Revisiting the J 1 - J 2 Heisenberg Model on a Triangular Lattice: Quasidegenerate Ground States and Phase Competition

arXiv:2603.08650 · cond-mat.str-el · Submitted 2026-03-09 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Quantum Radio. Generated commentary on the latest quantum physics and condensed matter papers.

Kai: I'm Kai, and with me are Mira and Lev, guest researcher.

Mira: Today's paper: "Revisiting the J 1 - J 2 Heisenberg Model on a Triangular Lattice".

Kai: Quasi-degenerate ground states in frustrated quantum magnets are central to understanding phase competition between magnetic orders and quantum spin liquids.

Mira: First, who's behind it and why it matters.

Paper summary: Kai: So Mira and I were just looking over this paper titled "Revisiting the J1 - J2 Heisenberg Model on a Triangular Lattice: Quasi-Degenerate Ground States and Phase Competition." The main point here is that they're looking at two nearly degenerate ground states in this spin model on a triangular lattice using large-scale matrix product state simulations. They are trying to figure out if these near degeneracies mean we're looking at some kind of topological order or just finite-size effects.

Mira: Exactly, Kai, the paper is centered on the thesis that when you look at these two nearly degenerate states, they show clear differences in how things behave locally on a cylinder. The authors are specifically examining the spin-one/two triangular-lattice J1-J2 Heisenberg model at a coupling ratio of J2/J1 = zero point one two five to see what that means for phase competition between magnetic orders and quantum spin liquids <ref:2603.08650#pg0,the spin-1/2 triangular-lattice J1-J2 Heisenberg model>.

Lev: From an error correction viewpoint, if these states are truly different, it means the low-energy physics isn't simply one topological sector of a gapped Z2 spin liquid, which would have implications for how we might approach those states on real hardware.

Kai: That’s right, Lev; they’re using state-of-the-art matrix product state simulations on YC6 cylinders to get these ground states. The paper points out that even though the two sectors are nearly orthogonal with a fidelity of order ten-four the local observables tell a story <ref:2603.08650#pg0>.

Mira: And what's really compelling is how they analyze the static and dynamical properties. They look at the equal-time structure factor, or ETSF, and then they look at the dynamical structure factor, or DSF. These measurements are used to probe whether these states are fundamentally distinct in their excitations even when they look similar on a surface level.

Lev: I wonder how those distinct spectral weights translate into actual error syndromes if you were trying to encode quantum information using one of these sectors. It sounds like the difference in spectral weight distribution is what matters for practical implementation.

Paper summary: Kai: Well, the paper finds that in the QSL candidate phase at J2/J1 = zero point one two five, the even sector shows a significant softening of certain features in its ETSF, which they link to it potentially being a U(one) Dirac spin liquid <ref:2603.08650#pg0,at J2/J1 = 0.125>. But then for the odd sector, the spectral weight is distributed more uniformly across the Brillouin zone.

Mira: That contrast is key because it directly challenges the idea that these two states are just different topological sectors of a gapped Z2 spin liquid, as they concluded in their paper "Revisiting the J one - J two Heisenberg Model on a Triangular Lattice: Quasi-Degenerate Ground States and Phase Competition <ref:2603.08650#pg0,Heisenberg Model on a Triangular Lattice: Quasi-Degenerate Ground States and Phase>."

Lev: So what does this mean for the theoretical landscape? If the odd sector shares features with a magnetically ordered phase at J2 = zero that suggests it might be closer to that classical state than we initially thought <ref:2603.08650#pg0>.

Kai: It suggests a whole different picture for how these competing phases interact in this specific parameter regime of the J1-J2 Heisenberg model. The paper shows that the two sectors are substantially different in their local observable properties on the YC6 cylinder geometry.

Mira: That is what they conclude, and it implies a careful reassessment of where we think these quasi-degenerate ground states originate within this phase competition area. It's not just about finding two states; it's about understanding the physical nature of each one.

Lev: If the odd sector is indeed closer to the one hundred twenty degree ordered phase, that has implications for how robust those stripe or QSL phases are when you introduce small perturbations, which is something critical for experimentalists trying to stabilize a desired state <ref:2603.08650#pg0>.

Kai: So, basically, this work on "Revisiting the J one - J two Heisenberg Model on a Triangular Lattice: Quasi-Degenerate Ground States and Phase Competition" isn't just about finding two states; it’s about showing that the distinction between them is physical and tied to different underlying quantum liquid behaviors <ref:2603.08650#pg0,Heisenberg Model on a Triangular Lattice: Quasi-Degenerate Ground States and Phase>.

Paper summary: Mira: Precisely, Kai; they use detailed analysis of both static and dynamical correlations to show these differences are more than just artifacts of the simulation setup. The paper calls for a new way to interpret this two-sector structure in this specific parameter regime.

Lev: From an experimental standpoint, knowing which sector is closer to the classical order helps us predict how difficult it will be to suppress or stabilize that order when we try to tune the magnetic fields or couplings in a real experiment.

Kai: And what's exciting for hardware guys is that this detailed understanding of the excitations gives us better targets for what we need to measure on our experimental setups. It tells us exactly what spectral weight signatures to look for in our measurements.

Mira: The implication is that we need a more nuanced theoretical framework when modeling these frustrated magnets, moving beyond simple topological classifications when dealing with competing orders near critical points like J2/J1 = zero point one two five as explored in this paper <ref:2603.08650#pg0>.

Lev: If the odd sector is indeed proximate to a one hundred twenty degree state, then understanding its low-energy physics might lead to new strategies for designing materials that exhibit specific types of quantum correlations under realistic conditions <ref:2603.08650#pg0>.

Kai: So, what we’re seeing here is a deeper look at how the magnetic ground state landscape looks when you introduce competing interactions on a triangular lattice. It's a lot to wrap your head around, but it’s really informative about the physics there.

Mira: Indeed, this paper provides strong evidence that quasi-degenerate ground states aren't always simple topological sectors; they can represent different physical regimes altogether depending on the parameter tuning.

Lev: And for those of us in error correction, this means we have a clearer picture of what kind of excitations we might encounter when trying to implement quantum codes in these frustrated systems.

Kai: We really hope this work helps guide our experimental efforts by telling us which features to focus on when probing these complex magnetic materials.

Mira: It certainly provides a necessary refinement to the models we use, showing that simple interpretations of near degeneracies aren't always sufficient when phase competition is involved.

Conclusion: Kai: So to wrap up this discussion on the "Revisiting the J one - J two Heisenberg Model on a Triangular Lattice: Quasi-Degenerate Ground States and Phase Competition" paper, we need to focus on what these authors actually found regarding those two nearly degenerate ground states.

Mira: Exactly, and the core of their work is showing that these two states aren't just different topological sectors of a simple spin liquid because they exhibit distinct behaviors in local observables like the static structure factor and dynamical correlations.

Lev: From my side, what this means for real hardware is that we have to be very careful about which sector we are actually targeting when trying to realize some kind of quantum state on a device, especially given how sensitive those spectral weight distributions are.

Kai: Right, so the authors push back against the simple idea that these quasi-degenerate states just signal a gapped Z2 spin liquid and instead suggest something more complex is happening in this specific parameter region.

Mira: They propose that one of these sectors might be better described as a U(one) Dirac spin liquid, while the other shares features with the magnetically ordered phase at zero J2, which is a significant distinction for our theoretical models.

Lev: If that odd sector truly shares low-energy features with the classical one hundred twenty-degree order, then it suggests that transition point we were looking at might be more nuanced than just a simple quantum critical point.

Kai: It really highlights how crucial it is to look beyond the surface level descriptions and dig into these detailed correlation functions to understand the actual physics of competing orders.

Mira: Their conclusion implies a necessary refinement in how we classify phases in frustrated magnets, moving away from overly simplistic topological categorizations when dealing with these competing interactions.

Lev: And for error correction, this means our simulations need to account for these distinct spectral properties because they directly impact the fidelity and coherence of any encoded information.

Kai: So the big picture here is that understanding these subtle differences between sectors gives us a much richer map of what happens in the J1-J2 model near its phase transitions.

Mira: Indeed, this paper provides strong evidence that quasi-degenerate ground states aren't always simple topological sectors; they can represent fundamentally different physical regimes depending on how you tune the couplings.

Lev: This opens up new avenues for theoretical exploration into how these competing orders interact dynamically, which is something we need to keep in mind as we design next-generation quantum simulators.

Arnold Sommerfeld Center for Theoretical Physics · Center for NanoScience · Munich Center for Quantum Science and Technology

cond-mat.str-el

Submitted: 2026-03-09

Updated: 2026-10-07

Comments: 9 pages, 9 figures

DOI: 10.1103/cljk-jcq3

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Importance score: 77/100

The gist: Quasi-degenerate ground states in frustrated quantum magnets are central to understanding phase competition between magnetic orders and quantum spin liquids.

Key concepts

Triangular-Lattice J1-J2 Heisenberg Model
This is a quantum model describing magnetic interactions on a triangular lattice with two types of coupling: J1 for nearest neighbors and J2 for next-nearest neighbors. This specific model is important because it can host three competing ground states depending on the ratio of these couplings.
Matrix Product State (MPS) Simulations
These are large-scale computational methods used to find the lowest energy state of a quantum system, especially for systems with limited width like cylinders. The study used MPS to simulate finite-size systems and identify near-degenerate ground states by examining their properties.
U(1) Dirac Quantum Spin Liquid
This is a specific type of quantum spin liquid candidate where the magnetic correlations behave similarly to a system with an emergent U(1) gauge field. The study suggests the even sector ground state exhibits spectral features consistent with this type of exotic liquid behavior.
Static Correlation Analysis (ETSF)
The equal-time structure factor measures how spins are correlated at a specific moment in time. By looking at the ETSF, researchers could compare the spatial patterns of correlations between the two ground states, revealing differences related to magnetic order or quantum liquid behavior.

Terminology

Summary

Quasi-degenerate ground states in frustrated quantum magnets are central to understanding phase competition between magnetic orders and quantum spin liquids. The research investigates these states in the spin-1/2 triangular-lattice J1-J2 Heisenberg model using large-scale matrix product state simulations to determine if the observed near degeneracies signal a topological order or merely finite-size effects.

The gist: Two nearly degenerate ground states on a width-6 cylinder at J2/J1 = 0.125 exhibit clear differences in local observables, suggesting they cannot be understood as merely topologically distinct sectors of a gapped Z2 spin liquid.

Model and Context

The study focuses on the spin-1/2 antiferromagnetic Heisenberg model on a triangular lattice (TLHAF) extended with a next-nearest-neighbor (NNN) coupling, defined by Equation (1):

H = J1 X⟨i,j⟩ Si · Sj + J2 X⟨⟨i,j⟩⟩ Si · Sj. This model is significant because it exhibits competing ground states: coplanar 120° magnetic long-range order for small J2/J1, a quantum spin liquid (QSL) candidate regime in the intermediate range of 0.07 ≲ J2/J1 ≲ 0.15, and collinear stripe order at larger values of coupling. The paper specifically examines the QSL candidate region at J2/J1 = 0.125, which coincides with the transition point of the corresponding classical model.

Ground State Identification on Cylinders

The investigation utilizes large-scale matrix product state (MPS) simulations on a YC6-36 cylinder geometry to study finite-width systems. The methodology involves obtaining ground-state MPS using single-site DMRG with controlled bond expansion, exploiting the non-Abelian SU(2) symmetry via the QSpace tensor library. A key observation in these simulations is the occurrence of a sudden drop in energy while performing the sweeping routine for D∗ = 2048, which is reproducible across independent runs. This behavior is noted as a robust indicator distinguishing between states: before the drop, ψeven⟩ has an energy per site higher than ψodd⟩ by 1.1%. Furthermore, the two states are nearly orthogonal with a fidelity F = ⟨ψevenψodd⟩ of order ∼ 10−4.

Static Correlation Analysis

The static properties were analyzed using the equal-time structure factor (ETSF), χ(k) = 1/Lm Xi,j e−ik·(ri−rⱼ)⟨Si · Sj⟩. In the 120° ordered phase (J2 = 0), the ETSF shows sharp isolated maxima at the corners of the Brillouin zone, consistent with long-range magnetic order. For the QSL candidate phase at J2/J1 = 0.125, results for both sectors are compared:

(i) The even sector ETSF exhibits a significant softening of the sharp features at K and M points, which is most consistent with the even sector potentially being a U(1) Dirac QSL.

(ii) The odd sector ETSF shows correlations that are broader near the corners of the Brillouin zone, though less prominently than in the even sector.

Dynamical Correlation Analysis

The dynamical structure factor (DSF), S(k, ω), was computed using the tangent space Krylov (TaSK) method to efficiently obtain high-resolution spectral data. Comparing ψeven⟩ and ψodd⟩ in the QSL regime reveals distinct spectral weight distributions:

= Spectral Weight Concentration:

The even sector's spectral weight is mostly concentrated at the K point. In contrast, for the odd sector, it is distributed more uniformly throughout the Brillouin zone, peaking at the K and M points with comparable intensity. This observation is consistent with expectations for a U(1) Dirac QSL.

Conclusion and Interpretation

The findings provide compelling evidence that the two sectors exhibit substantial differences in their locally observable properties on the YC6 cylinder. The paper concludes that this result rules out the mere emergence of quasi-degenerate ground states as a signature of a gapped Z2 QSL. Instead, it proposes an interpretation where:

  1. The even-sector ground state could be interpreted as the U(1) Dirac QSL.

  2. The odd-sector ground state shares many low-energy features with the magnetically ordered phase at J2 = 0.

The authors leave open questions regarding whether the true ground state is always in the odd sector or if it depends on system size and geometry, and whether the odd sector represents a distinct QSL nature or a proximate 120° state.

Improvements for AI systems

Here are specific improvements for AI systems based on the findings in this scientific paper, categorized by application area:

) 1. Enhanced Quantum Simulation and Material Discovery (Quantum Chemistry/Condensed Matter AI)

The core finding is that the even sector of a quasi-degenerate ground state in frustrated magnets (like the J1-J2 Heisenberg model) behaves like a U(1) Dirac spin liquid, while the odd sector shares features with magnetically ordered phases.

Improvements:

  • [Specific Improvement]: Develop and train Machine Learning models (specifically Neural Network Potentials or advanced Variational Autoencoders) that are explicitly parameterized to distinguish between the spectral signatures of the U(1) Dirac QSL (even sector) and proximate magnetic orders.

  • [Specific Improvement]: Implement a Sector Classifier module within quantum simulation pipelines that analyzes low-energy excitations (DSF/ETSF patterns). This module would be trained on high-fidelity MPS data to predict which topological sector is realized based on the computed correlation functions.

  • [Improved AI Capability]: The system can accurately predict whether a simulated frustrated material will exhibit a gapped Z2 spin liquid, a gapless U(1) Dirac QSL, or a proximate ordered phase purely from its low-energy spectral response data without needing full Hamiltonian diagonalization.

) 2. Topological Phase Identification and Classification (Topological AI)

The paper suggests that the even sector is consistent with a U(1) Dirac QSL, while the odd sector's features are complex/proximate.

Improvements:

  • [Specific Improvement]: Create a topological feature extraction algorithm that moves beyond simple order parameters to analyze high-dimensional correlation tensors (like the bond correlations shown in Table I) and momentum-space distribution of spectral weight. This algorithm should specifically look for the characteristic broad distribution across the Brillouin zone (U(1) Dirac signature).

  • [Specific Improvement]: Develop a Phase Transition Predictor that uses the relative fidelity between two nearly degenerate states as a proxy for topological order. If the fidelity drops sharply or exhibits specific scaling behavior during simulation sweeps, it flags a potential transition boundary between different QSL candidates.

  • [Improved AI Capability]: AI systems can be deployed to rapidly screen vast chemical/physical spaces (e.g., molecular Hamiltonians) and instantly classify whether a candidate material is likely to possess a gapless U(1) Dirac QSL or a gapped topological phase, drastically reducing the time required for experimental validation.

) 3. Accelerated Tensor Network Simulation (Computational Efficiency AI)

The paper validates the efficiency of using Tangent Space Krylov (TaSK) methods for computing dynamical structure factors in MPS representations, especially compared to standard time-evolution methods.

Improvements:

  • [Specific Improvement]: Develop a Krylov Basis Optimizer module that dynamically adjusts the number of kept SU(2) multiplets and the convergence criteria during DMRG sweeps based on real-time monitoring of energy density drops (as seen in Supplemental Material S-1). This prevents unnecessary computational expenditure when convergence is robust, while aggressively pursuing solutions when anomalies (like the energy drop at sweep 5) are detected.

  • [Specific Improvement]: Implement an adaptive spectral broadening scheme that dynamically adjusts the parameters of the Continued Fraction Expansion (CFE) based on the observed spectral features (e.g., sharpening broadening for sharp magnon modes vs. wider smoothing for continuum regions).

  • [Improved AI Capability]: Quantum chemistry or condensed matter simulation tools can achieve high-resolution dynamical correlation calculations—which are notoriously expensive—significantly faster by intelligently guiding the iterative MPS optimization process toward relevant physical states, leading to more accurate spectral function predictions in a fraction of the time.

) 4. Symmetry and Order Parameter Inference (Symmetry-Aware AI)

The paper highlights how different sectors possess distinct local order parameters (e.g., NN correlations in Table I).

Improvements:

  • [Specific Improvement]: Train Graph Neural Networks (GNNs) specifically designed for lattice structures that can ingest the full bond correlation matrices and output a probabilistic assessment of the ground state's symmetry class (e.g., predicting if it is closer to 120° order or isotropic/Dirac).

  • [Specific Improvement]: Integrate the Odd Sector Signature Detector into predictive models. Since the odd sector shares features with ordered phases, AI can learn to recognize these proximate order signatures in complex systems and flag them as potential metastable states that might be relevant under specific external perturbations.

  • [Improved AI Capability]: Materials informatics platforms can be used to search for materials whose ground state possesses a specific, desirable symmetry (e.g., a U(1) Dirac phase), guiding synthetic efforts toward target quantum phases rather than relying solely on traditional thermodynamic stability criteria.

Sources

Related papers