Bi-branes: Target space geometry for world sheet topological defects (Q927347)

The relevant geometric data for the target space description of world sheet topological defect are shown to be submanifolds in \(M_1\times M_2\), when two compatible conformal field theories corresponding to target spaces \(M_1\), \(M_2\) are given. Here, a collection of conformal field theories is said to be compatible if they share a chiral symmetry algebra including the Virasoro algebra. These submanifolds are called bi-branes. For example, if \(M_1= M_2= G\), a simply connected compact semisimple Lie group, then this submanifold is shown to be a biconjugacy class \[ {\mathcal B}_{h_1,h_2}= \{(g_1, g_2)\in G\times G\mid g_1= xh_1 y^{-1},\,g_2= xh_2 y^{-1}\text{ for some }x,y\in G\}. \] In the lattice version of the Ising model, a defect is produced changing the coupling along all bonds that cross a specified line from ferromagnetic to antiferromagnetic. Due to the \(\mathbb{Z}_2\)-gauge invariance of the Ising model, the defect comes with a well-defined rule for passing insertion in the bulk through the defect line. Hence there is a mixed fusion by which the topological defect acts on the conformal boundary condition (\S2). When \(M_1= M_2= G\), this fusion is analysed via the study of \({\mathfrak g}\oplus{\mathfrak g}\) modules \(H_\lambda\boxtimes H_{\lambda^+}\), where \({\mathfrak g}\) is the non-trivial central extension of the loop algebra of \(\widehat{\mathfrak g}\), the Lie algebra of \(G\), according to \textit{J. L. Cardy} [Boundary conditions, fusion rules and the Verlinde formula, Nuclear Phys. B 324, 581 (1989)] algebraic formual for fusion is derived ((9),(10)). On the other hand, by the Peter-Weyl theorem, \(\otimes_\lambda H_\lambda\boxtimes H_{\lambda^+}\) can be identified with a subspace of \({\mathcal F}(G)\), the space of functions on \(G\). Then, using Fourier transformation and applying the methods in [\textit{G. Felder}, \textit{J. Fröhlich}, \textit{J. Fuchs} and \textit{C. Schweigert}, J. Geom. Phys. 34, No.~2, 162--190 (2000; Zbl 1002.81042)], algebraic results are translated to analytic results and the boundary state for a given topological defect is shown to concentrate on \({\mathcal B}_{h_1,h_2}\) (\S2). The rest of the paper studies WZW bi-branes (\S3. The world volume of WZW bi-branes, \S4. The Wess-Zumino term in the presence of defects, and \S5. Fusion of bi-branes). They are natural extension of the classical theory of WZW models [cf. \textit{E. Witten}, Commun. Math. Phys. 92, 455--472 (1984; Zbl 0536.58012)]. But since a bi-brane is a submanifold of \(M_1\times M_2\) and the conformal invariance of theories with nonabelian currents require a non-trivial \(B\)-field background (Witten, loc. cit.), which is a Hermitian gerbe, not a bundle, instead of ordinary (co)homology theory and the geometry of fibre bundles, birelative (co)homology and geometry of gerbes are used. Singular birelative homology and birelative de Rham cohomology are explained in appendix A. It allows to consider bi-brane WZW action \[ S(\phi_1,\phi_2]= \int_{B_1} H_1+ \int_{B_2} H_2+ \int_D \overline\omega, \] ((45), (A.22)) to represent a birelative cohomology class. Here, \(\partial B_1\) is \(\phi_1(\Sigma_1)\cup p_1(\overline D)\), \(\overline D\) is the manifold \(D\) with opposite orientation, and \(\partial B_2= \phi_2(\Sigma_2)\cup p_2(D)\), where \(\phi_i: \Sigma_i\to M_i\), \(\Sigma= \Sigma\cup_S \Sigma_2\), \(\partial\Sigma_1= S\), \(\partial\Sigma_2= \overline S\), \(\phi_S: S\to M_1\times M_2\) and \(\partial D= \phi_S(S)\). In Appendix B, after explaining the gerbe module [cf. \textit{K. Gawȩdzi} and \textit{N. Reis}, Rev. Math. Phys. 14, No. 12, 1281--1334 (2002; Zbl 1033.81067)] and the gerbe bimodule, holonomy of gerbes in the presence of defects is explained, and show that \[ \exp(iS(\phi_1, \phi_2))= \text{hol}_{{\mathcal G}_1,{\mathcal G}_2,{\mathcal D}})\Sigma, S), \] as expected (B.29). In \S6, the last section, after discussing possibilities of extension of the results to more general classes of conformal field theories, such as theories of several bosons compactified on a torus and to orbifolds, the authors state the expectation, that defects, and more generally, the algebraic and categorical structure behind RCFT correlators, will enter in a CFT-inspired approach to the Langlands program [cf. \textit{A. N. Naupstin} and \textit{E. Witten}, Electric-magnetic duality and the geometric Langlands program, \url{arXiv:hep-th/0604151}].