Selmer groups for elliptic curves in \(\mathbb Z_l^d \)-extensions of function fields of characteristic \(p\) (Q848123)

Given an elliptic curve \(E\) over a number field \(K\), a crucial result in the Iwasawa theory for \(E\) is Mazur's ``control theorem'' [see \textit{B. Mazur}, Invent. Math. 18, 183--266 (1972; Zbl 0245.14015) or \textit{R. Greenberg}, Lect. Notes Math. 1716, 51--144 (1999; Zbl 0946.11027)]: for a prime \(\ell\) of good ordinary reduction for \(E\) -- meaning that \(E\) has good ordinary reduction at all places in \(K\) dividing \(\ell\) -- consider the cyclotomic \(\mathbb{Z}_\ell\)-extension \[ K_\infty=\bigcup_mK_m\supset\hdots\supset K_n\supset\hdots\supset K \] and let \(\Gamma=\text{Gal}(K_\infty/K)\). Then, for every \(n\geq 0\), the restriction map induces a homomorphism \(\text{Sel}_E(K_n)_\ell\to \text{Sel}_E(K_\infty)_\ell^{\Gamma^{\ell^n}}\) and Mazur's theorem states that the kernel and the cokernel of this homomorphism are finite and bounded independently of \(n\): here we denote by \(\text{Sel}_E(K_n)_\ell\subseteq H^1(K_n,E[\ell^\infty])\), for \(n\leq \infty\), the \(\ell\)-primary part of the full Selmer group of \(E/K_n\), sitting in the exact sequence \[ 0\rightarrow E(K_n)\otimes\mathbb{Q}_\ell/\mathbb{Z}_\ell\rightarrow \text{Sel}_E(K_n)_\ell\rightarrow {\text Ш}(E/K_n)[\ell^\infty]\rightarrow 0\;. \] Under suitable assumptions, the above theorem can be generalized to primes \(\ell\) of bad multiplicative reduction for \(E/K\) [see Greenberg, loc. cit.], in particular Proposition 3.6). The above theorem is the main tool for controlling the growth of Selmer groups along \(\mathbb{Z}_\ell\)-extensions and implies, among other things, that the Pontryagin dual of \(\text{Sel}_E(K_\infty)_\ell\) is finitely generated and torsion over the Iwasawa algebra \(\mathbb{Z}_\ell[[\Gamma]]\). The main result of the paper under review is an analogous control theorem for (non-isotrivial) elliptic curves over function fields of characteristic \(p\neq \ell\). More precisely, let \(F\) be a function field of characteristic \(p\) and constant field \(\mathbb{F}\) and consider a \(\mathbb{Z}_\ell^d\)-extension \(\mathcal{F}/F\) of \(F\) with Galois group \(\text{Gal}(\mathcal{F}/F)=\Gamma\): choose a sequence of finite sub-extensions \(\mathcal{F}\supset F_n\supset F\) such that \(\bigcup F_n=\mathcal{F}\), and denote by \(\Gamma_n\) the subgroup \(\text{Gal}(\mathcal{F}/F_n)\). Let now \(E/F\) be an elliptic curve such that \(j(E)\notin \mathbb{F}\) having good or split multiplicative reduction at all primes of \(F\). Then the authors study the kernels and the cokernels of the homomorphisms \(\text{Sel}_E(F_n)_r\to \text{Sel}_E(\mathcal{F})_r^{\Gamma_n}\) for all primes \(r\). They can prove that, if \(r\neq \ell\), these maps are isomorphisms (Theorem 4.16 for \(r\neq p\) and Theorem 5.1 for \(r=p\)); that, if \(r=\ell\) but the unique \(\mathbb{Z}_\ell\)-extension \(\mathbb{F}_p^{(\ell)}\) of \(\mathbb{F}_p\) is \textit{not} contained in \(\mathbb{F}\), then the kernels and the cokernels are finite and bounded independently of \(n\) (Theorem 4.5); and, finally, that if \(r=\ell\) and \(\mathbb{F}_p^{(\ell)}\subseteq \mathbb{F}\) but only finitely many places of \(F\) ramify in \(\mathcal{F}\) then the kernels are still finite of bounded order and the cokernels are cofinitely generated over \(\mathbb{Z}_\ell\), and even of bounded corank provided \(\Gamma\cong\mathbb{Z}_\ell\) (Theorem 4.8). As an application of these results, the authors state the following theorems, where \(\mathcal{S}=\text{Hom}_{\text{cont}}(\text{Sel}_E(\mathcal{F})_\ell,\mathbb{Q}_\ell/\mathbb{Z}_\ell)\) and \(\Lambda=\mathbb{Z}_\ell[[\Gamma]]\): Theorem 1.1. Assume that \(\mathbb{F}\) does not contain \(\mathbb{F}_p^{(\ell)}\). Then \(\mathcal{S}\) is a finitely generated \(\Lambda\)-module. Moreover, if \(\text{Sel}_E(F)_\ell\) is finite, then \(\mathcal{S}\) is \(\Lambda\)-torsion. Theorem 1.2. Assume that only finitely many primes of \(F\) are ramified in \(\mathcal{F}/F\) and that \(\mathbb{F}\) contains \(\mathbb{F}_p^{(\ell)}\). Then \(\mathcal{S}\) is a finitely generated \(\Lambda\)-module. Moreover, if: {\parindent=6mm \begin{itemize}\item[1.] the ramified primes are of good reduction for \(E\); \item[2.] for any ramified primes \(v\), \(E[\ell^\infty](F_v)\) is finite (\(F_v\) is the completion of \(F\) at \(v\)); \item[3.] \(\text{Sel}_E(F)_\ell\) is finite, \end{itemize}} then \(\mathcal{S}\) is \(\Lambda\)-torsion. For \(\ell=p\), this problem has been studied in [\textit{T. Ochiai} and \textit{F. Trihan}, Math. Proc. Camb. Philos. Soc. 146, No. 1, 23--43 (2009; Zbl 1156.14037)], mainly in the case when the extension comes from base-change of the constant field, in a more general setting by the authors of the paper under review in [\textit{A. Bandini} and \textit{I. Longhi}, Int. J. Number Theory 5, No. 2, 229--256 (2009; Zbl 1234.11071)], and in [\textit{K.-S. Tan}, Trans. Am. Math. Soc. 362, No. 8, 4433--4450 (2010; Zbl 1239.11122)] where higher dimensional abelian varieties are treated. Moreover, in [\textit{J. S. Ellenberg}, Compos. Math. 142, No. 5, 1215--1230 (2006; Zbl 1106.11021)], applications to the study of Mordell-Weil ranks are considered. The paper falls in five sections and an Appendix. The first one contains the statements of Theorems 1.1 and 1.2 together with an overall introduction to the work, while Section 2. fixes notations about field extensions and Selmer groups. Section 3. contains some auxiliary lemma, among which Lemma 3.4 giving a general result for bounding cohomology groups of discrete \(\Gamma\)-modules which turns out to be the main tool used in subsequent computations. Section 4. is concerned with the proof of the main result in case \(r\neq p\) and splits into four subsections: they are devoted, respectively, to the case \(r=\ell\) and \(\mathbb{F}_p^{(\ell)}\not\subset\mathbb{F}\); to the case \(r=\ell\) and \(\mathbb{F}_p^{(\ell)}\subset\mathbb{F}\); to certain applications of the main results to some kind of ``Main Conjecture''; and, finally, to the case \(r\neq \ell,p\). Lemma 4.2 gives an explanation, based on global class field theory, for the different behaviors of Selmer groups according to \(\mathbb{F}_p^{(\ell)}\) being or not contained in \(\mathbb{F}\). The last and very brief Section 5. is devoted to the case \(r=p\) and contains also a study of the structure of (the dual of) \(\text{Sel}_E(\mathcal{F})_r\) as module over the completed group ring \(\mathbb{Z}_r[[\Gamma]]\). Finally, in the Appendix, the authors discuss general results about \(\mathbb{Z}_\ell\)-extensions of fields and their interpretation through Kummer theory.