Hostname: page-component-cb9f654ff-kl2l2 Total loading time: 0 Render date: 2025-08-04T04:42:10.785Z Has data issue: false hasContentIssue false

On the Fitting ideals of anticyclotomic Selmer groups of elliptic curves with good ordinary reduction

Published online by Cambridge University Press:  14 July 2025

Chan-Ho Kim*
Affiliation:
Department of Mathematics and Institute of Pure and Applied Mathematics, https://ror.org/05q92br09 Jeonbuk National University , 567 Baekje-daero, Deokjin-gu, Jeonju, Jeollabuk-do 54896, Republic of Korea
Rights & Permissions [Opens in a new window]

Abstract

We give a short proof of the anticyclotomic analogue of the “strong” main conjecture of Kurihara on Fitting ideals of Selmer groups for elliptic curves with good ordinary reduction under mild hypotheses. More precisely, we completely determine the initial Fitting ideal of Selmer groups over finite subextensions of an imaginary quadratic field in its anticyclotomic $\mathbb {Z}_p$-extension in terms of Bertolini–Darmon’s theta elements.

Information

Type
Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (https://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2025. Published by Cambridge University Press on behalf of Canadian Mathematical Society

1 Introduction

1.1 The statement of the main result

Let E be an elliptic curve of conductor N over $\mathbb {Q}$ and $p \geq 5$ be a prime of good ordinary reduction for E such that

  1. (Im) the mod p Galois representation $\overline {\rho } : G_{\mathbb {Q}} = \mathrm {Gal}(\overline {\mathbb {Q}}/\mathbb {Q}) \to \mathrm {Aut}_{\mathbb {F}_p}(E[p]) \simeq \mathrm {GL}_2(\mathbb {F}_p)$ is surjective, and

  2. (Ram) $\overline {\rho }$ is ramified at every prime dividing N, so p does not divide Tamagawa factors of E.

Let K be an imaginary quadratic field of odd discriminant $-D_K < -4$ with $(D_K, Np) = 1$ such that

  1. (Spl) p splits in K, and

  2. (Na) $a_p(E) \not \equiv 1\ \pmod {p}$ .

Write

$$ \begin{align*}N = N^+ \cdot N^-,\end{align*} $$

where a prime divisor of $N^+$ splits in K and a prime divisor of $N^- $ is inert in K.

  1. (Def) Assume that $N^-$ is a square-free product of an odd number of primes.

Let $K_\infty $ be the anticyclotomic $\mathbb {Z}_p$ -extension of K and $K_n$ be the subextension of K in $K_\infty $ of degree $p^n$ for $n \geq 0$ . Let $\Lambda _n = \mathbb {Z}_p[\mathrm {Gal}(K_n/K)] \simeq \mathbb {Z}_p[X]/( (1+X)^{p^n} - 1 )$ be the finite layer Iwasawa algebra. Under (Def), denote by

$$ \begin{align*}\theta(E/K_n) = \sum_{\sigma \in \mathrm{Gal}(K_n/K)} a_{\sigma} \cdot \sigma \in \Lambda_n.\end{align*} $$

Bertolini–Darmon’s theta element of E over $K_n$ which interpolates the square-roots of $L(E/K, \chi , 1)$ for finite order characters $\chi $ on $\mathrm {Gal}(K_n/K)$ . It is reviewed in Section 2. For the natural projection map $\pi _{n, n-1} : \Lambda _{n} \to \Lambda _{n-1}$ , let $\nu _{n-1, n} : \Lambda _{n-1} \to \Lambda _n$ be the map defined by $\sigma \mapsto \sum _{\pi _{n, n-1}(\tau ) = \sigma } \tau $ . For $0 \leq m \leq n$ , we write $\nu _{m, n} = \nu _{n-1, n} \circ \nu _{n-2, n-1} \circ \cdots \circ \nu _{m, m+1} $ . Then we have the equality of ideals of $\Lambda _n$ (Lemma 2.1)

$$ \begin{align*}\left( \theta(E/K_n) , \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) \right) = \left( \nu_{m, n} \left( \theta(E/K_{m}) \right) : 0 \leq m \leq n \right) ,\end{align*} $$

and it is a principal ideal under (Na) (Lemma 2.2).

The goal of this article is to prove the following anticyclotomic analogue of the “strong” main conjecture of Kurihara [Reference Kurihara14, Conjecture 0.3], which refines the “weak” main conjecture of Mazur and Tate [Reference Mazur and Tate15, Conjecture 3].

Theorem 1.1 Under the assumptions mentioned above, i.e., (Im),(Ram),(Spl),(Na), and (Def), the theta elements over $K_m$ with $0 \leq m \leq n$ generate the initial Fitting ideal of dual Selmer groups over $K_n$ , i.e., we have equality of ideals of $\Lambda _n$

$$ \begin{align*}\left( \theta(E/K_n) , \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) \right)^2 = \mathrm{Fitt}_{\Lambda_n} \left( \mathrm{Sel}(K_n, E[p^\infty])^\vee \right) ,\end{align*} $$

which is indeed a principal ideal, where $\mathrm {Sel}(K_n, E[p^\infty ])$ is the classical Selmer group of $E[p^\infty ]$ over $K_n$ and $(-)^\vee $ means the Pontryagin dual.

Since theta elements interpolate the square-roots of twisted Rankin–Selberg L-values, it is natural that the square of the ideal generated by theta elements appears in the equality.

The strategy of our proof follows that given in [Reference Kim and Kurihara11] and we also add some details on the “p-destabilization” process and on the comparison of various anticyclotomic Selmer groups of elliptic curves.

2 Bertolini–Darmon’s theta elements and anticyclotomic p-adic L-functions

We quickly review the construction of Gross points of conductor $p^n$ , theta elements, and anticyclotomic p-adic L-functions. See [Reference Chida and Hsieh4, Reference Chida and Hsieh5, Reference Kim10] for details.

2.1 Gross points

Let K be the imaginary quadratic field of odd discriminant $-D_K < - 4$ . Define

$$ \begin{align*}\vartheta := \dfrac{ D_K - \sqrt{-D_K} }{ 2 } \end{align*} $$

so that $\mathcal {O}_K = \mathbb {Z} + \mathbb {Z}\vartheta $ . Let $B_{N^-}$ be the definite quaternion algebra over $\mathbb {Q}$ of discriminant $N^-$ . Then there exists an embedding $\Psi :K \hookrightarrow B_{N^-}$ [Reference Vignéras20]. More explicitly, we choose a K-basis $(1,J)$ of $B_{N^-}$ so that $B_{N^-} = K \oplus K \cdot J$ such that $\beta := J^2 \in \mathbb {Q}^\times $ with $\beta <0$ , $J \cdot t = \overline {t} \cdot J$ for all $t \in K$ , $\beta \in \left ( \mathbb {Z}^\times _q \right )^2$ for all $q \mid pN^+$ , and $\beta \in \mathbb {Z}^\times _q$ for all $q \mid D_K$ . Fix a square root $\sqrt {\beta } \in \overline {\mathbb {Q}}$ of $\beta $ . For a $\mathbb {Z}$ -module A, write $\widehat {A} = A \otimes \widehat {\mathbb {Z}}$ . Fix an isomorphism

$$ \begin{align*}i := \prod i_q : \widehat{B}^{(N^-)}_{N^-} \simeq \mathrm{M}_2(\mathbb{A}^{(N^-\infty)})\end{align*} $$

as follows:

  • For each finite place $q \mid N^+p$ , the isomorphism $i_q : B_{N^-,q} \simeq \mathrm {M}_2(\mathbb {Q}_q)$ is defined by

    where $\mathrm {trd}$ and $\mathrm {nrd}$ are the reduced trace and the reduced norm on B, respectively.
  • For each finite place $q \nmid pN^+$ , the isomorphism $i_q : B_{N^-,q} \simeq \mathrm {M}_2(\mathbb {Q}_q)$ is chosen so that $i_q \left ( \mathcal {O}_K \otimes \mathbb {Z}_q \right ) \subseteq \mathrm {M}_2(\mathbb {Z}_q) $ .

Under the fixed isomorphism i, for any rational prime q, the local Gross point ${\varsigma _q \in B^\times _{N^-,q}}$ is defined as follows:

  • $\varsigma _q := 1$ in $B^\times _{N^-,q}$ for $q \nmid pN^+$ .

  • $\varsigma _q := \frac {1}{\sqrt {D_K}}\cdot \left ( \begin {matrix} \vartheta & \overline {\vartheta } \\ 1 & 1 \end {matrix} \right ) \in \mathrm {GL}_2(K_{\mathfrak {q}}) = \mathrm {GL}_2(\mathbb {Q}_q) $ for $q \mid N^+$ with $q = \mathfrak {q} \overline {\mathfrak {q}}$ in $\mathcal {O}_K$ .

  • $\varsigma ^{(n)}_p = \left ( \begin {matrix} \vartheta & -1 \\ 1 & 0 \end {matrix} \right ) \cdot \left ( \begin {matrix} p^n & 0 \\ 0 & 1 \end {matrix} \right ) \in \mathrm {GL}_2(K_{\mathfrak {p}}) = \mathrm {GL}_2( \mathbb {Q}_{p} )$ where $p = \mathfrak {p}\overline {\mathfrak {p}}$ splits in K.

Let $\widehat {\Psi } : \widehat {K} \hookrightarrow \widehat {B}_{N^-}$ be the adelic version of $\Psi $ . We define $x_n : \widehat {K}^\times \to \widehat {B}^\times _{N^-}$ by $x_n(a) = \widehat {\Psi }(a) \cdot \varsigma ^{(n)} := \widehat {\Psi }(a) \cdot \left ( \varsigma ^{(n)}_p \times \prod _{q \neq p} \varsigma _q \right )$ . The collection $\left \lbrace x_n(a) : a \in \widehat {K}^\times \right \rbrace $ of points is called the Gross points of conductor $p^n$ on $\widehat {B}^\times _{N^-}$ . The fixed embedding $K \hookrightarrow B_{N^-}$ also induces an optimal embedding of $\mathcal {O}_n = \mathbb {Z}+p^n \mathcal {O}_K$ into the Eichler order $B_{N^-} \cap \varsigma ^{(n)}\widehat {R}_{N^+}(\varsigma ^{(n)})^{-1}$ where $R_{N^+}$ is the Eichler order of level $N^+$ under the fixed isomorphism i.

2.2 Theta elements

Let $f(z) = \sum _{n \geq 1} a_n q^n \in S_{2}(\Gamma _0(N))$ be the cuspidal newform of weight two with rational Fourier coefficients corresponding to E via the modularity theorem [Reference Breuil, Conrad, Diamond and Taylor3]. Let $\phi _f : B^\times _{N^-} \backslash \widehat {B}^{\times }_{N^-} / \widehat {R}^\times _{N^+} \to \mathbb {C}$ be the Jacquet–Langlands transfer of f. Since $B^\times _{N^-} \backslash \widehat {B}^{\times }_{N^-} / \widehat {R}^\times _{N^+}$ is a finite set and f is a Hecke eigenform, we are able to and do normalize

$$ \begin{align*}\phi_f : B^\times_{N^-} \backslash \widehat{B}^{\times}_{N^-} / \widehat{R}^\times_{N^+} \to \mathbb{Z}_p,\end{align*} $$

such that the image of $\phi _f $ does not lie in $p\mathbb {Z}_p$ . This integral normalization is related to the congruence ideals [Reference Kim9, Reference Kim and Ota12, Reference Pollack and Weston16]. Let

$$ \begin{align*}\widetilde{\theta}_n(E/K) = \sum_{[a] \in \mathcal{G}_n} \phi_f (x_n(a) ) \cdot [a] \in \mathbb{Z}_p[\mathcal{G}_n],\end{align*} $$

where $\mathcal {G}_n = K^\times \backslash \widehat {K}^\times / \widehat {\mathcal {O}}^\times _n $ and $[a]$ is the image of $a \in \widehat {K}^\times $ in $\mathcal {G}_n$ . Then Bertolini–Darmon’s theta element $\theta (E/K_n)$ of E over $K_n$ is defined by the image of $\widetilde {\theta }_n(E/K)$ in $\Lambda _n$

where the map is naturally induced from the quotient map $\mathcal {G}_n \to \mathrm {Gal}(K_n/K)$ . It is known that $\theta (E/K_n)$ interpolates “an half of” $L(E, \chi , 1)$ where $ \chi $ runs over characters on $\mathrm {Gal}(K_n/K)$ . See [Reference Chida and Hsieh5, Remark (iii) after Theorem A] for the precise meaning of “an half of”. Because $\theta (E/K_n)$ depends on the choice of Gross points, $\theta (E/K_n)$ is well-defined only up to multiplication by $\mathrm {Gal}(K_n/K)$ .

2.3 p-adic L-functions

Let $\alpha , \beta $ be the roots of the Hecke polynomial $X^2 - a_pX + p$ of f at p. Since f is ordinary at p, one of them, say $\alpha $ , is a p-adic unit.

The p -stabilization $f_\alpha \in S_{2}(\Gamma _0(Np))$ of f is defined by

$$ \begin{align*}f_\alpha (z) = f(z) - \beta \cdot f(pz)\end{align*} $$

whose $U_p$ -eigenvalue is $\alpha $ . Then the theta element of $f_\alpha $ over $K_n$ is characterized by the following relation:

(2.1) $$ \begin{align} \theta(f_\alpha/K_n) = \dfrac{1}{\alpha^n} \cdot \left( \theta(E/K_n) - \dfrac{1}{\alpha} \cdot \nu_{n-1, n}( \theta(E/K_{n-1}) ) \right). \end{align} $$

It is known that theta elements of E satisfies the three term relation (e.g., [Reference Darmon and Iovita6, Lemma 2.6])

(2.2) $$ \begin{align} \pi_{n+1, n} \left( \theta(E/K_{n+1}) \right) = a_p \cdot \theta(E/K_n) - \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) \end{align} $$

and the theta elements of $f_\alpha $ satisfy the norm compatibility

(2.3) $$ \begin{align} \pi_{n+1, n} \left( \theta(f_\alpha/K_{n+1}) \right)= \theta(f_\alpha/K_{n}). \end{align} $$

Let $\iota $ be the involution on $\Lambda _n$ defined by inverting group-like elements, so we have

$$ \begin{align*}\iota( \sum_{\sigma \in \mathrm{Gal}(K_n/K)} a_{\sigma} \cdot \sigma ) = \sum_{\sigma \in \mathrm{Gal}(K_n/K)} a_{\sigma} \cdot \sigma^{-1}.\end{align*} $$

We define the anticyclotomic p -adic L -function of E by

$$ \begin{align*}L_p(E/K_\infty) = \varprojlim_n\left( \theta(f_\alpha/K_n) \cdot \iota( \theta(f_\alpha/K_n) ) \right) \in \Lambda = \varprojlim_n \Lambda_n.\end{align*} $$

This element is well-defined. The functional equation for Bertolini–Darmon’s theta elements yields the equality of ideals of $\Lambda $ (e.g., [Reference Bertolini and Darmon1, Proposition 2.13], [Reference Bertolini and Darmon2, Lemma 1.5])

(2.4) $$ \begin{align} \left( \theta(f_\alpha/K_n) \right) = \left( \iota( \theta(f_\alpha/K_n) ) \right). \end{align} $$

We prove two useful lemmas.

Lemma 2.1 We have an equality of ideals of $\Lambda _n$

$$ \begin{align*}\left( \theta(E/K_n) , \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) \right) = \left( \nu_{m, n} \left( \theta(E/K_{m}) \right) : 0 \leq m \leq n \right) .\end{align*} $$

Proof From the three term relation (2.2), we have

$$ \begin{align*}\nu_{n-1, n} \left( \pi_{n, n-1} \left( \theta(E/K_{n}) \right) \right) = a_p \cdot \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) - \nu_{n-2, n} \left( \theta(E/K_{n-2}) \right)\end{align*} $$

for $n \geq 2$ . Since $\nu _{n-1, n} \left ( \pi _{n, n-1} \left ( \theta (E/K_{n}) \right ) \right ) = f_n \cdot \theta (E/K_{n})$ for some $f_n \in \Lambda _n$ , we have

$$ \begin{align*}\nu_{n-2, n} \left( \theta(E/K_{n-2}) \right) \in \left( \theta(E/K_n) , \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) \right) \subseteq \Lambda_n.\end{align*} $$

In the same manner, we can obtain

$$ \begin{align*}\nu_{n-3, n-1} \left( \theta(E/K_{n-3}) \right) \in \left( \theta(E/K_{n-1}) , \nu_{n-2, n-1} \left( \theta(E/K_{n-2}) \right) \right) \subseteq \Lambda_{n-1}.\end{align*} $$

By taking $\nu _{n-1, n}$ , we have

$$ \begin{align*} \nu_{n-3, n} \left( \theta(E/K_{n-3}) \right) & \in \left( \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) , \nu_{n-2, n} \left( \theta(E/K_{n-2}) \right) \right) \\ & \subseteq \left( \theta(E/K_{n}), \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) \right) \\ & \subseteq \Lambda_{n}. \end{align*} $$

By applying this argument recursively, the conclusion follows.

Lemma 2.2 Under (Spl) and (Na), we have an equality of ideals of $\Lambda _n$

$$ \begin{align*}\left( \theta(E/K_n) , \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) \right) = \left( \theta(f_\alpha/K_{n}) \right) .\end{align*} $$

Proof By the definition of the p-stabilization (2.1), we have one inclusion $\supseteq $ . Hence, we focus on the opposite inclusion. By the interpolation formula of the anticyclotomic p-adic L-functions [Reference Chida and Hsieh5, Theorem A] under (Spl), we have the comparison of (the square-roots of) L-values

$$ \begin{align*}\theta(f_\alpha/K) = \left( 1 - \dfrac{1}{\alpha} \right) \cdot \theta(E/K).\end{align*} $$

Under (Na), we have equality in $\mathbb {Z}_p$

$$ \begin{align*}\left( 1 - \dfrac{1}{\alpha} \right)^{-1} \cdot \theta(f_\alpha/K) = \theta(E/K),\end{align*} $$

so we have $(\theta (E/K)) \subseteq (\theta (f_\alpha /K))$ . In fact, they are the same ideal. From (2.1) and (2.3), we have

$$ \begin{align*} \theta(f_\alpha/K_1) & = \dfrac{1}{\alpha} \cdot \left( \theta(E/K_1) - \dfrac{1}{\alpha} \cdot \nu_{0, 1}( \theta(E/K) ) \right) \\ & = \dfrac{1}{\alpha} \cdot \left( \theta(E/K_1) - \dfrac{1}{\alpha} \cdot \nu_{0, 1}( \left( 1 - \dfrac{1}{\alpha} \right)^{-1} \cdot \theta(f_\alpha/K) ) \right) \\ & = \dfrac{1}{\alpha} \cdot \left( \theta(E/K_1) - \dfrac{1}{\alpha} \cdot \left( 1 - \dfrac{1}{\alpha} \right)^{-1} \cdot \nu_{0, 1}( \pi_{1,0} ( \theta(f_\alpha/K_1)) ) \right) \\ & = \dfrac{1}{\alpha} \cdot \left( \theta(E/K_1) - \dfrac{1}{\alpha} \cdot \left( 1 - \dfrac{1}{\alpha} \right)^{-1} \cdot f_1 \cdot \theta(f_\alpha/K_1) \right) \end{align*} $$

for some $f_1 \in \Lambda _1$ . This shows that $\theta (E/K_{1}) = g_{1} \cdot \theta (f_\alpha /K_{1})$ for some $g_{1} \in \Lambda _{1}$ .

We suppose that $\theta (E/K_{n-1}) = g_{n-1} \cdot \theta (f_\alpha /K_{n-1})$ for some $g_{n-1} \in \Lambda _{n-1}$ .

$$ \begin{align*} \theta(f_\alpha/K_n) & = \dfrac{1}{\alpha^n} \cdot \left( \theta(E/K_n) - \dfrac{1}{\alpha} \cdot \nu_{n-1, n}( \theta(E/K_{n-1}) ) \right) \\ & = \dfrac{1}{\alpha^n} \cdot \left( \theta(E/K_{n}) - \dfrac{1}{\alpha} \cdot \nu_{n-1, n}( g_{n-1} \cdot \theta(f_\alpha/K_{n-1}) ) \right) \\ & = \dfrac{1}{\alpha^n} \cdot \left( \theta(E/K_n) - \dfrac{1}{\alpha} \cdot g_{n-1} \cdot \nu_{n-1, n}( \pi_{n,n-1} ( \theta(f_\alpha/K_n)) ) \right) \\ & = \dfrac{1}{\alpha^n} \cdot \left( \theta(E/K_1) - \dfrac{1}{\alpha} \cdot g_{n-1} \cdot f_{n} \cdot \theta(f_\alpha/K_n) \right) \end{align*} $$

for some $f_n \in \Lambda _n$ . This shows that $\theta (E/K_{n}) = g_{n} \cdot \theta (f_\alpha /K_{n})$ for some $g_{n} \in \Lambda _{n}$ . By induction, we have inclusion

$$ \begin{align*}(\theta(E/K_n)) \subseteq (\theta(f_\alpha/K_n)),\end{align*} $$

so we also have

$$ \begin{align*}( \nu_{n-1, n} \left( \theta(E/K_{n-1}) \right) ) \subseteq ( \nu_{n-1, n} \left( \theta(f_\alpha/K_{n-1}) \right)).\end{align*} $$

Since $\nu _{n-1, n} \left ( \theta (f_\alpha /K_{n-1}) \right ) = f_{n} \cdot \theta (f_\alpha /K_n)$ , we have

$$ \begin{align*}( \nu_{n-1, n} \left( \theta(f_\alpha/K_{n-1}) \right)) \subseteq ( \theta(f_\alpha/K_{n}) ).\end{align*} $$

The conclusion follows.

3 Comparison of Selmer groups

3.1 Local properties of Galois representations

Let $\rho : G_{\mathbb {Q}} \to \mathrm {Aut}_{\mathbb {Q}_p}(V) = \mathrm {GL}_2(\mathbb {Q}_p)$ be the two-dimensional Galois representation associated with E.

  • Since E is good ordinary at p, we have

    $$ \begin{align*}\rho \vert_{ G_{\mathbb{Q}_p} } \sim \begin{pmatrix} \chi^{-1}_{\alpha} \cdot \chi_{\mathrm{cyc}} & * \\ 0 & \chi_{\alpha} \end{pmatrix}\end{align*} $$
    where $\chi _{\alpha }$ is the unramified character sending the arithmetic Frobenius at p to  $\alpha $ .
  • For $\ell $ dividing N exactly, we also have

    $$ \begin{align*}\rho \vert_{ G_{\mathbb{Q}_\ell} } \sim \begin{pmatrix} \pm \chi_{\mathrm{cyc}} & * \\ 0 & \pm \mathbf{1} \end{pmatrix} .\end{align*} $$

For a rational prime v dividing $N^-p$ , we consider the following subspaces:

  • For $v= p$ , let $F^+V \subseteq V$ be the subspace on which the inertia subgroup $I_v$ acts by $\chi _{\mathrm {cyc}}$ .

  • For a rational prime v dividing $N^-$ , $F^+V \subseteq V$ be the subspace on which the inertia subgroup $I_v$ acts by $\chi _{\mathrm {cyc}}$ or $\chi _{\mathrm {cyc}} \tau _v$ where $\tau _v$ is the non-trivial unramified quadratic character of $G_{\mathbb {Q}_v}$ .

Let L be an algebraic extension of K. For a prime w of L dividing $N^-p$ , we define the ordinary local condition of V at w by

$$ \begin{align*}\mathrm{H}^1_{\mathrm{ord}}(L_w, V) = \mathrm{ker} \left( \mathrm{H}^1(L_w, V) \to \mathrm{H}^1(L_w, V/F^+V) \right).\end{align*} $$

Denote by $T = \varprojlim _k E[p^k]$ the p-adic Tate module of E, so we have $T \otimes _{\mathbb {Z}_p} \mathbb {Q}_p = V$ , and by $E[p^\infty ] = \varinjlim _k E[p^k]$ the p-power torsion points of E. Then the same local conditions for T, $T/p^kT$ , $E[p^\infty ]$ , and $E[p^k]$ are defined by propagation.

3.2 $N^-$ -ordinary (residual) Selmer groups

Let $\Sigma $ be the finite set of places of $\mathbb {Q}$ consisting of the places dividing $Np\infty $ , and $K_\Sigma $ be the maximal extension of K unramified outside $\Sigma $ . We write

  • $\Sigma ^+ \subseteq \Sigma $ to be the subset of $\Sigma $ consisting of the places not dividing $p\infty $ which split in $K/\mathbb {Q}$ , and

  • $\Sigma ^- \subseteq \Sigma $ to be the subset of $\Sigma $ consisting of the places not dividing $p\infty $ which are inert in $K/\mathbb {Q}$ .

For a place w of $K_\infty $ , we write $w \in \Sigma ^{\pm }$ if w divides a rational prime $\ell $ contained in $\Sigma ^{\pm }$ , respectively. For every $k \geq 1$ , we define the $N^-$ -ordinary (and $N^+$ -strict) Selmer group of $E[p^k]$ $\mathrm {Sel}_{N^-}(K_\infty , E[p^k])$ by the kernel of the map

$$ \begin{align*}\mathrm{H}^1(K_\Sigma/ K_\infty, E[p^k]) \to \prod_{w \nmid \Sigma^+} \mathrm{H}^1(K_{\infty, w}, E[p^k]) \times \prod_{w \in \Sigma^- \textrm{ or } w \vert p} \dfrac{\mathrm{H}^1(K_{\infty,w}, E[p^k])}{\mathrm{H}^1_{\mathrm{ord}}(K_{\infty,w}, E[p^k])},\end{align*} $$

and define $\mathrm {Sel}_{N^-}(K_\infty , E[p^\infty ]) = \varinjlim _k \mathrm {Sel}_{N^-}(K_\infty , E[p^k])$ . This is the Selmer group used in the bipartite Euler system argument [Reference Bertolini and Darmon2, Definition 2.8].

3.3 Minimal and Greenberg Selmer groups

We follow the convention of [Reference Pollack and Weston16, Section 3.1]. The minimal Selmer group $\mathrm {Sel}_{\mathrm {min}}(K_\infty , E[p^\infty ])$ of $E[p^\infty ]$ is defined by the kernel of the map

$$ \begin{align*}\mathrm{H}^1(K_\infty, E[p^\infty]) \to \prod_{w \nmid p} \mathrm{H}^1(K_{\infty, w}, E[p^\infty]) \times \prod_{w \vert p} \dfrac{\mathrm{H}^1(K_{\infty,w}, E[p^\infty])}{\mathrm{H}^1_{\mathrm{ord}}(K_{\infty,w}, E[p^\infty])},\end{align*} $$

and the Greenberg Selmer group $\mathrm {Sel}_{\mathrm {Gr}}(K_\infty , E[p^\infty ])$ of $E[p^\infty ]$ is defined by the kernel of the map

$$ \begin{align*}\mathrm{H}^1(K_\infty, E[p^\infty]) \to \prod_{w \nmid p} \mathrm{H}^1(I_{\infty, w}, E[p^\infty]) \times \prod_{w \vert p} \dfrac{\mathrm{H}^1(K_{\infty,w}, E[p^\infty])}{\mathrm{H}^1_{\mathrm{ord}}(K_{\infty,w}, E[p^\infty])},\end{align*} $$

where $I_{\infty , w}$ is the inertia subgroup of $G_{K_{\infty ,w}}$ . Under (Ram), $\overline {\rho }$ is ramified at every prime dividing N, so p does not divide any Tamagawa factors. Then by using [Reference Pollack and Weston16, Lemma 3.4], we have an isomorphism

(3.1) $$ \begin{align} \mathrm{Sel}_{\mathrm{min}}(K_\infty, E[p^\infty]) \simeq \mathrm{Sel}_{\mathrm{Gr}}(K_\infty, E[p^\infty]). \end{align} $$

3.4 The comparison

We recall the final displayed equation in the proof of [Reference Pollack and Weston16, Proposition 3.6]:

where w runs over the primes of $K_\infty $ dividing $N^+$ . The local conditions at primes dividing $N^-$ of minimal Selmer groups and $N^-$ -ordinary Selmer groups coincide since such primes split completely in $K_\infty /K$ . Thus, we have inclusion

$$ \begin{align*}\mathrm{Sel}_{N^-}(K_\infty, E[p^k]) \subseteq \mathrm{Sel}_{\mathrm{min}}(K_\infty, E[p^\infty])[p^k]\end{align*} $$

which is of finite index and is independent of k.

Proposition 3.1 If $\mathrm {Sel}_{N^-}(K_\infty , E[p^\infty ])$ is $\Lambda $ -cotorsion with vanishing of $\mu $ -invariant, then $\mathrm {Sel}_{\mathrm {min}}(K_\infty , E[p^\infty ])$ is also $\Lambda $ -cotorsion with vanishing of $\mu $ -invariant.

Proof We have $\mathrm {Sel}_{N^-}(K_\infty , E[p^\infty ])[p] = \mathrm {Sel}_{N^-}(K_\infty , E[p])$ since $N^-$ -ordinary Selmer groups of $E[p^\infty ]$ are defined as the injective limit of $N^-$ -ordinary Selmer groups of $E[p^k]$ . By the assumption, $\mathrm {Sel}_{N^-}(K_\infty , E[p])$ is finite as noted in the proof of [Reference Kim, Pollack and Weston13, Corollary 2.3]. Since the inclusion $\mathrm {Sel}_{N^-}(K_\infty , E[p]) \subseteq \mathrm {Sel}_{\mathrm {min}}(K_\infty , E[p^\infty ])[p]$ is of finite index, $\mathrm {Sel}_{\mathrm {min}}(K_\infty , E[p^\infty ])[p]$ is also finite. By the same reasoning, the conclusion follows.

Proposition 3.2 Under (Im), if $\mathrm {Sel}_{\mathrm {min}}(K_\infty , E[p^\infty ])$ is $\Lambda $ -cotorsion, then $\mathrm {Sel}_{\mathrm {min}}(K_\infty , E[p^\infty ])$ has no proper $\Lambda $ -submodule of finite index. Thus, we have

$$ \begin{align*} \mathrm{char}_{\Lambda} \mathrm{Sel}_{\mathrm{min}}(K_\infty, E[p^\infty]) & = \mathrm{Fitt}_{\Lambda} \mathrm{Sel}_{\mathrm{min}}(K_\infty, E[p^\infty]) ,\\ \mathrm{Sel}_{\mathrm{min}}(K_\infty, E[p^\infty]) & \simeq \mathrm{Sel}_{N^-}(K_\infty, E[p^\infty]). \end{align*} $$

Proof This follows from [Reference Greenberg and Viola7, Proposition 4.14], which covers the cyclotomic case actually, but the argument generalizes to our setting as mentioned in the proof of [Reference Pollack and Weston16, Proposition 3.6].

The following corollary follows from (3.1) and the above two propositions.

Corollary 3.3 Under (Im) and (Ram), if $\mathrm {Sel}_{N^-}(K_\infty , E[p^\infty ])$ is $\Lambda $ -cotorsion with vanishing of $\mu $ -invariant, we have isomorphisms

$$ \begin{align*}\mathrm{Sel}_{N^-}(K_\infty, E[p^\infty]) \simeq \mathrm{Sel}_{\mathrm{min}}(K_\infty, E[p^\infty]) \simeq \mathrm{Sel}_{\mathrm{Gr}}(K_\infty, E[p^\infty]).\end{align*} $$

4 The proof of the main theorem via Iwasawa theory

We first gather some tools from Iwasawa theory and give a proof of Theorem 1.1.

4.1 Iwasawa theory

The anticyclotomic main conjecture for $(E,p, K)$ is now completely known for our setting.

Theorem 4.1 Under (Im),(Ram),(Spl),(Na), and (Def), we have the following statements.

  1. (1) $L_p(E/K_\infty )$ is non-zero.

  2. (2) $\mu (L_p(E/K_\infty ))= 0 $ .

  3. (3) $\mathrm {Sel}_{N^-}(K_\infty , E[p^\infty ])$ is $\Lambda $ -cotorsion with vanishing of $\mu $ -invariants.

  4. (4) $\left ( L_p(E/K_\infty ) \right ) = \mathrm {char}_\Lambda \left ( \mathrm {Sel}(K_\infty , E[p^\infty ])^\vee \right )$ .

Proof

  1. (1) It is proved in [Reference Vatsal18].

  2. (2) It is proved in [Reference Vatsal19].

  3. (3) This follows from (1), (2), and the Euler system divisibility

    $$ \begin{align*}\left( L_p(E/K_\infty) \right) \subseteq \mathrm{char}_\Lambda \left( \mathrm{Sel}_{N^-}(K_\infty, E[p^\infty])^\vee \right)\end{align*} $$
    obtained from the bipartite Euler system argument [Reference Bertolini and Darmon2, Reference Pollack and Weston16]. Condition (Na) is implicitly used in the Euler system argument. See [Reference Kim, Pollack and Weston13, Assumption 1.1 and Remark 1.4] for this issue.
  4. (4) By using (3) and Corollary 3.3, we can identify $\mathrm {Sel}_{N^-}(K_\infty , E[p^\infty ])$ with the minimal Selmer group. By [Reference Greenberg and Viola7, Proposition 2.1], the minimal Selmer group also coincides with the classical Selmer group. The opposite divisibility

    $$ \begin{align*}\left( L_p(E/K_\infty) \right) \supseteq \mathrm{char}_\Lambda \left( \mathrm{Sel}(K_\infty, E[p^\infty])^\vee \right)\end{align*} $$
    follows from [Reference Skinner and Urban17]. Condition (Spl) is needed only for this last statement to invoke [Reference Skinner and Urban17].

Corollary 4.2 Under (Im),(Ram),(Na), and (Def), the classical Selmer group $\mathrm {Sel}(K_\infty , E[p^\infty ])$ has no proper $\Lambda $ -submodule of finite index; thus,

$$ \begin{align*}\mathrm{char}_{\Lambda} \mathrm{Sel}(K_\infty, E[p^\infty]) = \mathrm{Fitt}_{\Lambda} \mathrm{Sel}(K_\infty, E[p^\infty]) .\end{align*} $$

Proof By Theorem 4.1.(3), $\mathrm {Sel}_{N^-}(K_\infty , E[p^\infty ])$ is $\Lambda $ -cotorsion. By Proposition 3.1, $\mathrm {Sel}_{\mathrm {min}}(K_\infty , E[p^\infty ])$ is also $\Lambda $ -cotorsion. The conclusion follows from Proposition 3.2 and the identification of the minimal Selmer group and the classical Selmer group.

We recall the control theorem.

Proposition 4.3 (Control theorem)

Let . Under (Im), (Na), and (Def), the restriction map

$$ \begin{align*}\mathrm{Sel}(K_n, E[p^\infty]) \to \mathrm{Sel}(K_\infty, E[p^\infty])[\omega_n]\end{align*} $$

is injective with the finite cokernel whose size is bounded independently of n. If we further assume (Ram), then it is an isomorphism.

Proof See [Reference Chida and Hsieh4, Proposition 1.9] with the identifications of Selmer groups in Corollaries 3.3 and 4.2.

4.2 The proof of Theorem 1.1

By the anticyclotomic main conjecture (Theorem 4.1), we have

$$ \begin{align*}\left( L_p(E/K_\infty) \right) = \mathrm{char}_{\Lambda} \left( \mathrm{Sel}(K_\infty, E[p^\infty])^\vee \right) .\end{align*} $$

By Corollary 4.2, the above equality becomes

$$ \begin{align*}\left( L_p(E/K_\infty) \right) = \mathrm{Fitt}_{\Lambda} \left( \mathrm{Sel}(K_\infty, E[p^\infty])^\vee \right)\end{align*} $$

Under the quotient map $\Lambda \to \Lambda _n = \Lambda / \omega _n$ , it becomes

$$ \begin{align*}\left( \left( \theta(f_\alpha/K_n) \cdot \iota( \theta(f_\alpha/K_n) ) \right) \right) = \mathrm{Fitt}_{\Lambda_n} \left( \left( \mathrm{Sel}(K_\infty, E[p^\infty]) [\omega_n]\right)^\vee \right)\end{align*} $$

since Fitting ideals are compatible with base change. By using the functional equation of theta elements (2.4) and the control theorem (Proposition 4.3), we have

$$ \begin{align*}\left( \theta(f_\alpha/K_n) \right)^2 = \mathrm{Fitt}_{\Lambda_n} \left( \mathrm{Sel}(K_n, E[p^\infty])^\vee \right) .\end{align*} $$

Theorem 1.1 now follows from Lemma 2.2, and the ideal is principal thanks to the above equality.

Acknowledgement

I would like to thank Rob Pollack and Masato Kurihara very much. Rob Pollack first suggested me to think about this problem when I was a Ph.D. student and I was able to solve it after having the discussions with Masato Kurihara. In particular, I learned the idea of the proof of Lemma 2.2 on the p-destabilization process from the discussion with Masato Kurihara at his home, January 3, 2017. The main result of this article partially refines that of [Reference Kim8]. We are grateful to thank the referee for pointing out inaccuracies and unrefined writings. The explanation is greatly improved in the revised version following the referee’s suggestions.

Funding

C.-H.K. was partially supported by research funds for newly appointed professors of Jeonbuk National University in 2024, and by Global-Learning & Academic research institution for Master’s Ph.D. Students, and Postdocs (LAMP) Program of the National Research Foundation of Korea (NRF) funded by the Ministry of Education (Grant No. RS-2024-00443714).

References

Bertolini, M. and Darmon, H., Heegner points on Mumford–Tate curves . Invent. Math. 126(1996), no. 3, 413456.Google Scholar
Bertolini, M. and Darmon, H., Iwasawa’s main conjectures for elliptic curves over anticyclotomic ${\mathbb{Z}}_p$ -extensions . Ann. Math. (2) 162(2005), no. 1, 164.Google Scholar
Breuil, C., Conrad, B., Diamond, F., and Taylor, R., On the modularity of elliptic curves over $\mathbb{Q}$ : wild 3-adic exercises . J. Amer. Math. Soc. 14(2001), no. 4, 843939.Google Scholar
Chida, M. and Hsieh, M.-L., On the anticyclotomic Iwasawa main conjecture for modular forms . Compos. Math. 151(2015), no. 5, 863897.Google Scholar
Chida, M. and Hsieh, M.-L., Special values of anticyclotomic $L$ -functions for modular forms . J. Reine Angew. Math. 741(2018), 87131.Google Scholar
Darmon, H. and Iovita, A., The anticyclotomic main conjecture for elliptic curves at supersingular primes . J. Inst. Math. Jussieu 7(2008), no. 2, 291325.Google Scholar
Greenberg, R., Iwasawa theory for elliptic curves . In: Viola, C. (ed.), Arithmetic theory of elliptic curves (Cetraro, 1997), volume 1716 of Lecture Notes in Mathematics, Springer-Verlag, Berlin, 1999, pp. 51144. Centro Internazionale Matematico Estivo (C.I.M.E.), Florence, Lectures from the 3rd C.I.M.E. Session held in Cetraro, July 12–19, 1997.Google Scholar
Kim, C.-H., An anticyclotomic Mazur–Tate conjecture for modular forms. Preprint, unpublished, arXiv:1612.03743.Google Scholar
Kim, C.-H., Anticyclotomic Iwasawa invariants and congruences of modular forms . Asian J. Math. 21(2017), no. 3, 499530.Google Scholar
Kim, C.-H., Overconvergent quaternionic forms and anticyclotomic $p$ -adic $L$ -functions . Publ. Mat. 63(2019), no. 2, 727767.Google Scholar
Kim, C.-H. and Kurihara, M., On the refined conjectures on Fitting ideals of Selmer groups of elliptic curves with supersingular reduction . Int. Math. Res. Not. IMRN 2021(2021), no. 14, 1055910599.Google Scholar
Kim, C.-H. and Ota, K., On the quantitative variation of congruence ideals and integral periods of modular forms . Res. Math. Sci. 10(2023), no. 22, 34.Google Scholar
Kim, C.-H., Pollack, R., and Weston, T., On the freeness of anticyclotomic Selmer groups of modular forms . Int. J. Number Theory 13(2017), no. 6, 14431455.Google Scholar
Kurihara, M., On the Tate Shafarevich groups over cyclotomic fields of an elliptic curve with supersingular reduction I . Invent. Math. 149(2002), 195224.Google Scholar
Mazur, B. and Tate, J., Refined conjectures of the “Birch and Swinnerton-Dyer type” . Duke Math. J. 54(1987), no. 2, 711750.Google Scholar
Pollack, R. and Weston, T., On anticyclotomic $\mu$ -invariants of modular forms . Compos. Math. 147(2011), 13531381.Google Scholar
Skinner, C. and Urban, E., The Iwasawa main conjectures for ${GL}_2$ . Invent. Math. 195(2014), no. 1, 1277.Google Scholar
Vatsal, V., Uniform distribution of Heegner points . Invent. Math. 148(2002), no. 1, 148.Google Scholar
Vatsal, V., Special values of anticyclotomic $L$ -functions . Duke Math. J. 116(2003), no. 2, 219261.Google Scholar
Vignéras, M.-F., Arithmétique des algèbres de quaternions, volume 800 of Lecture Notes in Mathematics, Springer, Berlin, 1980.Google Scholar