Positive line bundles on arithmetic surfaces (Q2366454): Difference between revisions

The paper, which is part of the author's doctoral thesis, contains two main results: a Nakai-Moishezon theorem on an arithmetic surface, result which was conjectured by Szpiro, and an analogue of a conjecture of Bogomolov about the discreteness of algebraic points on an algebraic curve. More precisely, let \(X\) be an arithmetic surface and \(\overline L=(L,\| \|)\) a Hermitian line bundle on \(X\) \((L\) invertible sheaf on \(X\), \(\| \|\) a continuous Hermitian metric on \(L_ \mathbb C\), invariant under the complex conjugation of \(C_ \mathbb C)\). A nonzero section \(\ell\) of \(L\) on \(X\) is ``strictly effective'' if \(\|\ell\|(x)<1\) for all \(x\in X_ \mathbb C\). One says that \(\overline L\) is ``ample'' if \(L\) is ample, the curvature form \(\omega(\overline L)\) of \(\overline L\) is semipositive, and there is a basis of \(\Gamma(L^ n)\) over \(\mathbb Z\) consisting of strictly effective sections for all sufficiently large \(n\). One says that \(\overline L\) is ``positive'' if \(\omega(\overline L)\) is semipositive, \(\overline L\cdot\overline L\) (the intersection number of Deligne) is positive, and \(\deg(\overline L| D):=\deg[\text{ div}\;\ell]-\sum_{x\in D_ \mathbb C}\log\|\ell\|(x)\) \((\ell\) a section of \(L| D)\) is positive for any integral divisor \(D\) on \(X\). The Nakai-Moishezon arithmetic theorem proved by the author says that: \(\overline L\) is ample if and only if \(\overline L\) is positive. The analogue of Bogomolov's conjecture has the following statement: Let \(C\to \mathbb G^ n_ m\) \((\mathbb G_ m=\) the multiplicative group) be an embedding of a curve defined over a number field \(K\). Assume that \(C\) is not a translate of a subgroup of \(G^ n_ m\). Then \(C(\overline K)\) is discrete under a certain semipositive distance function \(d_ \infty\) on \(\mathbb G^ n_ m(\overline{\mathbb Q})\) defined in terms of a canonical height function on the projective line \(\mathbb P^ 1(\overline{\mathbb Q})\). The proofs use results of \textit{G. Faltings} [Ann. Math. (2) 119, 387--424 (1984; Zbl 0559.14005)] and a theorem of \textit{G. Tian} [J. Differ. Geom. 32, No. 1, 99--130 (1990; Zbl 0706.53036)] on Fubini-Study metrics.