High Energy Modular Bootstrap, Global Symmetries and Defects
Sridip Pal, Zhengdi Sun
TL;DR
The paper develops a Tauberian-enhanced modular bootstrap for 2D unitary CFTs with topological defect lines, proving Cardy-like growth bounds for defect and symmetry sectors. It shows that all irreps of a finite global symmetry G appear in the untwisted sector with growth ∝ d_α^2/|G| times the seed density ρ_0(Δ), and extends to twisted sectors in the non-anomalous case, as well as to continuous U(1) charges with explicit Δ-dependent bounds. In the Schwarzian (JT gravity) limit, the results reproduce the expected bulk gauge-field factor d_α^2/|G|, and the framework covers Virasoro primaries for c>1 and large-c theories. The methodology relies on projecting partition functions onto sectors, using S-transforms to relate defect and symmetry channels, and applying vector-valued Tauberian arguments to obtain sharp, order-one bounded growth across sectors. Practical implications include holographic completeness intuition and quantitative sector-wise density bounds in theories with finite and continuous symmetries, plus potential extensions to other modular-function contexts.
Abstract
We derive Cardy-like formulas for the growth of operators in different sectors of unitary $2$ dimensional CFT in the presence of topological defect lines by putting an upper and lower bound on the number of states with scaling dimension in the interval $[Δ-δ,Δ+δ]$ for large $Δ$ at fixed $δ$. Consequently we prove that given any unitary modular invariant $2$D CFT symmetric under finite global symmetry $G$ (acting faithfully), all the irreducible representations of $G$ appear in the spectra of the untwisted sector; the growth of states is Cardy like and proportional to the ''square'' of the dimension of the irrep. In the Schwarzian limit, the result matches onto that of JT gravity with a bulk gauge theory. If the symmetry is non-anomalous, the result applies to any sector twisted by a group element. For $c>1$, the statements are true for Virasoro primaries. Furthermore, the results are applicable to large c CFTs. We also extend our results for the continuous $U(1)$ group.
