Autoequivalences of blocks and a conjecture of Zassenhaus (Q1901019)

The authors deal with the Zassenhaus conjecture for certain finite simple groups. For a finite group \(G\) its integral group ring is denoted by \(\mathbb{Z} G\). The Zassenhaus conjecture claims that \(\mathbb{Z} G=\mathbb{Z} H\) as augmented algebras implies that \(a\cdot G\cdot a^{-1}=H\) for a unit \(a\in\mathbb{Q} G\). The authors verify this conjecture for \(\text{PSL}(2,p)\) with \(p\) a rational prime, for the five simple Mathieu groups, for \(A_9\) and for the sporadic Thompson group. For simple groups, the Zassenhaus conjecture is equivalent to the existence of a Zassenhaus factorization of augmented automorphisms of \(\mathbb{Z} G\): Given an augmented automorphism \(\alpha\) of \(\mathbb{Z} G\), then it has a factorization \(\alpha=\rho\cdot\gamma\), where \(\rho\in\text{Aut}(G)\) and \(\gamma\) is a central automorphism of \(\mathbb{Z} G\); i.e. it is conjugation with a unit in \(\mathbb{Q} G\) centralizing \(\mathbb{Z} G\). The proof is an interesting interplay between ordinary and modular representation theory: \(\alpha\) has a Zassenhaus factorization if and only if it acts like a group automorphism on the character table of \(G\). In order to verify this in the special examples modular representation theory is invoked: Let \(B\) be a block of \({\mathfrak k} G\) of cyclic defect (\(\mathfrak k\) must be large enough). If \(\alpha : B\to B\) is an automorphism fixing a module corresponding to an end of the Brauer-tree, then it fixes all finitely generated modules of \(B\). (Reviewer's remark: This result is definitely false if one replaces \(\mathfrak k\) by a large enough \(p\)-adic ring of integers.) This observation, together with an analysis of the character tables and decomposition matrices allows one to verify the Zassenhaus conjecture for the above mentioned groups.