For an elliptic curve E/QE/\mathbb{Q} and an Artin representation ρ\rho, the equivariant Birch and Swinnerton-Dyer conjecture predicts that the order of vanishing of the Hasse-Weil-Artin LL-function L(E,ρ,s)L(E,\rho,s) at s=1s=1 equals the dimension of the ρ\rho-isotypic component of the Mordell-Weil group of EE over the field cut out by ρ\rho. Darmon and Rotger (J. Amer. Math. Soc. 30, 2017) proved the rank-zero case for ρ=ρ1⊗ρ2\rho=\rho_1\otimes\rho_2 a tensor product of two odd irreducible two-dimensional Artin representations with inverse determinants — a setting that includes ring class characters of real quadratic fields and A5A_5-type representations and lies beyond the reach of Heegner points — and gave conditional rank-two lower bounds depending on non-vanishing of certain non-classical Garrett-Hida pp-adic LL-values (values of a three-variable pp-adic LL-function outside its range of classical interpolation). A three-variable Λ\Lambda-adic family of diagonal-cycle classes has since been constructed by the authors, and anticyclotomic Euler systems of diagonal cycles now yield Bloch-Kato results on the imaginary-dihedral side; but the central questions remain open, namely finiteness of the ρ\rho-isotypic Selmer and Tate-Shafarevich groups in analytic rank zero outside the imaginary-dihedral case, the unconditional rank-two statement including the upper bound, and systematic non-vanishing of the non-classical pp-adic LL-values.

Background

Let EE be an elliptic curve over Q\mathbb{Q} and let ρ\rho be an Artin representation, i.e. a continuous representation of the absolute Galois group of Q\mathbb{Q}, with finite image, over a finite extension L/QL/\mathbb{Q}. The Hasse-Weil-Artin LL-function L(E,ρ,s)L(E,\rho,s) is the LL-function of the twist of EE by ρ\rho. If H/QH/\mathbb{Q} is the finite Galois extension cut out by ρ\rho and E(H)ρE(H)^\rho denotes the ρ\rho-isotypic component of the Mordell-Weil group E(H)E(H) of HH-rational points, the equivariant Birch and Swinnerton-Dyer (BSD) conjecture predicts

ran(E,ρ)=?r(E,ρ),r_{\mathrm{an}}(E,\rho) \stackrel{?}{=} r(E,\rho),

where ran(E,ρ)=ords=1L(E,ρ,s)r_{\mathrm{an}}(E,\rho)=\mathrm{ord}_{s=1}L(E,\rho,s) is the analytic rank and r(E,ρ)=dim⁡LE(H)ρr(E,\rho)=\dim_L E(H)^\rho is the algebraic rank; the classical BSD conjecture is the case ρ=1\rho=\mathbf{1}. The Tate-Shafarevich group \Sha(E/H)\Sha(E/H), the group of everywhere-locally-solvable principal homogeneous spaces of EE modulo the trivial ones, measures the failure of the local-to-global principle and is conjectured to be finite; its ρ\rho-isotypic part is controlled by the Bloch-Kato Selmer group Selp(E,ρ)=Hfin1(Q,Vp(E)⊗Vρ)\mathrm{Sel}_p(E,\rho)=H^1_{\mathrm{fin}}(\mathbb{Q},V_p(E)\otimes V_\rho), where Vp(E)V_p(E) is the pp-adic Tate module of EE and Vρ=ρ⊗QQpV_\rho=\rho\otimes_{\mathbb{Q}}\mathbb{Q}_p: the subgroup of the Galois cohomology H1(Q,Vp(E)⊗Vρ)H^1(\mathbb{Q},V_p(E)\otimes V_\rho) of classes unramified away from the level and crystalline at pp (the standard pp-adic Hodge-theoretic local condition at pp).

Darmon and Rotger studied this conjecture for ρ=ρ1⊗ρ2\rho=\rho_1\otimes\rho_2, where ρ1,ρ2\rho_1,\rho_2 are odd (i.e. complex conjugation acts with determinant −1-1), irreducible, two-dimensional Artin representations with det⁡(ρ1)=det⁡(ρ2)−1\det(\rho_1)=\det(\rho_2)^{-1}, so that ρ\rho is self-dual. By modularity, EE, ρ1\rho_1, and ρ2\rho_2 are attached to newforms f,g,hf,g,h (normalized Hecke-eigen cusp forms) of weights 2,1,12,1,1, and L(E,ρ,s)L(E,\rho,s) identifies with the Garrett-Rankin triple-product LL-function L(f,g,h,s)L(f,g,h,s) — the degree-eight LL-function of the triple product of the Galois representations attached to ff, gg, and hh — whose analytic continuation and functional equation are known. When the level NfN_f of ff is coprime to NgNhN_gN_h, all local root numbers of L(f,g,h,s)L(f,g,h,s) — the local constants whose product is the sign of the functional equation — are +1+1, so ran(E,ρ)r_{\mathrm{an}}(E,\rho) is even. If one factor, say ρ2\rho_2, is induced from a ring class character of an imaginary quadratic field (a character of the Galois group of a ring class field, the natural abelian extension attached to an order of the quadratic field), this is the classical Heegner setting: Heegner points, i.e. special points on modular curves arising from the theory of complex multiplication, combined with the Gross-Zagier formula (which expresses heights of Heegner points as first derivatives of LL-functions) and Kolyvagin's method of Euler systems (norm-compatible families of global cohomology classes converted by descent into bounds on Selmer groups), give the rank-zero statement. These tools fail when the quadratic field is real — the conjectural replacement, the analytically defined Stark-Heegner points, are not even known to be algebraic — and for exotic projective images such as A5A_5 (the alternating group on five letters).

The method of the source rests on the Gross-Kudla-Schoen diagonal cycle, the modified diagonal on the triple product of three modular curves, a homologically trivial algebraic cycle whose arithmetic reflects the triple product LL-function; on its pp-adic Abel-Jacobi images; and on their variation in Hida families, which are pp-adic analytic families of ordinary modular forms interpolating a fixed eigenform at all classical weights. The authors construct a one-variable Λ\Lambda-adic cohomology class κ(f,gh)\kappa(\mathbf{f},\mathbf{g}\mathbf{h}) with the weight-two form ff held fixed; its specializations at the unbalanced point of weights (2,1,1)(2,1,1) produce four generalized Kato classes in H1(Q,Vp(E)⊗Vρ)H^1(\mathbb{Q},V_p(E)\otimes V_\rho), so named because Kato's Euler system of Beilinson elements is the case where g,hg,h are weight-one Eisenstein series. An explicit reciprocity law relates the Bloch-Kato dual exponential of these classes (the pp-adic counterpart of the classical exponential map, evaluating Galois cohomology classes in the de Rham cohomology of Vp(E)⊗VρV_p(E)\otimes V_\rho) to the Garrett-Hida pp-adic LL-function, the pp-adic analytic function on the triple product of weight spaces interpolating central critical values L(fx,gy,hz,c)L(f_x,g_y,h_z,c) built by pairing Hida families against pp-adic test vectors. The main theorems are: Theorem A, if L(E,ρ,1)≠0L(E,\rho,1)\neq 0 then E(H)ρ=0E(H)^\rho=0; and Theorems B and D, if L(E,ρ,1)=0L(E,\rho,1)=0 and one of four non-classical pp-adic LL-values — values of the Garrett-Hida pp-adic LL-function at points outside its range of classical interpolation — is non-zero, then dim⁡LpSelp(E,ρ)≥2\dim_{L_p}\mathrm{Sel}_p(E,\rho)\ge 2, where Lp=L⊗QQpL_p=L\otimes_{\mathbb{Q}}\mathbb{Q}_p is the coefficient field of the Selmer group, with two linearly independent classes constructed in the Selmer group. The proof of Theorem A does not bound the Selmer or Tate-Shafarevich group, and the paper states that extending Kato's and Lei-Loeffler-Zerbes' finiteness refinements to the diagonal-cycle setting 'would be of great interest'.

Problem Statement

Let E/QE/\mathbb{Q} be an elliptic curve and let ρ1,ρ2:GQ→GL2(L)\rho_1,\rho_2:G_{\mathbb{Q}}\to\mathrm{GL}_2(L) be odd, irreducible, two-dimensional Artin representations over a finite extension L/QL/\mathbb{Q} with det⁡(ρ1)=det⁡(ρ2)−1\det(\rho_1)=\det(\rho_2)^{-1}; set ρ=ρ1⊗ρ2\rho=\rho_1\otimes\rho_2, and let H/QH/\mathbb{Q} be the finite Galois extension cut out by ρ\rho. Let f,g,hf,g,h be the newforms of weights 2,1,12,1,1 attached to E,ρ1,ρ2E,\rho_1,\rho_2, assume (Nf,NgNh)=1(N_f,N_gN_h)=1 so that all local root numbers of L(E,ρ,s)=L(f,g,h,s)L(E,\rho,s)=L(f,g,h,s) are +1+1, fix an odd prime p∤NfNgNhp\nmid N_fN_gN_h at which EE is ordinary (i.e. the trace of Frobenius ap(E)=p+1−∣E(Fp)∣a_p(E)=p+1-|E(\mathbb{F}_p)| is a pp-adic unit), and write Selp(E,ρ)=Hfin1(Q,Vp(E)⊗Vρ)\mathrm{Sel}_p(E,\rho)=H^1_{\mathrm{fin}}(\mathbb{Q},V_p(E)\otimes V_\rho) for the Bloch-Kato Selmer group. Establish the remaining cases of the equivariant Birch and Swinnerton-Dyer conjecture for these Hasse-Weil-Artin LL-functions, namely:

  1. In analytic rank zero, prove that L(E,ρ,1)≠0L(E,\rho,1)\neq 0 implies not only E(H)ρ=0E(H)^\rho=0 (Darmon-Rotger's Theorem A) but also the vanishing — and in particular the finiteness — of Selp(E,ρ)\mathrm{Sel}_p(E,\rho), equivalently the finiteness of the ρ\rho-isotypic Tate-Shafarevich group \Sha(E/H)ρ[p∞]\Sha(E/H)^\rho[p^\infty], for every ρ=ρ1⊗ρ2\rho=\rho_1\otimes\rho_2 outside the imaginary-dihedral case, i.e. in the settings where no Heegner-point Euler system exists; this entire class is the required scope, and the cases named below are examples of it rather than an exhaustive list: when ρ1\rho_1 or ρ2\rho_2 is induced from a ring class character of a real quadratic field, or when ρ\rho has exotic (e.g. A5A_5) projective image.

  2. In analytic rank two, prove unconditionally that ran(E,ρ)=r(E,ρ)r_{\mathrm{an}}(E,\rho)=r(E,\rho) when L(E,ρ,s)L(E,\rho,s) vanishes to exact order two at s=1s=1: remove the non-vanishing hypothesis on the four non-classical Garrett-Hida pp-adic LL-values in Theorems B and D of the source, and prove the upper bound dim⁡LpSelp(E,ρ)=2\dim_{L_p}\mathrm{Sel}_p(E,\rho)=2, where Lp=L⊗QQpL_p=L\otimes_{\mathbb{Q}}\mathbb{Q}_p is the coefficient field of the Selmer group.

The single objective underlying both deliverables is the equivariant BSD conjecture ran(E,ρ)=r(E,ρ)r_{\mathrm{an}}(E,\rho)=r(E,\rho), together with the finiteness of the associated Selmer and Tate-Shafarevich groups predicted by the Shafarevich-Tate conjecture, for the class of Artin representations ρ=ρ1⊗ρ2\rho=\rho_1\otimes\rho_2 as above — the setting introduced by Darmon-Rotger's diagonal-cycle method.

References

  1. Henri Darmon and Victor Rotger, Diagonal cycles and Euler systems II: The Birch and Swinnerton-Dyer conjecture for Hasse-Weil-Artin L-functions, Journal of the American Mathematical Society 30 (2017), no. 3, 601-672. DOI 10.1090/jams/861, https://doi.org/10.1090/jams/861
  2. Henri Darmon and Victor Rotger, Diagonal cycles and Euler systems I: A p-adic Gross-Zagier formula, Annales scientifiques de l'École normale supérieure 47 (2014), no. 4, 779-832. DOI 10.24033/asens.2227, https://doi.org/10.24033/asens.2227
  3. Henri Darmon, Alan Lauder, and Victor Rotger, Stark points and p-adic iterated integrals attached to modular forms of weight one, Forum of Mathematics, Pi 3 (2015), e8. DOI 10.1017/fmp.2015.7, https://doi.org/10.1017/fmp.2015.7
  4. Henri Darmon and Victor Rotger, Elliptic curves of rank two and generalised Kato classes, Research in the Mathematical Sciences 3 (2016), no. 1, Paper no. 27. DOI 10.1186/s40687-016-0074-9, https://doi.org/10.1186/s40687-016-0074-9
  5. Raúl Alonso, Francesc Castella, and Óscar Rivero, The diagonal cycle Euler system for GL_2 x GL_2, Journal of the Institute of Mathematics of Jussieu 24 (2025), no. 5, 1591-1653. DOI 10.1017/s1474748023000221, https://doi.org/10.1017/s1474748023000221
  6. Francesc Castella and Kim Tuan Do, Diagonal cycles and anticyclotomic Iwasawa theory of modular forms, Journal of the European Mathematical Society (2025). DOI 10.4171/jems/1735, https://doi.org/10.4171/jems/1735 (preprint arXiv:2303.06751, https://arxiv.org/abs/2303.06751)
  7. Francesc Castella and Ming-Lun Hsieh, On the nonvanishing of generalised Kato classes for elliptic curves of rank 2, Forum of Mathematics, Sigma 10 (2022), e12. DOI 10.1017/fms.2021.85, https://doi.org/10.1017/fms.2021.85
  8. Henri Darmon and Victor Rotger, p-adic families of diagonal cycles, preprint (2022). arXiv:2207.01314, https://arxiv.org/abs/2207.01314
  9. Henri Darmon and Victor Rotger, Stark-Heegner points and diagonal classes, preprint (2022). arXiv:2207.01310, https://arxiv.org/abs/2207.01310
  10. Massimo Bertolini, Henri Darmon, Victor Rotger, Marco Adamo Seveso, and Rodolfo Venerucci, Heegner points, Stark-Heegner points, and diagonal classes, Astérisque 434, Société Mathématique de France, Paris (2022). DOI 10.24033/ast.1173, https://doi.org/10.24033/ast.1173
  11. Raúl Alonso, Kâzım Büyükboduk, Antonio Cauchi, and Antonio Lei, Trito-non-ordinary Iwasawa theory of diagonal cycles, preprint (2025). arXiv:2511.11511, https://arxiv.org/abs/2511.11511
  12. Shilin Lai and Christopher Skinner, Anti-cyclotomic Euler system of diagonal cycles, preprint (2024). arXiv:2408.01219, https://arxiv.org/abs/2408.01219
  13. Kim Tuan Do, Anticyclotomic Euler system over biquadratic fields, preprint (2024). arXiv:2409.19819, https://arxiv.org/abs/2409.19819
  14. Michele Fornea and Zhaorong Jin, Hirzebruch-Zagier classes and rational elliptic curves over quintic fields, Mathematische Zeitschrift 308 (2024), no. 1, Paper no. 18. DOI 10.1007/s00209-024-03564-y, https://doi.org/10.1007/s00209-024-03564-y
  15. Yifeng Liu, Hirzebruch-Zagier cycles and twisted triple product Selmer groups, Inventiones mathematicae 205 (2016), no. 3, 693-780. DOI 10.1007/s00222-016-0645-9, https://doi.org/10.1007/s00222-016-0645-9
  16. Yifeng Liu, Yichao Tian, Liang Xiao, Wei Zhang, and Xinwen Zhu, Survey on bounding Selmer groups for Rankin-Selberg motives, preprint (2025). arXiv:2509.16881, https://arxiv.org/abs/2509.16881
  17. Kazuya Kato, p-adic Hodge theory and values of zeta functions of modular forms, in Cohomologies p-adiques et applications arithmétiques (III), Astérisque 295, Société Mathématique de France, Paris (2004), pp. 117-290. DOI 10.24033/ast.639, https://doi.org/10.24033/ast.639 (MR 2104361)
  18. Antonio Lei, David Loeffler, and Sarah Livia Zerbes, Euler systems for Rankin-Selberg convolutions of modular forms, Annals of Mathematics (2) 180 (2014), no. 2, 653-771. DOI 10.4007/annals.2014.180.2.6, https://doi.org/10.4007/annals.2014.180.2.6