On the Birch and Swinnerton-Dyer conjecture (Q5896427)

The main result is: Let \(E\) be an elliptic curve with complex multiplication. If \(L(E/\mathbb Q,s)\) has a zero of odd order at \(s=1\), then either the Mordell-Weil group \(E(\mathbb Q)\) has rank at least one, or the \(p\)-primary components of the Tate-Shafarevich group are infinite for all primes \(p\) (\(\neq 2,3)\) where at which \(E\) has ordinary good reduction. This complements a theorem of \textit{J. Coates} and \textit{A. Wiles} [Invent. Math. 39, 223--251 (1977; Zbl 0359.14009)] and provides further evidence for the Birch--Swinnerton-Dyer conjectures. The author records a variant of these conjectures suggested by Coates in case the Tate-Shafarevich groups have finite order. The proof is via a Hecke \(L\)-series associated with a Grössencharacter \(\psi\). The \(L\)-series obeys a functional equation \(L(s)=wL(2-s)\). Both cases \(w=\pm 1\) are needed in the proof. In particular the author shows that \(L(\Psi^{2k+1},k+1)=0\) for only finitely many \(k\) when \(\Psi^{2k+1}\) is a character with \(w=1\). This non-vanishing theorem leads via some intricate work with the anti-cyclotomic \(p\)-extension of the complex multiplication field to a proof that the Selmer groups are infinite in the case \(w=-1\) and thence to the main theorem.