Dying Dyons Don't Count

TL;DR

This work resolves the ambiguity in microscopic dyon counting for N=4, D=4 theories by linking the pole structure of the Siegel modular form to two-centered bound-state decays and proposing two contour prescriptions that restore S-duality in a moduli-dependent or moduli-independent (immortal) manner. The authors derive explicit jump formulas at walls of marginal stability and connect them to macroscopic Denef-type stability conditions, showing that large-charge attractor moduli eliminate two-centered bound states and yield an immortals-counting contour. The moduli-dependent contour precisely crosses the poles corresponding to decays, producing duality-invariant degeneracies across moduli space, while the attractor contour provides a universal, charge-based counting for large black holes. The results offer a coherent picture of dyon spectra, wall-crossing, and duality, with potential extensions to CHL models and insights for higher-derivative corrections.

Abstract

The dyonic 1/4-BPS states in 4D string theory with N=4 spacetime supersymmetry are counted by a Siegel modular form. The pole structure of the modular form leads to a contour dependence in the counting formula obscuring its duality invariance. We exhibit the relation between this ambiguity and the (dis-)appearance of bound states of 1/2-BPS configurations. Using this insight we propose a precise moduli-dependent contour prescription for the counting formula. We then show that the degeneracies are duality-invariant and are correctly adjusted at the walls of marginal stability to account for the (dis-)appearance of the two-centered bound states. Especially, for large black holes none of these bound states exists at the attractor point and none of these ambiguous poles contributes to the counting formula. Using this fact we also propose a second, moduli-independent contour which counts the "immortal dyons" that are stable everywhere.

Figures (2)

• Figure 1: (a) The Siegel upper-half plane for the modular form $\Phi$ is the future light-cone in the Minkowski space ${\mathbb R}^{1,2}$, and we consider the space of all contours to be a sheet of hyperboloid inside this light-cone, with all the points on the hyperboloid having a large distance from the origin. (b) A pole corresponding to an element $\gamma\in SL(2,{\mathbb Z})$ is a plane in ${\mathbb R}^{1,2}$ which always intersects the hyperboloid along a hyperbola.
• Figure 2: In this figure we show how the pole located at $\nu=0$ contributes to the degeneracy formula for contours with $\mathrm{Im}\nu>0$. (a) For charges with $P\cdot Q < 0$, one can deform the contour to the upper infinity of the cylinder where the integrand goes to zero without hitting the pole. (b) For charges with $P\cdot Q > 0$, one can deform the contour to the lower infinity of the cylinder, and by doing so pick up the residue of the pole.