Birational geometry of irreducible holomorphic symplectic tenfolds of O'Grady type (Q2118191)

In this paper the authors study the birational geometry of irreducible holomorphic symplectic manifolds (ihs) deformation equivalent to the O'Grady ten-dimensional example. One of the main tool to study the geometry of such manifolds is to consider the second integral cohomology group \(H^{2}(X,\mathbb{Z})\) that is endowed with a lattice structure due to a non-degenerate symmetric bilinear form define on it, called the Beuaville-Bogomolov-Fujiki form. Lagrangian fibrations are strictly related to irreducible holomorphic symplectic manifolds, more precisely there are maps from an irreducible symplectic manifolds to a non-trivial normal variety such that all fibers are lagrangian subvarieties. In Section 2 the authors provide a converse of this fact up to birationality. More precisely in Theorem 2.2 they say that if \(X\) is an ihs manifold of \(OG_{10}\) type and if there exists a primitive non-trivial isotropic line bundle such that its class belongs to the boundary of the birational Kähler cone, then there exist an ihs manifold \(Y\) and a bimeromorphic map \(Y \dashrightarrow X\) and a lagrangian fibration \(Y \to \mathbb{P}^{5}\), such that the line bundle on \(X\) is the pullback of the hyperplane section of \(\mathbb{P}^{5}\) pulled back along the bimeromorphic map. As a consequence of this result they get in Corollary 2.4 a proof of the weak splitting property conjectured by Beauville, i.e. for a manifold of \(OG_{10}\) type, if \(DCH(X) \subset CH_{\mathbb{Q}}(X)\) is the subalgebra generated by divisor classes, then the restriction of the cycle class map \(cl_{|DCH(X)}\colon DCH(X)\ \to H^{*}(X,\mathbb{Q})\) is injective. Moreover is Sections 3, 4 and 5 the authors provide a lattice-theoretic descriptions of the ample and movable cone (Theorem 3.2 and Theorem 5.5). They start with a descriptions in terms of square and divisibilities in the Beauville-Bogomolov-Fujiki lattice of prime exceptional divisors (Proposition 3.1) and of wall divisors (Proposition 5.4), that are the divisors such that their orthogonal complements cut the birational Kähler cone and the Kähler cone respectively. Recall that the birational Kähler cone of \(X\) is the union of the Kähler cones of all smooth birational models of \(X\) which are still ihs manifolds. The movable cone is contained in a prime exceptional chamber, that is, in the convention of the paper, the chamber of all divisors whose pairing (using the Beauville-Bogomolov-Fujiki form) is non negative with prime exceptional divisors. To get the characterization above the authors provides at first a geometric construction of wall divisors and then they study the birational transformations associated and they prove that these examples are the all possibilities exploiting a result of Ch.Lehn and Pacienza and a result of Meachan and Zhang. In Section 6, as an application, the authors study the symplectic compactifications of the twisted intermediate jacobians fibrations associated to a smooth cubic 4-fold as constructed by Saccà. They provide sufficient conditions to have uniqueness of such compactifications, in particular the non-trivial requirement is that the fibres of the natural fibration \(X \to \mathbb{P}^{5}\) have to be irreducible.