On spherical twisted conjugacy classes. (Q1928553)

Let \(G\) be a simple algebraic group over an algebraically closed field \(k\), and let \(\theta\) be an automorphism of \(G\) arising from a symmetry of the Dynkin diagram of \(G\). Given \(x\in G\), the `\(\theta\)-twisted conjugacy class' of \(x\) in \(G\) is the set \(\{gx\theta(g)^{-1}\mid g\in G\}\). The aim of this paper is to characterise (under some restrictions on \(G\), \(\theta\) and \(k\)) the spherical \(\theta\)-twisted conjugacy classes in \(G\); that is, those classes which break up into a finite number of \(B\)-orbits for a Borel subgroup \(B\) of \(G\). In order to state the main result of the paper, we need one further bit of terminology. We can think of the automorphism \(\theta\) also as an automorphism of the Weyl group \(W\) of \(G\), and an element \(w\in W\) is called a `twisted involution' if \(w\theta(w)=1\). Then the main result of the paper is the following: if \(G\) is simply-connected, \(\theta\) arises from an involution of the Dynkin diagram, and \(k\) has odd characteristic which is good for \(G\), then the spherical \(\theta\)-twisted conjugacy classes in \(G\) are precisely those which only intersect Bruhat cells corresponding to twisted involutions in \(W\). The author also proves a dimension formula for these conjugacy classes, which replicates a known result in characteristic zero, and gives examples to show that this result does not capture the situation for the triality automorphism when \(G\) has type \(D_4\).