Uniform estimates for sums of Fourier coefficients of cusp forms (Q1046922): Difference between revisions

Let \(f\) be a normalized Hecke eigen cusp form of even weight \(k\) for the full modular group \(\text{SL}_2(\mathbb Z)\) with Fourier expansion \(f(z)=\sum^\infty_{n=1}\lambda(n)n^{(k-1)/2}e(nz)\). Then \(\lambda(n)\) is real, \(| \lambda(n)| \leq \sum_{d| n}1\) by Deligne's theorem, and the associated \(L\)-function has an Euler product \[ L(f,s)=\sum^\infty_{n=1}\lambda(n)n^{-s}=\prod_p(1-\lambda(p)p^{-s}+p^{-2s})^{-1}=\prod_p(1-\alpha_p p^{-s})^{-1}(1-\alpha'_p p^{-s})^{-1}. \] The associated symmetric square \(L\)-function is \[ L(\text{Sym}^2f,s)=\zeta(2s)\sum^\infty_{n=1}\lambda(n^2)n^{-s}=\prod_p(1-\alpha^2_pp^{-s})^{-1}(1-p^{-s})^{-1}(1-\alpha^{\prime 2}_p p^{-s})^{-1}. \] Let \(c(n)\) denote its coefficients, i.e., \(L(\text{Sym}^2f,s)=\sum^\infty_{n=1}c(n)n^{-s}\). \textit{O. M. Fomenko} [Zap. Nauchn. Semin. POMI 314, 247--256 (2004); translation in J. Math. Sci., New York 133, No. 6, 1749--1755 (2006; Zbl 1094.11018)] proved that \[ C(x)=\sum_{n\leq x}c(n)\ll x^{1/2}(\log x)^2\quad\text{and}\quad S(x)=\sum_{n\leq x}\lambda(n^2)\ll x^{1/2}(\log x)^3, \] where the implied constants depend on \(f\). \textit{A. Sankaranarayanan} [Lith. Math. J. 46, 459--474 (2006; Zbl 1162.11337)] established the bound \(S(x)\ll x^{3/4}(\log x)^{9/2}\log \log x\) which holds uniformly for all \(f\) with weights \(k\ll x^{1/3}(\log x)^{22/3}\). In the paper under review the author proves bounds which hold uniformly for all \(f\) with arbitrary weights \(k\), \[ \begin{aligned} C(x) & \ll x^{3/5}(\log x)^{22/5}+k^{3/2}(\log x)^5,\\ C(x)&\ll x^{1/3}k(\log x)^{19/2}+x^{5/9}(\log x)^{23/3},\\ S(x)&\ll x^{1/2}k^{3/4}(\log x)^{19/2}\log\log x+ x^{3/5}(\log x)^{42/5}(\log \log x)^{4/5}, \end{aligned} \] where the implied constants are absolute. Similar results are established also for Maass cusp forms which are eigenfunctions of the Laplace operator and which are eigenfunctions of the Hecke operators. Here an additional difficulty is that one does not know whether the Fourier coefficients satisfy the Ramanujan-Petersson conjecture.