Differential operators on quantized flag manifolds at roots of unity (Q436149): Difference between revisions

Via the Beilinson-Bernstein localization theorem, sheaves of differential operators on the flag manifold play a pivotal role in the representation theory of semi-simple Lie algebras. Using the framework of noncommutative algebraic geometry, \textit{V. A. Lunts} and \textit{A. L. Rosenberg} [Sel. Math., New Ser. 5, No. 1, 123--159 (1999; Zbl 0928.16020)] defined an abelian category that can reasonably be called the category of quasi-coherent \(D\)-modules on the quantum flag manifold. It was shown by the author [Math. Z. 250, No. 2, 299--361 (2005; Zbl 1082.17010)] and also by \textit{E. Backelin} and \textit{K. Kremnizer} [Adv. Math. 203, No. 2, 408--429 (2006; Zbl 1165.17304)] that an analogue of the Beilinson-Bernstein theorem holds for quantum groups, at least when the deformation parameter is transcendental. The paper under review considers instead the case where the deformation parameter is a root of unity. Let \(G\) be a simple, simply connected algebraic group over \(\mathbb{C}\) and \(\mathcal{B}\) the associated flag manifold. Let \(\zeta\) be a primitive \(\ell\)th root of unity in \(\mathbb{C}\). Though there is no genuine space that can be called the quantum flag manifold \(\mathcal{B}_{\zeta}\), the abelian category of quasi-coherent \(\mathcal{D}_{\mathcal{B}_{\zeta}}\)-modules on this space is defined to be the quotient of the category of graded modules of a certain graded algebra \(A_{\zeta}\) by the Serre subcategory of all torsion modules. Lusztig's Frobenius morphism \(\mathrm{Fr} : U_{\zeta}(\mathfrak{g}) \rightarrow U(\mathfrak{g})\) on the quantum group allows one to construct an explicit sheaf \(\mathrm{Fr}_* \mathcal{D}_{\mathcal{B}_{\zeta}}\) of \(\mathcal{O}_{\mathcal{B}}\)-algebras on the flag manifold such that the category of quasi-coherent \(\mathrm{Fr}_* \mathcal{D}_{\mathcal{B}_{\zeta}}\)-modules is precisely the pushforward under \(\mathrm{Fr}\) of the category of quasi-coherent \(\mathcal{D}_{\mathcal{B}_{\zeta}}\)-modules. The first main result of this paper explicitly describes the spectrum of the center of the sheaf \(\mathrm{Fr}_* \mathcal{D}_{\mathcal{B}_{\zeta}}\). It is shown to be a certain smooth, symplectic variety \(\mathcal{V}\). One can consider \(\mathrm{Fr}_* \mathcal{D}_{\mathcal{B}_{\zeta}}\) as a sheaf of algebras on \(\mathcal{V}\). The second main result says that \(\mathrm{Fr}_* \mathcal{D}_{\mathcal{B}_{\zeta}}\) is in fact an Azumaya algebra on \(\mathcal{V}\) of rank \(\ell^N\), where \(N\) denotes the number of positive roots in the root system associated to \(G\). There is a projective morphism \(\delta\) from the space \(\mathcal{V}\) to a certain singular affine variety, denoted \(K \times_{H/W} H\) in the article. The final main result of the article says that the Azumaya algebra \(\mathrm{Fr}_* \mathcal{D}_{\mathcal{B}_{\zeta}}\) is split when restricted to each fiber of the morphism \(\delta\). In order to show this, the author constructs an explicit splitting bundle. He notes that a similar argument can be use to show that \(\mathrm{Fr}_* \mathcal{D}_{\mathcal{B}_{\zeta}}\) in fact splits on the formal neighborhood of any fiber of \(\delta\). For Part II, see [the author, Nagoya Math. J. 214, 1--52 (2014; Zbl 1311.14047)].