Wall crossing and M-theory (Q555213): Difference between revisions

The authors compute the BPS state partition function \(\mathcal{Z}_{BPS}\) of a single D6 brane, which is bound to any number of D0 and D2 branes, wrapped around a Calabi-Yau 3-fold \(X\) with no compact 4-cycles in all chambers of the Kähler moduli space. The resulting formula is \(\mathcal{Z}_{BPS}=\mathcal{Z}_{top}^2|\text{ chamber}\), where \(\mathcal{Z}_{top}\) is the A-model topological string partition function generating the Gromov-Witten invariants of \(X\), and the restriction is to an explicitly described subspace of states distinguished by the chamber of the moduli. The formula generalizes known counts for the conifold and generalized conifold geometries. Examples of the resolved conifold and a related non-toric variety are worked out explicitly. The authors also show that their formula gives an M-theory derivation of the ``semi-primitive'' wall crossing formula of Denef-Moore. Computation starts by lifting the D6 brane to the Taub-NUT space with the unit charge, D0 and D2 branes are represented by M2 branes and gravitons in it respectively. In the Taub-NUT geometry single particle BPS states can be counted directly, while a special choice of the background moduli ensures that at far away separation all bound states reduce to the single particle ones. The absence of compact 4-cycles in the Calabi-Yau is crucial for this reduction to work, but the 3-fold does not have to be toric. The BPS states are organized into a free field Fock space with generators corresponding to the spinning M2 branes in the M-theory compactified down to 5 dimensions.