On quartic double fivefolds and the matrix factorizations of exceptional quaternionic representations (Q2027736)

The author considers a particular family of varieties: quartic double fivefolds in \(\mathbb{P}(1,1,1,1,1,1,2)\). The reason for considering such a family is that this is one of the possible examples of complete intersection of Calabi-Yau type in a weighted projective space by \textit{A. Iliev} and \textit{L. Manivel} [J. Pure Appl. Algebra 219, No. 6, 2225--2244 (2015; Zbl 1314.14085)]. The reason for studying Fano manifolds of Calabi-Yau type is their natural appearance inside the mirror symmetry world. By a series of papers, such manifolds have similar geometric and cohomological properties to quintic threefolds. In this paper, the author mainly focuses on the generic quartic double fivefolds, for which he showed the following two theorems: Theorem [\textit{R. Abuaf}, Ann. Inst. Fourier 70, No. 4, 1403--1430 (2020; Zbl 07351862), Theorem 1.5]. The generic quartic double fivefold can be represented, in a finite number of ways, as a double cover of \(\mathbb{P}^5\) ramified along a linear section of the \(\mathrm{Spin}_{12}\)-invariant quartic \(\mathcal{Q}_{\mathrm{Spin}_{12}}\subset\mathbb{P}^{31}\). Theorem [\textit{R. Abuaf}, Ann. Inst. Fourier 70, No. 4, 1403--1430 (2020; Zbl 07351862), Theorem 1.6]. The 3 dimensional Calabi-Yau category associated to the derived category of the generic quartic double fivefold contains a rank 6 spherical vector bundle. Theorem 1.5 is an analogue of a result in [\textit{A. Iliev} and \textit{L. Manivel}, Proc. Lond. Math. Soc. (3) 108, No. 2, 517--540 (2014; Zbl 1304.14053)], where Iliev and Manivel showed that a generic cubic hypersurface in \(\mathbb{P}^8\) can be represented in a finite number of ways as linear sections of the \(E_6\)-invariant cubic hypersurface in \(\mathbb{P}^{26}\). This can be thought as the analogue of the fact that the quintic threefold has a finite number of determinatal representations by Beauville in [\textit{A. Beauville}, Mich. Math. J. 48, 39--64 (2000; Zbl 1076.14534)]. On the other hand, theorem 1.6 can be thought as the analogue of the similar property for generic cubic sevenfolds in [\textit{A. Iliev} and \textit{L. Manivel}, Proc. Lond. Math. Soc. (3) 108, No. 2, 517--540 (2014; Zbl 1304.14053)] and quintic threefolds in [\textit{P. Seidel} and \textit{R. Thomas}, Duke Math. J. 108, No. 1, 37--108 (2001; Zbl 1092.14025)]. The proof of theorem 1.5 follows the existing scheme in [\textit{A. Beauville}, Mich. Math. J. 48, 39--64 (2000; Zbl 1076.14534); \textit{A. Iliev} and \textit{L. Manivel}, Proc. Lond. Math. Soc. (3) 108, No. 2, 517--540 (2014; Zbl 1304.14053)]. The author studied the pullback of partial derivatives of equations of a \(\mathrm{Spin}_{12}\)-invariant quartic to \(\mathbb{C}[z_1,\dots,z_6]\) and used Macaulay2 to compute the dimension of the space generated by them. To show theorem 1.6, he used the Vinberg type II decomposition of the Lie algebras \(\mathfrak{e}_6\) and \(\mathfrak{e}_7\) and then used some particular matrix factorizations of \(P_{\mathrm{SL}_6}(y)+x\) and some particular quartic double fivefolds. As an application of the existence of the spherical vector bundle in theorem 1.6, he showed that the homological unit, which is an anaologue of the algebra \(H^{\boldsymbol{\cdot}}(O_X)\) for a triangulated category, inside the Calabi--Yau category of dimension three inside derived category of a generic quartic double fivefold is \(\mathbb{C}\oplus\mathbb{C}[3]\).