Violating the All-or-Nothing Picture of Local Charges in Non-Hermitian Bosonic Chains

This paper refutes the "all-or-nothing" expectation for local commuting charges in non-Hermitian bosonic chains by constructing explicit counterexamples with selective k-local charges, thereby demonstrating the limited universality of the Grabowski–Mathieu integrability test and providing a complete classification of such models.

The Gist

Imagine you are a detective trying to solve a mystery about a very long, complex machine made of quantum particles. This machine is a "bosonic chain," which you can think of as a row of infinite buckets (sites) where particles (bosons) can hop between neighbors.

For decades, physicists have believed in a simple rule of thumb about these machines, which we'll call the "All-or-Nothing Law."

The Old Rule: The "All-or-Nothing" Law

The law said: "If you find one secret key (a local charge) that keeps the machine running smoothly without changing its state, you will find infinite keys. The machine is perfectly solvable (integrable). If you can't find even one key, the machine is chaotic and unsolvable."

It was like a light switch: either the whole house is lit (perfectly ordered), or it's pitch black (chaotic). There was no middle ground.

The New Discovery: The "Broken Switch"

In this paper, the authors (Mizuki Yamaguchi and Naoto Shiraishi) built a specific type of machine where this rule breaks. They found machines that have some keys, but not all of them.

Think of it like a lockbox that has a working key for the first drawer, but the second drawer is jammed, the third works, the fourth is missing, and the fifth works again. It's a "partially working" system.

The Two Strange Machines They Found

The authors constructed two specific examples of these "broken" machines using non-Hermitian physics (a fancy way of saying the machine can gain or lose energy, like a leaky bucket).

1. The "One-and-Done" Machine (Type N+)

• The Analogy: Imagine a music box that plays a perfect melody for the first three notes. You try to find the key to play the fourth note, but it's missing. You try for the fifth, sixth, and seventh notes, and they are all missing too.
• The Science: They found a system with a 3-site key (a rule that works for three buckets at a time) but no keys for 4, 5, 6, or any larger number of buckets.
• The Lesson: Just because you found a small key doesn't mean the whole machine is solvable. The "All-or-Nothing" law is wrong.

2. The "Missing Link" Machine (Type C-)

• The Analogy: Imagine a train with many cars. The engine works, and cars 1, 2, and 3 are connected. But then, car 4 is missing. The train skips right over car 4 and connects car 5 to car 6, and so on, all the way to the end. The train is mostly connected, but that one specific link is gone.
• The Science: They found a system with keys for 3 buckets, and then keys for 5, 6, 7, 8... all the way up. But there is no key for 4 buckets.
• The Lesson: You can have a system that is "almost" perfectly solvable, but with a very specific, weird gap in the middle.

Why This Matters

1. The "3-Note" Test is Flawed
Previously, scientists had a test called the Grabowski–Mathieu test. It was like saying, "If you can find a 3-note melody that repeats, the whole song is solvable."
The authors proved this test is not universal. You can find that 3-note melody and still have a broken machine. You can't just check the smallest piece and assume the rest follows.

2. The "Bottom-Up" Detective Work
Usually, physicists find a solution first (like a magic formula) and then look for the keys. These authors did it backward. They started by asking, "What if we force a key to exist here? What happens?" This "bottom-up" approach allowed them to discover new, previously unknown machines that are perfectly solvable, which no one had found before.

3. The "Odd-Even" Surprise
They also found a machine that behaves differently depending on whether the total number of buckets is even or odd.

• If the chain has an even number of buckets, it has infinite keys.
• If the chain has an odd number of buckets, it has zero keys.
  It's like a door that only opens if the number of people in the room is even. This "parity sensitivity" is a completely new phenomenon.

The Big Picture

In the world of quantum physics, we often look for patterns to simplify complex problems. This paper shows that the universe is a bit more mischievous than we thought. In systems where particles can be created or destroyed (bosons) and where energy isn't perfectly conserved (non-Hermitian), the rules of order are much more complex.

Instead of a simple "All or Nothing" switch, we now know there is a whole spectrum of "Partial Integrability." Some systems are half-solvable, some are almost-solvable with a glitch, and some are only solvable if the room size is even.

It's a reminder that in the quantum world, sometimes the middle ground is the most interesting place to be.

Technical Summary

Here is a detailed technical summary of the paper "Violating the All-or-Nothing Picture of Local Charges in Non-Hermitian Bosonic Chains" by Mizuki Yamaguchi and Naoto Shiraishi.

1. Problem Statement

The paper addresses the fundamental question of quantum integrability in one-dimensional interacting many-body systems.

• The "All-or-Nothing" Empirical Rule: Historically, it has been widely believed that local commuting charges (conserved quantities) in integrable systems exhibit an "all-or-nothing" behavior. Specifically, if a system possesses a local charge of range $k$ (supported on $k$ consecutive sites), it is expected to possess such charges for all $k \geq 3$. Conversely, non-integrable systems possess no such charges beyond trivial global symmetries.
• The Grabowski–Mathieu Test: Based on this rule, Grabowski and Mathieu proposed that the existence of a 3-local charge is a sufficient diagnostic for complete integrability. If a 3-local charge exists, the system is integrable; if not, it is non-integrable.
• The Gap: While this picture holds for many spin systems, it remains unverified and less explored in bosonic chains, particularly those with non-Hermitian Hamiltonians. The unbounded local Hilbert space of bosons and the non-normality of non-Hermitian operators make standard diagnostic tools (like level statistics) difficult to apply.

Core Question: Does the "all-or-nothing" picture of local charges hold universally for translationally invariant bosonic chains with symmetric nearest-neighbor hopping and general (possibly non-Hermitian) on-site terms?

2. Methodology

The authors employ a bottom-up, direct analytic approach to classify integrability without relying on the existence of an R-matrix or Bethe ansatz solution.

• Model Definition: They analyze a translationally invariant bosonic chain of length $N$ with periodic boundary conditions:
  $$H = \sum_{i=1}^N (b^\dagger_i b_{i+1} + b_i b^\dagger_{i+1} + g_i)$$
  where $g$ is a general on-site operator (potentially non-Hermitian) expanded as $g = \sum c_{xy} (b^\dagger)^x b^y$.
• Charge Definition: A $k$-local charge $Q_k$ is an operator commuting with $H$ ($[Q_k, H]=0$) supported on at most $k$ sites, with at least one term of exactly length $k$.
• Expansion and Linear Algebra:
  1. Basis Expansion: They expand the candidate charge $Q_k$ and the commutator $[Q_k, H]$ in a basis of symbol sequences (e.g., $(b^\dagger)^x b^y$).
  2. Step-by-Step Constraints: The condition $[Q_k, H] = 0$ is treated as a system of linear equations for the expansion coefficients. They solve this iteratively:
     • Step 1: Enforce that the $(k+1)$-local part of the commutator vanishes. This constrains the form of $Q_k$ to specific boundary terms.
     • Step 2: Enforce that the $k$-local part vanishes. This further restricts the coefficients and relates them to the on-site term $g$.
     • Step 3 (and beyond): Enforce the vanishing of lower-order local terms. If the system becomes inconsistent at any step, no charge exists.
• Symmetry Decomposition: The analysis is split into uniform sectors (translation eigenvalue $+1$) and staggered sectors (translation eigenvalue $-1$).
• Displacement Transformation: To simplify the classification, they utilize the displacement operator $D(z) = \exp(z b^\dagger - z^* b)$ to eliminate certain coefficients in $g$ without changing the integrability properties.

3. Key Contributions and Results

A. Violation of the All-or-Nothing Picture

The authors explicitly construct partially integrable counterexamples where the hierarchy of local charges is broken. They identify four distinct types of behavior in the uniform sector:

1. Type C (Completely Integrable): Charges exist for all $k \geq 3$.
2. Type N (Non-integrable): No charges exist for any $k \geq 3$.
3. Type N+ (Partially Integrable): A 3-local charge exists, but no $k$-local charges exist for $k \geq 4$.
   • Example: $g = c_{40}(b^\dagger)^4 + c_{11}n$ (with specific constraints).
4. Type C− (Partially Integrable): A 3-local charge exists, and charges exist for all $k \geq 5$, but no 4-local charge exists.
   • Example: $g = c_{30}(b^\dagger)^3 + c_{11}n + c_{01}b$ (with specific constraints).

Significance: These results prove that the existence of a 3-local charge does not guarantee the existence of higher-range charges. Consequently, the Grabowski–Mathieu test based solely on 3-local charges is not universally applicable to non-Hermitian bosonic chains.

B. Exhaustive Classification

The paper provides a complete classification of the model family based on the coefficients of the on-site term $g$:

• Uniform Sector: The system is classified into Types C, N, N+, or C− depending on the specific algebraic relations between coefficients (e.g., $c_{30}(c_{11}+2) = 4c_{40}c_{01}$).
• Staggered Sector: A separate classification reveals Types SC (all $k$), SN (none), and SN+ (only $k=3$).
• New Integrable Models: The bottom-up approach identifies several new completely integrable models (Type C) that were previously unknown, including specific forms of on-site terms involving cubic and quartic bosonic interactions.

C. Even-Odd Sensitivity

The authors discover a model (Type N $\cap$ Type SC) where the existence of local charges depends drastically on the parity of the system size $N$.

• For even $N$, the system possesses infinitely many staggered local charges.
• For odd $N$, it possesses none.
  This highlights a unique sensitivity in non-Hermitian bosonic systems not typically seen in standard spin chains.

D. The Bose-Hubbard Model

The paper rigorously proves that the standard Bose-Hubbard model (with interaction $U \neq 0$) falls into Type N (non-integrable) in both uniform and staggered sectors, confirming the absence of nontrivial local charges for $U \neq 0$.

4. Significance and Implications

• Refinement of Integrability Diagnostics: The work challenges the universality of the "3-local charge implies integrability" heuristic. It suggests that for certain classes of systems (specifically non-Hermitian bosons), the "decisive range" $M$ for integrability tests might be higher (e.g., $M=5$ for Type C− models).
• New Integrable Physics: By discovering new integrable models via a bottom-up approach, the paper expands the landscape of exactly solvable systems. It raises the question of whether these new models (especially Type C−) admit a standard exact solution (like an R-matrix) or represent a genuinely new class of integrability.
• Role of Non-Hermiticity and Bosons: The results suggest that the violation of the all-or-nothing rule is driven more by the infinite-dimensional local Hilbert space of bosons than by non-Hermiticity itself, though the non-Hermitian sector is where these partial integrabilities manifest most clearly.
• Methodological Advance: The paper demonstrates the power of direct analytic methods (expanding commutators) to classify integrability and discover new models, offering an alternative to the traditional top-down approach (finding a solution first, then charges).

Conclusion

Yamaguchi and Shiraishi provide a rigorous map of integrability for non-Hermitian bosonic chains. They demonstrate that the landscape of local charges is far richer than previously thought, featuring "partially integrable" phases where the charge hierarchy is broken at specific ranges (e.g., missing $k=4$). This work necessitates a re-evaluation of integrability tests in bosonic and non-Hermitian systems and uncovers new exactly solvable models.