A Gross-Zagier formula for function fields

Publication date 31 August 1999

Drinfeld modular curves Gross-Zagier formula Heegner points

Drinfel'd modules; higher-dimensional motives, etc. (11G09) Arithmetic aspects of modular and Shimura varieties (11G18) Modular and Shimura varieties (14G35)

Review text

Let \(K = \mathbb F_q(T)\) be the field of rational functions over a finite field \(\mathbb F_q\), with ``ring of integers the polynomial ring \(\mathbb F_q[T]\). Let further \(K_{\infty} = \mathbb F_q((T^{-1}))\) be the completion of \(K\) at its infinite place and \(C_{\infty}\) the completed algebraic closure of \(K_{\infty}\). There is an obvious analogy (including some non-obvious aspects) between the data \((\mathbb F_q[T],K,K_{\infty}, C_{\infty})\) and its number-theoretic counter part \((\mathbb Z,\mathbb Q, \mathbb R, \mathbb C)\). The analogy is supplemented through the Drinfeld upper half-plane \(\Omega := C_{\infty}-K_{\infty}\) (playing the part of the classical complex upper half-plane) acted upon by \(\Gamma = {\text GL}(2,A)\). The analytic space \(\Gamma'\setminus\Omega\) (where \(\Gamma'\) is a subgroup of \(\Gamma\) of finite index) comes from an affine algebraic curve defined over a finite extension of \(K\). More specifically, let \(\Gamma_0(N)\) be the Hecke congruence subgroup of \(\Gamma\) associated with a non-constant monic polynomial \(N \in \mathbb F_q[T]\), and put \(X_0(N)\) for the smooth compactification of the curve \(\Gamma_0(N)\setminus \Omega\). It classifies isomorphism classes of rank-two Drinfeld \(\mathbb F_q[T]\)-modules with some level structure. Moreover, through some sort of ``Shimura-Taniyama-Weil conjecture, proved in [\textit{E.-U. Gekeler} and \textit{M. Reversat}, J. Reine Angew. Math. 476, 27-93 (1996; Zbl 0848.11029)], \(X_0(N)\) uniformizes all the elliptic curves \(E\) over \(K\) with conductor \((N)\cdot\infty\) and which have split multiplicative reduction at \(\infty\). Let now \(E/K\) be such a curve. It determines both a Drinfeld modular form on \(\Omega\) and an automorphic form \(f = f_E\), which may be regarded as a certain function on the edges of the graph \(\Gamma_0(N)\setminus{\mathcal T}\), where \({\mathcal T}\) is the Bruhat-Tits tree of \(\text{PGL}(2,K_{\infty})\). Let, on the other hand, \(L\) be a separable quadratic extension of \(K\) with ring of integers \(O_L\). From the modular interpretation of \(X_0(N)\), it is now straightforward how to define Heegner points on \(X_0(N)\) corresponding to Drinfeld modules with complex multiplication by \(O_L\). Through some ``Weil uniformization \(\pi:X_0(N) \rightarrow E\), such a Heegner point yields an \(L\)-valued point \(P_L\) of \(E\). Since the original work of Gross and Zagier that relates the height of Heegner points on elliptic Weil curves \(E/\mathbb Q\) with the arithmetic of special values of \(L(E,s)\), the \(L\)-function of \(E\), people have wondered whether a formula similar to the one in [\textit{B. H. Gross} and \textit{D. B. Zagier}, Invent. Math. 84, 225-320 (1986; Zbl 0608.14019)] also holds in the present function field case. An explicit formula of that type is given in the article (Theorem 4.2), along with a sketch of proof. Details are given in some previous papers of the same authors and (mainly) in the long and important article [the authors, Doc. Math. 5, 365-444 (2000; Zbl 1012.11039)]. In the present paper, the authors restrict to the case where the following assumptions hold: (a) \(\text{char}(\mathbb F_q) \not= 2\); (b) \(N \in \mathbb F_q[T]\) is square-free; (c) \(L = K(\sqrt{D})\) with \(D \in \mathbb F_q[T]\) irreducible of odd degree. As a consequence (Corollary 4.1) the validity of the Birch and Swinnerton-Dyer conjecture results for the curves \(E/K\) meeting the above assumptions for which the derivative of \(L(E,s)L(E_D,s)\) at \(s = 1\) doesn't vanish. (Note this is an infinite class of elliptic curves.) Here \(E_D\) is the twist of \(E\) by \(L = K(\sqrt{D})\). Whereas assumptions (b) and (c) are made for technical purposes only and may be removed (which of course will involve some changes in the final formula), (a) seems to be essential. This is since the argument makes heavy use of some calculations [the first author, Manuscr. Math. 88, No. 3, 387-407 (1995; Zbl 0855.11021)] involving quadratic forms over \(K\) and their theta series. Seen on a large enough scale, the proof strategy for the main result Theorem 4.2 is similar to Gross-Zagier's strategy: height calculations at finite and infinite places via intersection theory and Green's functions, ``holomorphic projection from the space of functions on the edges of the tree \({\mathcal T}\) to the space of automorphic forms, etc. But since many of the ingredients in Gross-Zagier's paper are unavailable in the function field case, the authors had to invent many own devices, e.g. theta series for quadratic forms over \(K\), or Green functions on \({\mathcal T}\). Although only a tiny part of the author's work in the field, the well-written present paper exhibits the nature of results and methods that can be found in [Doc. Math., J. DMV 5, 365--444 (2000; Zbl 1012.11039)] and further papers the authors' paper (loc. cit.). For the entire collection see [Zbl 0903.00035].

Recommendations

Cited in

(22)

This page was built for publication: A Gross-Zagier formula for function fields

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4247786)