Hida families, \(p\)-adic heights, and derivatives (Q624786)

Let \(f\) be a normalized newform of even weight \(k\geq 2\) and level \(N\), and let \(p\nmid 2N\) be a prime at which \(f\) is ordinary. Then \(f\) belongs to a Hida family \({\mathcal F}\) and there is a representation \(\rho_{\mathcal F}: G_{\mathbb Q} =\text{Gal}(\overline{\mathbb Q}/\mathbb Q))\to \Aut_{\mathbf H}({\mathcal T})\), where \(\mathbf H\) is a complete local domain that is finite and flat over \(\Lambda=\mathbb Z_p[[Y]]\) and \({\mathcal T}\) is a rank two module over \(\mathbf H\). Let \(\Gamma_c=\text{Gal}(\mathbb Q_{\infty}/\mathbb Q)\) be the Galois group of the \(\mathbb Z_p\)-extension of \(\mathbb Q\), let \(\kappa^{\text{univ}}: G_{\mathbb Q} \to \Gamma_c\to \Lambda^{\times}\) be the universal character, and let \(\tilde{\rho}_{\mathcal F}=\rho_{\mathcal F}\otimes \kappa^{\text{univ}}\), which has values in \(\Aut_{\mathbf H[[\Gamma_c]]}(\widetilde{\mathcal T})\), where \(\widetilde{\mathcal T} = {\mathcal T}\hat{\otimes}_{\mathbb Z_p} \Lambda\). Work of \textit{K. Kitagawa} in [p-adic monodromy and the Birch and Swinnerton-Dyer conjecture (1991, Boston), Contemp. Math. 165, 81--110 (1994; Zbl 0841.11028)] yields \({\mathcal L} \in \Lambda_{\mathbf H} =\mathbf H[[\Gamma_c]]\) that is the \(p\)-adic \(L\)-function for \({\mathcal F}\). If \(f\) has Nebentypus \(\omega^{2-k}\), there is a principal ideal \(\Theta\subseteq \Lambda_H\) corresponding to the locus where \(\tilde{\rho}_{\mathcal F}\) is self-dual. Let \(T=\widetilde{\mathcal T}/\Theta\widetilde{\mathcal T}\) and \(\widetilde{\mathcal T}^* = \text{Hom}_{\Lambda_{\mathbf H}}(\widetilde{\mathcal T}, \Lambda_H)(1)\). Similarly define \(T^*\). There is a certain cohomology class \(z(1)\in H^1(\mathbb Q, \widetilde{\mathcal T}^*)\) arising from Kato's Euler system. Assume that the image of \(z(1)\) under the corestriction map \(\text{Sel}(\mathbb Q, \widetilde{\mathcal T}^*) \to \text{Sel}(\mathbb Q, T^*)\) is not in the \(\mathbf H\)-torsion submodule of \(\text{Sel}(\mathbb Q, T^*)\). Assume also that Plater's height pairing is non-degenerate, that the sign of the functional equations of the \(L\)-functions of the forms belonging to \({\mathcal F}\) is \(-1\), and that \(\rho_{\mathcal F}\) is residually irreducible. The main result of the paper is that then \(\text{ord}_{\Theta}({\mathcal L}) =1\). Under additional assumptions, a converse is proved, namely that if \(\text{ord}_{\Theta}({\mathcal L})=1\) then the image of \(z(1)\) is not \(\mathbf H\)-torsion. The main ingredient of the proofs is a generalization of a formula of \textit{K. Rubin} [in: Algebra and number theory. Proceedings of a conference held at the Institute of Experimental Mathematics, University of Essen, 1992. Berlin: de Gruyter, 247--266 (1994; Zbl 0829.11034)] relating the \(p\)-adic height pairing with the local Tate pairing.