On the independence of Heegner points in the function field case (Q710490)

Let \(E\) be an elliptic curve over the global function field \(k\) with split multiplicative reduction at a fixed place \(\infty\) of \(k\). (This condition may always be achieved after a possible finite extension of k, provided that \(j(E)\) is non-constant.) Due to work by a number of people (Drinfeld, Deligne, Jacquet-Langlands, ...), there exists a ``Weil uniformization'' \(X_0(\mathfrak n )\rightarrow E\) through a certain Drinfeld modular curve \(X_0(\mathfrak n)\), see [\textit{E.-U. Gekeler} and \textit{M. Reversat}, J. Reine Angew. Math. 476, 27--93 (1996; Zbl 0848.11029)]. As a consequence, we dispose of Heegner points on \(X_0(\mathfrak n)\) and on \(E\) which ought to have properties similar to those in the ``classical'' case. Let \(K_1,\ldots K_r\) be finitely many imaginary quadratic number fields, let \(E\) be an elliptic curve over \(\mathbb Q\) of conductor \(N\), provided with a Weil uniformization through \(X_0(N)\), and let \(P_i\), (\(1 \leq i \leq r\)) be Heegner points in \(E(\overline{\mathbb Q})\) attached to the \(K_i\). (This requires that the \(K_i\) satisfy a certain Heegner condition with respect to \(N\).) \textit{M. Rosen} and \textit{J. H. Silverman} showed in [J. Number Theory 127, No. 1, 10--36 (2007; Zbl 1151.11024)] that the \(P_i\) are independent in \(E(\overline{\mathbb Q})\)/torsion, provided that \(E\) has no complex multiplication and the odd parts of the class numbers of the \(K_i\) are sufficiently large (larger than a constant that depends only on \(E\) and the chosen Weil uniformization). The authors of the present article show a similar statement in the function field context (Theorem 1). Here an ``imaginary'' quadratic extension \(K\) of \(k\) is one where the chosen place \(\infty\) remains prime, and Heegner points on \(X_0(\mathfrak n)\) (thus on \(E\)) are described by Drinfeld modules with some supplementary structure (instead of elliptic curves). A principal difference with the classical case is the requirement that the ``prime-to \(2p\)-part'' (where \(p = \text{char}(k)\)) of the class numbers of the rings of integers \(O_i\) of the involved imaginary quadratic function fields \(K_i\) must be sufficiently large. The proof uses ideas similar to those in [Rosen-Silverman, loc. cit.]. In the paper's final section, restricting to the case where \(k = \mathbb F_q(t)\), the authors discuss the frequency by which imaginary quadratic extensions \(K\) of \(k\) fulfill the requirements of Theorem 1, and find that there are at least infinitely many. Despite the sophisticated and technically demanding character of the paper, it can be read with pleasure, since it is well-thought and transparently organized.