On reachable elements and the boundary of nilpotent orbits in simple Lie algebras (Q2387398)

Let \(G\) be a connected complex simple algebraic group with Lie algebra \({\mathfrak g}\). If \(e\in {\mathfrak g}\), then \({\mathfrak g}_e\) is the centralizer of \(e\) and \(Ge\subset{\mathfrak g}\) is the \(G\)-orbit (conjugacy class) of \(e\). \textit{A. G. Elashvili} and \textit{G. Grélaud} [C. R. Acad. Sci. Paris, Sér. I 317, 445--447 (1993; Zbl 0814.17005)] gave a classification of elements of \({\mathfrak g}\) having the property \(e\in [{\mathfrak g}_e, {\mathfrak g}_e]\). Such elements (and their orbits) are called reachable in the present paper. Clearly, each reachable element is nilpotent. The boundary of \(Ge\) is the subvariety \(\partial(Ge)=\overline{Ge}\setminus Ge\). It was noticed by D.~Vogan that the reachable orbits in the classical Lie algebras are exactly the nilpotent orbits having the property \( \text{codim}_{\overline{Ge}}\partial(Ge)\geq 4. \) A direct verification based on known classifications shows that this coincidence remains true for exceptional Lie algebras as well. The author gives an a priori proof of this coincidence for the nilpotent elements of height 2. Some general estimates for the boundary of an arbitrary nilpotent orbit is given and various properties of reachable elements are investigated.