A refined non-vanishing of the p p -adic logarithm of a rational point on an abelian variety
Ashay Burungale, Christopher Skinner, Xin Wan · 2026
Read A refined non-vanishing of the p p -adic logarithm of a rational point on an abelian variety by Ashay Burungale, Christopher Skinner, Xin Wan (2026) 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-2603-20886.txt.
Introduction
Burungale, Skinner, and Wan prove refined non-vanishing results for p-adic logarithms of rational points on abelian varieties. Their work is inspired by a formula due to Bertolini, Darmon, and Prasanna. The main cases involve abelian varieties of GL2 type. These are tied to Hilbert modular newforms and Heegner points. The key tool is the p-adic analytic subgroup theorem. The results address when non-torsion points stay nonzero under the p-adic logarithm map.
Catalog record
| CREATOR | Ashay Burungale, Christopher Skinner, Xin Wan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 50248 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.20886
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
A refined non-vanishing of the p p -adic logarithm of a rational point on an abelian variety Ashay Burungale, Christopher Skinner, Xin Wan Abstract . Inspired by a beautiful formula of Bertolini, Darmon, and Prasanna – the oft-termed BDP formula – we address questions about the non-vanishing of non-torsion points under p p -adic logarithms of abelian varieties. We largely consider situations most applicable to GL 2 {\mathrm{GL}}_{2} -type abelian varieties associated with Hilbert modular newforms and Heegner points. Not surprisingly, the main tool employed is the p p -adic analytic subgroup theorem. 1. Introduction 1.1. The BDP formula In [ 3 ] Darmon, in joint work with Bertolini and Prasanna, proved a beautiful formula that expresses a special value of a certain p -adic L -function of a modular newform as (the square of) the image under a p -adic Abel–Jacobi map of a generalized Heegner cycle. In the special case that the newform f has weight 2 , level N , and trivial Nebentypus, this formula takes the shape: L_{p}(f,\mathbf{N}_{K})=(1-a_{p}(f)+p^{-1})^{2}(\log_{\omega_{B_{f}}}(P_{f}))^{2} (1.1) (cf. [ 4 , Thm. 3.12] ). This has come to be known as the BDP formula . Here, p\nmid 2N is a prime, K/\mathbb{Q} is an imaginary quadratic field in which all primes dividing Np split (so K satisfies a ‘Heegner hypothesis’ relative to f ), and L_{p}(f,-) is a continuous function of a p -adic space of ( p -adic) Hecke characters over K that takes values in {\mathbb{C}}_{p} . The function L_{p}(f,-) interpolates the algebraic parts of special L -values L(f,\chi^{-1},0) for \chi in a particular set of Hecke characters over K . The norm character \mathbf{N}_{K} belongs to the p -adic space of characters but does not belong to the set with interpolated special L -values. On the other side of the formula, B_{f} is an abelian variety over \mathbb{Q} in the isogeny class associated with f that is realized as a quotient \Phi_{f}:J_{1}(N)\twoheadrightarrow B_{f} and P_{f}\in B_{f}(K) is the Heegner point. The differential \omega_{B_{f}} is the unique differential in \Omega^{1}_{B_{f}}\otimes_{{\mathbb{Q}}}\overline{{\mathbb{Q}}} that pulls back under \Phi_{f} to the differential \omega_{f}=2\pi if(\tau)d\tau in \Omega^{1}_{J_{1}(N)}\otimes_{{\mathbb{Q}}}\overline{{\mathbb{Q}}}=\Omega^{1}_{X_{1}(N)}\otimes_{{\mathbb{Q}}}\overline{{\mathbb{Q}}} . The logarithm \log_{\omega_{B_{f}}} is the p -adic logarithm associated with the differential. Hidden in all this is the choice of an embedding \iota_{p}:\overline{{\mathbb{Q}}}\hookrightarrow\overline{{\mathbb{Q}}}_{p} via which the algebraic parts of the special L -values are viewed as belonging to {\mathbb{C}}_{p} and via which the differential \omega_{f} is viewed as belonging to \Omega^{1}_{B_{f}}\otimes_{{\mathbb{Q}}}\overline{{\mathbb{Q}}}_{p} ; there is a p -adic L -function and a formula for each choice of \iota_{p} . The BDP formula ( 1.1 ) was inspired by an earlier, similar formula of Rubin for CM elliptic curves [ 23 ] , and in [ 4 ] Darmon and his collaborators gave a new proof Rubin’s formula based on ( 1.1 ). As was beautifully explained in [ 1 ] , this formula can also be seen as a special case of Perrin-Riou’s p -adic Beilinson conjectures and is closely analogous to p -adic regulator formulas for the p -adic L -functions of Kubota–Leopoldt and Katz, which involve p -adic logarithms of circular and elliptic units, respectively. The BDP formula has also played a central role in progress toward the Birch–Swinnerton-Dyer conjecture for elliptic curves, especially in conjunction with the Gross–Zagier formula, the methods of Kolyvagin, and Iwasawa theory (see in particular [ 24 , 2 , 26 , 15 , 6 , 7 , 8 , 9 ] ). The question addressed in this note arises naturally from the BDP formula and wish to extend the results for elliptic curve to the abelian varieties associated with a weight 2 newform. 1.2. The question Suppose the Heegner point P_{f}\in B_{f}(K) is non-torsion, which happens if and only if the L -function L(f/K,s) has a zero of order one at s=1 by the Gross–Zagier formula [ 14 , 27 ] . Then it is natural to ask: Is its p -adic logarithm \log_{\omega_{f}}(P_{f}) non-zero? Indeed, there are good reasons coming from Iwasawa theory to expect that this is so. For example, the p -adic height of P_{f} appearing in Kobayashi’s p -adic Gross–Zagier formula [ 16 , 5 ] (for the case that f is non-ordinary at p ) can only be non-zero if \log_{\omega_{f}} is. Other reasons arise from considering the Iwasawa-theoretic main conjecture associated with p -adic L -function L_{p}(f,-) appearing in the BDP formula ( 1.1 ), in which case the answer ‘yes’ can be connected to the truth of the expected BSD formula for the derivative at s=1 of the complex L -function L(f,s) (see [ 24 , 15 ] ). Moreover, in the case that B_{f} is an elliptic curve E , the answer is clearly yes: E(K) injects into E(\overline{{\mathbb{Q}}}_{p}) and the kernel of the p -adic logarithm \log_{\omega_{E}}:E(\overline{{\mathbb{Q}}}_{p})\rightarrow\overline{{\mathbb{Q}}}_{p} is just the torsion points. The subtleties appear when B_{f} is not an elliptic curve, as we will explain below. In [ 24 , Lem. 2.2.2] , one of us (C.S.) implicity asserted a positive answer to this question (this has no impact on the main results of [ 24 ] ). Another of us (X.W.) later pointed out that no proof was given. In this note we supply a proof that the answer to this question is ‘yes.’ Perhaps not surprisingly, our proof is based on results from p -adic transcendence theory – the p -adic analytic subgroup theorem, in particular. 1.3. The case of general B_{f} To simplify notation, we will now write A for B_{f} . Let \iota_{\infty}:\overline{{\mathbb{Q}}}\hookrightarrow{\mathbb{C}} be a fixed embedding. Then via \iota_{\infty} , the Fourier coefficients of f generate a totally real number field F\subset\overline{{\mathbb{Q}}} , the dimension of A equals the degree [F:{\mathbb{Q}}] of F , and the endomorphism ring {\mathrm{End}}_{\mathbb{Q}}(A) contains an order {\mathcal{O}} of F . Let t_{A} be the tangent space of A and recall that the space of differentials \Omega^{1}_{A} is canonically identified with the dual of t_{A} . In particular, \Omega=\Omega^{1}_{A}\otimes_{{\mathbb{Q}}}\overline{{\mathbb{Q}}} is the \overline{{\mathbb{Q}}} -dual of V=t_{A}\otimes_{{\mathbb{Q}}}\overline{{\mathbb{Q}}} ; given \omega\in\Omega we write \ell_{\omega}:V\rightarrow\overline{{\mathbb{Q}}} for the corresponding linear map. The action of {\mathcal{O}} on A induces an action of F on \Omega^{1}_{A} and hence an action of F\otimes_{{\mathbb{Q}}}\overline{{\mathbb{Q}}} on \Omega . Let \Sigma_{F}=\{\sigma:F\hookrightarrow\overline{{\mathbb{Q}}}\} be the set of embeddings of F into \overline{{\mathbb{Q}}} . The decomposition F\otimes_{{\mathbb{Q}}}\overline{{\mathbb{Q}}}=\oplus_{\sigma\in\Sigma_{F}}\overline{{\mathbb{Q}}}e_{\sigma} , where e_{\sigma}^{2}=e_{\sigma} and the action of F on e_{\sigma} is via the embedding \sigma , induces a decomposition \Omega=\oplus_{\sigma\in\Sigma_{F}}\Omega_{\sigma}, with \Omega_{\sigma}=e_{\sigma}\Omega . For each \sigma\in\Sigma_{F} , let f^{\sigma} be the newform whose Fourier coefficients are obtained from f by applying \sigma . Then \omega_{f^{\sigma}} is a \overline{{\mathbb{Q}}} -basis for \Omega_{\sigma} . Let p be a prime and let \iota_{p}:\overline{{\mathbb{Q}}}\hookrightarrow\overline{{\mathbb{Q}}}_{p} be a fixed embedding. Let V_{p}=t_{A}\otimes_{\overline{{\mathbb{Q}}}}\overline{{\mathbb{Q}}}_{p} . The p -adic logarithm of A is a locally analytic homomorphism \log_{A}:A(\overline{{\mathbb{Q}}}_{p})\rightarrow V_{p} whose kernel is the group of torsion points. The logarithm is {\mathcal{O}} -invariant. The embedding \iota_{p} induces an inclusion A(\overline{{\mathbb{Q}}})\hookrightarrow A(\overline{{\mathbb{Q}}}_{p}) as well as an identification V_{p}=V\otimes_{\overline{{\mathbb{Q}}}}\overline{{\mathbb{Q}}}_{p} . In particular, each \ell_{\omega} , \omega\in\Omega extends to a \overline{{\mathbb{Q}}}_{p} -linear map \ell_{\omega}:V_{p}\rightarrow\overline{{\mathbb{Q}}}_{p} . We set \log_{\omega}:A(\overline{{\mathbb{Q}}}_{p})\rightarrow\overline{{\mathbb{Q}}}_{p} to be the composition \log_{\omega}=\ell_{\omega}\circ\log_{A}. If \omega=\omega_{f^{\sigma}} , then this is the p -adic logarithm appearing in the BDP formula. Note that the definition of \log_{\omega} depends on the choice of \iota_{p} . Replacing