A rank zero p p -converse to a theorem of Gross–Zagier, Kolyvagin and Rubin Ashay A. Burungale and Ye Tian Abstract . Let E E be a CM elliptic curve defined over ℚ {\mathbb{Q}} and p p a prime. We show that corank ℤ p ​ Sel p ∞ ​ ( E / ℚ ) = 0 ⟹ ord s = 1 ​ L ​ ( s , E / ℚ ) = 0 {\mathrm{corank}}_{{\mathbb{Z}}_{p}}{\mathrm{Sel}}_{p^{\infty}}(E_{/{\mathbb{Q}}})=0\implies{\mathrm{ord}}_{s=1}L(s,E_{/{\mathbb{Q}}})=0 for the p ∞ p^{\infty} -Selmer group Sel p ∞ ​ ( E / ℚ ) {\mathrm{Sel}}_{p^{\infty}}(E_{/{\mathbb{Q}}}) and the complex L L -function L ⁡ ( s , E / ℚ ) L(s,E_{/{\mathbb{Q}}}) . Along with Smith’s work on the distribution of 2 ∞ 2^{\infty} -Selmer groups, this leads to the first instance of the even parity Goldfeld conjecture: For 50 % 50\% of the positive square-free integers n n , we have ord s = 1 ​ L ​ ( s , E / ℚ ( n ) ) = 0 , {\mathrm{ord}}_{s=1}L(s,E^{(n)}_{/{\mathbb{Q}}})=0, where E ( n ) : n ​ y 2 = x 3 − x E^{(n)}:ny^{2}=x^{3}-x is a quadratic twist of the congruent number elliptic curve E : y 2 = x 3 − x E:y^{2}=x^{3}-x . 1. Introduction A p -converse to a theorem of Gross–Zagier, Kolyvagin and Rubin was recently proved [ 3 ] for CM elliptic curves in the rank one case for primes p of good ordinary reduction. In this article we prove such a p -converse theorem for CM curves in the rank zero case for any prime p . As a consequence of the 2 -converse and Smith’s work [ 21 ] , we prove the first instance of the even parity Goldfeld conjecture. 1.0.1. p -converse Let E be an elliptic curve defined over {\mathbb{Q}} . For a prime p , the p^{\infty} -Selmer group {\mathrm{Sel}}_{p^{\infty}}(E_{/{\mathbb{Q}}}) encodes the arithmetic of E via the exact sequence 0\rightarrow E({\mathbb{Q}})\otimes_{{\mathbb{Z}}}{\mathbb{Q}}_{p}/{\mathbb{Z}}_{p}\rightarrow{\mathrm{Sel}}_{p^{\infty}}(E_{/{\mathbb{Q}}})\rightarrow\hbox{\cyr X}(E_{/{\mathbb{Q}}})[p^{\infty}]\rightarrow 0. Let L(s,E_{/{\mathbb{Q}}}) be the associated Hasse–Weil L -function. The Birch and Swinnerton-Dyer (BSD) conjecture connects the arithmetic of E to the analytic properties of L(s,E_{/{\mathbb{Q}}}) . The implication {\mathrm{ord}}_{s=1}L(s,E_{/{\mathbb{Q}}})=r\leq 1\implies{\mathrm{rank}}_{{\mathbb{Z}}}E({\mathbb{Q}})=r,\#\hbox{\cyr X}(E_{/{\mathbb{Q}}})<\infty is a fundamental result towards the BSD conjecture due to Coates–Wiles [ 4 ] , Gross–Zagier [ 7 ] , Kolyvagin [ 15 ] and Rubin [ 16 ] . We refer to the implication (1.1) {\mathrm{corank}}_{{\mathbb{Z}}_{p}}{\mathrm{Sel}}_{p^{\infty}}(E_{/{\mathbb{Q}}})=r\implies{\mathrm{ord}}_{s=1}L(s,E_{/{\mathbb{Q}}})=r as the rank r p -converse theorem. In the early 90’s Rubin’s seminal work [ 17 ] led to the rank zero p -converse theorem for CM elliptic curves E_{/{\mathbb{Q}}} and primes (1.2) p\nmid\#\mathcal{O}_{K}^{\times} for K the CM field. More recently, the p -converse has been revived by Skinner [ 19 ] and Zhang [ 24 ] . Our main result is the following. Theorem 1.1 . Let E be an elliptic curve defined over an imaginary quadratic field K , with complex multiplication by an order of K . Let p be a prime. Then {\mathrm{corank}}_{{\mathbb{Z}}_{p}}{\mathrm{Sel}}_{p^{\infty}}(E_{/K})=0\implies{\mathrm{ord}}_{s=1}L(s,E_{/K})=0. In particular, if E descends to {\mathbb{Q}} , then ( 1.1 ) holds. In the main text the p -converse is proved for any CM newform (cf. Theorem 3.1 ). 1.0.2. Goldfeld’s conjecture Let E be an elliptic curve defined over {\mathbb{Q}} , and E^{(n)} its quadratic twist associated to n\in{\mathbb{Z}}\setminus\{0\} . In 1979 Goldfeld [ 6 ] conjectured that {\mathrm{ord}}_{s=1}L(s,E^{(n)}_{/{\mathbb{Q}}})=0 for 50\% of the fundamental 1 1 1 Goldfeld also conjectured that {\mathrm{ord}}_{s=1}L(s,E^{(n)}_{/{\mathbb{Q}}})=1 for 50\% of the fundamental discriminants n . discriminants n (the even parity Goldfeld conjecture). The first instance: Theorem 1.2 . The even parity Goldfeld conjecture is true for the congruent number elliptic curve E:y^{2}=x^{3}-x , that is 2 2 2 Note that \epsilon(E^{(n)})=+1 precisely when n\equiv 1,2,3\ \mathrm{mod}\ {8} [ 2 ] . , for a density one subset of the positive square-free integers n\equiv 1,2,3\ \mathrm{mod}\ {8} , {\mathrm{ord}}_{s=1}L(s,E^{(n)}_{/{\mathbb{Q}}})=0. For perspectives on the distribution of analytic ranks of elliptic curves, one may refer to Katz and Sarnak [ 13 , §4] . Under the GRH, Heath-Brown [ 9 ] proved the non-vanishing as in Theorem 1.2 for a positive proportion of the twists. Smith [ 20 ] proved that {\mathrm{ord}}_{s=1}L(s,E^{(n)}_{/{\mathbb{Q}}})=0 for at least 41.9\% of the positive square-free integers n\equiv 1,2,3\ \mathrm{mod}\ {8} . The mod 2 method [ 20 ] builds on [ 22 ] , [ 23 ] , [ 8 ] . Our approach to Theorem 1.2 also yields the following. Proposition 1.3 . Let E be an elliptic curve defined over an imaginary quadratic field K , with CM by an order of K . Suppose that 3 is not inert in K . Then for at least 50\% of t\in K^{\times}/(K^{\times})^{2} , {\mathrm{ord}}_{s=1}L(s,E^{(t)}_{/K})=0. This is an evidence towards Goldfeld’s conjecture over imaginary quadratic fields [ 14 , Conj. 7.12] , which posits {\mathrm{ord}}_{s=1}L(s,E^{(t)}_{/K})=0 for a density one subset of t\in K^{\times}/(K^{\times})^{2} since \epsilon(E^{(t)}_{/K})=+1 for any t . 1.0.3. About the proofs Theorem 1.2 is based on a 2 -adic method. It is a consequence of the rank zero 2 -converse Theorem 1.1 , and the distribution of {\mathrm{corank}}_{{\mathbb{Z}}_{2}}{\mathrm{Sel}}_{2^{\infty}}(E^{(n)}_{/{\mathbb{Q}}}) for positive square-free n\equiv 1,2,3\ \mathrm{mod}\ {8} due to Smith [ 21 ] (cf. Theorem 3.3 ). Note that the prime p=2 is fundamental to [ 21 ] , and is excluded by the results of [ 17 ] , [ 18 ] . Our approach to the p -converse is based on Iwasawa theory of zeta elements and is uniform for any prime p . For primes p of non-ordinary reduction, the existence of a desirable p -adic L -function is still elusive. We consider main conjectures in terms of the {\mathrm{GL}}_{1/K} and {\mathrm{GL}}_{2/{\mathbb{Q}}} -zeta elements, namely elliptic units and Beilinson–Kato elements. Some of the following notation is not followed in the rest of the article. Let T be the p -adic Tate module of E_{/{\mathbb{Q}}} . To recall Kato’s main conjecture, let {\mathbb{Q}}_{\infty} be the cyclotomic {\mathbb{Z}}_{p} -extension of {\mathbb{Q}} and \Lambda={\mathbb{Z}}_{p}[\![{\mathrm{Gal}}({\mathbb{Q}}_{\infty}/{\mathbb{Q}})]\!] . Let z_{E}\in H^{1}({\mathbb{Q}},T)\otimes_{{\mathbb{Z}}_{p}}{\mathbb{Q}}_{p} be the associated Beilinson–Kato element. For the Iwasawa cohomology H^{1}({\mathbb{Z}}[\frac{1}{p}],T\otimes_{{\mathbb{Z}}_{p}}\Lambda) , let {\bf{z}}_{E}\in H^{1}({\mathbb{Z}}[\frac{1}{p}],T\otimes_{{\mathbb{Z}}_{p}}\Lambda)\otimes_{{\mathbb{Z}}_{p}}{\mathbb{Q}}_{p} be the \Lambda -adic deformation of z_{E} due to Kato. Let X_{{\mathrm{st}}}(E_{/{\mathbb{Q}}_{\infty}}) be the attached strict Selmer group. Kato’s main conjecture predicts: (i) X_{{\mathrm{st}}}(E_{/{\mathbb{Q}}_{\infty}}) is \Lambda -torsion and (ii) we have \xi\big(H^{1}({\mathbb{Z}}[\frac{1}{p}],T\otimes_{{\mathbb{Z}}_{p}}\Lambda)\otimes{\mathbb{Q}}_{p}/\Lambda\cdot{{\bf{z}}_{E}}\big)=\xi(X_{{\mathrm{st}}}(E_{/{\mathbb{Q}}_{\infty}})), an equality of ideals in \Lambda\otimes_{{\mathbb{Z}}_{p}}{\mathbb{Q}}_{p} for \xi(\cdot) the \Lambda\otimes_{{\mathbb{Z}}_{p}}{\mathbb{Q}}_{p} -characteristic ideal. A key observation is that the rank zero p -converse is a consequence of Kato’s main conjecture. A principal result is a proof of the CM case of Kato’s main conjecture (cf. Theorem 2.6 ). It is approached via an equivariant main conjecture for the CM field involving elliptic units. This relies on a link between the Beilinson–Kato elements and elliptic units due to Kato [ 12 , §15] . As for the equivariant main conjecture up to tensoring with {\mathbb{Q}}_{p} , a result of Johnson-Leung and Kings [ 11 ] is employed. Our approach to the p -converse differs from that of Rubin [ 17 ] . It is uniform for any prime p , whereas [ 17 , §11] distinguishes the ordinary and non-ordinary primes, the latter being more complex. The Beilinson–Kato elements do not appear in [ 17 ] , while they are elemental to the present approach. The hypothesis ( 1.2 ) seems essential for the Euler system argument in [ 17 ] . The proof of Kato’s main conjecture, Theorem 2.6 , has other consequences. It gives rise to a rank one 2 -converse theorem for CM curves, which is utilised in the proof of a recent result of Alpoge–Bhargava–Shnidman: a positive proportion of the integers are the sum of two rational cubes. After tensoring with {\mathbb{Q}}_{p} , the main conjectures bypass intricacies which remain to be resolved, especially for the prime p=2 (cf. [ 5 ] ). For the rational version, the strategy of [ 11 ] was inspired by Kato’s pioneering work on the equivariant Tamagawa number conjecture (ETNC). We hope that Theorem 1.2 is just a glimpse of concrete applications of the general framework of the ETNC. It will be interesting to exhibit instances of the even parity Goldfeld conjecture for non-CM curves. Since the results of [ 21 ] apply, it is natural to seek the rank zero 2 -converse for non-CM curves. In the near future we plan to consider instances of Kato’s main conjecture for 2 -ordinary non-CM curves. 1.0.4. Organisation Section 2 presents the CM case of Kato’s main conjecture. We begin with a result towards the equivariant main conjecture for imaginary quadratic fields § 2.1 and then prove Kato’s main conjecture for CM newforms § 2.2 . Section 3 describes the proofs of main results. The p -converse theorem § 3.1 is followed by results towards the Goldfeld conjecture § 3.2 . 2. Kato’s main conjecture The principal result of this section is Theorem 2.6 . 2.0.1. Notation Let c\in{\mathrm{Gal}}({\mathbb{C}}/{\mathbb{R}}) be the complex conjugation. Let \overline{{\mathbb{Q}}} be a fixed algebraic closure of {\mathbb{Q}} . For a subfield F\subset\overline{{\mathbb{Q}}} , let G_{F}={\mathrm{Gal}}(\overline{{\mathbb{Q}}}/F) . For a place v of F , let \overline{F}_{v} denote a fixed algebraic closure of F_{v} and G_{F_{v}}={\mathrm{Gal}}(\overline{F}_{v}/F_{v}) . Choose an F -linear embedding \overline{{\mathbb{Q}}}\hookrightarrow\overline{F}_{v} , which identifies G_{F_{v}} as a subgroup of G_{F} . Let p be a prime. Let \mu_{p^{\infty}} be the set of p -power roots of unity in \overline{{\mathbb{Q}}}^{\times} . Let (\zeta_{p^{n}})_{n\geq 1} be a compatible system of primitive p^{n} -th roots of unity. Let \epsilon:G_{{\mathbb{Q}}}\rightarrow{\mathbb{Z}}_{p}^{\times} be the p -adic cyclotomic character. Fix embeddings \iota_{\infty}:\overline{{\mathbb{Q}}}\hookrightarrow{\mathbb{C}} and \iota_{p}:\overline{{\mathbb{Q}}}\hookrightarrow{\mathbb{C}}_{p} . For a ring R , a prime ideal {\mathfrak{q}}\subset R and an R -module M , let M_{{\mathfrak{q}}} denote the localisation. 2.1. Equivariant main conjecture We describe a result towards the equivariant main conjecture for imaginary quadratic fields. For details, one may refer to [ 17 ] , [ 11 ] , and especially [ 12 , §15] . Let K be an imaginary quadratic field. Fix an embedding \iota:K\hookrightarrow{\mathbb{C}} . 2.1.1. Setup Let K^{{\mathrm{ab}}} be the maximal abelian extension of K in {\mathbb{C}} . For a non-zero ideal {\mathfrak{a}} of \mathcal{O}_{K} , let K({\mathfrak{a}})\subset K^{{\mathrm{ab}}} be the ray class field of conductor {\mathfrak{a}} . Fix a prime p and a non-zero ideal {\mathfrak{f}} . Let K(p^{\infty}{\mathfrak{f}})=\bigcup_{n}K(p^{n}{\mathfrak{f}}),\text{ }G_{p^{\infty}{\mathfrak{f}}}={\mathrm{Gal}}(K(p^{\infty}{\mathfrak{f}})/K). Note that G_{p^{\infty}{\mathfrak{f}}}\simeq{\mathbb{Z}}_{p}^{2}\times\Delta_{{\mathfrak{f}}} for \Delta_{{\mathfrak{f}}} a finite group. Let {\mathfrak{h}}^{1}=\varprojlim_{K^{\prime}}(\mathcal{O}_{K^{\prime}}[1/p]^{\times}\otimes_{{\mathbb{Z}}}{\mathbb{Z}}_{p})\text{, }{\mathfrak{h}}^{2}=\varprojlim_{K^{\prime}}{\mathrm{Cl}}_{K^{\prime}}(p), where K^{\prime} varies over finite extensions of K contained in K(p^{\infty}{\mathfrak{f}}) , {\mathrm{Cl}}_{K^{\prime}}(p) denotes the p -primary part of the ideal class group {\mathrm{Cl}}_{K^{\prime}} and the inverse limit is with respect to the norm maps. Recall that {\mathfrak{h}}^{1} and {\mathfrak{h}}^{2} are finitely generated modules over the three-dimensional semi-local ring {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!] . Moreover, {\mathfrak{h}}^{1} is a torsion-free {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!] -module such that {\mathfrak{h}}^{1}_{{\mathfrak{q}}} is of dimension one for any height zero prime ideal {\mathfrak{q}} of {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!] , and {\mathfrak{h}}^{2} is a torsion {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!] -module. Let {\mathfrak{z}}\subset{\mathfrak{h}}^{1} be the submodule of elliptic units as in [ 12 , §15.5] . Then {\mathfrak{h}}^{1}/{\mathfrak{z}} is a torsion {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!] -module. For a {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!] -module M , put M_{{\mathbb{Q}}}=M\otimes_{{\mathbb{Z}}}{\mathbb{Q}} . For a ({\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!]\otimes_{{\mathbb{Z}}}{\mathbb{Q}}) -torsion module N , let \xi(N) be the characteristic ideal. 2.1.2. Equivariant main conjecture The following result towards the equivariant main conjecture for imaginary quadratic fields is due to Johnson-Leung and Kings [ 11 ] . It builds on [ 17 ] and [ 10 ] . Theorem 2.1 . Let the notation be as above. Then we have \xi({\mathfrak{h}}^{2}_{{\mathbb{Q}}})=\xi(({\mathfrak{h}}^{1}/{\mathfrak{z}})_{{\mathbb{Q}}}), an equality of ideals in {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!]\otimes{\mathbb{Q}} . Proof. Let {\mathfrak{p}} be a height one prime ideal of {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!] such that p\notin{\mathfrak{p}} . Then we have {\mathrm{length}}_{{\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!]_{{\mathfrak{p}}}}{\mathfrak{h}}^{2}_{{\mathfrak{p}}}={\mathrm{length}}_{{\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!]_{{\mathfrak{p}}}}(({\mathfrak{h}}^{1}/{\mathfrak{z}})_{{\mathfrak{p}}}), after 3 3 3 One may to [ 12 , §15.6] for a description of the underlying Iwasawa modules in terms of Iwasawa cohomology. The formulation in Theorem 2.1 is then identical with the one in [ 11 , §7.2] . Johnson-Leung and Kings [ 11 , §7.2] . ∎ Remark 2.2 . For height one primes {\mathfrak{p}} of {\mathbb{Z}}_{p}[\![G_{p^{\infty}{\mathfrak{f}}}]\!] with p\in{\mathfrak{p}} , a result towards the equivariant main conjecture is proved in [ 11 , Thm. 5.7 & §5.5] . The conjecture is still open for p non-split in K . 2.2. Kato’s main conjecture For background, one may refer to [ 12 , §12 & §15] . 2.2.1. Elliptic newforms Let f\in S_{k}(\Gamma_{1}(N)) be an elliptic newform with q -expansion coefficents \{a_{n}(f)\}_{n\in{\mathbb{N}}} and F={\mathbb{Q}}(\{a_{n}(f)\}_{n}) the Hecke field. For a place \lambda of F above p , let F_{\lambda} be the completion and O_{\lambda} its integer ring. Let V_{F_{\lambda}}(f) be the two-dimensional G_{{\mathbb{Q}}} -representation associated to f as in [ 12 , (8.3)] . 2.2.2. Cyclotomic tower For n\in{\mathbb{Z}}_{\geq 0} , let G_{n}={\mathrm{Gal}}({\mathbb{Q}}(\zeta_{p^{n}})/{\mathbb{Q}})\text{, }G_{\infty}=\varprojlim_{n}G_{n}={\mathrm{Gal}}({\mathbb{Q}}(\zeta_{p^{\infty}})/{\mathbb{Q}}). The cyclotomic character induces an isomorphism \kappa:G_{\infty}\simeq{\mathbb{Z}}_{p}^{\times} , and let \sigma_{c} be the inverse image of c\in{\mathbb{Z}}_{p}^{\times} . Recall that {\mathbb{Z}}_{p}^{\times}\simeq{\mathbb{Z}}_{p}\times\Delta for \Delta\simeq\begin{cases}{\mathbb{Z}}/(p-1)&\text{if $p$ odd}\\ {\mathbb{Z}}/2&\text{else}.\end{cases} Let \Lambda=O_{\lambda}[\![G_{\infty}]\!] be a two dimensional complete semi-local ring. 2.2.3. Iwasawa cohomology Let T\subset V_{F_{\lambda}}(f) be a G_{{\mathbb{Q}}} -stable O_{\lambda} -lattice. For q\in{\mathbb{Z}} , let {\mathbb{H}}^{q}(T)=\varprojlim_{n}H^{q}({\mathbb{Z}}[\zeta_{p^{n}},1/p],T) be a \Lambda -module, where H^{q} denotes the étale cohomology as in [ 12 , 8.2] . Let {\mathbb{H}}^{q}(V_{F_{\lambda}}(f))={\mathbb{H}}^{q}(T)\otimes{\mathbb{Q}} be a (\Lambda\otimes{\mathbb{Q}}) -module. The following basic result is due to Kato [ 12 , Thm 12.4] . Theorem 2.3 . Let f\in S_{k}(\Gamma_{1}(N)) be an elliptic newform and p a prime. Then: (1) {\mathbb{H}}^{2}(V_{F_{\lambda}}(f)) is a torsion (\Lambda\otimes{\mathbb{Q}}) -module and (2) {\mathbb{H}}^{1}(V_{F_{\lambda}}(f)) is a free (\Lambda\otimes{\mathbb{Q}}) -module of rank one. 2.2.4. Beilinson–Kato elements The following pivotal existence of zeta elements for the p -adic Galois representation associated to an elliptic newform is due to Kato [ 12 , Thm. 12.5] . Theorem 2.4 . Let f\in S_{k}(\Gamma_{1}(N)) be an elliptic newform and p a prime. (1) There exists a non-zero F_{\lambda} -linear map V_{F_{\lambda}}(f)\rightarrow{\mathbb{H}}^{1}(V_{F_{\lambda}}(f));\text{ $\gamma\mapsto{\bf{z}}_{\gamma}(f)$}. (2) Let Z(f) be the (\Lambda\otimes{\mathbb{Q}}) -submodule of {\mathbb{H}}^{1}(V_{F_{\lambda}}(f)) generated by {\bf{z}}_{\gamma}(f) for all \gamma\in V_{F_{\lambda}}(f) . Then {\mathbb{H}}^{1}(V_{F_{\lambda}}(f))/Z(f) is a torsion (\Lambda\otimes{\mathbb{Q}}) -module. Remark 2.5 . For a characterising property of the map \gamma\mapsto{\bf{z}}_{\gamma}(f) in terms of the underlying critical L -values, one may refer to [ 12 , Thm. 12.5 (1)] . 2.2.5. CM modular forms A Hecke character \psi over K has infinity type (m,n) if the restriction of \psi to the archimedean part {\mathbb{C}}^{\times} of the idele class group over K is of the form z\mapsto z^{m}\cdot\bar{z}^{n}, where \bar{\cdot}\in{\mathrm{Gal}}(K/{\mathbb{Q}}) is induced by the embedding \iota and the complex conjugation c . Let f\in S_{k}(\Gamma_{1}(N)) be a CM newform. Then, we have L(s,f)=L(s,\psi) for a Hecke character \psi over the underlying CM field K with infinity type (1-k,0) , and L(s,\cdot) the complex L -function. We follow arithmetic normalisation for the complex L -functions. 2.2.6. Kato’s main conjecture Theorem 2.6 . Let f\in S_{k}(\Gamma_{1}(N)) be a CM newform and p a prime. Let F be the Hecke field and \lambda a prime above p . Let V_{F_{\lambda}}(f) be the associated Galois representation and {\mathbb{H}}^{q}(V_{F_{\lambda}}(f)) the Iwasawa cohomology. Let Z(f)\subset{\mathbb{H}}^{1}(V_{F_{\lambda}}(f)) be the (\Lambda\otimes{\mathbb{Q}}) -submodule of Beilinson–Kato elements. Then \xi({\mathbb{H}}^{2}(V_{F_{\lambda}}(f)))=\xi({\mathbb{H}}^{1}(V_{F_{\lambda}}(f))/Z(f)), an equality of ideals in \Lambda\otimes{\mathbb{Q}} . Proof. In [ 12 , §15] it is shown that the CM case of Kato’s main conjecture [ 12 , Conj. 12.10] in \Lambda\otimes{\mathbb{Q}} is a consequence 4 4 4 Being in a setup of (\Lambda\otimes{\mathbb{Q}}) -modules, the hypotheses of [ 12 , Lem. 15.13] hold. of the equivariant main conjecture for the underlying imaginary quadratic field up to tensoring with {\mathbb{Q}} . (Especially see [ 12 , 15.6] .) This is based on a link between the Beilinson–Kato elements and elliptic units [ 12 , (15.16.1)] . The proven case [ 12 , Thm. 15.2] of the equivariant main conjecture due to Rubin [ 17 ] is employed in [ 12 , Prop. 15.17] . In the present setting we instead utilise Theorem 2.1 . ∎ Remark 2.7 . A finer analysis of [ 12 , §15] is essential to relate the integral versions of the main conjectures in the above proof, which we hope to study in the near future. 3. Proofs of main results The central result is Theorem 3.1 , which leads to Theorem 1.1 , Theorem 1.2 and Proposition 1.3 . 3.1. p -converse Theorem 3.1 . Let f\in S_{k}(\Gamma_{1}(N)) be a CM newform of even weight and p a prime. Let F be the Hecke field and \lambda a prime above p . Let V_{F_{\lambda}}(f) be the associated Galois representation and H^{1}_{\rm{f}}({\mathbb{Q}},V_{F_{\lambda}}(f)(k/2)) the Bloch–Kato Selmer group. Then H^{1}_{\rm{f}}({\mathbb{Q}},V_{F_{\lambda}}(f)(k/2))=0\implies{\mathrm{ord}}_{s=k/2}L(s,f)=0. Proof. The following is based on Theorem 2.6 . Let T\subset V_{F_{\lambda}}(f) be a G_{{\mathbb{Q}}} -stable O_{\lambda} -lattice. Since H^{1}_{\rm{f}}({\mathbb{Q}},V_{F_{\lambda}}(f)(k/2))=0 , by the exact sequence [ 12 , (14.9.3)] , note that (3.1) H^{2}({\mathbb{Z}}[1/p],T(k/2))\otimes{\mathbb{Q}}=0. Let \pm=(-1)^{k/2-1} and pick \gamma\in V_{F_{\lambda}}(f) such that \gamma^{\pm}\neq 0 . Let z be the image of the Beilinson–Kato element {\bf{z}}_{\gamma}(f) (cf. Theorem 2.4 ) under the map (3.2) {\mathbb{H}}^{1}(V_{F_{\lambda}}(f))\simeq{\mathbb{H}}^{1}(V_{F_{\lambda}}(f)(k/2))\rightarrow H^{1}({\mathbb{Z}}[1/p],T(k/2))\otimes{\mathbb{Q}}. By ( 3.1 ) and the Euler–Poincaré formula of Tate [ 12 , (14.9.5)] , \dim_{F_{\lambda}}H^{1}({\mathbb{Z}}[1/p],T(k/2))\otimes{\mathbb{Q}}=1. Let {\mathfrak{q}} be the kernel of the O_{\lambda} -homomorphism \Lambda\rightarrow O_{\lambda} mapping \sigma_{c} to c^{-k/2} for any c\in{\mathbb{Z}}_{p}^{\times} (cf. § 2.2.2 ). Note that the map ( 3.2 ) factors through {\mathbb{H}}^{1}(V_{F_{\lambda}}(f))_{{\mathfrak{q}}}/{\mathfrak{q}}{\mathbb{H}}^{1}(V_{F_{\lambda}}(f))_{{\mathfrak{q}}}, and {\mathbb{H}}^{2}(V_{F_{\lambda}}(f))_{{\mathfrak{q}}} vanishes by ( 3.1 ). Hence Theorem 2.6 implies that z is an F_{\lambda} -basis of H^{1}({\mathbb{Z}}[1/p],T(k/2))\otimes{\mathbb{Q}} . (See also [ 12 , p. 242] .) Since H^{1}_{\rm{f}}({\mathbb{Q}},V_{F_{\lambda}}(f)(k/2)) vanishes, it follows that (3.3) {\mathrm{loc}}_{p}(z)\notin H^{1}_{\rm{f}}({\mathbb{Q}}_{p},V_{F_{\lambda}}(f)(k/2)) for {\mathrm{loc}}_{p}:H^{1}({\mathbb{Q}},V_{F_{\lambda}}(f)(k/2))\rightarrow H^{1}({\mathbb{Q}}_{p},V_{F_{\lambda}}(f)(k/2)) the localisation and H^{1}_{\rm{f}}({\mathbb{Q}}_{p},V_{F_{\lambda}}(f)(k/2)) the Bloch–Kato subgroup. Finally, ( 3.3 ) and Kato’s explicit reciprocity law [ 12 , Thm. 12.5 (1)] imply that {\mathrm{ord}}_{s=k/2}L(s,f)=0. ∎ Remark 3.2 . This deduction of the rank zero p -converse from Kato’s main conjecture holds for any elliptic newform. Proof of Theorem 1.1 . Let \psi:{\mathbb{A}}_{K}^{\times}/K^{\times}\rightarrow K^{\times} be the Hecke character associated to E_{/K} . Let f and \overline{f} be the CM newforms associated to \psi and \overline{\psi} , the corresponding Hecke field being a subfield of K . Note that L(s,E_{/K})=L(s,\psi)\cdot L(s,\overline{\psi})=L(s,f)\cdot L(s,\overline{f}) . In the same vein one has {\mathrm{Sel}}_{p^{\infty}}(E_{/K})^{\vee}\otimes K\simeq H^{1}_{\rm{f}}({\mathbb{Q}},V_{K_{\lambda}}(f))\oplus H^{1}_{\rm{f}}({\mathbb{Q}},V_{K_{{\lambda}}}(\overline{f})), where the superscript ‘ \vee ’ denotes the Pontryagin dual, and V_{K_{\lambda}}(\cdot)=V_{F_{\lambda}}(\cdot)\otimes_{F}K . Hence, by Theorem 3.1 , {\mathrm{corank}}_{{\mathbb{Z}}_{p}}{\mathrm{Sel}}_{p^{\infty}}(E_{/K})=0\implies{\mathrm{ord}}_{s=1}L(s,E_{/K})=0. If E descends to {\mathbb{Q}} , then {\mathrm{corank}}_{{\mathbb{Z}}_{p}}{\mathrm{Sel}}_{p^{\infty}}(E_{/K})=2\cdot{\mathrm{corank}}_{{\mathbb{Z}}_{p}}{\mathrm{Sel}}_{p^{\infty}}(E_{/{\mathbb{Q}}}) and {\mathrm{ord}}_{s=1}L(s,E_{/K})=2\cdot{\mathrm{ord}}_{s=1}L(s,E_{/{\mathbb{Q}}}) , which concludes the proof of Theorem 1.1 . 3.2. Goldfeld’s conjecture 3.2.1. Over the rationals The proof of Theorem 1.2 is partly based on the following progress towards the distribution of the associated 2^{\infty} -Selmer ranks due to Smith [ 21 , Cor. 1.2] . Theorem 3.3 . Let E:y^{2}=x^{3}-x be the congruent number elliptic curve. Then for a density one subset of the square-free positive integers n\equiv 1,2,3\ \mathrm{mod}\ {8} , {\mathrm{corank}}_{{\mathbb{Z}}_{2}}{\mathrm{Sel}}_{2^{\infty}}(E^{(n)}_{/{\mathbb{Q}}})=0. Remark 3.4 . For earlier results, one may refer to [ 22 ] , [ 23 ] , [ 20 ] and references therein. Proof of Theorem 1.2 . This is a consequence of Theorem 3.3 and the 2 -converse Theorem 1.1 . 3.2.2. Over imaginary quadratic fields The proof of Proposition 1.3 utilises the following [ 1 , Thm. 2.7] . Theorem 3.5 . Let E be an elliptic curve defined over an imaginary quadratic field K , with CM by an order of K . Suppose that 3 is not inert in K . Then for at least 50\% of t\in K^{\times}/(K^{\times})^{2} , {\mathrm{corank}}_{{\mathbb{Z}}_{3}}{\mathrm{Sel}}_{3^{\infty}}(E^{(t)}_{/K})=0. Proof of Proposition 1.3 . This is a consequence of Theorem 3.5 and the 3 -converse Theorem 1.1 . Remark 3.6 . (i) For any n\in{\mathbb{Z}}\setminus\{0\} , the cube sum elliptic curve E_{n}:x^{3}+y^{3}=2n satisfies the hypotheses of Proposition 1.3 , with K={\mathbb{Q}}(\zeta_{3}) . (ii) A positive proportion of the quadratic twists in Proposition 1.3 do not descend to {\mathbb{Q}} . The proposition yields (perhaps) the first examples of elliptic curves E over a number field F different from {\mathbb{Q}} for which a weak form of the Goldfeld conjecture holds: the analytic rank being zero for a positive proportion of the quadratic twists of E over F which do not descend to {\mathbb{Q}} .