Quantum many-body scars with tunable entanglement and Hamiltonian inverse design from ZX-calculus

AI summary

Overview

Research area: Quantum many-body physics and quantum information, at the intersection of diagrammatic quantum languages (ZX-calculus), entanglement structure of many-body states, and the theory of quantum many-body scars and parent Hamiltonians.

Technical level: Advanced. The abstract assumes familiarity with entanglement scaling laws (area law, logarithmic, power law), frustration-free parent Hamiltonians, level-spacing statistics, and the ZX-calculus rewrite system.

Scope (one sentence): The paper constructs fractal many-body states from ZX-diagrams based on Sierpiński triangle and carpet geometries, bounds and tunes their entanglement, and uses graphical identities to inverse-design frustration-free Hamiltonians that host one such state as an exact quantum many-body scar in a chaotic spectrum.

What This Paper Is About

ZX-calculus and similar diagrammatic languages are well established for manipulating quantum circuits — simplifying, verifying, and compiling them — but they have barely been used as a tool for building many-body states and the Hamiltonians that have those states as eigenstates. This paper asks whether the graphical structure of a ZX-diagram can be promoted into a design principle: the diagram's connectivity dictates the state's entanglement, and the diagram's local rewrite identities can certify that a chosen Hamiltonian annihilates the state exactly. The goal is an explicit demonstration that ZX-calculus can serve as a framework for Hamiltonian inverse design, with quantum many-body scars as the test case.

Key Contributions

1. Two families of fractal many-body states from ZX-diagrams. The paper introduces state families whose diagrams are based on the Sierpiński triangle and the Sierpiński carpet, treating the diagrams as state constructors rather than circuit descriptions.

2. Connectivity-derived entanglement bounds and tunable actual scaling. The underlying graph connectivity imposes atypical subvolume-law minimum-cut upper bounds on entanglement — at most logarithmic for the triangle family and sublinear power-law for the carpet family. Free parameters in the diagram construction then tune the realized scaling: area-law or logarithmic for the triangle, logarithmic or power-law for the carpet.

3. Frustration-free Hamiltonians with diagrammatic terms for the triangle family. Combining parent-Hamiltonian methods, ZX-calculus insight, and local ZX identities that certify exact annihilation of the target state, the paper produces frustration-free Hamiltonians whose terms admit simple representations in the same diagrammatic language as the states.

4. A scar embedded in a chaotic spectrum. A local deformation of that Hamiltonian produces chaotic level statistics while keeping the fractal ZX state in the bulk of the energy spectrum as an exact quantum many-body scar, establishing the graphical workflow as a route to Hamiltonian inverse design.

Main Findings

• Fractal states are diagrammatically natural: Sierpiński-triangle and Sierpiński-carpet ZX-diagrams yield well-defined families of many-body states whose structure is fixed by the diagram's connectivity.

• Entanglement is bounded by geometry, not by dynamics: The connectivity imposes subvolume-law minimum-cut upper bounds — logarithmic for the triangle family, sublinear power-law for the carpet family — so these states cannot reach volume-law entanglement.

• The bound is not the whole story — the actual scaling is tunable: Adjusting the diagram's free parameters realizes area-law or logarithmic scaling in the triangle family, and logarithmic or power-law scaling in the carpet family. The abstract presents both options as demonstrated; no further numerical detail is given in it.

• Local observables keep a fractal footprint: The states retain fractal-like spatial structure in local observables, which the paper identifies as the reason they are natural candidates for atypical eigenstates in otherwise thermalizing systems.

• Exact annihilation is certified graphically: Local ZX identities supply the proof that the constructed Hamiltonian terms annihilate the target state, giving frustration-free parent Hamiltonians expressed in the same diagrammatic language as the state.

• Scars survive a chaotic deformation: A local deformation of the triangle-family Hamiltonian yields chaotic level statistics while the fractal state remains an exact eigenstate in the bulk of the spectrum — the defining signature of a quantum many-body scar.

• ZX-calculus as an inverse-design framework: The paper's overarching claim is that scars, their local annihilators, and the chaotic Hamiltonians embedding them can all be constructed and related through a set of graphical identities.

Methodology in Plain English

The approach is constructive and graphical rather than numerical. Start with a ZX-diagram whose wiring mimics a fractal pattern (the Sierpiński triangle or carpet). Read that wiring as a recipe for a many-body state. Because the diagram's connectivity restricts how quickly information can spread across it, cut-based arguments give an upper bound on how entangled the resulting state can be — a bound set by the pattern, not by any fine-tuning. The diagram also carries free parameters; varying them changes the state's actual entanglement scaling within that bound. Next, borrow the standard "parent Hamiltonian" idea — find local terms that the target state perfectly satisfies — but use graphical identities to prove the annihilation exactly, term by term. Each Hamiltonian term can then be drawn in the same diagram language as the state. Finally, add a local perturbation designed to make the spectrum look thermal (chaotic level repulsion) without destroying the special state, which then sits in the middle of the spectrum as a scar.

Why This Matters

Research impact. The work proposes a new constructive route to quantum many-body scars: instead of discovering a scar in a model and then explaining it, one specifies the desired state graphically and derives the Hamiltonian from diagrammatic identities. It also links entanglement-scaling design directly to graph connectivity, and shows that diagrammatic calculi — traditionally a verification and compilation tool — can function as a generative design language for states and Hamiltonians.

Potential applications (the abstract does not discuss applications; these are plausible directions the construction suggests):

• Designing Hamiltonians with prescribed entanglement structure for quantum simulation of non-thermalizing dynamics.
• Providing exactly solvable reference states and annihilator sets for benchmarking numerical methods near scar physics.
• Supplying test models for studying how much entanglement a scar can carry, and whether scars can persist with area-law or power-law entanglement.
• Extending graphical certification of eigenstates into compiler-style toolchains for Hamiltonian synthesis.

Industry relevance. Quantum-computing software and compiler groups already build on ZX-calculus and related diagrammatic methods; a demonstration that the same diagrams can generate states and Hamiltonians broadens the toolset toward Hamiltonian-level design. The relevance is currently methodological and long-term rather than an immediate engineering result, since the abstract presents a theoretical construction with no hardware or algorithmic benchmark claims.

Future Directions

• Extending the Hamiltonian construction to the carpet family. The abstract states frustration-free Hamiltonians and scar embedding for the triangle family specifically; an obvious open question is whether the carpet family admits an analogous parent Hamiltonian and scar construction, given its different entanglement bound.

• Generalizing beyond the two fractals. Other ZX-diagram topologies, other self-similar patterns, and higher-dimensional or non-planar geometries could yield further state families with different entanglement bounds.

• Characterizing the scar's properties in detail. The abstract reports the scar exists in the bulk with chaotic level statistics; questions of robustness to perturbations, the nature of the scarred eigenstate's entanglement and observables at finite size, and the dynamics of initialized states remain open.

• Turning the identities into a systematic design procedure. A broader question is whether the graphical identities used here can be organized into a repeatable inverse-design calculus — an algorithm-like pipeline from a target diagram to a Hamiltonian — rather than a case-by-case construction.

Target Audience

Researchers in quantum many-body theory and quantum information working on quantum many-body scars, entanglement scaling, or parent-Hamiltonian constructions; members of the ZX-calculus and quantum-compilation community interested in applications beyond circuit optimization; and graduate students or advanced readers comfortable with diagrammatic quantum formalisms who want a concrete example of graphical methods driving Hamiltonian design.

Authors’ abstract

Diagrammatic languages such as ZX-calculus provide compact and intuitive descriptions of quantum processes and have become established tools for circuit simplification, verification, and compilation. However, their potential as a framework for constructing many-body states and the Hamiltonians that host them remains largely unexplored. Here, we introduce families of fractal many-body states obtained from ZX-diagrams based on the Sierpiński triangle and Sierpiński carpet. By construction, the underlying graph connectivity imposes atypical subvolume-law minimum-cut upper bounds on entanglement: at most logarithmic scaling for the triangle family and sublinear power-law for the carpet family. By adjusting the free parameters of the ZX-diagram construction, we can tune the actual entanglement scaling: we demonstrate both area-law and logarithmic scaling for the triangle, and both log law and power-law for the carpet. Additionally, their local observables retain fractal-like spatial structure, identifying these states as natural candidates for atypical eigenstates in otherwise thermalizing systems. For the triangle family, we combine parent-Hamiltonian methods, insights from ZX-calculus, and local ZX identities that certify exact annihilation of the target state, producing frustration-free Hamiltonians whose terms admit simple representations in the same diagrammatic language as the states themselves. We then construct a local deformation that produces chaotic level statistics while embedding the fractal ZX state in the bulk of the energy spectrum as an exact quantum many-body scar. Our results demonstrate, through this explicit construction, that ZX-calculus can serve as a framework for Hamiltonian inverse design, in which quantum many-body scars, their local annihilators, and the chaotic Hamiltonians embedding them can be constructed and related through a set of graphical identities.