Endotrivial modules for the general linear group in a nondefining characteristic. (Q484354)

Let \(k\) be an algebraically closed field of prime characteristic \(p\) and \(G\) be a finite group with order divisible by \(p\). A (finite-dimensional) \(kG\)-module \(M\) is said to be endotrivial if its endomorphism algebra \(\text{End}_k(M)\) is isomorphic (as a \(kG\)-module) to the direct sum of the trivial module \(k\) and a projective module. In other words, the endomorphism algebra is trivial in the stable module category. One can define an equivalence relation on endotrivial modules so that the set \(T(G)\) of equivalence classes of such modules forms a group, which is known to be finitely generated and abelian. A longstanding open problem is to determine \(T(G)\) for arbitrary finite groups. In this paper, the authors focus on groups that are ``almost'' special or general linear in a nondefining characteristic. More precisely, the authors consider finite groups \(G\) with \(\text{SL}(n,q)\subseteq G\subseteq\text{GL}(n,q)\) where \(q\) is not divisible by \(p\). The authors determine \(T(G/Z)\) for any central subgroup \(Z\subset G\) under the assumption that the Sylow \(p\)-subgroup of \(G\) is abelian. Note that any endotrivial \(k(G/Z)\)-module must inflate to an endotrivial \(kG\)-module. The authors first investigate the torsion free portion of \(T(G)\) for a group \(G\) as above. If the Sylow \(p\)-subgroup of \(G\) is cyclic, then this is trivial. Otherwise, it is shown to be of dimension one (and a generator is given), independent of whether the Sylow subgroup is abelian or not. It remains then to determine the torsion portion. The case when the Sylow subgroup is cyclic follows from known results, and the bulk of the paper is devoted to dealing with the non-cyclic (but still abelian) case. Via Clifford theory, the proof is essentially reduced to the case of \(\text{SL}(n,q)\), and the latter case is broken into a number of sub-cases based on the relationship between \(n\), \(q\), and \(p\). A key tool is recent work of \textit{P. Balmer} [J. Eur. Math. Soc. (JEMS) 15, No. 6, 2061-2079 (2013; Zbl 1291.20009)] which gives a combinatorial method for computing the kernel of the restriction map \(T(G)\to T(H)\) for a subgroup \(H\subset G\). The proof also makes use of Young modules for \(\text{GL}(n,q)\) which arise as indecomposable summands of the module obtained by inducing the trivial module from a Levi subgroup to \(\text{GL}(n,q)\).