Walter's basic theorem for fusion systems (Q827053)

One of the milestones in the classification of the finite simple groups is \textit{J. H. Walter}'s proof of the B-Conjecture together with a characterization of the Chevalley groups [The \({\mathfrak B}\)-conjecture; characterization of Chevalley groups. Providence, RI: American Mathematical Society (AMS) (1986; Zbl 0592.20017)]. To simplify things in this review we will restrict ourself to simple groups \(G\). A 2-component \(L\) of a group \(G\) is a subnormal perfect subgroup such that \(L/O(L)\) is quasisimple. A component is a 2-component with \(L\) with \(O(L) = 1\). The B-conjecture basically says that for \(G\) simple the 2-components of \(C_G(i)\), \(i\) an involution, are components. One step in proving this is Theorem III in Walter's paper above, which in the paper under review is called Walter's basic theorem. To state this and also the result of the paper under review we need some definitions. By \(\mathrm{Chev}(p)\) denote the quasisimple Chevalley groups and their twisted analogues over a field of characteristic \(p\). In the paper, \(\mathrm{Chev}^\ast(p)\) is \(\mathrm{Chev}(p)\) without some specific small groups. Now assume that \(G\) is simple and some inductive properties hold (for example the B-conjecture holds in all proper sections), and assume that there is some involution \(i\in G\) such that \(C_G(i)/O(C_G(i))\) has some component \(L\) of type \(\mathrm{Chev}^\ast(p)\). Then either \(G\) itself is in \(\mathrm{Chev}^\ast(p)\) or we have a restricted set of components \(L\) as \(\mathrm{SL}_2(q)\), for \(q = 5,7,9\), \(\hat{A}_n\), or a central extension of \(\mathrm{PSL}_3(4)\). There is a program started by Michael Aschbacher to classify the simple saturated 2-fusion systems and then use this result for a new classification of the finite simple groups. The advantage is that normal subgroups of odd order do not play any role in fusion systems, in particular a proof of the B-conjecture will not show up. In this paper, the author proves the analogue of theorem III above in the world of fusion systems. Hence, it is not considered as a step in proving a B-conjecture it is a step in classifying some 2-fusion systems related to Chevalley groups. First, the analogue of \(\mathrm{Chev}^\ast(p)\) is \(\mathrm{Chev}^\ast[m]\), which is the class of 2-fusion systems of groups \(G(q) \in \mathrm{Chev}^\ast(p)\) for \(q = p^f\) and \(p\) some prime such that \(m\) is the 2-share of \(q^2-1\). We also write \(\mathrm{SL}_2[m]\), \(\mathrm{Spin}_7[m]\) for the corresponding 2-fusion systems for \(\mathrm{SL}_2(q)\) or \(\mathrm{Spin}_7(q)\). Further, \(\mathrm{Chev}[\mathrm{large}]\) consists of all of \(\mathrm{Chev}^\ast[m]\), \(m \geq 8\), again with a few exceptions for \(m = 8\). Also in the world of 2-fusion systems one may speak about components of centralizers of involutions. Now, the author proves a version of Walter's theorem, which reads as follows: Let \(\mathcal{F}\) be a saturated fusion system on a finite 2-group \(S\). Assume that components of centralizers of involutions are in the class of 2-fusion systems of the know quasisimple fusion systems (this is an inductive hypothesis). If there is some component, which is in \(\mathrm{Chev}[\mathrm{large}]\), then there is a fully centralized involution \(t\) and an intrinsic component \(\mathcal L\) of \(C_{\mathcal F}(t)\), where \(\mathcal L\) is of type \(\mathrm{SL}_2[m]\), \(\mathrm{Spin}_7[m]\), \(m \geq 8\), \({\mathcal L}/Z({\mathcal L})\) is the 2-fusion system of \(L_3(4)\), or \({\mathcal L}\) is the 2-fusion system of \(\hat{A}_n\), \(n \geq 8\). The first class did not show up in Walter's result due to Aschbacher's classical involution theorem, which yields that \(G \in\mathrm{Chev}(p)\) in Walter's theorem. In the fusion system world for \(\mathcal F\) simple there is the corresponding result due to Aschbacher on quaternion fusion packets, which gives that for \(\mathcal L\) of type \(\mathrm{SL}_2[m]\) we receive that \(\mathcal F\) is of type \(\mathrm{Chev}^\ast[m]\). Moreover, the case of \(\mathrm{Spin}_7[m]\) did not show up in Walter's paper. The occurrence in the theorem of fusion systems is due to the existence of the exotic Benson-Solomon systems.