Zeros of period polynomials for symmetric power L L -functions
Robert Dicks, Hui Xue · 2025
Read Zeros of period polynomials for symmetric power L L -functions by Robert Dicks, Hui Xue (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-2501-18024.txt.
Introduction
Robert Dicks and Hui Xue work in number theory, on where the zeros of certain polynomials fall. Their study appeared in twenty twenty-five. The polynomials come from modular forms, and their values encode key numbers of an attached L-function.
For the classic case, all zeros lie on one neat circle, a pattern much like the famous Riemann hypothesis. The authors build a matching polynomial for the symmetric powers of the L-function, when the level is squarefree. Their version keeps the same symmetry between two linked points. They then prove the circle result holds when the weight is large enough. For very large weights, they also show the result holds once the level grows large.
Catalog record
| CREATOR | Robert Dicks, Hui Xue |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2501-18024
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Zeros of period polynomials for symmetric power L L -functions Robert Dicks, Hui Xue Abstract . Suppose that k k and N N are positive integers. Let f f be a newform on Γ 0 ( N ) \Gamma_{0}(N) of weight k k with L L -function L f ( s ) L_{f}(s) . Previous works have studied the zeros of the period polynomial r f ( z ) r_{f}(z) , which is a generating function for the critical values of L f ( s ) L_{f}(s) and has a functional equation relating z z and − 1 / N z -1/Nz . In particular, r f ( z ) r_{f}(z) satisfies a version of the Riemann hypothesis: all of its zeros are on the circle of symmetry { z ∈ ℂ : | z | = 1 / N } \{z\in\mathbb{C}\ :\ |z|=1/\sqrt{N}\} . In this paper, for a positive integer m m , we define a natural analogue of r f ( z ) r_{f}(z) for the m th m^{\operatorname{th}} symmetric power L L -function of f f when N N is squarefree. Our analogue also has a functional equation relating z z and − 1 / N z -1/Nz . We prove the corresponding version of the Riemann hypothesis when k k is large enough. Moreover, when k > 2 ( log 2 ( 13 e 2 π / 9 ) + m ) + 1 k>2(\operatorname{log}_{2}(13e^{2\pi}/9)+m)+1 , we prove our result when N N is large enough. 2020 Mathematics Subject Classification 11F11, 11F67 1. Introduction For positive integers k and N , let S_{k}(N) denote the complex vector space of cusp forms of weight k on \Gamma_{0}(N) . Let f\in S_{k}(N) be a newform with Fourier expansion f=\sum_{n\geq 1}a(n)q^{n} (see Section 2 for details). The L -function for f , which is given by (1.1) L_{f}(s):=\sum^{\infty}_{n=1}\frac{a(n)}{n^{s}}=\left(\prod_{p\mid N}\frac{1}{1-a(p)p^{-s}}\right)\prod_{p\nmid N}\frac{1}{1-a(p)p^{-s}+p^{k-1-2s}}, has an analytic continuation to the complex plane. Moreover, the completed L -function L^{*}_{f}(s):=\left(\frac{\sqrt{N}}{2\pi}\right)^{s}\Gamma(s)L_{f}(s) satisfies the functional equation (1.2) L^{*}_{f}(s)=\epsilon_{f}L^{*}_{f}(k-s) for some \epsilon_{f}\in\{\pm 1\} . The completed L -function can be written as (1.3) L^{*}_{f}(s)=N^{s/2}\int^{\infty}_{0}f(iy)y^{s}\frac{dy}{y}, which means that it arises as a “period integral” of f (see [ KZ01 ] for more information on period integrals). The period polynomial of f is the degree k-2 polynomial r_{f}(z):=\int^{i\infty}_{0}f(\tau)(\tau-z)^{k-2}d\tau. Using ( 1.3 ) and expanding (\tau-z)^{k-2} , we can write r_{f}(z) as (1.4) r_{f}(z)=i^{k-1}N^{-\frac{k-1}{2}}\sum^{k-2}_{n=0}{\binom{k-2}{n}}(\sqrt{N}iz)^{n}L^{*}_{f}(k-1-n) or as r_{f}(z)=-\frac{(k-2)!}{(2\pi i)^{k-1}}\sum^{k-2}_{n=0}\frac{(2\pi iz)^{n}}{n!}L_{f}(k-1-n). Thus, r_{f}(z) is a generating function for the critical values L_{f}(1) , L_{f}(2) , \dots , L_{f}(k-1) . These values are crucially important in arithmetic geometry and number theory [ JMOS16 ] [ LMT17 ] . For example, they appear in the context of the Bloch-Kato conjecture for special values of L -functions which generalizes the Birch and Swinnerton-Dyer conjecture relating Tate-Shafarevich groups of elliptic curves to analytic properties of modular L -functions (see [ BK90 ] for more details). Moreover, by virtue of the functional equation ( 1.2 ) and ( 1.4 ), we see that (1.5) r_{f}(z)=-\epsilon_{f}i^{k}(\sqrt{N}z)^{k-2}r_{f}\left(-\frac{1}{Nz}\right). Jin, Ma, Ono, and Soundararajan [ JMOS16 ] showed that r_{f}(z) satisfies an analogue of the Riemann hypothesis: all of its zeros are located on the circle \{z\in\mathbb{C}\ :\ |z|=1/\sqrt{N}\} , the circle of symmetry for the functional equation ( 1.5 ). In this paper, for a positive integer m , we define an analogue of r_{f}(z) for the m^{\operatorname{th}} symmetric power L -function and prove the corresponding version of the Riemann hypothesis when N is squarefree and when either k is large enough, or N is large enough assuming that k>2(\operatorname{log}_{2}(13e^{2\pi}/9)+m)+1 . Stating our result requires introducing notation. Suppose that k is a positive integer, that N is a squarefree positive integer and that f=\sum_{n\geq 1}a(n)q^{n}\in S_{k}(N) is a newform. For each prime p , we denote by \alpha_{p} and \beta_{p} the roots of the polynomial X^{2}-a(p)X+p^{k-1} when p\nmid N ; otherwise, we define \alpha_{p}=a(p) and \beta_{p}=0 . Let m be a positive integer. The m^{\operatorname{th}} symmetric power L -function of f is defined by (1.6) L_{m,f}(s):=\prod_{p}\prod_{0\leq i\leq m}(1-\alpha^{m-i}_{p}\beta^{i}_{p}p^{-s})^{-1}, which converges absolutely for {\rm Re}(s)>\frac{m(k-1)}{2}+1 [ Xia16 , pg. 224 ] . Note that when m=1 we have L_{1,f}(s)=L_{f}(s) . Newton and Thorne [ NT21 , Theorem A ] showed that L_{m,f}(s) has an analytic continuation to the whole complex plane. Moreover, the completed m^{\operatorname{th}} symmetric power L -function L_{m,f}^{*}(s):=N^{\frac{ms}{2}}\gamma_{m}(s)L_{m,f}(s), satisfies the functional equation (1.7) L_{m,f}^{*}(s)=\epsilon_{m,f}L^{*}_{m,f}(m(k-1)+1-s) for some \epsilon_{m,f}\in\{\pm 1\} . Here [ Zag77 , pg. 118 ] (1.8) \gamma_{m}(s):=\begin{cases}(2\pi)^{-rs}\displaystyle\prod^{r-1}_{j=0}\Gamma(s-j(k-1))&\text{if }m=2r-1,\\ \pi^{-\frac{s}{2}}\Gamma\left(\frac{s}{2}-\left\lfloor\frac{r(k-1)}{2}\right\rfloor\right)\gamma_{2r-1}(s)&\text{if }m=2r.\end{cases} Zagier defines \gamma_{m}(s) only for N=1 , but it does not depend on N (see e.g. [ Xia16 , (1.6) ] , where the gamma factor differs from ours by a constant multiple). The main object of our investigation is (1.9) R_{m,f}(z):=i^{m(k-1)}N^{-\frac{m(k-1)}{2}}{\sum}^{{m(k-1)-1}}_{n=0}{\binom{m(k-1)-1}{n}}(\sqrt{N}iz)^{n}L^{*}_{m,f}(m(k-1)-n), which is our analogue of r_{f}(z) . In fact, it follows from ( 1.4 ) that R_{1,f}(z)=r_{f}(z) ; moreover the polynomial R_{m,f}(z) satisfies a functional equation similar to ( 1.5 ) (see Lemma 2.3 below). With this notation, our main result is as follows. Theorem 1.1 . Suppose that k , m and N are positive integers such that m>1 and N is squarefree. Let f\in S_{k}(N) be a newform. Then R_{m,f}(z) has all of its zeros on the circle \{z\in\mathbb{C}\ :\ |z|=1/\sqrt{N}\} if k is large enough. The result also holds for large enough N if k>2(\operatorname{log}_{2}(13e^{2\pi}/9)+m)+1 . Remark . (1) When m=1 , the result is true for all k and N , see [ JMOS16 , Theorem 1.1 ] . (2) Löbrich, Ma, and Thorner [ LMT17 ] have done related work for special values of motivic L -functions. For a pure motive \mathcal{M} over \mathbb{Q} of rank d\geq 2 , global conductor N\in\mathbb{Z}^{+} and odd weight w , they define a polynomial p_{\mathcal{M}}(z) which resembles r_{f}(z/i\sqrt{N}) by using the completed motivic L -function \Lambda(s,\mathcal{M}) and the Hodge numbers for \mathcal{M} . Assuming that the (usual) motivic L -function L(s,\mathcal{M}) coincides with the L -function of an algebraic tempered cuspidal symplectic representation of \operatorname{GL}_{d}(\mathbb{A}_{\mathbb{Q}}) , they prove for a large number of cases that the zeros are on the unit circle and are equidistributed. Moreover, their work has applications to odd symmetric power L -functions of cuspidal newforms. More precisely, when f is a non-CM newform of squarefree level N\geq 13 and weight k\geq 6 with integral coefficients whose L -function coincides with that of an algebraic tempered cuspidal symplectic representation of \operatorname{GL}_{n+1}(\mathbb{A}_{\mathbb{Q}}) , then all of the zeros of their polynomial are on the unit circle and they are equidistributed as n or N goes to \infty (they have more complicated results in weights 2 and 4 ). Our work differs from theirs in that our definition is slightly simpler (the key difference being the lack of an analogue of Hodge numbers in our definition) and we have no restriction on the Fourier coefficients of the forms under our consideration. (3) Computations in Sage for N=1 show that the zeros of R_{2,f}(z) are very close to the unit circle (if not on it). It is natural to suspect that Theorem 1.1 holds in this case and to investigate whether or not it holds for other small values of k and N .