Interacting Chern insulator transition on the sphere: revealing the Gross–Neveu–Yukawa criticality

Zhi-Qiang Gao, Taige Wang, Dung-Hai Lee · 2025

Read Interacting Chern insulator transition on the sphere: revealing the Gross–Neveu–Yukawa criticality by Zhi-Qiang Gao, Taige Wang, Dung-Hai Lee (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-15338.txt.

Introduction

Zhi-Qiang Gao, Taige Wang, and Dung-Hai Lee study a 2025 physics question about how a material shifts between two phases of a Chern insulator. With strong particle links, the shift point may match a known class called Gross-Neveu-Yukawa. They test the case with two fermion types.

Placing the fermions on a sphere lets the team use full symmetry and skip a known parity flaw. Exact state lists for small systems gave energy gaps that match scale values from field theory work. The match held for known low-scale terms and also showed new higher terms not seen before.

Catalog record

CREATORZhi-Qiang Gao, Taige Wang, Dung-Hai Lee
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE72 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-15338

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Interacting Chern insulator transition on the sphere: revealing the Gross–Neveu–Yukawa criticality Zhi-Qiang Gao, Taige Wang, Dung-Hai Lee Abstract In two dimensions a Chern insulator can be tuned from a topological state to an ordinary band insulator without breaking any symmetry, offering a clean example of “beyond-Landau” criticality. We show that electron interactions drive this transition to the Gross–Neveu–Yukawa (GNY) fixed point, a benchmark problem of broad interest across high-energy and condensed-matter physics. By solving the minimal Dirac model on the surface of a sphere, we obtain the full spectrum of critical excitations from systems containing only a few electrons. The leading critical exponents match state-of-the-art bootstrap bounds, and several previously unseen collective modes appear. Our results demonstrate that placing a Dirac cone on a curved geometry provides a surprisingly simple and accurate setup for strongly interacting topological transitions. Figure 1: Mean-field phase diagram in the (h,u_{0}) plane. Topological (h<0) and trivial (h>0) phases meet via a continuous transition for u<u_{c} (red line), described by a free-fermion CFT, which becomes first-order (dashed line) for u>u_{c} . The point u=u_{c} (blue dot) is a GNY transition where fermions and the mass operator both get gapless. Introduction. Quantum phases can differ in topology even when they share all microscopic symmetries. Because such changes of topology fall outside the Landau framework based on symmetry breaking, they have become touchstones for modern theories of phase transitions. A widely discussed two–dimensional example is the passage from a trivial band insulator to a Chern insulator: as one tunes a control parameter, the bulk gap collapses while a single chiral edge channel and a quantised Hall response switch on [ 1 , 2 ] . The essential low-energy degrees of freedom are those of a single Dirac cone. Introducing the shorthand \bm{\sigma}\cdot\nabla=\sigma_{x}\partial_{x}+\sigma_{y}\partial_{y} , we collect kinetic, mass, and interactions into one Hamiltonian \tilde{H}=\int d^{2}x\Bigl[\psi^{\dagger}\bigl(-i\,\bm{\sigma}\cdot\nabla+h\sigma_{z}\bigr)\psi-\frac{u_{0}}{2}\bigl(\psi^{\dagger}\sigma_{z}\psi\bigr)^{2}\Bigr]. (1) Changing the sign of the single-particle mass h changes the Chern number by \pm 1 and produces a free-fermion conformal field theory (CFT) at h=0 . For weak interaction u_{0} the quartic term is irrelevant, so the transition remains in that free-Dirac class. At stronger coupling u_{c} the system develops a spontaneous mass \langle\psi^{\dagger}\sigma_{z}\psi\rangle\neq 0 ; the point where this ordering collides with the band-inversion line is a multicritical fixed point that belongs to the Gross–Neveu–Yukawa (GNY) universality class, as sketched in Fig. 1 . A Hubbard–Stratonovich decoupling of Eq. ( 1 ) maps the lattice model exactly onto the GNY action (see End Matter for details), \mathcal{L}_{\text{GNY}}=\bar{\psi}_{\alpha}\bigl(i\gamma^{\mu}\partial_{\mu}-g\phi\bigr)\psi_{\alpha}+\frac{1}{2}(\partial_{\mu}\phi)^{2}+\frac{t}{2}\phi^{2}+\frac{\lambda}{4}\phi^{4}, (2) where a single Dirac fermion \psi is Yukawa-coupled to an Ising-like scalar field \phi . The GNY point is of considerable interest to both high-energy and condensed-matter communities, including chiral symmetry breaking in QCD [ 3 ] , Kekulé ordering in graphene [ 4 ] , and nodal superconductivity [ 5 ] . Analytical and numerical studies—conformal bootstrap [ 6 , 7 , 8 ] , \epsilon expansions [ 9 ] , large- N techniques [ 10 , 11 ] , and sign-problem-free quantum Monte Carlo (QMC) [ 12 , 13 , 14 , 15 ] —agree on the leading critical exponents yet disagree on many sub-leading operators, leaving the full spectrum unsettled. A faithful numerical treatment of an interacting single Dirac cone is notoriously difficult for four distinct reasons. First, any local, time-reversal-symmetric lattice regularization suffers from the parity anomaly [ 16 ] . Standard approaches enlarge the flavor number [ 13 , 14 ] , break time-reversal symmetry [ 17 ] , or adopt non-local hoppings [ 12 ] , each of which can obscure the intended universality class. Second, the critical interacting Dirac fermion is intrinsically strongly correlated. It cannot be reached perturbatively from the free Dirac fermion CFT, as the four-fermion interaction is irrelevant at the free fixed point. Third, without SO(3) symmetry, most microscopic calculations can only extract critical exponents from order parameters and correlation functions instead of using state-operator correspondence to read out scaling dimensions of full operator content of the CFT [ 18 ] . They also typically suffer from relatively large numerical error due to the lack of enough symmetry. Finally, even on the sphere that preserves SO(3) , maintaining the linear dispersion of the Dirac fermion is nontrivial: standard spectrum-based fuzzy sphere constructions [ 18 , 19 , 20 , 21 , 22 , 23 , 24 , 25 ] are not suitable for Dirac fermions, since the monopole at the sphere center generates Landau levels instead of the target linear spectrum. Here we sidestep all these obstacles by formulating the interacting Dirac cone directly on the sphere, which generalizes the fuzzy sphere framework to a critical point of interacting Dirac fermion. Curvature supplies the necessary spin connection, while the overall geometry—with its full rotational symmetry—greatly reduces finite-size effects. Exact diagonalization of systems containing fewer than ten electrons already yields excitation spectra whose level spacings translate directly into operator dimensions. The leading values coincide with the best bootstrap bounds, and several higher-dimension primaries, including an almost-conserved charged current, appear for the first time. Spherical geometry therefore offers a compact and transparent numerical microscope for the Gross–Neveu–Yukawa fixed point and for Dirac criticality more broadly. Figure 2: (a) Single-particle energy levels \mathcal{E}_{jm}^{\ell} of free Dirac fermions on the sphere. Levels up to j\leq j_{\mathrm{max}}=5/2 are shown. (b) Rescaled order parameter \langle\tilde{m}^{2}\rangle of the ground state across the paramagnetic-to-ferromagnetic transition as a function of the contact interaction U_{0} (see main text for details). Dirac fermions on \mathbb{S}^{2} . Setting the radius of the sphere to 1, the Dirac operator takes the form D=i\gamma^{\mu}\nabla_{\mu}, (3) where \nabla_{\mu}=\partial_{\mu}+\Omega_{\mu} is the covariant derivative for spinor fields, and \Omega_{\mu} is the spin connection. On \mathbb{S}^{2} , this connection is constant, leading to [ 26 ] \displaystyle D=\bm{\sigma}\cdot\mathbf{L}+1=\mathbf{J}^{2}-\mathbf{L}^{2}+\frac{1}{4}, (4) where \mathbf{L} denotes the orbital angular momentum, and \mathbf{J}=\mathbf{L}+\tfrac{1}{2}\bm{\sigma} is the total angular momentum. From this operator, we construct the free Dirac Hamiltonian H_{0}=\int d\Omega\,\psi^{\dagger}(\Omega)D(\Omega)\psi(\Omega)+h\psi^{\dagger}(\Omega)\sigma_{z}\psi(\Omega). (5) Besides the spatial \mathrm{SO}(3) and global charge \mathrm{U(1)} symmetries, this Hamiltonian also features three discrete \mathbb{Z}_{2} symmetries. When h=0 , it possesses a time-reversal \mathcal{T} symmetry that spontaneously breaks across the transition. It also admits charge conjugation \mathcal{C} (though \mathcal{C} does not commute with the global \mathrm{U(1)} ). A spacetime parity \mathcal{P} symmetry is also present, but it does not commute with \mathrm{SO}(3) . We elaborate on how \mathcal{C} , \mathcal{T} , and \mathcal{P} act on the fermionic operators in the Supplementary Material [ 27 ] . Crucially, preserving these symmetries at the GNY criticality ensures that only one relevant operator remains. In particular, maintaining \mathcal{T} in the interacting regime confines the system to the h=0 line of Fig. 1 , enabling the multicritical point to be approached by tuning just a single parameter u_{0} . The single-particle eigenstates of D are the spinor spherical harmonics \mathcal{Y}_{jm}^{\ell} , satisfying D\,\mathcal{Y}_{jm}^{\ell}=\mathcal{E}_{jm}^{\ell}\mathcal{Y}_{jm}^{\ell},\quad\mathcal{E}_{jm}^{\ell}=\pm\bigl(j+\frac{1}{2}\bigr)\text{ for }\ell=j\mp\frac{1}{2}, (6) with m=-j,-j+1,\dots,j and j=\tfrac{1}{2},\tfrac{3}{2},\dots . We show the single-particle spectrum \mathcal{E}_{jm} in Fig. 2 (a). These solutions are related to the ones in Ref. 28 by \Upsilon_{l,m}^{\pm}=V(\Omega)\,\mathcal{Y}_{lm}^{\,l\mp 1/2} , where V(\Omega) generates a local coordinate transformation such that V^{\dagger}(\Omega)\bigl(\hat{\mathbf{n}}(\Omega)\cdot\bm{\sigma}\bigr)V(\Omega)=\sigma_{z} . In the remainder of the paper, we shall treat j_{\rm max} as playing the role of finite system size. This is justified by the

Read the complete text