Moduli spaces of stable pairs (Q1692383)

The theory of stable pairs in the derived category of coherent sheaves is initialed by \textit{R. Pandharipande} and \textit{R. P. Thomas} [Invent. Math. 178, No. 2, 407--447 (2009; Zbl 1204.14026)] to study curve counting invariants on Calabi-Yau 3-folds, the so-called the PT invariants, which are closely related to the GW invariants and the DT invariants [\textit{R. Pandharipande} and \textit{R. P. Thomas}, Lond. Math. Soc. Lect. Note Ser. 411, 282--333 (2014; Zbl 1310.14031)]. In the article under review, the author gives a more general definition of stable pairs, constructs a moduli space of stable pairs over a smooth projective variety, provides its deformation-obstruction theory and shows in the surface case the existence of virtual fundamental class. Let \((X, \mathcal{O}_X(1))\) be a polarized smooth projective variety of dimension \(n\) over an algebraically closed field \(k\). Fix a coherent sheaf \(E_0\). Let \(P\) be a polynomial of degree \(d\) and \(\delta\) be \(0\) or a polynomial with positive leading coefficient. A pair \((E,\alpha)\) (of type \(P\)) on \(X\) consists a coherent sheaf \(E\) (with Hilbert polynomial \(P\)) and a morphism \(\alpha: E_0 \to E\). The Hilbert polynomial of a pair \((E,\alpha)\) is defined by adding the Hilbert polynomial of \(E\) an extra term \(\epsilon(\alpha)\delta\), where \(\epsilon(\alpha)\) is either \(1\) if \(\alpha\neq 0\) or \(0\) otherwise. The \(\delta\)-(semi)stability of a pair \((E,\alpha)\) is defined by the purity assumption on \(E\) together with the slope-inequality comparing with its sub-pairs in the usual way, where the slope function is the reduced Hilbert polynomial of pair. A geometric understanding of such stability in the case \(\deg \delta\geq \deg P\) is that, \((E,\alpha)\) is \(\delta\)-stable if and only if \(E\) is pure and \(\dim \mathrm{coker}\alpha < \deg P\). In particular, if \(E_0\) is the structure sheaf of a 3-folds, and \(E\) is an \(1\)-dimensional sheaf, this recovers the original \textit{PT-stable pair}, i.e. (1) the sheaf \(E\) is pure and (2) the cokernel of \(\alpha\) is \(0\)-dimensional. The construction of moduli space \(S_{E_0}(P,\delta)\) parameterizing \(\delta\)-semistable pairs \((E,\alpha)\) is given by a GIT quotient, which is very similar to the case of moduli space of sheaves. The difference is that the author works in a subset of a product \(\mathbb{P}\times Q\), where \(Q\) is the Grothendieck's Quot scheme packaging information of \(E\), and \(\mathbb{P}\) is a projective space packaging information of \(\alpha\). To make the construction work, the author proves that the family of \(\delta\)-semistable pairs is bounded in Section 3. The definition of \(\delta\)-(semi)stability could thus be understood from the GIT-(semi)stability and the Hilbert-Mumford numerical criterion, just as in the sheaf case, see Lemma 4.1. The construction of the moduli space for the case \(\deg \delta < \deg P\) has been previously worked out by [\textit{M. Wandel}, Manuscr. Math. 147, No. 3--4, 477--500 (2015; Zbl 1326.14019)]. The author provides the deformation-obstruction theory for stable pairs. Let \(0\to K \to B \mathop{\to}\limits^{\sigma} A \to 0\) be a small extension (i.e. \(\mathfrak{m}_B K=0\)) in the category of local Artinian \(k\)-algebras. Denote \(I^{\bullet} =\{E_{0}\mathop{\to}\limits^{\alpha} E\}\) the cochain complex concentrating at degree \(0\) and \(1\). Suppose \(\alpha_A: E_0\otimes_k A \to E_A\) is a morphism over \(X_A=X\times_{\mathrm{Spec} k}\mathrm{Spec} A\) extending \(\alpha\), where \(E_A\) is a coherent sheaf flat over \(A\). Then there is a class \[ \mathrm{ob}(\alpha_A, \sigma)\in \mathrm{Ext}^1(I^{\bullet}, E\otimes K), \] such that there exists a flat extension of \(\alpha_A\) over \(X_B\) if and only if \(\mathrm{ob}(\alpha_A, \sigma)=0\). If extensions exist, the space of extensions is a torsor under \(\mathrm{Hom}(I^{\bullet}, E\otimes K)\). The construction of \(\mathrm{ob}(\alpha_A, \sigma)\) uses two locally free resolutions of \(E_0\) and \(E_A\) respectively. When \(X\) is a surface, and \(E_0\) is torsion free, the author shows that if \(\deg \delta \geq \deg P=1\) then the moduli space \(S_{E_0}(P,\delta)\) admits a virtual fundamental class, i.e. there is a two-term complex of locally free sheaves resolving the deformation and obstruction sheaves. The author also studies the special case when \(X\) is a \(K3\) surface, and generalizes a result of \textit{T. Kawai} and \textit{K. Yoshioka} [Adv. Theor. Math. Phys. 4, No. 2, 397--485 (2000; Zbl 1013.81043)] on the generating series of topological Euler characteristics of the moduli spaces of particular stable pairs. It is worth mentioning that the original notion of PT-stable pairs is captured purely homologically and generalized to higher rank case by \textit{J. Lo} [Math. Res. Lett. 19, No. 4, 873--885 (2012; Zbl 1267.14026)], based on Bayer's polynomial Bridgeland stability [\textit{A. Bayer}, Geom. Topol. 13, No. 4, 2389--2425 (2009; Zbl 1171.14011)]. And Lo's notion of stable pairs is used by Y. Toda [\textit{Y. Toda}, ``Hall algebras in the derived category and higher rank DT invariants'', \url{arXiv:1601.07519}] to study PT/DT correspondence on Calabi-Yau 3-folds, and by \textit{G. Oberdieck} and \textit{J. Shen} [``Curve counting on elliptic Calabi-Yau threefolds via derived categories'', \url{arXiv:1608.07073}], with the further notion \(\pi\)-stable pairs, to study the modularity of PT-invariants on certain elliptic Calabi-Yau 3-folds.