Supersingular primes and Bogomolov property

Soumyadip Sahu · 2025

Read Supersingular primes and Bogomolov property by Soumyadip Sahu (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2504-13498.txt.

Introduction

Soumyadip Sahu's 2025 paper proves a discreteness property for the fields built from an elliptic curve's torsion points. The Bogomolov property says, roughly, that points of small height cannot crowd together; only finitely many sit below any positive bound. Sahu works with an elliptic curve over a number field that has at least one real embedding, and adjoins all its torsion points over a finite extension.

The result extends earlier work of Habegger. The enlarged field has the property. The curve's canonical height on points over that field obeys a matching discreteness law. The engine of the proof is a general criterion due to Plessis, which reduces everything to finding one supersingular prime with the right features. The paper is a compact number theory argument aimed at specialists in the field.

Catalog record

CREATORSoumyadip Sahu
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-13498

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Supersingular primes and Bogomolov property Soumyadip Sahu Abstract Let E E be an elliptic curve over a number field K K with at least one real embedding and L L be a finite extension of K K . We generalize a result of Habegger to show that L ⁡ ( E tor ) L(E_{\text{tor}}) , the field generated by the torsion points of E E over L L , has the Bogomolov property. Moreover, the Néron-Tate height on E ⁡ ( L ⁡ ( E tor ) ) E\big(L(E_{\text{tor}})\big) also satisfies a similar discreteness property. Our main tool is a general criterion of Plessis that reduces the problem to the existence of a supersingular prime for E E satisfying certain conditions. MSC2020: Primary 11G50, Secondary 11G05 Keywords: Bogomolov property, supersingular primes 1 Statement of the result The absolute logarithmic Weil height (abbreviated as height ) of an algebraic number is a nonnegative real number [ 5 ] that measures its arithmetic complexity. By Northcott’s theorem, there are only finitely many algebraic numbers with bounded degrees and bounded heights. Kronecker’s theorem tells us that the height of an algebraic number is zero if and only if it is zero or a root of unity. As a consequence, given a number field, there exists \epsilon>0 so that the height of each element of the field, except zero and roots of unity, is larger than \epsilon . This phenomenon inspires the following definition. We say that an algebraic extension L/\mathbb{Q} has the Bogomolov property if there exists \epsilon>0 such that each element of L^{\times} having height <\epsilon is a root of unity. The property’s name is motivated by a similar conjecture of Bogomolov regarding points of small height on algebraic curves of genus \geq 2 . In recent years, many number theorists have investigated the Bogomolov property for a diverse class of infinite algebraic extensions arising from arithmetic and geometric sources [ 1 ] . For example, Amoroso and Zannier [ 3 ] , building upon earlier work of Amoroso and Dvornicich, showed that K^{\text{ab}} , the maximal abelian extension of a number field K , has the Bogomolov property. One can also generalize the notion of Bogomolov property to the elliptic curves by replacing Weil height with the Néron-Tate height. In more detail, let E be an elliptic curve defined over a number field K . Given an algebraic extension L/K , we say that the group E(L) has the Bogomolov property if there exists \epsilon>0 so that each point on E(L) having Néron-Tate height <\epsilon is a torsion point. Several authors have studied the elliptic Bogomolov property to illustrate the similarities between the arithmetic of the torus and the elliptic curve. For instance, Silverman [ 14 ] showed that if E is an elliptic curve over a number field K , then E(K^{\text{ab}}) has the Bogomolov property. Our point of interest is a striking result of Habegger [ 9 ] which proves both the toric and elliptic Bogomolov property for the field generated by the torsion points of an elliptic curve. We fix an algebraic closure \overline{\mathbb{Q}} of \mathbb{Q} and work with the subfields of \overline{\mathbb{Q}} . Set \mu_{\infty} to be the collection of roots of unity in \overline{\mathbb{Q}} . For an elliptic curve E over a number field K , let E_{\text{tor}} denote the torsion subgroup of E(\overline{\mathbb{Q}}) . Now suppose L/K is an algebraic extension. Then L(E_{\text{tor}}) is the field generated over L by the coordinates of the torsion points of E for a K -rational Weierstrass model. This field does not depend on the choice of the Weierstrass model. Theorem 1.1 . (Habegger) Let E be an elliptic curve defined over \mathbb{Q} . Then \mathbb{Q}(E_{\emph{tor}}) has the Bogomolov property. Moreover, the group E\big(\mathbb{Q}(E_{\emph{tor}})\big) possesses the elliptic Bogomolov property. Amoroso and Terracini [ 2 ] explored the possibility of formulating Theorem 1.1 for the infinite extensions attached to a compatible family of \ell -adic Galois representations. In particular, one can ask if the following generalization of Habbeger’s result holds: Conjecture . Let E be an elliptic curve over a number field K . Then K(E_{\text{tor}}) and E(K(E_{\text{tor}})) have the Bogomolov property. Note that if L/K is a finite extension, then one can view E as an elliptic curve over L to obtain a statement for L(E_{\text{tor}}) from this conjecture. Therefore, the conjecture above is equivalent to the following stronger version: Conjecture 1.2 . Let E be an elliptic curve over a number field K and L/K be a finite extension. Then L(E_{\text{tor}}) and E(L(E_{\text{tor}})) have the Bogomolov property. If L is an algebraic extension with the Bogomolov property and F/L is a finite extension, then F need not have the Bogomolov property; see [ 1 ] . For an elliptic curve with complex multiplication, Conjecture 1.2 is an easy consequence of the aforementioned results of Amoroso-Zannier and Silverman regarding the maximal abelian extension (Proposition 2.2 ). The main point of Habegger’s work is to deal with the non-CM case by careful use of Galois theory and equidistribution theorems. In a subsequent development, Frey [ 8 ] generalized Habegger’s arguments to establish the field extension part of Conjecture 1.2 for the elliptic curves over \mathbb{Q} . Her work also provides an explicit lower bound for the Weil height in terms of a suitable supersingular prime of the elliptic curve. More recently, Plessis [ 10 ] improved Frey’s work to derive a very general criterion for the Bogomolov property of the field generated by torsion points of an elliptic curve over a number field which plays a key role in our work. Let L/K be a finite extension of number fields. Given a finite place v of K , we write e_{w}(L|K) , resp. f_{w}(L|K) , for the ramification index, resp. residue degree, of a place w of L with w\mid v . Moreover, set d_{v}(L)=\max_{w\mid v}[L_{w}:K_{v}]. We use h(\cdot) to denote the Weil height of an algebraic number. Theorem 1.3 . (Plessis) Let E be an elliptic curve defined over a number field K . Suppose that L/K is a finite Galois extension. Now suppose v is a place of K lying over a rational prime p so that (i) E has a supersingular reduction at v and j_{E}\not\equiv 0,1728(\emph{mod}v) , (ii) p>\max\{3,2d_{v}(L)\} and the image of the canonical Galois representation (1.1) \emph{Gal}\big(L(E[p])/L\big)\to\emph{GL}_{2}(\mathbb{Z}/p\mathbb{Z}) contains \emph{SL}_{2}(\mathbb{Z}/p\mathbb{Z}) , (iii) e_{v}(K\mid\mathbb{Q})=1 and f_{v}(K\mid\mathbb{Q})\leq 2 . Then, for all \alpha\in L(E_{\emph{tor}})^{\times}\backslash\mu_{\infty} , we have h(\alpha)\geq\frac{1}{4^{p^{2}d_{v}(L)}+1}\Big(\frac{\log p}{d_{v}(L)(40\sqrt{2}+2)[K:\mathbb{Q}]p^{2p^{2}d_{v}(L)+2}}\Big)^{2+\frac{4}{p^{p^{2}d_{v}(L)/4}-2}}. Moreover, if (iv) v is unramified in L , (v) the representation in (ii) is surjective, then E\big(L(E_{\emph{tor}})\big) has the Bogomolov property. An obvious obstruction to directly applying this theorem is to find a supersingular prime satisfying the desiderata. Nevertheless, Elkies [ 7 ] proved that if the base field K has at least one real embedding, then E is supersingular at infinitely many primes of K . Plessis uses this result to generalize Theorem 1.1 to the elliptic curves over a real quadratic field. The goal of this short note is to demonstrate that the third condition in Theorem 1.3 poses no difficulty at all if we study the problem for an elliptic curve over its minimal field of definition, namely \mathbb{Q}(j_{E}) , where j_{E} is the

Read the complete text