Stable Entanglement in a PT-Symmetric Non-Hermitian Double Jaynes–Cummings Model

Quantum entanglement is famously delicate: disturb the particles, allow information to leak away, or let the surrounding environment interact with them, and the correlations that make entanglement useful can rapidly disappear. A theoretical study now suggests that carefully balanced gain and loss may offer a surprising route to preserving those correlations. In work published in Quantum Information Processing, Bao-gang Zhu, Ze-kai Tian, Yi-Lin Yang, Zhong-Xiao Man and colleagues analyze a non-Hermitian double Jaynes–Cummings model in which two atom–cavity systems are governed by parity–time, or (\mathcal{P}\mathcal{T}), symmetry. Their calculations reveal sharply different entanglement dynamics on either side of a symmetry-breaking transition. In one regime, entanglement repeatedly vanishes and returns. In another, it can settle toward a nonzero value under suitable conditions, suggesting a mechanism for stabilizing quantum correlations in systems that are intrinsically open rather than perfectly isolated.

The Jaynes–Cummings model is one of quantum optics’ foundational descriptions. It captures the interaction between a two-level atom—an idealized quantum system with a ground state and an excited state—and a quantized electromagnetic field inside a cavity. When the atom and field are close to resonance, energy can oscillate between them: an excitation in the atom becomes a photon, and the photon can later re-excite the atom. These coherent exchanges are known as Rabi oscillations. A double Jaynes–Cummings model contains two such atom–cavity pairs, allowing researchers to study how quantum correlations move among two atoms and two field modes. The new work adds non-Hermitian terms to this arrangement. Rather than describing a closed system with a Hermitian Hamiltonian, whose energy eigenvalues are conventionally real, the model includes effective gain and loss—mathematical representations of amplification, dissipation or controlled coupling to external reservoirs.

Non-Hermitian physics does not mean that quantum mechanics has simply been discarded. In practical models, a non-Hermitian Hamiltonian often acts as an effective description of a subsystem that exchanges energy or particles with its environment. A lossy cavity, for example, can be represented by a term that removes amplitude, while an externally driven or amplified component can contribute an opposing gain term. If these processes are arranged with the right spatial or modal relationship, the system may possess (\mathcal{P}\mathcal{T}) symmetry. Here, parity reverses the relevant spatial or structural coordinate, while time reversal changes the direction of dynamical evolution and complex conjugates quantities such as the wavefunction. A (\mathcal{P}\mathcal{T})-symmetric system can therefore balance loss in one part against gain in another. Below a critical interaction strength or gain–loss threshold, its eigenvalues may remain real, corresponding to the unbroken symmetry phase. Beyond that threshold, eigenvalues generally form complex-conjugate pairs, and the system enters the (\mathcal{P}\mathcal{T})-symmetry-breaking phase.

Zhu and colleagues investigate how that transition reshapes the evolution of an initially entangled state. Their analysis follows both the entanglement between the atomic subsystems and the atomic population inversion, a quantity that measures the difference between excited- and ground-state populations. In ordinary cavity quantum electrodynamics, these observables are closely linked to the exchange of excitations between atoms and photons. As photons are absorbed and emitted, atomic populations oscillate, while correlations can be transferred from atoms to fields and back again. The researchers use the model’s dynamical equations to track these processes over time while varying the coupling constant, which controls the strength of the interaction between the atoms and their cavity modes. The result is not a single universal pattern: changing the coupling can move the system from coherent, symmetry-preserving behavior into a regime dominated by non-Hermitian amplification and attenuation.

In the (\mathcal{P}\mathcal{T})-symmetric phase, the atom–photon interaction remains sufficiently balanced to produce Rabi oscillations. The population inversion changes periodically, reflecting the repeated conversion of atomic excitation into photonic excitation and back again. Entanglement, however, follows a more dramatic trajectory. The calculations show episodes of entanglement sudden death, in which the measured quantum correlation falls to zero over a finite interval, followed by entanglement sudden birth, when the correlation reappears. These effects are not necessarily signs that the underlying quantum state has been destroyed permanently. In a multipartite system, entanglement can migrate between different pairs or become temporarily hidden in correlations involving the cavity fields. When the dynamics return some of that correlation to the atomic pair, the atoms can become entangled again. The result resembles a quantum relay in which information repeatedly changes location rather than simply fading away.

The terminology “sudden death” can sound more absolute than it is. In quantum-information theory, entanglement is a property of a chosen partition of a system. If researchers examine the two atoms while ignoring the photons, they calculate a reduced density matrix by tracing out the field degrees of freedom. The resulting atomic state may be separable even while the full atom–field state remains entangled. Measures such as concurrence, often used for two-qubit systems, quantify the strength of the remaining two-atom correlation. A concurrence of zero means that the selected atomic pair has no entanglement according to that measure; it does not imply that every quantum correlation in the complete four-part system has disappeared. The predicted alternation between sudden death and sudden birth therefore highlights how energy exchange, decoherence-like effects and subsystem selection interact. In the model, non-Hermitian gain and loss modify these exchanges without eliminating the possibility of later revival.

The most striking behavior appears after the coupling strength pushes the system into the (\mathcal{P}\mathcal{T})-symmetry-breaking phase. There, the researchers find that entanglement can evolve toward a nonzero value in certain parameter ranges rather than repeatedly collapsing to zero. The population inversion displays a related qualitative change. Instead of maintaining the same simple periodic pattern associated with balanced Rabi exchange, it can develop behavior that reflects the complex eigenvalues of the effective Hamiltonian. In linear non-Hermitian dynamics, an imaginary component of an eigenvalue corresponds to exponential growth or decay of a mode. Physical implementations must ultimately account for normalization and for the reservoirs that create the gain and loss, but within the effective model these modes can select which components of the quantum state dominate at long times. That mode selection appears to be central to the persistence of a nonzero entanglement signal.

The study also identifies the coupling constant as a control knob for the transition between the two dynamical regimes. In the language of the model, increasing the atom–field interaction changes the balance between coherent exchange and the non-Hermitian terms, eventually carrying the system from unbroken to broken (\mathcal{P}\mathcal{T}) symmetry. Such transitions are often associated with exceptional points, parameter values at which eigenvalues and their corresponding eigenvectors coalesce. Near an exceptional point, small changes in system parameters can produce disproportionately large changes in the spectrum and in the time evolution. The source material does not report an experimental observation of an exceptional point in this particular double Jaynes–Cummings setup, nor does it provide a laboratory device or measured data. Instead, the work offers a theoretical map of how symmetry, coupling and initial entanglement could jointly determine the fate of quantum correlations. Its significance lies in identifying stable behavior within a framework usually associated with loss and instability.

That possibility could matter for quantum technologies, although the study is not a demonstration of a working quantum memory or processor. Entanglement is a resource for quantum communication, sensing and computation, but maintaining it in real devices requires managing unavoidable interactions with the environment. Conventional strategies suppress noise through isolation, error correction or engineered reservoirs. A (\mathcal{P}\mathcal{T})-symmetric strategy would take a different approach: instead of treating gain and loss solely as enemies, it would shape them so that the system’s preferred dynamical modes preserve useful correlations. The double Jaynes–Cummings model provides a compact theoretical test bed for that idea because it includes both discrete quantum emitters and quantized light. Future work will need to determine whether the predicted nonzero entanglement survives realistic noise, fluctuations in the gain and loss rates, imperfect resonance, thermal photons and the complications of implementing balanced amplification without adding extra quantum noise. For now, the calculations suggest that in an open quantum world, entanglement may not need perfect isolation to endure—it may instead require carefully designed imperfection.