Perverse coherent sheaves and the geometry of special pieces in the unipotent variety. (Q1002242)

The paper develops a theory of ``\(S_2\)-ification'' in order to address a conjecture of Lusztig concerning `special pieces'. First we look at the ``\(S_2\)-ification'', where \(S_2\) refers to one of Serre's conditions for normality. Let \(X\) be a scheme of finite type over a Noetherian base scheme \(S\) admitting a dualizing complex, and let \(U\subset X\) be an open set whose complement has codimension at least 2. One assumes an affine group scheme \(G\) to be acting on everything. The authors extend the Deligne-Bezrukavnikov theory of perverse coherent sheaves [not to be confused with ordinary constructible perverse sheaves] by showing that a coherent intermediate extension (or intersection cohomology) functor from perverse sheaves on \(U\) to perverse sheaves on \(X\) may be defined for a much broader class of perversities than has previously been known. One also introduces a derived category version of the coherent intermediate extension functor. Under suitable hypotheses, one introduces a construction (called ``\(S_2\)-extension'') in terms of perverse coherent sheaves of algebras on \(X\) that takes a finite morphism to \(U\) and extends it in a canonical way to a finite morphism to \(X\). More specifically, one uses that the coherent intermediate extension (for the ``\(S_2\)-perversity'') of a sheaf of algebras is again a sheaf of algebras. Thus a new scheme may be constructed as the total Spec of a coherent intermediate extension of the push forward to \(U\) of a structure sheaf. In particular, this construction gives a canonical ``\(S_2\)-ification'' of appropriate \(X\). The construction also has applications to the ``Macaulayfication'' problem, and it is particularly well-behaved when \(X\) is Gorenstein. We now turn to the main goal. It is to address a conjecture of Lusztig on the geometry of special pieces (certain subvarieties of the unipotent variety of a reductive algebraic group). The conjecture asserts in part that each special piece is the quotient of some variety (previously unknown for the exceptional groups and in positive characteristic) by the action of a certain finite group. One uses \(S_2\)-extension to give a uniform construction of the desired variety. Because of its construction the variety must be the correct one. To prove the conjecture in full one must still show the variety is smooth. Partial results in this direction are given.