The Two Orbital, Interacting Hatano-Nelson Model

This paper investigates the interacting, spinful two-chain Hatano-Nelson model to map out the conditions for a purely real spectrum, analyze the relationship between periodic and open boundary condition modes via winding numbers, and validate the non-Hermitian description's ability to qualitatively capture the low-filling dynamics of such non-equilibrium systems.

The Gist

Imagine you are watching a crowded dance floor where two groups of dancers (let's call them "Up" and "Down" spinners) are trying to move around. In a normal, fair world, if a dancer steps left, they are just as likely to step right. But in this paper, the authors are studying a biased dance floor.

Here, the rules of the dance are "non-Hermitian." This is a fancy physics word that essentially means the floor is slippery in one direction and sticky in the other. If a dancer tries to move forward, they glide easily; if they try to move backward, they get stuck. This creates a "wind" that pushes everyone in one direction.

The Setup: Two Parallel Dance Floors

The authors set up a scene with two parallel dance floors (chains) connected by bridges.

• Floor A has a wind blowing to the right.
• Floor B has a wind blowing to the left.
• The dancers can jump between the floors via bridges (interchain hopping).
• The dancers also dislike being on top of each other (repulsion), so they try to avoid sharing the same spot.

The big question the paper asks is: Under what conditions does this chaotic, wind-blown system settle down into a stable, predictable state?

In physics terms, they are looking for a "Real Spectrum."

• Complex Numbers (The Chaos): If the system is unstable, the dancers' movements are like a swirling vortex. Their energy has a "twist" (imaginary part) that makes them grow or shrink uncontrollably, like a feedback loop in a microphone.
• Real Numbers (The Stability): If the system is stable, the dancers move in a predictable, rhythmic way. Their energy is "real" and steady.

The Key Findings (The "Recipe" for Stability)

The authors discovered that you can turn this chaotic, wind-blown mess into a stable, rhythmic dance by adjusting three main knobs:

1. The Bridge Strength (Interchain Hopping)
Imagine the bridges between the two floors are weak. The wind on Floor A pushes dancers to the right edge, and the wind on Floor B pushes them to the left edge. They get stuck at the walls (this is called the "Skin Effect").

• The Fix: If you make the bridges strong, the dancers can easily jump back and forth between the floors. This mixing cancels out the wind's bias. It's like if two people are being pushed in opposite directions but are holding hands tightly; they stop moving and just stand still.
• The Result: Once the bridges are strong enough, the chaotic "imaginary" energy disappears, and the system becomes stable (Real Spectrum).

2. The Dancer's Grumpiness (Interaction/Repulsion)
What if the dancers really hate being on top of each other?

• The Twist: If they are very grumpy (strong repulsion), they form tight pairs (called "doublons"). These pairs are heavy and hard to push around.
• The Result: Being grumpy actually helps stability, but only if the bridges are also strong. If the bridges are weak, the grumpy pairs get stuck at the edges just like the solo dancers. But if the bridges are strong, the heavy pairs can move freely, and the system stays stable.

3. The Wind Speed (Non-Hermiticity)
If the wind is too strong, even strong bridges might not be enough to stop the chaos. The authors mapped out exactly how strong the bridges need to be depending on how strong the wind is and how grumpy the dancers are.

The "Skin Effect" Analogy

One of the coolest things they found is the Skin Effect.
Imagine a crowd of people in a hallway where everyone is forced to walk to the right. Eventually, everyone piles up against the right wall, leaving the left side empty.

• In this two-floor system, because the winds are opposite, the "Up" dancers pile up on the right wall, and the "Down" dancers pile up on the left wall.
• The paper shows that if the dancers are grumpy (interacting), they pile up even tighter against the walls.

The "Real World" Test: Will it actually happen?

The authors didn't just do math; they simulated what happens if you actually run this system in real life, where things can fall apart (dissipation).

• They asked: "If we start with dancers everywhere, will they eventually pile up at the walls, or will they just vanish?"
• The Answer: Yes! Even in a messy, real-world scenario where dancers can leave the floor (jump operators), the system still shows that "piling up at the walls" behavior for a while before everyone eventually leaves. This proves that the weird "skin effect" isn't just a mathematical trick; it's a real physical phenomenon you could see in experiments.

Why Does This Matter?

This isn't just about electrons. This math applies to:

• Light: Lasers and optical fibers where light moves differently in one direction.
• Dark Matter: Using arrays of sensors to detect invisible particles.
• New Materials: Designing materials that conduct electricity in one direction but not the other.

In a nutshell:
The paper shows that if you have two systems fighting against each other (opposite winds), you can stop the chaos and create a stable, predictable world by connecting them strongly and letting the particles interact. It's a recipe for turning a chaotic, one-way street into a stable, two-way highway.

Technical Summary

1. Problem Statement

The paper investigates the interplay between non-Hermiticity (specifically asymmetric hopping) and electron-electron interactions in a multi-orbital system. While the single-orbital, one-dimensional Hatano-Nelson (HN) model is well-understood—exhibiting a transition from complex eigenvalues (delocalized) to real eigenvalues (localized) upon introducing disorder or interactions—its behavior in a multi-chain geometry with spinful fermions remains less explored.

The specific challenges addressed are:

• How does the presence of on-site Hubbard repulsion ($U$) affect the reality of the energy spectrum in a non-Hermitian system?
• How does interchain coupling ($V_0$) between two HN chains with opposite non-reciprocity ($\delta_A = -\delta_B$) influence the spectral properties?
• What is the relationship between the topological winding number (under Periodic Boundary Conditions, PBC) and the skin effect (under Open Boundary Conditions, OBC) in the interacting regime?
• Can the non-Hermitian description of such systems be validated through Lindbladian dynamics (open quantum system evolution)?

2. Methodology

The authors employ a combination of analytical derivations and numerical simulations:

• Model Hamiltonian: A two-chain (ladder) lattice model with $L$ sites per chain.
  • Intra-chain hopping: Non-Hermitian with asymmetric amplitudes $t \pm \delta_\alpha$, where $\delta_A = -\delta_B = \delta$.
  • Inter-chain hopping: Hermitian with strength $V_0$.
  • Interaction: On-site Hubbard repulsion $U$ between spin-up and spin-down fermions.
  • Sector: The study focuses on the two-particle sector with one spin-up and one spin-down electron ($N_\uparrow = N_\downarrow = 1$).
• Numerical Techniques:
  • Exact Diagonalization (ED): Used to compute the full complex energy spectrum for system sizes up to $L=100$.
  • Winding Number Calculation: A spectral winding number $W(E)$ is computed by threading magnetic fluxes through the rings. Crucially, the authors use an opposite-flux pattern ($\phi_A = -\phi_B$) to match the reversed non-reciprocity of the chains.
  • Skin Mode Analysis: Under OBC, the spatial density profiles of right eigenstates are analyzed to determine localization lengths.
  • Lindbladian Dynamics: The system's time evolution is simulated using the quantum trajectory method to solve the full Lindblad master equation, comparing "no-jump" effective Hamiltonian dynamics with full dissipative dynamics.

3. Key Contributions and Results

A. Phase Diagrams for Real Spectra

The authors map out the conditions under which the energy spectrum becomes purely real (a signature of PT-symmetry unbroken phase or localization).

• Non-interacting Limit ($U=0$): The spectrum becomes purely real when the interchain coupling satisfies $V_0 \ge 2\delta$. This corresponds to the condition where the two bands merge at exceptional points.
• Interacting Regime ($U \neq 0$):
  • Weak Interaction: The critical coupling required for a real spectrum shifts to $V_0 \approx 4t$. This scale arises from the separation of the two-particle scattering continua in the Hermitian limit.
  • Strong Interaction: A high-energy branch of eigenvalues (doublon-like states) emerges near energy $E \approx U$. This branch initially possesses a large imaginary component. To drive the entire spectrum (including this high-energy branch) to the real axis, $V_0$ must be significantly larger than $U$ ($V_0 \gg U$).
  • Conclusion: Interactions do not simply suppress non-Hermiticity; they create new spectral branches that require stronger interchain hybridization to stabilize into a real spectrum.

B. Topology and Winding Numbers

• Point-Gap Topology: In the regime where the high-energy doublon branch is spectrally isolated (enclosing a point gap around $E \approx U$), the system exhibits a non-trivial winding number.
• Quantization: Using the opposite-flux pattern, the total winding number is found to be $W=4$.
• Spin Decomposition: In the two-particle sector, this total winding decomposes additively into spin-resolved contributions: $W_\uparrow = W_\downarrow = 2$. This additivity is a non-trivial result for an interacting system.
• Breakdown: At large $V_0$, the isolated point gap collapses, and the winding number becomes trivial, correlating with the spectrum becoming entirely real.

C. Skin Modes and Boundary Conditions

• Skin Effect: Under OBC, the eigenstates associated with the isolated doublon branch exhibit the non-Hermitian skin effect.
• Opposite Accumulation: Due to the reversed non-reciprocity ($\delta_A = -\delta_B$), the charge density accumulates at opposite edges of the two chains (left edge for chain A, right edge for chain B).
• Localization Length: The skin localization length $\xi$ decreases (localization strengthens) as either the non-reciprocity $\delta$ or the interaction strength $U$ increases.

D. Dynamics and Stability (Lindbladian Evolution)

• No-Jump Evolution: Simulating the system using only the effective non-Hermitian Hamiltonian (ignoring quantum jumps) fails to reproduce clean skin-mode profiles for broad initial states. The dynamics are oscillatory and do not clearly show edge accumulation.
• Full Lindbladian Dynamics: When including jump operators (dissipation), the system exhibits transient edge accumulation consistent with the skin effect on time scales comparable to the system's lifetime. However, particle loss eventually drives the system to the vacuum state.
• Significance: This confirms that the spectral and topological features of the effective non-Hermitian model are physically observable in open quantum systems, provided the observation window is within the dissipative lifetime.

4. Significance

• Generalization of HN Physics: The work extends the understanding of the Hatano-Nelson model from single chains to interacting, multi-orbital systems, revealing that interactions fundamentally alter the criteria for spectral reality.
• Topological-Interaction Interplay: It demonstrates that topological invariants (winding numbers) can remain well-defined and quantized even in the presence of strong interactions, provided the spectral gap structure is preserved.
• Experimental Relevance: The findings provide a theoretical framework for interpreting experiments in synthetic dimensions (e.g., optical frequency modes in ring resonators) and optomechanical sensor arrays where non-reciprocal interactions and many-body effects coexist.
• Dynamical Stability: By linking the static spectral properties to time-dependent Lindbladian evolution, the paper bridges the gap between formal non-Hermitian Hamiltonian theory and the physical reality of open quantum systems.

In summary, the paper establishes that while interchain coupling can restore a real spectrum in interacting non-Hermitian systems, the threshold for this transition is non-monotonic and heavily dependent on the interaction strength, with distinct topological and dynamical signatures emerging in the process.