On higher rank coisotropic \(A\)-branes (Q2428140)

A decade ago \textit{A. Kapustin} and \textit{D. Orlov} [J. Geom. Phys. 48, No. 1, 84--99 (2003; Zbl 1029.81058)] realized that topological \(A\)-branes on a Calabi-Yau manifold \(M\) are not always supported on Lagrangian submanifolds, but may be represented by coisotropic submanifolds with extra structure. Their methods, derived from mirror symmetry, utilized boundary conditions on \(N=(2,2)\) supercurrents, and did not allow to go beyond rank one gauge groups. In contrast, this paper uses world-sheet analysis of supersymmetric non-linear sigma-models to tackle gauge fields of higher rank. The paper begins by reproducing the rank one results of Kapustin and Orlov via the new approach. Specifically, a coasotropic submanifold \(C\) carries the characteristic foliation \(\mathcal{F}\) induced by its conormal bundle, and its \(A\)-brane structure equips it with a line bundle and a connection on it. This connection induces a complex structure \(\widehat{J}\) on the bundle transverse to \(\mathcal{F}\). Therefore, the dimension of \(C\) can only be \(n+2k\), where \(2n\) is the (real) dimension of the entire manifold \(M\). In addition to non-perturbative effects from instantons, familiar from the Lagrangian \(A\)-branes, coisotropic ones also carry perturbative geometry induced by a non-commutative \(\widehat{J}\)-holomorphic product \(*\). This perturbative geometry is the only \(A\)-brane structure present on space-filling branes on \(M\) of even complex dimension, while it completely disappears on Lagrangian branes. The novel part of the paper deals with stacks of coisotropic \(A\)-branes and tachyon profiles of brane-antibrane pairs. The author splits off the non-abelian part of the gauge field and identifies additional structures by imposing the \(N=2_A\) supersymmetry. It turns out that a stack of \(A\)-branes is given by a complex of transversely \(\widehat{J}\)-holomorphic vector bundles (sheaves) with connections that are of type \((1,1)\) and vanish along the leaves of \(\mathcal{F}\). The differential \(Q\) is non-commutative in the sense of the above mentioned product, \(Q*Q=0\). This insures consistency of tachyon condensation after the \(A\)-twist is performed, which requires that the complex localize on a coisotropic submanifold again after the twist. The author remarks that the most general \(A\)-branes will involve complexes holomorphic with respect to different complex structures supported on different submanifolds, but at present it is unclear how to define holomorphy of morphisms in such generality, or how to compose them.