This paper studies a version of the kagome antiferromagnet designed to model recently discovered materials. The model, called the 3J model, has three unequal nearest-neighbor exchange couplings on the kagome lattice to mimic the weak distortion present in candidate materials such as YCu3(OH)6Br2[Br1−x(OH)x]. The main result is that, as the three couplings are varied, the authors find an intermediate quantum spin liquid (QSL) phase sitting between two magnetically ordered phases. They also report that this QSL is not the same as the spin liquid thought to appear at the fully symmetric (isotropic) kagome point, but rather a distinct phase stabilized by the bond anisotropy in the 3J model.

To reach these conclusions the team combined several state-of-the-art numerical tools. For ground-state properties they used large-scale density-matrix renormalization group (DMRG), a well-established numerical method for strongly correlated quantum systems, together with neural quantum states (NQS), which use neural networks (here a Vision Transformer architecture) to represent quantum wavefunctions. For finite temperatures they used advanced tensor-network methods (XTRG and tanTRG) to compute thermal observables. The calculations were performed on cylindrical geometries that capture the kagome lattice physics while remaining computationally tractable, and the complementarity of methods helped cross-check the findings.

A key link to experiment comes from the calculated specific heat. In the magnetically ordered phase the authors find a shoulder in the quantity specific heat divided by temperature (C/T) at a temperature that is a fraction of the dominant coupling scale Jhex. That shoulder disappears inside the QSL phase. The paper reports that this pattern—C/T with a shoulder in ordered samples and no shoulder in QSL candidate samples—matches measurements on several experimental samples, including an ordered YCu3(OH)6Cl compound and QSL-candidate YCu3(OH)6Br2[Br1−x(OH)x] and LuCu3(OH)6Br2[Br1−x(OH)x] samples. This makes a direct connection between the microscopic 3J model and experimental thermodynamics.