Zeta-polynomials for modular form periods (Q728224)

Let \(f\) be a new cusp-form of level \(N\) and even weight \(k\geq 4\) and let \[ \Lambda(f,s):= (2\pi)^{-s} N^{s/2} \Gamma(s) L(f,s) \] be the completed \(L\)-function associated to \(f\); suppose, moreover, the \(L\)-function \(L(f,s)\), associated to \(f\), be normalised so that \[ \Lambda(f,s)= \varepsilon(f) \Lambda(f,k-s) \] for \(s\in\mathbb{C}\), with \(\varepsilon(f)\in\{\pm 1\}\). For \(m\in\mathbb{Z}\), \(m\geq 0\), the authors define \[ M_f(m):= \sum^{k-2}_{j=0} (j!(k-2- j)!)^{-1} \Lambda(f,j+1) j^m \] and \[ Z_f(s):= \varepsilon(f) \sum^{k-2}_{h=0} (-s)^h \sum^{k-2-h}_{m=0} (m+ h)!(m! h!)^{-1}{\mathfrak s}(k-2, m+h) M_f(m), \] where \({\mathfrak s}(n,m)\) stands for a so-called signed Stirling number, and prove that \(Z_f(s)\) is a zeta-polynomial in the sense by \textit{Yu. I. Manin} [Izv. Math. 80, No. 4, 751--758 (2016); Izv. Ross. Akad. Nauk, Ser. Mat. 80, No. 4, 123--130 (2016; Zbl 1417.11081)], that is \[ Z_f(s)=\varepsilon(f) Z_f(1-s)\text{ and }Z_f(s)= 0\Rightarrow\text{Re}\,(s)= 1/2 \] for \(s\in\mathbb{C}\). Assuming the Bloch-Kato Tamagawa number conjecture, the authors relate the special values \(Z_f(-n)\), \(n\in\mathbb{Z}\), \(n\geq 0\), of their polynomials to the corresponding invariants of the motive \({\mathcal M}_f\) associated to the modular form \(f\).