Nef line bundles on algebraic surfaces (Q1902059): Difference between revisions

This paper contains three results. The first one describes projective semipolarized surfaces, that is, projective surfaces endowed with a nef line bundle, up to birational equivalence. It is shown that any semipolarized surface \((X, L)\) is birationally equivalent to a smooth semipolarized surface \((Y, M)\) such that either \(M+ K_Y\) is nef, or \(Y= \mathbb{P}^2\) and \(M={\mathcal O}, {\mathcal O} (1), {\mathcal O} (2)\), or \((Y, M)\) is a scroll over a smooth curve, or \(Y\) is a \(\mathbb{P}^1\)-bundle over a smooth curve and \(L\) is the pullback of a line bundle of non-negative degree. The second and third theorems are devoted to a more complete, precise and very interesting description of the pair \((Y, M)\) associated as above to a semipolarized normal surface \((X, L)\) whose sectional genus is 0 or 1. Here, the sectional genus is the integer \(g\) given by the equation \(2g- 2= (\omega_X+ L)L\) where \(\omega_X\) is the canonical sheaf of \(X\). The proofs are clear and hold on \textit{S. Mori}'s characterization of the extremal rational curves in a smooth projective surface [Ann. Math., II. Ser. 116, 133-176 (1982; Zbl 0557.14021)]. The base field is of arbitrary characteristic \(p\geq 0\). In the case of characteristic zero similar results for surfaces and threefolds are due to \textit{T. Fujita} [Nagoya Math. J. 115, 105-123 (1989; Zbl 0699.14002)]. As the author says, the paper is inspired by earlier results in the case of complex smooth polarized surfaces due to \textit{A. Lanteri} and \textit{M. Palleschi} [J. Reine Angew. Math. 352, 15-23 (1984; Zbl 0535.14003)].