Numerically anticanonical divisors on Kato surfaces (Q2347259)

From MaRDI portal
scientific article
Language Label Description Also known as
English
Numerically anticanonical divisors on Kato surfaces
scientific article

    Statements

    Numerically anticanonical divisors on Kato surfaces (English)
    0 references
    0 references
    27 May 2015
    0 references
    A Kato surface is a minimal compact complex surface \(S\) with positive second Betti number \(b_2(S)>0\) containing a global spherical shell, that is, there is an open subset \(V\) biholomorphic to a neighborhood of \(S^3 \subset \mathbb{C}^2\) such that \(S\backslash V\) is connected. A Kato surface admits exactly \(b=b_2(S)\) rational curves some of which form a cycle. An intermediate Kato surface is a Kato surface satisfying that there is at least one such curve which is not contained in a cycle. In this paper, the authors treat intermediate Kato surfaces. For an intermediate Kato surface \(S\), the rational curves \(D_1, \ldots, D_b\) form a basis for \(H^2(S, \mathbb{Q})\), and the first Chern class of \(S\), denoted by \(-K \in H^2(S, \mathbb{Z})\backslash \{0\}\), is presented as a linear combination of the curves: \(-K=\sum_{i=1}^b d_iD_i\) with \(d_i \in \mathbb{Q}\). Further, \(\mathrm{index}(S)=1\) is a necessary condition for \(S\) to admit a twisted holomorphic vector field or a twisted anticanonical section; see [\textit{G. Dloussky} and \textit{K. Oeljeklaus}, Ann. Inst. Fourier 49, No. 5, 1503--1545 (1999; Zbl 0978.32021)], and it is also a necessary condition for \(S\) to admit a bi-Hermitian metric; see [\textit{V. Apostolov} et al., Proc. Lond. Math. Soc. (3) 79, No. 2, 414--428 (1999; Zbl 1035.53061)], where the index of \(S\) is the least positive integer \(d\) such that \(-dK\) is represented by an effective divisor: \(dd_i \in \mathbb{N} \cup \{0\}\) for all \(i\). The main result is that for an intermediate Kato surface \(S\), the authors write down explicitly all the rational coefficients \(d_i\), using the Dloussky sequence and the dual graph of \(S\), and inductively defined multilinear forms. In particular, they give topological obstructions for the existence of a numerical anticanonical divisor, or equivalently \(\mathrm{index}(S)=1\). Further, they present some applications. They discuss a relation with the results due to \textit{G. Dloussky} [Ann. Fac. Sci. Toulouse, Math. (6) 20, No. 1, 15--69 (2011; Zbl 1219.14022)] on the discriminant \(k\) of the singularity obtained by contracting to a point the maximal curve of \(S\): They show how \(k\) is related to \(\mathrm{index}(S)\). They show some relations among the multiplicities of the Dloussky sequence, the contracting germ of \(S\), and the dimension of the logarithmic moduli space of the pair of \(S\) and the maximal curve, using the results due to \textit{K. Oeljeklaus} and \textit{M. Toma} [Math. Ann. 341, No. 2, 323--345 (2008; Zbl 1144.32004)].
    0 references
    0 references
    non Kähler surfaces
    0 references
    Kato surfaces
    0 references
    holomorphic vector fields
    0 references
    0 references
    0 references