Strongly real 2-blocks and the Frobenius-Schur indicator. (Q2504134): Difference between revisions

Let \(G\) be a finite group of even order and let \(B\) be a 2-block of irreducible complex characters of \(G\). We say that \(B\) is a real block if it contains the complex conjugate of each of its irreducible constituents. It is known that \(B\) is real if and only if it contains a real-valued irreducible complex character. It is also a general fact that any 2-block has at least one real defect class. In the case of real 2-blocks, more is true, however, since we may define the concept of an extended defect group of such a block, this being a Sylow 2-subgroup of the extended centralizer of any element in a real defect class of the real block. It is a theorem that all such Sylow 2-subgroups are conjugate, so that the extended defect group of a real 2-block is well defined up to conjugacy. We say that a (non-principal) real 2-block is strongly real if an extended defect group of the block splits over each defect group. In view of the conjugacy theorem mentioned above, this is the same as saying that each element in any real defect class of the block is inverted by an involution. Let \(\Omega\) denote the subset of involutions in \(G\). \(G\) permutes the elements of \(\Omega\) by conjugation and we let \(\mathbb{C}\Omega\) denote the complex permutation module derived from this action. We also recall the concept of the Frobenius-Schur indicator \(\nu(\chi)\) of an irreducible complex character \(\chi\). We have \(\nu(\chi)=0\) if \(\chi\) is not real-valued, \(\nu(\chi)=1\) if \(\chi\) is the character of a real representation, and \(\nu(\chi)=-1\) if \(\chi\) is real-valued but is not the character of a real representation. The author proves the following remarkable theorem. Let \(B\) be a non-principal real 2-block of \(G\). Then the following are equivalent: (a) \(B\) is strongly real. (b) \(\sum\nu(\chi)\chi(1)>0\), the sum extending over all irreducible characters in \(B\). (c) \(B\) contains an irreducible constituent of the permutation module \(\mathbb{C}\Omega\). The proof of this theorem uses modular representation theory and exploits the author's key observation that the modular group algebra of \(G\) is a module for the wreath product \(G\wr C_2\) of \(G\) by a group of order 2. He is then able to extend J. A. Green's techniques of investigating 2-blocks of \(G\) by means of this larger group. An interesting consequence of his main theorem is the following. Let \(B\) be a non-principal real 2-block of \(G\). Let \(E\) be an extended defect group of \(B\) and let \(D\) be a defect group of \(B\) contained in \(E\) (with \(|E:D|=2\)). The reviewer proved that \(B\) contains a real-valued irreducible character \(\chi\) of height 0 and Frobenius-Schur indicator \(\nu(\chi)\) equal to \(-1\) if and only if \(E/D'\) does not split over \(D/D'\), where \(D'\) is the derived group of \(D\). Then if \(B\) is not strongly real, but \(E/D'\) does split over \(D/D'\), \(B\) contains an irreducible character of height greater than 0 whose Frobenius-Schur indicator equals \(-1\). It is uncommon to find sufficient conditions to predict the existence of characters of height greater than 0 in blocks of characters.