Families of elliptic curves in \(\mathbb P^3\) and Bridgeland stability (Q2414689)

\textit{T. Bridgeland} stability [Ann. Math. (2) 166, No. 2, 317--345 (2007; Zbl 1137.18008)] has shed light on the structure of moduli spaces in recent years. \textit{B. Xia} used it [Trans. Am. Math. Soc. 370, No. 8, 5535--5559 (2018; Zbl 1390.14059)] to analyze the Hilbert scheme of twisted cubic curves in $\mathbb P^3$, recovering results of \textit{R. Piene} and \textit{M. Schlessinger} [Am. J. Math. 107, 761--774 (1985; Zbl 0589.14009)]. Here the authors carry out a similar analysis for the Hilbert scheme of elliptic quartic curves $C$: letting $v = \text{ch}({\mathcal I}_C)$ be the Chern character, they use the tilt stability theory for threefolds of \textit{A. Bayer} et al. [J. Algebr. Geom. 23, No. 1, 117--163 (2014; Zbl 1306.14005)] to construct a path $\gamma: [0,1] \to \text{Stab} (\mathbb P^3)$ to the space of Bridgeland stability conditions along which the moduli spaces of stable complexes with Chern character $v$ vary from the empty space to the Hilbert scheme of twisted cubics. \par Their path passes through six chambers while crossing five walls. The first chamber yields no semistable objects, but after the first wall crossing one encounters the Grassmann variety $M_1=\mathbb G (1,9)$ parametrizing pencils of quadrics. Crossing the second wall yields the blow up $M_2 \to M_1$ along the locus of non-ideal sheaves with the exceptional divisor generically corresponding to ideals of unions of a plane cubic and a line meeting at a point. After the third wall crossing they find the moduli space $M_3$ having two irreducible components $M_3^1$ (the blow up of $M_2$ along the incidence variety of length two subschemes of $\mathbb P^3$) and $M_3^2$ (a $\mathbb P^{14}$-bundle over $\text{Hilb}^2 (\mathbb P^3) \times (\mathbb P^3)^*$ parametrizing the closure of unions of plane quartics with two points). The fourth moduli space $M_4$ has two irreducible components $M_4^1=M_3^2$ and $M_4^2$, which is birational to $M_3^2$ but the newly introduced points correspond to plane quartics and two points where exactly one of the points is in the plane of the quartic. Crossing the fifth wall yields $M_5 = \text{Hilb}^{4t} (\mathbb P^3)$ which has two components $M_5^1$, the 16 dimensional closure of smooth elliptic quartics and $M_5^2$, the 23 dimensional closure of plane quartics union two points, which meet transversely dimension 15. The deformation theory where components meet requires many Ext group calculations, some of which are carried out by the software Macaulay2. \par It follows from the descriptions that the Hilbert component $M_5^1$ of elliptic quartics is a double blow up of the Grassmann variety $\mathbb G (1,9)$ along smooth centers, recovering a result of \textit{I. Vainsencher} and \textit{D. Avritzer} [Lond. Math. Soc. Lect. Note Ser. 179, 47--58 (1992; Zbl 0774.14024)]. From the standpoint of Mori theory, they use double blow up description to compute the cone of effective divisors for $M_5^1 = \text{Hilb}^{4t} (\mathbb P^3)$, showing that it is generated by the divisorial loci corresponding to the family $E_1$ of plane cubics union a line, the family $E_2$ corresponding to plane quartics with two nodes and one embedded point at each node, and the family $\Delta$ corresponding to nodal elliptic quartics.