Dynamics and stability of U(1) spin liquids beyond mean-field theory: Triangular-lattice J 1 - J 2 Heisenberg model

Quantum spin liquids (QSLs) are long-range entangled phases of frustrated magnets exhibiting fractionalized spin excitations, and this work extends parton mean-field theory to describe their dynamical response and stability in the triangular-lattice J1–J2 Heisenberg model without free parameters.

The gist

We calculate the dynamical spin structure factor within a self-consistent MFT+RPA theory without free parameters, finding that in addition to spinon continua a new sharp mode dominates the finite-frequency response, which can be thought of as a spinon exciton bound state induced by the interactions.

Model and Approach

The study focuses on the exchange-frustrated spin-half Heisenberg model on the triangular lattice:

1. The Hamiltonian is defined by nearest-neighbor coupling J1 and next-nearest neighbor coupling J2:

H = X Σij Jij S⃗i · S⃗j, where Jij is J1 on nearest-neighbor bonds and J2 on next-nearest, as shown in Fig. 1(a).

2. The physical Hilbert space is projected by requiring one fermion per site via fiαfiα = 1.

3. The low-energy spectrum of the emergent QSL phase is described by parton mean-field theory (MFT) [1, 2].

Interactions and Dynamical Response

The paper extends MFT to include fluctuations of the QSL and competing spin order parameters:

(1) The physical Hilbert space must be projected out by requiring one fermion per site, via fiαfiα = 1.

(2) The model with interactions and filling constraint is analytically intractable. One then describes the low-energy spectrum of the emergent QSL phase by resorting to MFT [1, 2].

The dynamical spin structure factor S(q, É) is calculated using the fluctuation-dissipation theorem:

(4) S(q, É) = 1/N Σij e−iq·(i−j) Z Σdt eiωtSzi(t)Szj(0).

The interacting susceptibility is treated using the Random Phase Approximation (RPA):

(35) Ç(q, É) = Ç0(q, É) · inv 1 − Uq · Ç0(q, É).

This leads to the identification of a sharp paramagnon mode at low energy: "The most striking effect of the spinon interactions is to create a sharp paramagnon mode coming down to low energy; this spinon exciton, a quasiparticle with quantum numbers of a magnon arising as a two-spinon bound state, displays a magnetoroton-like minimum around K."

Phase Diagram and Transitions

Tuning the ratio J2/J1 allows for quantitative predictions of phase boundaries:

(55) We can derive expressions for the critical points that yield r120c = 0.07158 and rstripec = 0.1800.

The sharp mode condenses at zero energy below J2c = 0.0716 J1, signaling a continuous transition to 120-degree magnetic order.

The resulting phase diagram and dynamical structure factor are in remarkable agreement with state-of-the-art numerical calculations for the model’s ground state [11, 13, 16] and time-dependent response [15–17].

Gauge Fluctuations and Continuum Theory

The authors investigate the effect of gauge field fluctuations beyond first-order MFT:

(69) The relevant gauge excitations for this purpose are the spin-triplet monopoles [17, 18], which transform as spin operators Sαi ≃ 3Φ[ieiK·ri Φα+3(ri) + h.c.] with momentum K at the corner of the Brillouin zone.

The monopole contribution to the dynamical structure factor is found to be:

(82) S(q, É) = Im c32Φ(K − q2 − É2 + i¸) 3/2−∆Φ.

The gauge fluctuations modify the nature of phase transitions:

1. The 120-degree ordering transition is in the chiral Heisenberg Gross–Neveu-QED3 universality class [45].

2. Gauge fluctuations modify the bound state propagator, leading to a hybridization of the paramagnon and monopoles, where "the frequency-dependent propagator leads to a hybridization and quasiparticle continuum in the previously sharp mode.

Based on the provided scientific paper, here are specific improvements that could be made to AI systems, along with what those improved systems could achieve:

) 1. Improved Quantum Phase Transition Prediction via Self-Consistent RPA-based Models

The paper introduces a self-consistent Parton Mean-Field + Random Phase Approximation (MFT+RPA) theory to calculate dynamical spin structure factors and phase diagrams for the U(1) Dirac QSL on the triangular lattice, incorporating spinon interactions and gauge field fluctuations without free parameters.

2. What the improved AI system can do:

The improved system could accurately predict quantum critical points (QCPs) in complex frustrated magnetic materials (like those modeled by the J1–J2 Heisenberg model) with high fidelity, specifically predicting:

• The exact phase boundaries between ordered phases (e.g., 120-degree order vs. stripe order).

• The critical exchange ratios (e.g., predicting the transition point at exactly 0.0716 for the U(1) Dirac QSL).

• The energy dispersion of emergent quasiparticles, such as spinon exciton bound states (paramagnons), and how these modes disperse as a function of parameters like J2/J1.

3. Improved Chiral Spin Liquid Characterization via Effective Field Theory

The paper extends the method to study Chiral QSLs by coupling the model to a scalar spin chirality term, leading to an effective QED3 theory with monopole degrees of freedom.

4. What the improved AI system can do:

The improved system could perform first-principles predictions on materials exhibiting time-reversal symmetry breaking (chiral magnets), specifically:

• Determine the stability boundaries for chiral spin liquid states based on competing magnetic orders (e.g., tetrahedral order).

• Predict the critical ratios where continuous transitions occur versus first-order transitions.

• Identify the universality class of these transitions (e.g., predicting that the 120-degree transition falls into the Chiral Heisenberg Gross–Neveu–QED3 universality class).

5. Improved Dynamical Response Simulation via Hybridization Mapping

The system incorporates a sophisticated mechanism to handle the interaction between spinon excitations and gauge field fluctuations by modeling their hybridization (paramagnon mode renormalization).

6. What the improved AI system can do:

The improved system could simulate and predict the complex spectral functions measured in Inelastic Neutron Scattering (INS) experiments, including:

• The broadening and shifting of sharp excitation modes due to coupling with a continuum.

• The precise frequency dependence of quasiparticle lifetimes, which is crucial for distinguishing between different phases.

7. Improved Gauge Field Dynamics Modeling via Monopole Contributions

The theory explicitly includes the contribution of emergent gauge field fluctuations, specifically spin-triplet monopoles, which are identified as the relevant low-energy excitations in 2D QED3.

8. What the improved AI system can do:

The improved system could model and predict how long-range magnetic correlations manifest in dynamical measurements by accounting for topological excitations (monopoles), allowing for a quantitative description of the spinon exciton signature observed in experiments like INS.

9. Improved Predictive Power for Candidate Materials

The framework is designed to be general enough to include extensions like spin anisotropy, interlayer coupling, and disorder.

10. What the improved AI system can do:

The improved system could serve as a powerful tool for inverse problem analysis in condensed matter physics, allowing researchers to input microscopic material parameters (like anisotropy or disorder) and predict the resulting dynamical response functions and phase diagrams for candidate materials (e.g., YbZn2GaO5).

Quantum spin liquids (QSLs) are long-range entangled phases of frustrated magnets exhibiting fractionalised spin excitations. In two dimensions, there is limited analytical understanding of their excitation spectra beyond parton mean-field theories, which fail to capture many features of the finite frequency dynamical response. We use a self-consistent mean-field theory combined with the random phase approximation (RPA) for the J 1 - J 2 Heisenberg model on the triangular lattice to describe the strong spinon-spinon interactions of the U(1) Dirac QSL. We obtain quantitative results for the dynamical spin structure factor and phase diagram compatible with comprehensive numerical efforts. We show the continuum response at the Brillouin zone corners is described by a spinon-exciton hybridised with monopole gauge excitations. We further show that the transition from the QSL to 120-degree coplanar order is in the QED 3 O(3)-GN universality class. We extend the method to chiral QSLs and XXZ anisotropy, and discuss its broad range of applicability to other models and for describing inelastic neutron scattering experiments in KYbSe 2.