A complex ball uniformization of the moduli space of cubic surfaces via periods of<i>K</i>3 surfaces (Q5711482)

The authors show that the moduli space of nodal cubic surfaces is isomorphic to a quotient of a 4-dimensional complex ball by an arithmetic subgroup of the unitary group given as corollary 9.3 of the main theorem 9.2 of this paper. A similar uniformization is given for different covers of the moduli space corresponding to geometric markings of the Picard Group or a choice of a line on the surface. Their construction and most of their results in this direction are related to the work of \textit{D. Allcock} et al. [J. Algebr. Geom. 11, 659--724 (2002; Zbl 1080.14532)] and \textit{K. Matsumoto} et al. [ J. Algebr. Geom. 3, 741--775 (2003; Zbl 1048.14027)]. Their paper is organized as follows. In section 2 they study stable cubic surfaces, referred here as nodal cubic surfaces. In section 2.1 they define markings of nodal cubic surfaces and introduce the moduli space of marked nodal cubic surfaces \({M^m}_{\text{ncub}}\), which admits a natural compactification and discuss different constructions of this compactification in sections 2.8--2.10. For a nodal cubic surface and a line on it they define in section 3 a pair of binary forms, of degree 2 and 5 modulo the action of SL(2). Using this they prove in corollary 3.8 to proposition 3.6 that the moduli space of cubic surfaces together with a choice of a line on it is a rational variety. In section 4, the authors define a \(K3\) surface \(X_{S,l}\), which by theorem 4.9 depends only on \(S\), henceforth denoted as \(X_S\). They relate the \(K3\) surface \(X_S\) to the associated cubic fourfold with a plane to the cubic threefold \(V\) corresponding to \(S\) by Allcock et al. [loc.cit ] in section 4.15 and to a curve called by the authors the ``Matsumoto-Terasoma'' curve \(C\) in remark 4.16. In section 5 the authors show that the Picard lattice of a generic \(X_S\) is isomorphic to the lattice \(M= U \oplus{ A_2}^{\oplus 5} \). The lattice of transcendental cycles is isomorphic to the lattice \( A_2(-1) {\oplus}{ A_2}^{\oplus 4} \) as shown in proposition 5.3. In section 6 they study the moduli space of \(M\)-polarized \(K3\) surfaces \((X, \phi:M \rightarrow \text{Pic}(X))\). They show that by explicitly describing an isometry \(\rho\) on the orthogonal complement of \(\phi(M)\) in \( H^2(X_S,Z)\) that the image of the period map of the surfaces \(X_S\) lies in the fixed locus of a certain automorphism of order 3 on the period space of \(M\)-polarized \(K3\) surfaces. They construct in this way in section 6.14 the moduli space \( {K3^m}_{M, \rho} \) of \((M, \rho)\)-polarized \(K3\) surfaces as a quotient of \(B\) (see section 6.7 for the explicite definition). In section 7, they establish a natural \(W(E_6)\)-equivariant isomorphism from the moduli space of marked nonsingular cubic surfaces \({M^m}_{\text{cub}}\) onto an open subset \({K3^m}_{M, \rho}- {\Delta}^m \) in the theorem of subsection 7.2. In section 8 they study in detail the geometry of the discriminant locus \(\Delta^m \). In section 9 they show that the \(W(E_6)\) -equivariant isomorphism from \({M^m}_{\text{cub}}\) onto \({K3^m}_{M, \rho}- {\Delta^m} \) can be extended to an equivariant isomorphism from \({M^m}_{\text{ncub}}\) to \({K3^m}_{M, \rho}\) in theorem 9.2 showing further that the quotient \( {K3^m}_{M, \rho}/ W(D_5)\) and \({M^m}_{\text{ncub}}/W(D_5)\) are isomorphic in equation 9.1. Finally in section 10, they compare the Hodge structure on the \(K3\) surface \(X_S\) with the principally polarized Hodge structure on \(H^1(P, Z) \) where \(P\) is the intermediate Jacobian of a cubic threefold associated to the cubic surface \(S\).