Framed sheaves on projective stacks (Q488954)

The paper under review constructs moduli spaces of framed sheaves over projective stacks. These objects are related to Donaldson's framed \(\mathrm{SU}(r)\)-instantons on the four-dimensional real space and play a key role in physics. Different extensions of the work of Donaldson to other noncompact \(4\)-dimensional Riemannian manifolds have been widely studied, where a suitable compactification is needed as well as a convenient way to relate the data in the compatifying locus. To overcome some of the difficulties it turns out to be useful to change the base space to a projective Deligne-Mumford stack. Let \(\mathcal{X}\) be a normal projective irreducible stack of dimension \(d\) over an algebraically closed field of characteristic zero, endowed with a coarse moduli scheme \(\pi: \mathcal{X}\rightarrow X\) and a polarization \((\mathcal{G},\mathcal{O}_{X}(1))\) on \(\mathcal{X}\). The notion of polarization for stacks is a generalization of the one of very ample line bundles for projective varieties, and allows us to relate the sheaves on the stack to those on the moduli scheme. A framed sheaf is defined as a pair \((\mathcal{E},\phi_{\mathcal{E}})\) where \(\mathcal{E}\) is a coherent sheaf on \(\mathcal{X}\) and \(\phi_{\mathcal{E}}\) is a morphism from \(\mathcal{E}\) to a fixed coherent sheaf \(\mathcal{F}\), which is called the framing. Fixing a polynomial \(P\) of degree \(d\) and a polynomial \(\delta\) of degree \(d-1\), Theorem 1.1 shows the existence of an algebraic stack \(\mathfrak{M}^{\mathrm{s(s)}}\) of finite type over \(k\) which admits a good moduli space \(\pi: \mathfrak{M}^{\mathrm{s(s)}}\rightarrow M^{\mathrm{s(s)}}\) which is a projective scheme, and such that the associated contravariant functor \([\mathfrak{M}^{\mathrm{s(s)}}]\) (which takes any \(k\)-scheme \(S\) to the set of isomorphism classes of objects in \(\mathfrak{M}^{\mathrm{s(s)}}(S)\)) is isomorphic to the moduli functor of \(\delta\)-(semi)stable framed sheaves on \(\mathcal{X}\) with Hilbert polynomial \(P\). This construction is made by using Geometric Invariant Theory but adapting the arguments to the stack setting, essentially proving the boundedness of the class of \(\delta\)-semistable objects and constructing a Quot-scheme parametrizing them. Theorem 1.2 calculates the tangent space and the obstruction to the smoothness of the moduli space in terms of hyper-Ext groups. When \(\mathcal{X}\) is a projective irreducible orbifold of dimension \(2\), the authors consider framings along \(1\)-dimensional smooth integral closed substacks \(\mathcal{D}\). They are proven to be \(\delta\)-stable for a suitable choice of polynomial \(\delta\), hence there exists a quasi-projective scheme which is a fine moduli space for them (Theorem 1.3). Last part of the paper applies the previous results to the theory of \(2\)-dimensional orbifolds which are toric root stacks (Theorem 1.4). The paper finishes with some appendixes. The first three of them show technical results about coherent sheaves on projective stacks: a semicontinuity theorem, a Serre duality theorem and a characterization of the dual of a coherent sheaf. Appendix \(D\) applies the techniques and results of the paper to study in detail the example of framed sheaves on stacky Hirzebruch surfaces.