WZW orientifolds and finite group cohomology (Q2517957)

To define Feynman amplitudes of classical fields on the orientifold theory unambiguously, all inequivalent \(\Gamma\)-equivalent structures on the level-\(k\) gerbes on compact simply connected groups \(G\) are classified [cf. \textit{E. Witten}, Commun. Math. Phys. 92, 455--472 (1984; Zbl 0536.58012)]. Here \(\Gamma=\mathbb{Z}_2\ltimes Z\) is the orientifold groups, where \(\mathbb{Z}_2\) acts on \(G\) by a twisited inversion \(g\to(\zeta g)^{-1}\), \(\zeta\) is an element of the center of \(G\), and \(Z\) is a subgroup of the center of \(G\). The classification of the \(\Gamma\)-equivariant structures on the level-\(k\) gerbe on \(G\) descends to the classification of the Jandl structure on the induced gerbe on the quotient group \(G/Z\) [cf. \textit{U. Schreiber, C. Schweigert} and \textit{K. Waldorf}, Commun. Math. Phys. 274, 31--64 (2007; Zbl 1148.53057)]. The authors emphasize different Jandl structures giving rise to different orientifold extensions of the \(Z\)-orbifold theory, i.e. to different unoriented close-string theories with the target space \(G/\Gamma\). In this paper, this classification problem is reduced to the computation of \(H^p(\Gamma, U(1)_\varepsilon)\), \(p=2,3\). Here \(U(1)_\varepsilon\) means \(U(1)\) is considered with the action \(\lambda \to\lambda^{-1}\) of those elements of \(\Gamma\) which reverse the sign of \(H\), the 3-form curvature of the gerbe. Then by case by case computations, existence of even number induced Jandl structures is shown. It is also shown in all cases except for \(G=\text{Spin}(8n)\) and \(Z=\mathbb{Z}_2\times \mathbb{Z}_2\), the only obstructions to the existence of a \(\Gamma\)-equivariant structure with the trivial twist element \(\zeta=1\) are the ones that obstruct the existence of a \(Z\)-equivariant structure. In the exceptional case, \(Z\)-equivarinat structures exist for all integer levels \(k\) [cf. \textit{K. Gawȩdzki} and \textit{N. Reis}, J. Geom. Phys. 50, 28--55 (2004; Zbl 1067.22009)]. The outline of the paper is as follows: In \S2, a gerbe on a manifold \(M\) is defined by the local data \((B_i,A_{ij},g_{ijk})\) and its 3-form curvature \(H\) is defiend by \(\{dB_i\}\). When \(M\) is a connected and simply connected compact simple Lie group, gerbes are closely related to the WZW models, because there exists a gerbe such that \[ H= \frac{k}{12\pi}\text{tr}(g^{-1}dg)^3. \] Gerbes on \(M\) is classified by \(H^2(M,U(1))\) [\textit{K. Gawȩdzki}, Commun. Math. Phys. 258, 23--73 (2005; Zbl 1094.81047); \textit{M. K. Murray}, J. Lond. Math. Soc., II. Ser. 54, 403--416 (1996; Zbl 0867.55019 ); \textit{M. K. Murray} and \textit{D. Stevenson}, J. Lond. Math. Soc., II Ser. 62, 925--937 (2000; Zbl 1019.55009)]. As for non-abelian gerbe [cf. \textit{A. Asada}, Proc. Prospect Math. Sci., 13--40. World Sci. (1988; Zbl 0682.58044); Coll. Math. Soc. János Bolyai 56, 33--66 (1992; Zbl 0806.57014)]. Cohomology languages used in gerbe theory are also explained. Then gerbes on orbifolds are explained and assuming the action of \(\Gamma\) preserves \(H\), the first obstruction to the existence of a \(\Gamma\)-equivarint structure on a gerbe is given as a cohomology class in \(H^2(M,U(1))\). If this class vanishes, then the next obstruction is given as a cohomology class in \(H^2(\Gamma,H^1)\) and the last obstruction is given as a cohomology class in \(H^3(\Gamma,H^0)\). Here, \(H^i=H^i(M,U(1))\), \(i=0,1\). If \(H^1(\Gamma,H^1)=0\), then \(H^2(\Gamma,H^0)\) label inequivalent \(\Gamma\)-structures on a gerbe on \(M\). When \(H\) is not invariant by the action of \(\Gamma\), the coefficients \(U(1)\) in these cohomologies need to modify, and the coefficients \(U(1)_\varepsilon\) is derived. The Jandl structure, \(\Gamma\)-equivariant structure when \(\Gamma=\mathbb{Z}_2\) with \(\varepsilon(\pm 1)=\pm 1\), is also explained in this section (cf. \textit{U. Schreiber, C. Schweigert} and \textit{K. Waldorf} [Comm. Math. Phys. 274, 31--64 (2007; Zbl 1148.53057)]. Computations of \(H^p(\Gamma,U(1)_\varepsilon)\) is done by using Lyndon-Hochschild-Serre spectral sequence. This is explained in \S3. After these preparations, detailed studies of \(\Gamma\)-equivariant structures on the level-\(k\) gerbe on compact simply connected simple Lie groups are done in \S4. They are main part of this paper. Computations need to study of the effects of center and Dynkin diagram symmetry. After studying the casses \(G=SU(2)\), \(SU(3)\) and \(SU(4)\) as training examples, case by cases study are done. They begin from the case \(G=B_r=\text{Spin}(2r+1)\), then the cases \(G=C_r=Sp(2r)\), the most complicated case \(G=D_r=\text{Spin}(2r)\) where the case \(r=2s\) is discussed independently, and the cases \(G=E_6\) and \(E_7\) follow. As for other cases such as \(G=G_2,F_4, E_8\), computations are simple, because they have trivial center and no non-trivial Dynkin diagramm symmetry. Results are summarized as 3 pages table (\S6). The method of this paper based on a systematic approach to the classical orientifold theory. The authors say there is a study based on the sewing and modular invariance constraints for the crosscap states in the simple-current orbifolds of WZW theory which provides same results [\textit{I. Brunner} and \textit{K. Hor}, Notes on orientifolds of rational conformal field theories, JHEP0407(2004), 023].