For an elliptic curve E/QE/\mathbb{Q} of good ordinary reduction at an odd prime pp that splits in an imaginary quadratic field KK satisfying the Heegner hypothesis, Agboola–Castella formulated Birch–Swinnerton-Dyer-type conjectures for the two anticyclotomic pp-adic LL-functions Lp(f)=Lp(f)2L_{\mathfrak p}(f)=\mathscr L_{\mathfrak p}(f)^{2} and Lp(f)\mathscr L_{\mathfrak p}(f) attached to EE: the order of vanishing along the anticyclotomic Zp\mathbb{Z}_p-extension should be 2(max⁡{r+,r−}−1)2(\max\{r^+,r^-\}-1) and max⁡{r+,r−}−1\max\{r^+,r^-\}-1 respectively, where r±r^\pm are the ranks of the ±\pm-eigenspaces of E(K)E(K), with leading terms expressed through derived anticyclotomic pp-adic regulators, the order of the Tate–Shafarevich group, and Tamagawa factors. The source and later work prove only one inequality and the leading coefficient up to a pp-adic unit, conditionally on the Iwasawa–Greenberg main conjecture. This problem asks for a proof of the two conjectures in their exact form; their derived regulators are defined in the source under the maximal non-degeneracy conjecture of Mazur–Bertolini–Darmon for the derived pp-adic height filtration, which is itself open. As of 2026 the exact equalities, the rank-part equality, and the non-degeneracy conjecture remain open.

Background

Let E/QE/\mathbb{Q} be an elliptic curve of conductor NN with associated newform ff (the weight-2 cuspidal modular form attached to EE by the modularity theorem), let pp be an odd prime of good ordinary reduction for EE (ordinary means p∤ap(E)p\nmid a_p(E), where ap(E)=p+1−#E(Fp)a_p(E)=p+1-\#E(\mathbb{F}_p)), and let KK be an imaginary quadratic field. The Heegner hypothesis (Heeg)(\mathrm{Heeg}) requires that every prime factor of NN splits in KK; additionally (spl)(\mathrm{spl}) requires that p=ppˉp=\mathfrak p\bar{\mathfrak p} splits in KK. The anticyclotomic Zp\mathbb{Z}_p-extension K∞/KK_\infty/K is the unique Zp\mathbb{Z}_p-extension of KK — a Galois extension whose Galois group is isomorphic to the additive group of pp-adic integers Zp\mathbb{Z}_p — on which complex conjugation (the nontrivial automorphism of K/QK/\mathbb{Q}) acts by −1-1; one writes Γ=Gal(K∞/K)\Gamma=\mathrm{Gal}(K_\infty/K), Λ=Zp[[Γ]]≃Zp[[T]]\Lambda=\mathbb{Z}_p[[\Gamma]]\simeq\mathbb{Z}_p[[T]] for the Iwasawa algebra (the completed group ring of Γ\Gamma), and J=(T)J=(T) for its augmentation ideal; the source works over ΛO^:=Λ⊗^ZpO^\Lambda_{\hat{\mathcal O}}:=\Lambda\hat\otimes_{\mathbb{Z}_p}\hat{\mathcal O}, with O^\hat{\mathcal O} the completion of the ring of integers of the maximal unramified extension of Qp\mathbb{Q}_p, and denotes by JJ the augmentation ideal of either ring. Let r±r^\pm denote the rank of the ±\pm-eigenspace E(K)±E(K)^\pm under complex conjugation, so rank E(K)=r++r−\mathrm{rank}\,E(K)=r^++r^-. A Heegner point ypy_{\mathfrak p} on EE is the trace over its Galois orbit of the image of a CM point (a point with complex multiplication) on the modular curve X0(N)X_0(N), which classifies cyclic subgroups of order NN on elliptic curves, under a fixed modular parametrization X0(N)→EX_0(N)\to E; write ωE\omega_E for the Néron differential of EE (the invariant differential on the Néron model, normalized over Z\mathbb{Z}) and log⁡ωE\log_{\omega_E} for the formal pp-adic logarithm on the formal group of EE at p\mathfrak p, and let tt denote the Heegner-point index factor, measuring the failure of log⁡ωE(yp)\log_{\omega_E}(y_{\mathfrak p}) to generate the local logarithm lattice of E(K)+⊗ZpE(K)^+\otimes\mathbb{Z}_p. Bertolini–Darmon–Prasanna, as strengthened by Castella–Hsieh, constructed the anticyclotomic pp-adic LL-function Lp(f)∈ΛO^\mathscr L_{\mathfrak p}(f)\in\Lambda_{\hat{\mathcal O}}, interpolating square roots of the central critical values of the twists of f/Kf/K by certain infinite-order characters of Γ\Gamma; the trivial character lies outside the range of pp-adic interpolation, and for it the construction yields the pp-adic Gross–Zagier formula, expressing the value through log⁡ωE\log_{\omega_E} of a Heegner point, the period factor (1−ap(E)+pp)\big(\frac{1-a_p(E)+p}{p}\big), and Tamagawa-type constants. The source sets Lp(f):=Lp(f)2∈ΛO^L_{\mathfrak p}(f):=\mathscr L_{\mathfrak p}(f)^{2}\in\Lambda_{\hat{\mathcal O}}, the pp-adic LL-function interpolating the corresponding central critical values. In the cyclotomic setting the pp-adic height pairing (the Zp\mathbb{Z}_p-valued analogue of the Néron–Tate canonical height, built from the p\mathfrak p-adic logarithm) on E(K)⊗ZpE(K)\otimes\mathbb{Z}_p is expected to be non-degenerate; in the anticyclotomic setting it degenerates systematically whenever ∣r+−r−∣>1|r^+-r^-|>1. Bertolini–Darmon understood these degeneracies as giving rise to derived pp-adic height pairings: on the pro-pp Selmer group Sp(E/K)S_{\mathfrak p}(E/K) (the extended Mordell–Weil group in the terminology of Bertolini–Darmon), which fits into the exact sequence 0→E(K)⊗Zp→Sp(E/K)→Tp\Sha(E/K)→00\to E(K)\otimes\mathbb{Z}_p\to S_{\mathfrak p}(E/K)\to T_p\Sha(E/K)\to 0 with Tp\Sha(E/K)T_p\Sha(E/K) the pp-adic Tate module of the Tate–Shafarevich group \Sha(E/K)\Sha(E/K) of E/KE/K (so that Sp(E/K)≃E(K)⊗ZpS_{\mathfrak p}(E/K)\simeq E(K)\otimes\mathbb{Z}_p under the standing finiteness of \Sha(E/K)p∞\Sha(E/K)_{p^\infty}), the successive radicals of the height pairings define a descending filtration Sp(k)S_{\mathfrak p}^{(k)}, and the successive derived pairings produce an enhanced (derived) pp-adic regulator Regp,der\mathrm{Reg}_{p,\mathrm{der}}, an element of (Jϱ/Jϱ+1)⊗Q(J^\varrho/J^{\varrho+1})\otimes\mathbb{Q}, and a derived square-root regulator Regp,der1/2\mathrm{Reg}^{1/2}_{p,\mathrm{der}} (well-defined up to sign), generalizing the classical pp-adic regulator Regp\mathrm{Reg}_p (the determinant of the height pairing); in the source these regulators are defined under the non-degeneracy assumption Sp(p)=USp(E/K)S_{\mathfrak p}^{(p)}=US_{\mathfrak p}(E/K) recalled below. The Mazur–Bertolini–Darmon maximal non-degeneracy conjecture (Conjecture 2.9 of the source) predicts that the filtration has the largest possible ranks: rankZpSp(2)=∣r+−r−∣\mathrm{rank}_{\mathbb{Z}_p}S_{\mathfrak p}^{(2)}=|r^+-r^-| and Sp(3)=USp(E/K)S_{\mathfrak p}^{(3)}=US_{\mathfrak p}(E/K), where USp(E/K)=⋂nNormKn/KSp(E/Kn)US_{\mathfrak p}(E/K)=\bigcap_n\mathrm{Norm}_{K_n/K}S_{\mathfrak p}(E/K_n) is the group of universal norms (isomorphic to Zp\mathbb{Z}_p, by the theorem of Cornut–Vatsal that Heegner points become non-torsion over the anticyclotomic tower); equivalently, the graded ranks eke_k satisfy e1=2min⁡{r+,r−}e_1=2\min\{r^+,r^-\}, e2=∣r+−r−∣−1e_2=|r^+-r^-|-1, and ek=0e_k=0 for k≥3k\ge 3. Agboola–Castella proved, under mild hypotheses, one of the inequalities predicted by the rank part of their conjectures and the predicted leading coefficient formula up to a pp-adic unit on the algebraic side (Theorem 1.3), transferring it to the analytic side under the then-partially-proved Iwasawa–Greenberg main conjecture (an identity between the characteristic ideal of a Selmer group and a pp-adic LL-function) for Lp(f)L_{\mathfrak p}(f) (Corollary 3.4). Later work sharpened the surrounding ingredients: Burungale–Castella–Kim proved Perrin-Riou's Heegner point main conjecture, which supplies the Iwasawa–Greenberg main conjecture in the ordinary split setting; Castella–Hsu–Kundu–Lee–Liu gave an independent method removing the extra hypotheses of Agboola–Castella and extending the up-to-unit leading-coefficient result to all good primes (including supersingular ones, i.e. p∣ap(E)p\mid a_p(E), where they also prove the main conjecture under mild hypotheses), reducing the rank part to maximal non-degeneracy; Sano showed by the theory of derived Bockstein regulators that the Agboola–Castella conjecture follows from the Iwasawa–Greenberg main conjecture up to a pp-adic unit. Nevertheless, in every available result the leading-coefficient formula retains the pp-adic unit ambiguity, the order-of-vanishing statement is proved only as an inequality, and the maximal non-degeneracy conjecture itself is unproven. The natural extension of both conjectures to non-split (inert or ramified) primes pp and to the generalized Heegner hypothesis is left open by the source.

Problem Statement

Under the standing hypotheses of Agboola–Castella (2022) — E/QE/\mathbb{Q} an elliptic curve of conductor NN, pp an odd prime of good ordinary reduction for EE, KK an imaginary quadratic field of discriminant prime to NpNp satisfying (Heeg)(\mathrm{Heeg}) (every prime factor of NN splits in KK) and (spl)(\mathrm{spl}) (p=ppˉp=\mathfrak p\bar{\mathfrak p} splits in KK), the residual representation ρˉE,p\bar\rho_{E,p} of GQ=Gal(Qˉ/Q)G_\mathbb{Q}=\mathrm{Gal}(\bar{\mathbb{Q}}/\mathbb{Q}) on the pp-torsion E[p]E[p] surjective, p∤#E(Fp)p\nmid\#E(\mathbb{F}_p), the mod-pp representation ρˉE,p\bar\rho_{E,p} (equivalently E[p]E[p]) ramified at every prime ℓ∣N\ell|N, and #\Sha(E/K)p∞<∞\#\Sha(E/K)_{p^\infty}<\infty — prove the anticyclotomic pp-adic Birch–Swinnerton-Dyer conjectures for the two pp-adic LL-functions over K∞K_\infty in their exact form, where JJ is the augmentation ideal of ΛO^=Zp[[Gal(K∞/K)]]⊗^ZpO^\Lambda_{\hat{\mathcal O}}=\mathbb{Z}_p[[\mathrm{Gal}(K_\infty/K)]]\hat\otimes_{\mathbb{Z}_p}\hat{\mathcal O}, in which Lp(f)\mathscr L_{\mathfrak p}(f) and Lp(f)=Lp(f)2L_{\mathfrak p}(f)=\mathscr L_{\mathfrak p}(f)^{2} live, r±=rank E(K)±r^\pm=\mathrm{rank}\,E(K)^\pm, ρ=max⁡{r+,r−}−1\rho=\max\{r^+,r^-\}-1, cℓc_\ell is the Tamagawa number of E/QℓE/\mathbb{Q}_\ell (the index of E(Qℓ)E(\mathbb{Q}_\ell) in its subgroup of points with nonsingular reduction), Regp,der\mathrm{Reg}_{p,\mathrm{der}} is the derived pp-adic regulator and Regp,der1/2\mathrm{Reg}^{1/2}_{p,\mathrm{der}} the derived square-root regulator of Bertolini–Darmon type: (i) Conjecture 2.20: ordJLp(f)=2(max⁡{r+,r−}−1)\mathrm{ord}_J L_{\mathfrak p}(f)=2(\max\{r^+,r^-\}-1), and the image of Lp(f)L_{\mathfrak p}(f) in J2ρ/J2ρ+1J^{2\rho}/J^{2\rho+1} equals (1−ap(E)+pp)2⋅Regp,der⋅#\Sha(E/K)⋅∏ℓ∣Ncℓ2\big(\frac{1-a_p(E)+p}{p}\big)^{2}\cdot \mathrm{Reg}_{p,\mathrm{der}}\cdot \#\Sha(E/K)\cdot \prod_{\ell|N}c_\ell^{2}; (ii) Conjecture 2.21: ordJLp(f)=max⁡{r+,r−}−1\mathrm{ord}_J \mathscr L_{\mathfrak p}(f)=\max\{r^+,r^-\}-1, and the image of Lp(f)\mathscr L_{\mathfrak p}(f) in Jρ/Jρ+1J^{\rho}/J^{\rho+1} equals ±(1−ap(E)+pp)⋅Regp,der1/2⋅#\Sha(E/K)⋅∏ℓ∣Ncℓ\pm\big(\frac{1-a_p(E)+p}{p}\big)\cdot \mathrm{Reg}^{1/2}_{p,\mathrm{der}}\cdot \sqrt{\#\Sha(E/K)}\cdot \prod_{\ell|N}c_\ell. In the source these two conjectures are formulated using the derived regulators of its Definitions 2.10 and 2.16, whose definition presupposes the maximal non-degeneracy conjecture of Mazur–Bertolini–Darmon (Conjecture 2.9 of the source: rankZpSp(2)=∣r+−r−∣\mathrm{rank}_{\mathbb{Z}_p}S^{(2)}_{\mathfrak p}=|r^+-r^-| and Sp(3)=USp(E/K)S^{(3)}_{\mathfrak p}=US_{\mathfrak p}(E/K), equivalently the graded ranks e1=2min⁡{r+,r−}e_1=2\min\{r^+,r^-\}, e2=∣r+−r−∣−1e_2=|r^+-r^-|-1, ek=0e_k=0 for k≥3k\ge 3), which remains open; the exact equalities above are the target. Note that the currently known results establish only the inequality ordJLp(f)≥2ρ\mathrm{ord}_J L_{\mathfrak p}(f)\ge 2\rho and the above leading-term formulas up to a pp-adic unit, so removing that unit ambiguity is an integral part of the problem.

References

  1. Adebisi Agboola, Francesc Castella, 'On anticyclotomic variants of the p-adic Birch and Swinnerton-Dyer conjecture', Journal de théorie des nombres de Bordeaux 33 (2022), 629–658. DOI: 10.5802/jtnb.1174, https://doi.org/10.5802/jtnb.1174; arXiv:1910.08680, https://arxiv.org/abs/1910.08680
  2. Francesc Castella, Chi-Yun Hsu, Debanjana Kundu, Yu-Sheng Lee, Zheng Liu, 'Derived p-adic heights and the leading coefficient of the Bertolini–Darmon–Prasanna p-adic L-function', Transactions of the American Mathematical Society, Series B 12 (2025), 748–788. DOI: 10.1090/btran/227, https://doi.org/10.1090/btran/227; arXiv:2308.10474, https://arxiv.org/abs/2308.10474
  3. Takamichi Sano, 'Derived Bockstein regulators and anticyclotomic p-adic Birch and Swinnerton-Dyer conjectures', preprint (2023). arXiv:2308.08875, https://arxiv.org/abs/2308.08875
  4. Ashay Burungale, Francesc Castella, Chan-Ho Kim, 'A proof of Perrin-Riou's Heegner point main conjecture', Algebra & Number Theory 15 (2021), 1627–1653. DOI: 10.2140/ant.2021.15.1627, https://doi.org/10.2140/ant.2021.15.1627
  5. Daniel Macias Castillo, Takamichi Sano, 'On Selmer complexes, Stark systems and derived p-adic heights', preprint (2026). arXiv:2603.23978, https://arxiv.org/abs/2603.23978
  6. Yangyu Fan, Xin Wan, 'p-adic Waldspurger Formula for Non-split Primes and Converse of Gross–Zagier and Kolyvagin Theorem', preprint (2023). arXiv:2304.09806, https://arxiv.org/abs/2304.09806
  7. Matteo Longo, Maria Rosaria Pati, Stefano Vigni, 'Anticyclotomic Iwasawa main conjectures for modular forms', preprint (2026). arXiv:2603.22483, https://arxiv.org/abs/2603.22483
  8. Massimo Bertolini, Henri Darmon, Kartik Prasanna, 'Generalized Heegner cycles and p-adic Rankin L-series', Duke Mathematical Journal 162 (2013), no. 6. DOI: 10.1215/00127094-2142056, https://doi.org/10.1215/00127094-2142056
  9. Massimo Bertolini, Henri Darmon, 'Derived p-adic heights', American Journal of Mathematics 117 (1995), no. 6, 1517–1554. DOI: 10.2307/2375029, https://doi.org/10.2307/2375029
  10. Massimo Bertolini, Henri Darmon, 'Heegner points on Mumford-Tate curves', Inventiones Mathematicae 126 (1996), 413–456. DOI: 10.1007/s002220050105, https://doi.org/10.1007/s002220050105
  11. Francesc Castella, Ming-Lun Hsieh, 'Heegner cycles and p-adic L-functions', Mathematische Annalen 370 (2018), no. 1-2, 567–628. DOI: 10.1007/s00208-017-1517-3, https://doi.org/10.1007/s00208-017-1517-3
  12. Takamichi Sano, 'On the Tamagawa number conjecture for modular forms twisted by anticyclotomic Hecke characters', preprint (2025). arXiv:2510.01601, https://arxiv.org/abs/2510.01601