Pseudo-Hermiticity of the Nakajima-Zwanzig Projected Liouvillian in the Jaynes-Cummings Model

This paper resolves the long-standing anomaly of the purely real spectrum of the non-Hermitian Nakajima-Zwanzig projected Liouvillian in the Jaynes-Cummings model by demonstrating its pseudo-Hermiticity under a positive-definite metric, a structural property that persists through bath truncation and extends to the full Rabi model with re-entrant exceptional-point boundaries.

The Gist

Imagine you are trying to predict how a tiny, vibrating atom (like a light bulb filament) interacts with a sea of invisible waves (light). In the world of quantum physics, this is a messy business because the atom is never truly alone; it's constantly bumping into the environment.

To make sense of this, scientists use a mathematical "filter" called the Nakajima–Zwanzig projection. Think of this filter as a pair of sunglasses that blocks out the chaotic background noise so you can focus only on the atom's behavior. The mathematical engine that drives this filter is called the Projected Liouvillian (let's call it the "Engine").

Here is the puzzle the paper solves:

The Mystery: A Broken Mirror That Still Shows a Clear Image

Usually, when you look at a broken mirror (a non-symmetric mathematical object), the reflection gets distorted. In physics, if an "Engine" is broken (non-Hermitian), its internal gears (its spectrum) usually spin in a chaotic, complex way, making predictions difficult.

However, in a famous model called the Jaynes–Cummings model (which describes a simple atom and a single beam of light), scientists noticed something weird. Even though the Engine looked "broken" on the surface, its internal gears were spinning perfectly in a straight line (a purely real spectrum). It was like seeing a reflection in a shattered mirror that somehow still showed a perfect, undistorted face. For years, no one knew why this happened.

The Solution: The "Magic Frame" (Pseudo-Hermiticity)

The author, Kejun Liu, discovered that the Engine isn't actually broken; it's just wearing a special frame.

In math terms, this is called pseudo-Hermiticity.

• The Analogy: Imagine a wobbly, uneven table (the Engine). If you try to balance a ball on it, the ball rolls off (complex chaos). But, if you place a specific, custom-made mat under the table legs (the metric $\eta$), the table suddenly becomes perfectly level.
• The paper proves that for this specific atom-light model, there exists a "magic mat" (a positive-definite metric) that, when applied, makes the wobbly Engine behave like a perfect, stable machine. This explains why the gears spin in a straight line despite the Engine looking messy.

The Twist: It's Not Just One Big Block

You might think, "Maybe the Engine is just made of smaller, perfect blocks stuck together."
The paper says no.

• The author broke the Engine down into different sections (like different rooms in a house).
• Some rooms were perfectly symmetrical.
• But two specific rooms were actually quite wobbly and broken.
• The Miracle: Even though those two rooms were broken, the "magic mat" covered the entire house, holding everything together so the whole system still worked perfectly. This proves the stability is a deep, structural feature, not just a lucky accident of the building's layout.

The Deformation: Stretching the Model

The author then tested how strong this "magic mat" really is. They took the simple atom-light model and slowly stretched it into a more complex, messy model (the Rabi model) by adding extra, weird interactions.

• Phase 1 (Safe): At the start, the mat works perfectly.
• Phase 2 (The Danger Zone): As they stretched it, the mat got thin and the table wobbled. The gears started spinning in chaos (complex numbers appeared). This is a "danger zone" where the rules of the game break down.
• Phase 3 (Safe Again): Surprisingly, if they stretched it all the way to the end, the mat reappeared! The system became stable again, but this time it was held together by a different kind of symmetry (like a different type of glue).

This "re-entrant" behavior (Safe → Chaos → Safe) shows that the stability is a robust feature of the physics, protected by specific symmetries at the beginning and end of the process.

Why Does This Matter?

The paper concludes that this "magic mat" explains why certain mathematical rules (called Kramers–Kronig relations) work perfectly for this specific atom-light model. These rules are like the laws of cause-and-effect; they ensure that what happens in the future is logically connected to the past.

Because the Engine has this "pseudo-Hermitian" property, we know for sure that the memory of the atom's past interactions behaves in a predictable, oscillating way, rather than decaying into nonsense. This gives a solid structural reason why our standard tools for analyzing light and matter work so well in this specific scenario.

In short: The paper found a hidden "leveling mat" that explains why a messy quantum system behaves with perfect order, proving that this order is a fundamental feature of the model, not a fluke.

Technical Summary

1. Problem Statement

The paper addresses a long-standing anomaly in the theory of open quantum systems, specifically within the Nakajima–Zwanzig (NZ) projection-operator formalism.

• The Context: The NZ formalism describes non-Markovian dynamics via a memory kernel $K(t)$, generated by the projected Liouvillian operator $QLQ$ (where $Q = I - P$ is the projector onto the orthogonal complement of the system subspace, and $L = [H, \cdot]$ is the Liouvillian).
• The Anomaly: While $QLQ$ is generally non-Hermitian (especially when the bath reference state $\rho_B \neq I/d_B$) and typically possesses a complex spectrum, numerical surveys of the Jaynes–Cummings (JC) model revealed that $QLQ$ possesses a purely real spectrum across all accessible parameters.
• The Consequence: A purely real spectrum is crucial for the Hardy-space analyticity of the memory kernel, which validates the Kramers–Kronig (KK) dispersion relations. The community lacked a structural explanation for why the JC model (a canonical light-matter model) exhibited this real spectrum despite the operator's manifest non-Hermiticity.

2. Methodology

The author employs a combination of rigorous numerical linear algebra and symmetry analysis to resolve the anomaly:

• Numerical Construction: The projected Liouvillian $QLQ$ was constructed numerically for the single-mode JC Hamiltonian with varying cutoffs ($N_{max} = 3$ to $20$) and coupling strengths ($g = 0.1$ to $2.0$).
• Biorthonormal Diagonalization: The eigenvalues and eigenvectors of $QLQ$ were computed. Since $QLQ$ is non-Hermitian, a biorthonormal basis of left ($\langle l_n|$) and right ($|r_n\rangle$) eigenvectors was used.
• Metric Construction: Based on the theory of $\eta$-pseudo-Hermiticity (Mostafazadeh), a positive-definite metric operator $\eta$ was constructed using the left eigenvectors:
  $$\eta = \sum_{\lambda_n \neq 0} c_n |l_n\rangle\langle l_n|$$
  The authors verified if this $\eta$ satisfies the intertwining relation $(QLQ)^\dagger \eta = \eta (QLQ)$.
• Sector Analysis: The Hilbert space was decomposed into sectors based on the excitation number difference $\Delta N$. The Hermiticity of individual sectors was tested to determine if the global real spectrum was a trivial consequence of block-diagonal Hermiticity.
• Deformation Study: The JC Hamiltonian was continuously deformed toward the full Rabi model by adding counter-rotating terms ($H(\lambda) = H_{JC} + \lambda g (\sigma_+ a^\dagger + \sigma_- a^\dagger)$) to probe the robustness of the pseudo-Hermitian phase and identify phase boundaries (Exceptional Points).

3. Key Contributions

1. Resolution of the Anomaly: The paper proves that the real spectrum of $QLQ$ in the JC model is not a numerical artifact but a result of genuine pseudo-Hermiticity. A positive-definite metric $\eta$ exists that renders the operator Hermitian under a similarity transformation.
2. Demonstration of "Genuine" Pseudo-Hermiticity: The study refutes the hypothesis that the real spectrum arises simply because individual symmetry sectors are Hermitian.
   • While sectors with $\Delta N = \pm 1, \pm 3, \pm 4$ are individually Hermitian, the $\Delta N = 0$ and $\Delta N = \pm 2$ sectors are significantly non-Hermitian (with residuals of 1.70 and 5.06, respectively).
   • Despite these non-Hermitian sectors, the global operator remains pseudo-Hermitian, meaning the metric $\eta$ couples these sectors to preserve a real spectrum.
3. Discovery of a Re-entrant Phase: Under deformation to the Rabi model, the system exhibits a re-entrant pseudo-Hermitian phase:
   • Phase I ($0 \le \lambda \lesssim 0.39$): Real spectrum protected by U(1) symmetry (excitation number conservation).
   • Complex Window ($0.39 \lesssim \lambda \lesssim 0.87$): Spectrum becomes complex; the metric condition number diverges at Exceptional Points (EPs).
   • Phase II ($0.87 \lesssim \lambda \le 1$): Real spectrum restored, now protected by $Z_2$ parity symmetry of the full Rabi model.
4. Scaling and Convergence: The pseudo-Hermiticity persists in the bath-truncation limit ($N_{max} \to 20$, matrix dimensions up to $1764 \times 1764$), with intertwining residuals remaining below $10^{-11}$.

4. Key Results

• Spectral Reality: For all tested parameters, the maximum imaginary part of the eigenvalues is effectively zero ($< 10^{-14}$), confirming a purely real spectrum.
• Metric Properties:
  • The constructed metric $\eta$ is positive-definite on the non-trivial subspace.
  • The condition number $\kappa(\eta)$ grows as $d^{1.8}$ (where $d$ is the Hilbert space dimension) but remains finite, indicating structural stability.
  • The non-Hermiticity of the operator grows as $d^{0.77}$.
• Sector Non-Hermiticity: Table II in the paper explicitly shows that the global non-Hermitian content is entirely carried by the $\Delta N = 0$ and $\Delta N = \pm 2$ sectors, yet the global spectrum remains real due to the metric.
• Exceptional Points (EPs): The transition between the pseudo-Hermitian phases is marked by EPs. At the lower boundary ($\lambda \approx 0.39$ for $N_{max}=3$), the condition number diverges, signaling a second-order EP. However, at higher truncations ($N_{max} \ge 5$), this divergence vanishes, suggesting the EP becomes a "weak" coalescence in the thermodynamic limit, though the phase separation remains.

5. Significance

• Theoretical Foundation for KK Relations: The results provide a structural reason for the validity of Kramers–Kronig dispersion relations in the canonical Jaynes–Cummings model. The real spectrum ensures the memory kernel has a Cauchy spectral representation, placing it in the Hardy class. This contrasts with models like the spin-boson model, where complex eigenvalues lead to the failure of standard KK relations.
• New Classification of Open Systems: This work extends the concept of pseudo-Hermiticity (previously studied in PT-symmetric quantum mechanics) to projected Liouvillians in open quantum systems. It suggests a new classification program based on the interplay between Hamiltonian symmetries and the NZ projector structure.
• Physical Observability: The distinction between pseudo-Hermitian (purely oscillatory memory kernel) and non-pseudo-Hermitian (decaying memory kernel) dynamics is observable via non-Markovian process tomography.
• Experimental Relevance: The predicted re-entrant phase and the transition at the Exceptional Point could be probed in tunable circuit-QED platforms, offering a new avenue to study "causality phase transitions" and Liouvillian exceptional points.

In conclusion, the paper establishes that the Jaynes–Cummings model's memory kernel generator is structurally pseudo-Hermitian, a property protected by symmetry that ensures the physical consistency of the model's causal response functions, even in the presence of strong non-Hermitian components within specific symmetry sectors.