An exact-diagonalization study of a frustrated quantum spin model has found strong finite-size evidence for a gapless phase called a nodal Bose liquid. The signal comes from two main tests: the energy gap drops sharply as the simulated system grows, while the ground-state entanglement follows a distinctive size dependence. The calculation still stops short of establishing what happens in the infinite system.
How the calculation works
The primary model is a geometrically frustrated spin-1/2 chiral XYZ model on a triangular lattice. The analysis compares finite lattice tori with a large-N Z_N clock generalization and several Hamiltonian deformations. Its central computational device is dynamical splitting: the implementation diagonalizes two factor Hamiltonians independently, then adds their eigenvalues to reconstruct the full spectrum.
The model also has a symmetry-related feature that could easily be overread. Its line-symmetry algebra makes every energy level at least twofold degenerate, whether or not the spectrum has a gap. That exact repetition is a property of the model's algebra. By itself, it does not establish a gapped topological phase.
At large N, the corresponding clock model has a gapless ground state with three subsystem-symmetry-related nodal lines. That result provides the broad structure against which the N=2 calculation is compared, but it is not direct evidence for the N=2 model.
The finite-size case for gaplessness
On the square-torus exact-diagonalization scan, the reported gap fell by more than two orders of magnitude between L=3 and L=9. Fits made on logarithmic scales gave an approximate size-scaling exponent z of 4.2 for odd sizes and 4.3 for even sizes. Those values should not be treated as a settled dynamical exponent, because they drift when the fitting window changes.
The scope of the scan also matters. Global center-sector comparisons were exhaustive for L=3 through L=8. The L=9 point was a targeted estimate based on the uniform sector and competing sector patterns, rather than an exhaustive scan of every center sector.
A second diagnostic came from the ground state's entanglement across cuts. It increased nearly linearly with log L, and data from cylinder cuts closely followed the predicted chord form. The fitted effective coefficients were 2.48 for a transverse width of L_y=3 and 7.51 for L_y=5. Only two cylinder widths were available for that comparison, so the values identify a pattern without fixing a universal number precisely.
Taken together, the gap and entanglement calculations were interpreted as strong evidence for a nodal Bose-liquid ground state in the N=2 chiral XYZ model. The conclusion is based on finite-size numerical diagnostics. It does not establish the exact thermodynamic gap or exponent, and the exact level degeneracy does not by itself prove a gapped topological phase.
Deformations reveal competing possibilities
The researchers then varied the Hamiltonian to examine states near the original model. In the large-N split-preserving deformation, they report a first-order transition at x_c=3/4. The branch is inhomogeneous for x below that value and translationally invariant for x above it. Both branches are reported as gapless and nodal.
Finite-size results at N=2 showed that both the global gap and the gap restricted to the uniform sector decreased strongly with system size at the sampled deformation strengths. Neither indicated a conventional gap, supporting an extended gapless uniform branch. Controlled extrapolation was not available, however.
A separate split-breaking interpolation produced evidence for a possible change of phase. Its finite-size data were consistent with a transition near x=0.4 to 0.5 between the nodal liquid and a gapped two-copy Wen-plaquette phase. An entanglement diagnostic called the normalized Kitaev-Preskill combination stayed around 1.2 to 1.3 through x=0.4, rose across the interval where the gap opened, and approached 2. Because the low-x side is gapless, its finite-size value was not assigned a topological entanglement entropy.
What remains unsettled
The broader stability question remains open. Large-N renormalization-group analysis classified the leading subsystem-symmetry-breaking perturbations studied here as relevant, but the paper does not establish whether that conclusion carries over to N=2.
Uniform onsite-field spectra were suggestive of C3 breaking but inconclusive about spontaneous breaking or whether an arbitrarily weak field opens a gap. At larger field, the ground state became nondegenerate and approached a polarized state.
The result is therefore a claim about a specific Hamiltonian and finite calculations, not a settled description of all systems with similar symmetries. The thermodynamic gap and dynamical exponent, the translation-breaking pattern of the nonuniform branch, and the response to explicit symmetry breaking at N=2 all require further work. The study is an arXiv version 1 preprint dated 25 Aug 2026.
Paper data and sources
Original title: Dynamical splitting and a nodal Bose liquid in 2d chiral XYZ model
Authors: Tarun Grover
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-25
DOI: Not available
Original paper · Full text