A \(p\)-adic Waldspurger formula (Q2413076)

The first aim of the paper under review is to construct (Theorem 3.2.10), for any fixed rational prime \(p\), a \(p\)-adic \(L\)-function \(\mathcal{L}(A)\) which interpolates the central values of certain twists of the classical \(L\)-function \(L(s,\rho_A^{(\iota)})\) associated to an abelian variety \(A_{/F}\) which can be parametrised by a Shimura curve \(X\). Here \(F\) is a totally real number field, which comes with a fixed embedding \(F \hookrightarrow \mathbb{C}_p\), and \(\iota\) is any embedding \(\iota \colon \mathrm{End}(A) \hookrightarrow \mathbb{C}\). Moreover, the twists of \(L(s,\rho_A^{(\iota)})\) under consideration are given by characters \[ \chi \ \colon \ E^\times \backslash (\mathbb{A}_E^\infty)^\times \to \mathbb{C}_p^\times, \] where \(E\) is any fixed totally imaginary quadratic extension of \(F\) with the crucial property (Assumption 1.8.1) that the prime ideal \(\mathfrak{p} \subseteq \mathcal{O}_F\) associated to the embedding \(F \hookrightarrow \mathbb{C}_p\) splits in \(E\). Note in particular that the interpolation properties of the \(p\)-adic \(L\)-function work for any isomorphism \(\mathbb{C} \cong \mathbb{C}_p\), whose choice is necessary to twist \(L(s,\rho_A^{(\iota)})\) by such characters \(\chi\). The other main result of the paper under review is a \(p\)-adic analogue of the classical Waldspurger formula (Theorem 1.1.1). This \(p\)-adic formula (Theorem 3.3.2) relates the values of \(\mathcal{L}(A)\) to \(p\)-adic logarithms of certain cycles (called Heegner cycles) defined on the Shimura curve \(X\). From Theorem 3.3.2 one can also easily derive a relation (Theorem 3.4.4) between the values of the \(p\)-adic \(L\)-function and \(p\)-adic periods of certain functions \(X(\mathbb{C}_p) \to \mathbb{C}_p\), called \(p\)-adic Maass functions (Definition 3.4.1). In the previous discussion \(X\) is a Shimura curve of infinite level, associated to a given totally definite, incoherent quaternion algebra over the ring of rational adèles \(\mathbb{A} := \mathbb{A}_\mathbb{Q}\). Working at the infinite level allows the authors of the paper under review to construct (Theorem 2.3.17) the global Mellin transform of any given stable convergent modular form (Definition 2.3.10), using results of \textit{P. Schneider} and \textit{J. Teitelbaum} [Doc. Math. 6, 447--481 (2001; Zbl 1028.11069)]. This Mellin transform is an element of a certain distribuition algebra (Definition 3.2.5), and it interpolates the iterated action of the Atkin-Serre operator (Equation 2.18) at classical points in the weight space. This global Mellin transform is used to define the universal \(p\)-adic torus period (Definition 4.2.5), which is an element of the same distribution algebra. The \(p\)-adic \(L\)-function is then defined (Page 815) as the quotient of this universal \(p\)-adic torus period by the local period distribution (Definition 4.1.7). This local period distribution is also an element of the same distribution algebra where the Mellin transform lives, and it interpolates the twisted matrix coefficient integrals (Definition 4.1.4) appearing in the complex Waldspurger formula (Theorem 1.1.1). To show that the definition of the \(p\)-adic \(L\)-function is well posed, and in particular that the local period distribution is non-zero for sufficiently many test vectors, the authors use results from the work of \textit{J. B. Tunnell} [Am. J. Math. 105, 1277--1307 (1983; Zbl 0532.12015)] and \textit{H. Saito} [Compos. Math. 85, No. 1, 99--108 (1993; Zbl 0795.22009)]. Then the authors prove Theorem 3.3.2 by relating the values of the universal \(p\)-adic torus period to the \(p\)-adic logarithm of Heegner cycles (Proposition 4.3.6). The proof of Proposition 4.3.6 uses the results of Appendix A, which generalise the work of \textit{R. F. Coleman} [Ann. Math. (2) 121, 111--168 (1985; Zbl 0578.14038)]. Finally, the main result of Appendix B is a description (Theorem B.2.3) of the Kodaira-Spencer map associated to certain \(p\)-divisible groups endowed with the action of the ring of integers of a finite extension of \(\mathbb{Q}_p\), in terms of their Serre-Tate coordinates (Theorem B.1.1). This description generalises the one given by \textit{N. Katz} in Theorem 3.7.1 of [Surfaces algebriques, Semin. de geometrie algebrique, Orsay 1976-78, Lect. Notes Math. 868, 138-202 (1981; Zbl 0477.14007)]. The main results of the paper under review, which we have outlined in the previous paragraphs, generalise the ones proved by \textit{M. Bertolini}, \textit{H. Darmon} and \textit{K. Prasanna} in [Duke Math. J. 162, No. 6, 1033--1148 (2013; Zbl 1302.11043)]. The paper under review is also intimately related to the monograph [The Gross-Zagier formula on Shimura curves. Princeton, NJ: Princeton University Press (2013; Zbl 1272.11082)], authored by \textit{X. Yuan} and the second and third authors of the paper under review. In the reviewer's opinion, the paper under review is a mayor technical achievement which, combined with the work of \textit{L. Cai}, \textit{J. Shu} and \textit{Y. Tian} [Algebra Number Theory 8, No. 10, 2523--2572 (2014; Zbl 1311.11054)], will soon lead to new interesting developments in the study of the Birch e Swinnerton-Dyer (BSD) and Bloch-Kato conjectures for central values of \(L\)-functions.