guide for physicists interested in the reliability of mean‑field methods in non‑Hermitian quantum many‑body problems.