Generic singularities of nilpotent orbit closures (Q340388): Difference between revisions

Let \(G\) be a connected, simple algebraic group of adjoint type over \(\mathbb C\) with corresponding Lie algebra \(\mathfrak g\). The orbit \(\mathcal O\) of a nilpotent element in \(\mathfrak g\) under the adjoint action is called a nilpotent orbit, and its closure \(\overline{\mathcal O}\) is a union of finitely many nilpotent orbits. The closure inclusion is a partial order on the set of nilpotent orbits, and this article considers the generic singularities, the singularities of \(\overline{\mathcal O}\) at points of maximal orbits of its singular locus. \textit{H. Kraft} and \textit{C. Procesi} [Invent. Math. 62, 503--515 (1981; Zbl 0478.14040)] have determined the generic singularities for the classical types of Lie algebras, \textit{E. Brieskorn} [Actes Congr. internat. Math. 1970, 2, 279-284 (1971; Zbl 0223.22012)] and \textit{P. Slodowy} [``Simple singularities and simple algebraic groups''. Lect. Notes Math. 815 (1980; Zbl 0441.14002)] determined them for the whole nilpotent cones \(\mathcal N\) for \(\mathfrak g\) of any type. The goal of this paper is to determine the generic singularities of \(\overline{\mathcal O}\) when \(\mathfrak g\) is of exceptional type. A basic result is that the singular locus of \(\overline{\mathcal O}\) coincides with the boundary of \(\mathcal O\) in \(\overline{\mathcal O}\). To study generic singularities of \(\overline{\mathcal O}\), it is sufficient to study each maximal orbit \(\mathcal O^\prime\) in the boundary \(\overline{\mathcal O}\) of \(\mathcal O\). Such an \(\mathcal O^\prime\) is called a \textit{minimal degeneration} of \(\mathcal O\). The local geometry of \(\overline{\mathcal O}\) at \(e\in\mathcal O^\prime\) is determined by the intersection of \(\overline{\mathcal O}\) with a transverse slice to \(\mathcal O^\prime\) at \(e\) in \(\mathfrak g\). Such slices exists in all cases, given by the affine space \(\mathcal S_e=e+\mathfrak g^f\), called the \textit{Slodowy slice}, where \(e,f\) are the nilpotent parts of an \(\mathfrak{sl}_2 \)-triple, and \(\mathfrak g^f\) is the centralizer of \(f\) in \(\mathfrak g\). The local geometry is thus encoded in \(\mathcal S_{\mathcal O,e}=\overline{\mathcal O}\cap\mathcal S_e\) which is named a \textit{nilpotent Slodowy slice}. If \(\mathcal O^\prime\) is a minimal degeneration of \(\mathcal O\), \(\mathcal S_{\mathcal O,e}\) has an isolated singularity at \(e\), and the generic singularities of \(\overline{\mathcal O}\) can be determined by studying the various \(\mathcal S_{\mathcal O,e}\) as \(\mathcal O^\prime\) runs over all minimal degenerations and \(e\in\mathcal O^\prime\), the isomorphism type of \(\mathcal S_{\mathcal O,e}\) being independent of \(e\). The main result of the article is a classification of \(\mathcal S_{\mathcal O,e}\) up to algebraic isomorphism for each minimal degeneration \(\mathcal O^\prime\) of \(\mathcal O\) in the exceptional types. In a few cases, the result is able to determine the normalization of \(\mathcal S_{\mathcal O,e}\), and in another few cases, \(\mathcal S_{\mathcal O,e}\) is only determined up to local analytic isomorphism. The article uses the theory of symplectic varieties. By \textit{Y. Namikawa} [J. Reine Angew. Math. 539, 123--147 (2001; Zbl 0996.53050)], a normal variety is symplectic if and only if its singularities are rational, Gorenstein, and its smooth part carries a holomorphic symplectic form, and it is proved that the normalization of a nilpotent orbit \(\overline{\mathcal O}\) is a symplectic variety. Because the normalization \(\overline{\mathcal O}\) has rational Gorenstein singularities, so has the normalization \(\tilde{\mathcal S}_{\mathcal O,e}\), and it is a symplectic variety. A singularity of a symplectic variety is what is called a \textit{symplectic singularity}, and the authors claim that a better understanding of those is important for studying the conjecture that a Fano contact manifold is homogeneous. Thus it is important to find examples of symplectic singularities, and the study of the isolated symplectic singularity \(\tilde{\mathcal S}_{\mathcal O,e}\) contributes to this. The results in the article is motivated by representation theory. The geometry of the nilpotent cone \(\mathcal N\) was important in Springer's construction of Weyl group representations and the resulting Springer correspondence. It is proved that modular representation theory of the Weyl group of \(\mathfrak g\) is encoded in the geometry of \(\mathcal N\). Its decomposition matrix is a part of the decomposition matrix for equivariant perverse sheaves on \(\mathcal N\). Also, the authors remark that the reappearance of certain singularities in different nilpotent cones leads to equalities between parts of decomposition matrices. In the \(\text{GL}_n\)-case, the row and column removal rule for nilpotent singularities gives a geometric explanation for a similar rule for decomposition matrices of symmetric groups. The main results in the article concerns simple surface singularities and their symmetries. A finite subgroup \(\Gamma\subset\text{SL}_2(\mathbb C)\cong\text{Sp}_2\) acts on \(\mathbb C^2\) and the quotient variety is an affine symplectic variety with an isolated singularity at the image of \(0\), known as a simple surface singularity, a double point, a du Val singularity, or a Kleinian singularity. Up to conjugacy in \(\text{SL}_2(\mathbb C)\), such \(\Gamma\) are in one-to-one correspondence with the simply-laced, simple Lie-algebras over \(\mathbb C\). Thus the simple singularities can be denoted as \(A_k\), \(D_k\;(k\geq 4)\), \(E_6\), \(E_7\), \(E_8\), according to the associated simple Lie algebra. The article contains a proof of the fact that in dimension 2, an isolated symplectic singularity is equivalent to a simple surface singularity. Also, the more general case is considered. An automorphism of \(X=\mathbb C^2/\Gamma\) gives rise to a graph automorphism of the dual graph \(\Delta\) of \(X\). The authors ask when the action of \(\text{Aut}(\Delta)\) on the dual graph comes from an algebraic action on \(X\), and proves for which types of \(X\) this is true. The article contains the study of the regular nilpotent orbit, starting with the generic singularities of the nilpotent cone. In proving a conjecture of Grothendieck, Brieskorn and Slodowy [loc. cit.] described the generic singularities of the nilpotent cone \(\mathcal N\) of \(\mathfrak g\). In this particular case, \(\mathcal O\) is the regular nilpotent orbit, and so \(\overline{\mathcal O}\) equals \(\mathcal N\) with only one degeneration at the subregular nilpotent orbit \(\mathcal O^\prime\). Slodowy concludes that when \(e\in\mathcal O^\prime\), the slice \(\mathcal S_{\mathcal O,e}\) is algebraically isomorphic to a simple surface singularity. Also, when the Dynkin diagram of \(\mathfrak g\) is simply-laced, the Lie algebra associated to the simple surface singularity is \(\mathfrak g\). When it is not simply-laced, the singularity \(\mathcal S_{\mathcal O,e}\) belongs to a list given explicitly in the article. The authors explain an intrinsic realization of the symmetry of \(\mathcal S_{\mathcal O,e}\) when \(\mathfrak g\) is not simply-laced. Kraft and Procesi [loc. cit]. classified the generic singularities of nilpotent orbit closures for all the classical groups, up to smooth equivalence. The \textit{minimal singularities} are those corresponding to the equivalence classes of a particular class of singularities, explicitly defined by their orbit closures, and denoted \(a_k,b_k,c_k,d_k(k\geq 4),g_2,f_4,e_6,e_7,e_8\). The \textit{generic singularities} in the classical types are classified: An irreducible component of a generic singularity is either a simple surface singularity or a minimal singularity, up to smooth equivalence. When a generic singularity is not irreducible, then it is smoothly equivalent to a union of two simple surface singularities of type \(A_{2k-1}\) meeting transversally in the singular point, denoted \(2A_{2k-1}\). The main goal of the article is to describe the classification of generic singularities in the exceptional Lie algebras. The symmetry of minimal singularities and the generalization to the general case is essential. The exceptional Lie algebras introduce additional singularities, and and the non-normal cases is treated thoroughly. The main theorem classify generic singularities of nilpotent orbit closures in a simple Lie algebra of exceptional type, and graphs at the end of the article list precise results. The object of the article is very interesting, and its goal is justified. In addition to treating the subject in a very interesting way, the authors introduce new theories, and explain established ones in a good way, making this a good overview of the subject.