Mean squares of Quadratic twists of the Fourier coefficients of Modular forms
Peng Gao, Yuetong Zhao · 2025
Read Mean squares of Quadratic twists of the Fourier coefficients of Modular forms by Peng Gao, Yuetong Zhao (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-22256.txt.
Introduction
Peng Gao and Yuetong Zhao evaluate a smoothed sum involving quadratic twists of Fourier coefficients of a modular form, in their 2025 arXiv paper. The sum runs over positive odd square-free integers up to a bound, and inside it over coefficients twisted by a Kronecker symbol. Mean square estimates of this kind are a basic tool in analytic number theory, used to understand how modular form coefficients behave on average when twisted by quadratic characters. The authors obtain an asymptotic formula for the smoothed double sum.
Catalog record
| CREATOR | Peng Gao, Yuetong Zhao |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 27 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-22256
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Mean squares of Quadratic twists of the Fourier coefficients of Modular forms Peng Gao, Yuetong Zhao Abstract . In this paper, we evaluate asymptotically a smoothed version of the sum ∑ ∗ d ≤ X ∗ ( ∑ n ≤ Y λ f ( n ) ( 8 d n ) ) 2 , \displaystyle\sideset{}{{}^{*}}{\sum}_{d\leq X}\left(\sum_{n\leq Y}\lambda_{f}(n)\Big(\frac{8d}{n}\Big)\right)^{2}, where ( 8 d ⋅ ) \left(\frac{8d}{\cdot}\right) is the Kronecker symbol, ∑ ∗ \sideset{}{{}^{*}}{\sum} denotes a sum over positive odd square-free integers and λ f ( n ) \lambda_{f}(n) are Fourier coefficients of a given modular form f f . Mathematics Subject Classification (2010) : 11N37, 11L05, 11L40 Keywords : mean square, quadratic Dirichlet character, modular L -functions 1. Introduction Let f be a fixed holomorphic Hecke eigenform of weight \kappa\equiv 0\pmod{4} for the full modular group SL_{2}(\mathbb{Z}) . We write the Fourier expansion of f at infinity as (1.1) \displaystyle f(z)=\sum_{n=1}^{\infty}\lambda_{f}(n)n^{\frac{\kappa-1}{2}}e(nz), where e(z)=e^{2\pi iz} . It follows from Deligne’s proof [ 2 ] of the Weil conjecture that we have (1.2) \displaystyle|\lambda_{f}(n)|\leq d(n) where d(n) is the divisor function of n . Considerable research has been devoted to understanding the behavior of Fourier coefficients of cusp forms. For example, it is shown by J. L. Hafner and Ivić in [ 6 ] that \displaystyle\sum_{n\leq x}\lambda_{f}(n)\ll x^{1/3}. In [ 3 ] , P. Gao and L. Zhao studied bounds for moments of sums involving with \lambda_{f}(n) twisted by \chi_{8d}(n) for d running over odd, positive, square-free integers. Here we write \chi_{d} for the Kronecker symbol \big(\frac{d}{\cdot}\big) and we note that \chi_{8d} is a primitive Dirichlet character when d is odd and square-free. More precisely, for any real number m>0 , let \displaystyle S_{m}(X,Y;f)=:\sideset{}{{}^{*}}{\sum}_{\begin{subarray}{c}d\leq X\end{subarray}}\Big|\sum_{n\leq Y}\chi_{8d}(n)\lambda_{f}(n)\Big|^{m}, where \sideset{}{{}^{*}}{\sum} denotes a sum over positive odd square-free integers. Then it is shown in [ 3 , Theorem 1.6] that under the generalized Riemann hypothesis (GRH) that for m\geq 4 , one has (1.3) \displaystyle S_{m}(X,Y;f)\ll XY^{m/2}(\log X)^{\frac{m(m-3)}{2}+2}. The above result is achieved using the method developed by B. Szabó in [ 15 ] to apply sharp upper bounds for shifted moments of families of L -functions to obtain bounds for moments of the corresponding character sums. Here we also note that the establishment of sharp upper bounds for shifted moments of families of L -functions under GRH relies crucially on a method of K. Soundararajan [ 14 ] and its refinement by A. J. Harper [ 7 ] . Similar to ( 1.3 ), it is shown in [ 4 , Theorem 1.3] that under GRH, for any real m>\sqrt{5}+1 , (1.4) \displaystyle\sideset{}{{}^{*}}{\sum}_{\begin{subarray}{c}d\leq X\end{subarray}}\Big|\sum_{n\leq Y}\chi_{8d}(n)\Big|^{m}\ll XY^{m/2}(\log X)^{\frac{m(m-1)}{2}+1}. It is further proved by M. Munsch in [ 10 ] that for a smoothed version of the left-hand side expression above, one may replace under GRH the exponent \frac{m(m-1)}{2}+1 by \frac{m(m-1)}{2} of the right-hand side expression above for integers m\geq 2 and the resulting estimation is optimal. The result of Munsch is in accordance to an earlier result of M. V. Armon [ 1 , Theorem 2] , who obtained unconditionallly the optimal bound for the left-hand side expression of ( 1.4 ) for the case m=1 . Note more over that in [ 5 ] , the first named author evaluated the expression \displaystyle\sideset{}{{}^{*}}{\sum}_{\begin{subarray}{c}d\end{subarray}}\left|\sum_{n}\chi_{8d}(n)\Phi\left(\frac{n}{Y}\right)\right|^{2}\Psi\left(\frac{d}{X}\right), where \Phi and \Psi are smooth, compactly supported functions on {\mathbb{R}}_{+}=(0,\infty) , having all partial derivatives extending continuously to the boundary, and satisfying the partial derivative bounds (1.5) \displaystyle\begin{split}y^{j}\Phi^{(j)}\left(\frac{y}{Y}\right)\ll(1+\frac{y}{Y})^{-A},\quad x^{j}\Psi^{(j)}\left(\frac{x}{X}\right)\ll(1+\frac{x}{X})^{-B},\end{split} for any positive integers A,B . An asymptotic formula for the above expression is given in [ 5 , Theorem 1.1] , which is valid when X^{1/2}\ll Y\ll X^{1-\varepsilon} for any \varepsilon>0 . Motivated by the results in [ 5 ] and in [ 3 ] . It is the aim of this paper to evaluate asymptotically a smoothed version of S_{m}(X,Y;f) . For this, we fix two non-negative, smooth and compacted supported functions \Phi,\Psi as mentioned above and we define (1.6) \displaystyle S_{f}(X,Y;\Phi,\Psi)=\sideset{}{{}^{*}}{\sum}_{d}\Bigg(\sum_{n}\lambda_{f}(n)\chi_{8d}(n)\Phi\Big(\frac{n}{Y}\Big)\Bigg)^{2}\Psi\Big(\frac{d}{X}\Big). Our main result evaluates S_{f}(X,Y;\Phi,\Psi) asymptotically as follows. Theorem 1.1 . With the notation as above. We have for large X and Y and any \varepsilon>0 , (1.7) \displaystyle S_{f}(X,Y;\Phi,\Psi)=C_{0}(\Phi,\Psi)XY+O(X^{1/2+\varepsilon}Y^{3/2+\varepsilon}+XY^{1/2+\varepsilon}), where C_{0}(\Phi,\Psi) is given in ( 3.15 ), which is a constant depending on \Phi and \Psi only. We note that as often is the case, the most interesting case for evaluating S_{f}(X,Y;\Phi,\Psi) is when X and Y are of compatible size X\asymp Y . Here one checks that ( 1.7 ) yields a valid asymptotic formula when Y\ll X^{1-\varepsilon} for any \varepsilon>0 . Note also in view of ( 1.3 ) and the result of Theorem 1.1 also suggests that one has for all real m\geq 2 , \displaystyle S_{m}(X,Y;f)\ll XY^{m/2}(\log X)^{\frac{m(m-3)}{2}+1}. Our proof of Theorem 1.1 makes use of the Poisson summation formula established in [ 13 ] and follows closely the treatments developed by K. Soundararajan and M. P. Young [ 12 ] in their work on the second moment of quadratic twists of modular L -functions. 2. Preliminaries In this section, we gather a few auxiliary results needed in the proof of Theorem 1.1 . 2.1. Gauss sums For all odd integers m and all integers k , we define the Gauss-type sums G_{k}(m) , as did in Section 2.2 in [ 13 ] , such that (2.1) \displaystyle G_{k}(m)=\left(\frac{1-i}{2}+\left(\frac{-1}{m}\right)\frac{1+i}{2}\right)\sum_{a\mod m}\left(\frac{a}{m}\right)e\left(\frac{ak}{m}\right). We reserve the letter p for a prime number throughout the paper. Let also \varphi(m) be the Euler totient function of m . Our next result evaluates G_{k}(m) , which is taken from [ 13 , Lemma 2.3] . Lemma 2.2 . If (m_{1},m_{2})=1 then G_{k}(m_{1}m_{2})=G_{k}(m_{1})G_{k}(m_{2}) . Suppose that p^{a} is the largest power of p dividing k (put a=\infty if k=0 ). Then for b\geq 1 we have G_{k}(p^{b})=\left\{\begin{array}[]{cl}0&\mbox{if $b\leq a$ is odd},\\ \varphi(p^{b})&\mbox{if $b\leq a$ is even},\\ -p^{a}&\mbox{if $b=a+1$ is even},\\ (\frac{k/p^{a}}{p})p^{a}\sqrt{p}&\mbox{if $b=a+1$ is odd},\\ 0&\mbox{if $b\geq a+2$}.\end{array}\right. 2.3. Poisson Summation For any smooth function F , we define (2.2) \widetilde{F}(\xi)=\int\limits^{\infty}_{-\infty}\left(\cos(2\pi\xi x)+\sin(2\pi\xi x)\right)F(x)dx. We note the following Poisson summation formula from [ 13 , Lemma 2.6] . Lemma 2.4 . Let F be a smooth function compactly supported on {\mathbb{R}}_{+} . We have, for any odd integer n , \sum_{(d,2)=1}\left(\frac{d}{n}\right)F\left(\frac{d}{X}\right)=\frac{X}{2n}\left(\frac{2}{n}\right)\sum_{k}(-1)^{k}G_{k}(n)\widetilde{F}\left(\frac{kX}{2n}\right), where \widetilde{F} is defined in ( 2.2 ) and G_{k}(n) is defined in ( 2.1 ). 2.5. Symmetric square L -functions Recall that f is a fixed holomorphic Hecke eigenform of weight \kappa\equiv 0\pmod{4} for the full modular group SL_{2}(\mathbb{Z}) and that \lambda_{f}(n) is given in ( 1.1 ). The symmetric square L -function L(s,\operatorname{sym}^{2}f) of f is then defined for \Re(s)>1 by (see [ 8 , p. 137] and [ 8 , (25.73)] ) \displaystyle\begin{split}L(s,\operatorname{sym}^{2}f)=&\zeta(2s)\sum_{n\geq 1}\frac{\lambda_{f}(n^{2})}{n^{s}}=\prod_{p}\Big(1-\frac{\lambda_{f}(p^{2})}{p^{s}}+\frac{\lambda_{f}(p^{2})}{p^{2s}}-\frac{1}{p^{3s}}\Big)^{-1},\end{split} where \zeta(s) is the Riemann zeta function. It follows from a result of G. Shimura [ 11 ] that L(s,\operatorname{sym}^{2}f) is holomorphic at s=1 . Moreover, the corresponding completed symmetric square L -function (2.3) \displaystyle\Lambda(s,\operatorname{sym}^{2}f)= \displaystyle\pi^{-3s/2}\Gamma\Big(\frac{s+1}{2}\Big)\Gamma\Big(\frac{s+\kappa-1}{2}\Big)\Gamma\Big(\frac{s+\kappa}{2}\Big)L(s,\operatorname{sym}^{2}f) is entire and satisfies the functional equation \Lambda(s,\operatorname{sym}^{2}f)=\Lambda(1-s,\operatorname{sym}^{2}f) . 2.6. Analytical behaviors of a Dirichlet Series Let q be an integer and \alpha,\beta,\gamma\in\mathbb{C} . We define for any square-free k_{1} , (2.4) Z(\alpha,\beta,\gamma;q,k_{1})=\sum_{k_{2}=1}^{\infty}\sum_{(n_{1},2q)=1}\sum_{(n_{2},2q)=1}\frac{\lambda_{f}(n_{1})\lambda_{f}(n_{2})}{n_{1}^{\alpha}n_{2}^{\beta}k_{2}^{2\gamma}}\frac{G_{k_{1}k_{2}^{2}}(n_{1}n_{2})}{n_{1}n_{2}}, where G_{k}(m) be defined as in ( 2.1 ). We write L_{c}(s,\chi) for the Euler product of L(s,\chi) with the factors from p|c removed. Our next lemma, taken from [