🔢 mathematics

Parent Hamiltonians of Ergodic Matrix Product States

This paper investigates parent Hamiltonians for ergodic matrix product states (EMPS) defined by random tensors, demonstrating that under mild injectivity assumptions, these states serve as unique frustration-free ground states in the thermodynamic limit while establishing conditions for spectral gaps using the martingale method and illustrating scenarios where such gaps may or may not exist.

Original authors: Owen Ekblad, Eloy Moreno-Nadales, Eric B. Roon, Jeffrey H. Schenker

Published 2026-07-03

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Owen Ekblad, Eloy Moreno-Nadales, Eric B. Roon, Jeffrey H. Schenker

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: Building a House with Random Bricks

Imagine you are trying to build a very long, perfect wall (representing a quantum spin chain). In the world of physics, there is a special type of blueprint called a Matrix Product State (MPS). Think of an MPS as a set of instructions for laying down bricks.

Usually, these instructions are Translation-Invariant (TI). This means the blueprint is the same everywhere. If you look at the instructions for the first brick, they are identical to the instructions for the hundredth brick. It's like a factory stamping out identical Lego bricks and following the exact same pattern forever. In this standard world, we know that if you follow these instructions, the resulting wall is stable, and there is a specific "Parent Hamiltonian" (a set of physical laws or forces) that makes this wall the most stable, lowest-energy structure possible.

This paper asks: What happens if the instructions aren't identical?

The authors look at a scenario where the instructions change from site to site, but in a statistically consistent way. Imagine you have a bag of different colored bricks. You pull one out at random for every spot on the wall. You can't predict exactly which color is at spot #50, but you know the distribution of colors is the same everywhere. This is called an Ergodic Matrix Product State (EMPS).

The paper investigates the "Parent Hamiltonian" for these random walls. In physics, the Parent Hamiltonian is the set of rules that says, "This specific wall is the perfect, most stable state."

The Main Discoveries

1. The Wall is Still Unique (Theorem A)

In the standard, identical-brick world, we know the wall is the only perfect state for its set of rules. The authors prove that even with random bricks, as long as the randomness isn't "broken" (a technical condition called injectivity), the resulting wall is still the unique most stable state.

The Catch: In the standard world, the rules (the Parent Hamiltonian) usually only care about a brick and its immediate neighbors (finite-range). In this random world, the rules might need to look at a brick and its neighbor 100 spots away, or 1,000 spots away. The "reach" of the rules is random and can be very long.
The Analogy: Imagine a neighborhood where the rule for your house depends on your immediate neighbor (standard). In this new random neighborhood, the rule for your house might depend on the person living 50 houses down the street. The authors prove that even with these long-distance, random rules, the neighborhood still has one unique, perfectly stable configuration.

2. Is the Wall Sturdy? (The Spectral Gap)

Physicists care about whether a system has a spectral gap. Think of this as a "safety buffer."

Gapped (Sturdy): If you try to shake the wall, it takes a lot of energy to make it wobble. It snaps back quickly.
Gapless (Wobbly): If you try to shake it, it wobbles easily with very little energy. It's unstable.

In the standard, identical-brick world, these walls are almost always sturdy (gapped). However, the authors point out that in this random world, the wall can be wobbly (gapless). They found a specific example (a disordered version of a famous model called AKLT) where the randomness makes the wall unstable.

3. When Does the Wall Stay Sturdy? (Theorem B)

The authors didn't just say "it might be wobbly." They figured out exactly when it stays sturdy.

They used a mathematical tool called the Martingale Method (think of it as a way to track the "average" behavior of the randomness over time). They found that the wall is sturdy if the randomness "settles down" quickly enough.

The "Onset Length": Imagine you are looking at the wall. For the first few bricks, the randomness is chaotic and unpredictable. But after a certain distance (the "onset length"), the pattern becomes predictable and stable.
The Result: If this "chaotic zone" at the start is short and consistent everywhere, the wall is sturdy (gapped). If the chaotic zone keeps getting longer and longer as you move down the line, the wall becomes wobbly (gapless).

The "AKLT" Example

The paper uses a specific example called the Disordered AKLT model to show this in action.

The Standard AKLT: A perfectly ordered, sturdy wall.
The Disordered AKLT: They took the standard instructions and added random noise.
- If the noise is "mild" (the chaotic zone is short), the wall remains sturdy.
- If the noise is "wild" (the chaotic zone stretches out infinitely), the wall becomes wobbly.

This example proves that the difference between a sturdy and a wobbly wall in this random world comes down to how quickly the local randomness "calms down" into a predictable pattern.

Summary

The Problem: How do we describe the physical rules (Parent Hamiltonians) for quantum systems built with random, changing parts?
The Finding: Even with random parts, there is a unique, perfect state. However, the rules governing it can be very long-range (looking far down the line).
The Stability: These systems aren't always stable. They can be "wobbly" (gapless).
The Condition for Stability: The system is stable only if the local randomness "settles down" quickly. If the randomness stays chaotic for too long, the system loses its stability.

The paper essentially provides a map for understanding when these random quantum systems are solid and when they are fragile, filling a gap in our understanding of how disorder affects quantum matter.

Technical Summary: Parent Hamiltonians of Ergodic Matrix Product States

Problem Statement
Matrix Product States (MPS) are fundamental in condensed matter theory, typically characterized as unique, frustration-free, gapped ground states of local "parent Hamiltonians" in the translation-invariant (TI) regime. However, recent investigations have extended MPS to systems with on-site disorder, specifically Ergodic Matrix Product States (EMPS), where the defining tensors are sampled from an ergodic stochastic process rather than being constant. A critical discrepancy exists between the TI and ergodic regimes: while TI MPS generically possess gapped parent Hamiltonians, recent work [36] demonstrated that certain injective EMPS are ground states of gapless, finite-range interactions. This paper addresses the general properties of parent Hamiltonians for injective EMPS, investigating whether they are unique ground states, the nature of their interaction range, and the conditions under which they exhibit a spectral gap.

Methodology
The authors employ a combination of operator algebraic techniques, ergodic theory, and the martingale method.

Ergodic Quantum Processes: The framework utilizes ergodic quantum processes $(T, \phi)$ acting on the space of $D \times D$ matrices, where $T$ is an invertible measure-preserving transformation and $\phi$ is a unital completely positive (UCP) map. The EMPS is defined via a transfer apparatus $(E^{[k]})_{k \in \mathbb{Z}}$ derived from site-dependent random tensors.
Injectivity and Transfer Apparatus: The study relies on the concept of "injectivity," where the map $\Gamma^{[l,r]}$ (mapping auxiliary space matrices to physical space states) becomes injective for sufficiently large intervals. The authors establish the equivalence between finite injectivity length and the "eventual strict positivity" (ESP) of the transfer apparatus, a property ensuring exponential decay of correlations.
Construction of Parent Hamiltonians: Unlike the TI case where a fixed interaction range $R$ suffices, the authors construct parent Hamiltonians using "injectivity sequences" $(L_x)_{x \in \mathbb{Z}}$ . These sequences are random variables determining the support of local interaction terms. The construction ensures the Hamiltonian is locally finite-ranged but potentially infinite-ranged globally, depending on the unboundedness of the injectivity length.
Martingale Method for Spectral Gaps: To analyze the spectral gap, the authors adapt the martingale method (Nachtergaele [20]) to the disordered setting. This involves "stochastic regrouping" of the spin chain into blocks defined by the random injectivity lengths. They derive lower bounds on the spectral gap by tracking local statistical quantities, specifically the decay constants of the transfer map and the eigenvalues of the random density matrices $\rho_k$ .

Key Contributions and Results

Existence and Uniqueness of Parent Hamiltonians (Theorem A):
The authors prove that the thermodynamic limit of an injective EMPS is a pure state. Furthermore, there exists an ergodic, random, locally finite-ranged interaction $\Psi$ such that the EMPS is the unique frustration-free ground state of $\Psi$ .
- Significance: Unlike the TI case, the interaction range of $\Psi$ is not necessarily bounded. The support of local terms $h[k]$ is determined by a random length scale $R[k] - k$ , which may be unbounded. Despite this, the authors show that $\Psi$ generates a strongly continuous group of automorphisms on the quasi-local algebra, satisfying the necessary conditions for dynamics even without a uniform Lieb-Robinson bound.
Spectral Gap Conditions (Theorem B):
The paper establishes sufficient conditions for the parent Hamiltonian to possess a spectral gap. Using the martingale method, they derive a lower bound for the spectral gap of finite-volume restrictions of the Hamiltonian.
- The gap is guaranteed if the "onset length" (the scale at which exponential decay becomes visible) is uniformly bounded and the random density matrices $\rho_k$ are uniformly bounded away from zero.
- The bound depends on the decay constant $\mu$ of the transfer map, the worst-case onset length $\xi[k]$ , and the smallest eigenvalues of $\rho_k$ .
- The authors demonstrate that if the onset lengths are unbounded (even if the density matrices are well-behaved), the gap may vanish. This explains the mechanism behind the gapless examples found in [36].
Examples and Counter-Examples:
The paper provides a class of examples (disordered deformations of the AKLT model) to illustrate these findings.
- They show that a specific modification of the gapless example from [36] can be made gapped by ensuring the disorder parameters (sines and cosines of random angles) do not vanish too frequently, thereby satisfying the uniform boundedness of the onset length.
- Conversely, they confirm that if the disorder allows for arbitrarily long intervals where the transfer map fails to contract sufficiently, the spectral gap closes.

Significance and Scope
The paper clarifies the structural differences between translation-invariant and ergodic matrix product states regarding their parent Hamiltonians. It demonstrates that while injective EMPS always admit unique, frustration-free parent Hamiltonians, these Hamiltonians need not be finite-range or gapped. The work identifies the "unboundedness of the onset length" as a primary obstruction to the existence of a spectral gap in the ergodic regime.

The authors conclude by posing an open problem: the classification of the set of gapped injective EMPS. They suggest that while unbounded onset lengths are a known obstruction, it remains an open question whether other obstructions exist, particularly in independent and identically distributed (i.i.d.) cases. The paper does not propose new experimental realizations but provides a rigorous mathematical framework for understanding the stability of spectral gaps in disordered quantum spin chains modeled by tensor networks.