Notes on non-commutative Iwasawa theory (Q609470)

Let \(E\) be an elliptic curve over a number field \(k\) and fix an odd prime number \(p\). In studying the Iwasawa theory of \(E\) over a \(p\)-adic extension \(k_\infty/k\) one his interested in the Selmer group of \(E\) over \(k_\infty\) and tries to compare it, on the one hand, with some analytic \(p\)-adic \(L\)-function and, on the other hand, with the Selmer group of the elliptic curve over the base field (see [\textit{R. Greenberg}, ``Iwasawa theory for elliptic curves'', Lect. Notes Math. 1716, 51--144 (1999; Zbl 0946.11027)] for a wonderful introduction to the subject). In the paper under review, the authors study the structure of the Selmer group of \(E\) over \(k_\infty\) in a non-commutative setting, partially generalizing known results of [\textit{Y. Ochi} and \textit{O. Venjakob}, ``On the structure of Selmer groups over \(p\)-adic Lie extensions'', J. Algebr. Geom. 11, No. 3, 547--580 (2002; Zbl 1041.11041)]. More precisely, let \(k_\infty/k\) be a Galois extension with Galois group \(G\) and assume that 1) \(k_\infty/k\) is unramified outside a finite number of places and contains the cyclotomic extension \(k_{\text{cyc}}\) of \(k\): set, accordingly, \(\Gamma:=\text{Gal}(k_\infty/k_{\text{cyc}})\); 2) \(G\) is a solvable (compact) \(p\)-adic Lie group; 3) \(G\) is uniformly powerful in the notation of [\textit{J. D. Dixon, M. P. F. du Sautoy, A. Mann} and \textit{D. Segal}, Analytic pro-\(p\) groups. Revised and enlarged by Marcus du Sautoy and Dan Segal. 2nd ed. Cambridge Studies in Advanced Mathematics. 61. Cambridge: Cambridge University Press (1999; Zbl 0934.20001)]. Let now \(\Lambda(G)\) be the usual Iwasawa algebra of \(G\) defined as the projective limit \(\Lambda(G):=\varprojlim\mathbb{Z}_p[G/U]\) of the group rings of the quotients \(G/U\) for \(U\) running over the set of normal open subgroups of \(G\). It is well known that the Pontryagin dual \[ \text{Sel}(E/k_\infty)^\lor:=\text{Hom}_{\text{cont}}(\text{Sel}(E/k_\infty),\mathbb{Q}_p/\mathbb{Z}_p) \] of the \(p\)-Selmer group \[ \text{Sel}(E/k_\infty):=\text{ker}\Big(H^1(k_\infty,E[p^\infty])\rightarrow\prod_v H^1(K_{\infty,v}E)[p^\infty]\Big) \] is finitely generated over \(\Lambda(G)\) and the authors show the following Theorem 2.3. Let \(E/k\) be an elliptic curve which has good reduction at all primes above \(p\) and assume that \(k_\infty/k\) satisfies the above assumptions 1.-3. Then we have \(\text{rank}_{\Lambda(G)}\text{Sel}(E/k_\infty)^\lor=s(E/k)\) provided \(\text{rank}_{\Lambda(\Gamma)}\text{Sel}(E/k_{\text{cyc}})^\lor=s(E/k)\) where \[ s(E/k):=\sum_{\substack{ v\text{ supersingular}\\v\mid p}} [k_v:\mathbb{Q}_p]\;. \] As noticed in {Remark 2.4}, the above theorem holds if \(\text{Gal}(k_\infty/k_{\text{cyc}})\cong\mathbb{Z}_p\) by [\textit{Y. Hachimori} and \textit{O. Venjakob}, ``Completely faithful Selmer groups over Kummer extensions'', Doc. Math., J. DMV Extra Vol., 443--478 (2003; Zbl 1117.14046)] while the condition \(\text{rank}_{\Lambda(\Gamma)}\text{Sel}(E/k_{\text{cyc}})^\lor=s(E/k)\) holds when \(E\) is defined over \(\mathbb{Q}\) and \(k/\mathbb{Q}\) is abelian. The proof of the theorem is mainly a verification of the technical hypothesis needed to apply a result of Balister and Howson proved in [\textit{P. N. Balister} and \textit{S. Howson}, ``Note on Nakayama's lemma for compact \(\Lambda\)-modules'', Asian J. Math. 1, No. 2, 224--229 (1997; Zbl 0904.16019)]. The second half of the paper is devoted to the proof of the following result, which was already proven in a more intricate way in [\textit{Y. Ochi} and \textit{O. Venjakob}, loc. cit.], Theorem 5.5: {Theorem 3.2.} Assume that \(E/k\) has good reduction at all primes of \(k\) above \(p\) and that the following conditions hold: \parindent=8mm \begin{itemize}\item[(i)] \(\dim (G_v)\geq 2\) for every prime of bad reduction \(v\) for \(E/k\)\; \item[(ii)] All primes \(v\) above \(p\) are deeply ramified in \(k_\infty/k\) in the sense of [\textit{J. Coates} and \textit{R. Greenberg}, ``Kummer theory for abelian varieties over local fields'', Invent. Math. 124, No. 1--3, 129--174 (1996; Zbl 0858.11032)] (an equivalent condition is that \(k_{\infty,v}/k_v\) does not have a finite conductor); \item[(iii)] For every prime \(v\mid p\) of good ordinary reduction, either \(\text{dim}(G_v)>2\); or \(\text{dim}(G_v)=2\) but \(\tilde{E}_v(\kappa_{\infty,v})[p^\infty]\) is finite, where \(\tilde{E}_v\) is the reduction of \(E\pmod{v}\) and \(\kappa_{\infty,v}\) is the residue field of \(k_{\infty,v}\); \item[(iv)] \(H^2(k_S/k_\infty,E[p^\infty])=0\) where \(k_S\) denotes the maximal extension of \(k\) unramified outside the infinite primes, the primes above \(p\), the ramified primes in \(k_\infty/k\), and the primes where \(E\) has bad reduction; \item[(v)] A certain map \(\varphi\) defined in equation (2.1) is surjective. Then \(\text{Sel}(E/k_\infty)^\lor\) has no non-trivial pseudo-null \(\Lambda(G)\)-submodules (for the definition of pseudo-nullity used here, see {Definition 3.1}). The proof uses to a large extend results proven in [\textit{Y. Hachimori} and \textit{O. Venjakob}, loc. cit.]; the really new ingredient is {Proposition 3.5} stating that the quotient of a \(\Lambda(G)\)-module without non-trivial pseudo-null submodules by a reflexive submodule has itself no non-trivial pseudo-null submodules.\end{itemize}