Holography and Defect Conformal Field Theories
Oliver DeWolfe, Daniel Z. Freedman, Hirosi Ooguri
TL;DR
The paper constructs and analyzes a holographic dual for a defect conformal field theory realized by a D3/D5-brane system: AdS_5×S^5 with an AdS_4×S^2 defect is dual to 4D N=4 SU(N) SYM in R^4 coupled to a 3D N=4 defect hypermultiplet. It provides a complete defect action via the superspace boundary method, derives the KK spectrum of D5 open-string modes and matches them to gauge-invariant defect operators with integer conformal dimensions, and computes defect and mixed correlators from gravity to compare with weak-coupling field theory predictions. The study reveals plenty of structure from defect conformal symmetry, including one-point and mixed correlators, and discusses the (partial) quantum conformality results, highlighting differences from the standard N=4 SYM non-renormalization theorems. It also identifies open questions, such as proving non-Abelian conformality, exploring generalizations, and understanding backreaction and nonperturbative aspects of the dCFT holography.
Abstract
We develop both the gravity and field theory sides of the Karch-Randall conjecture that the near-horizon description of a certain D5-D3 brane configuration in string theory, realized as AdS_5 x S^5 bisected by an AdS_4 x S^2 "brane", is dual to N=4 Super Yang-Mills theory in R^4 coupled to an R^3 defect. We propose a complete Lagrangian for the field theory dual, a novel "defect superconformal field theory" wherein a subset of the fields of N=4 SYM interacts with a d=3 SU(N) fundamental hypermultiplet on the defect preserving conformal invariance and 8 supercharges. The Kaluza-Klein reduction of wrapped D5 modes on AdS_4 x S^2 leads to towers of short representations of OSp(4|4), and we construct the map to a set of dual gauge-invariant defect operators O_3 possessing integer conformal dimensions. Gravity calculations of <O_4> and <O_4O_3> are given. Spacetime and N-dependence matches expectations from dCFT, while the behavior as functions of lambda = g^2 N at strong and weak coupling is generically different. We comment on a class of correlators for which a non-renormalization theorem may still exist. Partial evidence for the conformality of the quantum theory is given, including a complete argument for the special case of a U(1) gauge group. Some weak coupling arguments which illuminate the duality are presented.
