Rational Cherednik algebras and Hilbert schemes. II: Representations and sheaves (Q2492110): Difference between revisions

The authors apply a \(\mathbb{Z}\)-algebra construction and good filtrations to study the representation theory of Cherednik algebras of type \(A_{n-1}\), denoted \(H_c\), and of its spherical subalgebras \(U_c=eH_ce\). This representation theory is related to \(\text{Hilb}(n)\) and to the resolution of singularities \(\tau:\text{Hilb}(n)\rightarrow\mathfrak h\oplus\mathfrak h^\ast/W.\) It is assumed throughout the article that \(c\neq\frac 12+\mathbb{Z}\), and for simplicity that \(c\notin\mathbb{Q}_{\leq 0}.\) The aim of the article is to relate \(H_c\)- and \(U_c\)-modules to sheaves over the Hilbert scheme of points in the plane. It is shown that finite-dimensional \(H_c\)- and \(U_c\)-modules form moduli for the sections of some sheaves on the Hilbert scheme of points on the plane. The Borel-Weil theorem describes the finite-dimensional simple modules of a complex semisimple Lie algebra \(\mathfrak g\) as sections of line bundles on the flag variety \(\mathfrak B=G/B,\) and some properties of \(U_c\) are similar to those of \(U(\mathfrak g)/P\) where \(P\) is a minimal primitive ideal in the enveloping algebra \(U(\mathfrak g),\) and it is the Beilinson-Bernstein equivalence of categories that provides the analogue for this approach. Both \(H_c\) and \(U_c\) can be filtered by degree of differential operators with associated graded rings \(\text{gr}H_c\cong\mathbb{C}[\mathfrak h\oplus\mathfrak h^\ast]\ast W\) and \(\text{gr}U_c\cong\mathbb{C}[\mathfrak h\oplus\mathfrak h^\ast]^W\) where \(\mathfrak h\subset \mathbb{C}^n\) is the reflection representation of the \(n\)th symmetric group \(W\). The role of the Springer resolution is played by the crepant resolution \(\tau:\text{Hilb}(n)\rightarrow\mathfrak h\oplus\mathfrak h^\ast/W\), where \(\text{Hilb}(n)\) is the subvariety of \(\text{Hilb}^n\mathbb{C}^2\) consisting of those ideals of \(\mathbb{C}[x,y]\) of colength \(n\) whose supports have centre of mass at the origin. The analogy with the Beilinson-Bernstein equivalence is then given by one of the main theorems of the article: There exists a filtered \(\mathbb{Z}\)-algebra \(B\) such that (1) \(U_c\)-mod, the category of finitely generated \(U_c\)-modules, is equivalent to \(B\)-{qgr}, the quotient category of finitely generated graded \(B\)-modules modulo those of finite length; (2) gr\(B\), the associated graded ring of \(B\), is isomorphic to the \(\mathbb{Z}\)-algebra associated with the homogeneous coordinate ring \(\bigoplus_{k\geq 0}H^0(\text{Hilb}(n),\mathcal{L}^k)\) for a certain ample line bundle \(\mathcal{L}\) on \(\text{Hilb}(n)\). This theorem shows that \(U_c\) can be regarded as a noncommutative deformation of a homogeneous coordinate ring of the Hilbert scheme \(\text{Hilb}(n)\), and it provides a recipe for passing from a left \(U_c\)-module \(M\) with a good filtration \(\Lambda\) to a coherent sheaf \(\Phi(M)=\Phi_{\Lambda}(M)\) on \(\text{Hilb}(n)\). This depends on the induced tensor product filtration. This also works for filtered \(H_c\)-modules \((N,\Lambda)\), resulting in a sheaf \(\widehat{\Phi}_{\Lambda}(N)\) also on \(\text{Hilb}(n)\). Under the hypotheses on \(c\), \(H_c\) and \(U_c\) are Morita equivalent. The basic technique of this article is to exploit the functors \(\Phi\) and \(\widehat{\Phi}\) to understand the representation theory of \(U_c\) and \(H_c\). For \(\mathcal{O}_c\), the category of finitely generated \(H_c\)-modules on which \(\mathbb{C}[\mathfrak h^\ast]\) acts locally nilpotently, the analogues of Verma modules over a simple complex Lie algebra are the standard modules \(\Delta_c(\mu)\) where \(\mu\) is an irreducible representation of the \(n\)th symmetric group \(W\). For any good filtration \(\Lambda\) on \(e\Delta_c(\mu)\), the associated variety is independent of \(\mu\). On the other hand, the associated variety of \(\widetilde{e\Delta_c(\mu)}\) is a subvariety of \(\text{Hilb}(n)\) which does depend on \(\mu\). This result is stated by introducing the characteristic cycle, \(\text{\textbf{Ch}}(\Delta_c(\mu))\) which counts the irreducible components of the characteristic variety of \(\hat{\Phi}(\Delta_c(\mu))\) in \(\text{Hilb}(n)\) with multiplicities. Another main result of the article is the following: Let \(\Delta_c(\mu)\) be the standard \(H_c\)-module corresponding to \(\mu\in \text{Irrep}(W)\). Then \[ \text{\textbf{Ch}}(\Delta_c(\mu))=\sum_{\lambda}K_{\mu\lambda} [Z_\lambda] \] where \(K_{\mu\lambda}\) are Kostka numbers and the \(Z_\lambda\)'s are the irreducible components of \(Z=\tau^{-1}(\mathfrak h/W).\) This introductory remarks gives just some of the important results of this article. It contains a lot more, not possible to give any meaning in such a short review as this. It should be said however that the definitions and theory needed to understand this theory is nicely and explicitly treated. Graded and filtered modules for \(\mathbb Z\)-algebras, rational Cherednik algebras and the Hilbert schemes mentioned above are both sections that are understandable without too much special knowledge. The two final sections treating the representations of \(H_c\), coherent sheaves on \(\text{Hilb}(n)\), finite-dimensional \(U_c\)-modules and their characteristic varieties is somewhat harder to completely understand.