Crystal Melting and Toric Calabi-Yau Manifolds
Hirosi Ooguri, Masahito Yamazaki
TL;DR
The paper constructs a crystal-melting statistical model that counts BPS bound states of D0 and D2 branes bound to a single D6 on arbitrary toric Calabi–Yau threefolds. This model is rooted in quiver quantum mechanics and brane tilings, where the crystal structure and melting rules are dictated by the path algebra $A$ and its F- and D-term constraints, encoded via $ heta$-stability and non-commutative Donaldson–Thomas theory. A one-to-one correspondence is established between molten crystal configurations and $ heta$-stable $A$-modules, thereby equating crystal melting configurations with BPS bound states and reproducing DT indices. The work also clarifies how the topological string partition function and crystal-melting counts relate through wall-crossing, offering a unified perspective on DT invariants, topological strings, and toric geometry with potential links to mirror symmetry and black-hole microstate physics.
Abstract
We construct a statistical model of crystal melting to count BPS bound states of D0 and D2 branes on a single D6 brane wrapping an arbitrary toric Calabi-Yau threefold. The three-dimensional crystalline structure is determined by the quiver diagram and the brane tiling which characterize the low energy effective theory of D branes. The crystal is composed of atoms of different colors, each of which corresponds to a node of the quiver diagram, and the chemical bond is dictated by the arrows of the quiver diagram. BPS states are constructed by removing atoms from the crystal. This generalizes the earlier results on the BPS state counting to an arbitrary non-compact toric Calabi-Yau manifold. We point out that a proper understanding of the relation between the topological string theory and the crystal melting involves the wall crossing in the Donaldson-Thomas theory.
