Home On sup-norm bounds part II: GL(2) Eisenstein series
Article
Licensed
Unlicensed Requires Authentication

On sup-norm bounds part II: GL(2) Eisenstein series

  • Edgar Assing EMAIL logo
Published/Copyright: May 7, 2019

Abstract

In this paper we consider the sup-norm problem in the context of analytic Eisenstein series for GL2 over number fields. We prove a hybrid bound which is sharper than the corresponding bound for Maaß forms. Our results generalize those of Huang and Xu where the case of Eisenstein series of square-free levels over the base field had been considered.

MSC 2010: 11F03; 11F70

Communicated by Freydoon Shahidi


A Averaging non-unitary Whittaker new vectors

In this appendix we extend [20, Proposition 2.9] to allow non-unitary principal series representations. This is needed to deal with the Whittaker expansion of Eisenstein series for general s. The computations in this appendix rely heavily on the explicit expressions for the constants ct,l(μ) (defined in [2, (1.6)]) given in [2, Lemma 2.2,2.3]. We will mostly stick to the notation of this paper, with some additions from [2]. Recall for example the matrices gt,l,vG(F𝔭) defined below [2, (1.4)] and the set

𝔛k={ξ:F×S1:ξ(ϖ𝔭)=1 and a(ξ)k}.

All the computations in this appendix are done in a fixed non-archimedean field F𝔭.

For the sake of exposition we consider three cases.

Lemma A.1.

Let πp=χ1χ2 be a principal series representation of G(Fp) with a(χ1)>a(χ2)=0. In this case we have a(ωπ)=a(χ1)=np=mp. For 0lnp and t=-(l+np)+r we have

v𝔬𝔭×|Wπ𝔭(gt,l,v)|2d×v{ζF𝔭(1)q𝔭-r(|χ1(ϖ𝔭r+n𝔭)|2+|χ1(ϖ𝔭-r-n𝔭)|2)if r0,0if r<0.

Proof.

By [2, (1.6)] and character orthogonality it is clear that

(A.1)St,l=v𝔬𝔭×|Wπ𝔭(gt,l,v)|2d×v=μ𝔛l|ct,l(μ)|2.

We insert the expressions for ct,l(μ) given in [2, Lemma 2.3] and consider several cases.

Case (1). The easiest situation occurs when l=0. In this case we have

St,0={q-r|χ1(ϖ-r-n𝔭)|2if r0,0if r<0.

Case (2). For 0<l<n𝔭, we observe that a(μπ𝔭)=n𝔭+a(μ) and obtain

St,l=μ𝔛l,t=-a(μωπ,𝔭)-lζF𝔭(1)2q𝔭-l|χ1(ϖl-n𝔭-a(μ))|2
{ζF𝔭(1)2𝔛lq𝔭-lmax(1,|χ1(ϖ-n𝔭)|2)if r=0,0else.

Recalling 𝔛l=ζF(1)-1q𝔭l yields St,lζF𝔭(1)max(1,|χ1(ϖ-n𝔭)|2).

Case (3). At last we look at l=n. One has

St,n=μ𝔛n𝔭,μωπ𝔭-1,t=-a(μωπ,𝔭)-n𝔭ζF𝔭(1)2q𝔭-n𝔭|χ1(ϖr)|2+δt=-n𝔭-1ζF𝔭(1)2q𝔭t|χ1(ϖ-t-2)|2+δt-n𝔭q𝔭-t-2n𝔭|χ1(ϖt+2n𝔭)|2q𝔭-r|χ1(ϖ𝔭r)|2.

To estimate the first sum, we write a(μωπ𝔭)=n𝔭-r. This corresponds precisely to t=-2n𝔭+r. The sum is empty for r<0. If r0, we use the trivial bound

(A.2){μ𝔛n𝔭{ωπ-1}:a(μωπ𝔭)=n𝔭-r}ζF𝔭(1)-1q𝔭n𝔭-r.

This yields

St,n{2ζF𝔭(1)q𝔭-r|χ1(ϖ𝔭r)|2if r0,0if r<0.

Combining these three estimates completes the proof. ∎

Lemma A.2.

Let πp=χ1χ2 with a(χ1)>a(χ2)>0. For 0lnp and t=-max(2l,mp+l,np)+r we have

v𝔬𝔭×|Wπ𝔭(gt,l,v)|2d×v{ζF𝔭(1)2q𝔭-r(|χ1(ϖ𝔭r+a(χ1))|2+|χ1(ϖ𝔭-r-a(χ1))|2)if r0,0if r<0.

Note that this covers also the analogous case 0<a(χ1)<a(χ2). We remark that the exponents appearing inside of χ1 were not optimized.

Proof.

For convenience we write a(χi)=ai. Note that in this situation n=a1+a2 and m=a1. The strategy is to start from (A.1) and insert the expressions from [2, Lemma 2.2]. Let us first deal with some easy cases.

Case (1). If 0l<a2, we have t=-n𝔭+r and

(A.3)St,l=μ𝔛l,t=-n𝔭ζF𝔭(1)2q𝔭-l|χ1(ϖ𝔭a2-a1)|2{ζF𝔭(1)2|χ1(ϖ𝔭a2-a1)|2if r=0,0else.

Here we used 𝔛l={μ𝔛l:a(μ)=l}ql.

Case (2). Similarly, if a1<ln, we have t=-2l+r and

(A.4)St,l=μ𝔛l,t=-2lζF𝔭(1)2q𝔭-l|χ1(ϖ𝔭0)|2{ζF𝔭(1)2if r=0,0else.

Case (3). Finally, if a2<l<a1, we have t=-(m𝔭+l)+r and

St,l=μ𝔛l,t=-a1-lζF𝔭(1)2q𝔭-l|χ1(ϖ𝔭l-a1)|2{ζF(1)2|χ1(ϖ𝔭l-a1)|2if r=0,0else.

Case (4). Next we consider l=a2. Note that this implies t=-n𝔭+r and a(μχ1)=a1 as well as a(μχ2)=a2-r. The explicit expressions for ct,l(μ) yield

St,a2=μ𝔛a2,μ|𝔬𝔭×χ2|𝔬𝔭×,a(μχ2)=a2-rζF𝔭(1)2q𝔭-a2|χ1(ϖ𝔭a2-a1-r)|2+δr=a2-1ζF𝔭(1)2q𝔭-r-2|χ1(ϖ𝔭a2-a1-r)|2+δra2q𝔭-r|χ1(ϖ𝔭-t-a1)|2q𝔭-r(|χ1(ϖ𝔭a2-a1-r)|2+|χ1(ϖ𝔭-t-a1)|2)
ζF𝔭(1)2q𝔭-r(|χ1(ϖ𝔭a2-a1-r)|2+|χ1(ϖ𝔭-t-a1)|2).

Here we used (A.2) to estimate the μ-sum.

Case (5). The last case to look at is l=a1. We have t=-2l+r, a(μχ2)=a1 and a(μχ1)=a1-r. Therefore,

St,a2=μ𝔛a1,μ|𝔬𝔭×χ1|𝔬𝔭×,a(μχ1)=a1-rζF𝔭(1)2q𝔭-a1|χ1(ϖ𝔭r)|2+δr=a1-1ζF𝔭(1)2q𝔭-r-2|χ1(ϖ𝔭r)|2+δra1q-r|χ1(ϖ𝔭t+a1)|2q-r(|χ1(ϖ𝔭r)|2+|χ1(ϖ𝔭t+a1)|2)
ζF𝔭(1)2q𝔭-r|χ1(ϖ𝔭r)|2+q𝔭-r|χ1(ϖ𝔭t+a1)|2.

This covers all cases and the proof is complete. ∎

Lemma A.3.

Let πp=χ1χ2 with a(χ1)=a(χ2)>0. For 0lnp and t=-max(2l,m+l,np)+r we have

v𝔬𝔭×|Wπ𝔭(gt,l,v)|2d×v{ζF𝔭(1)2rq𝔭-r2(|χ1(ϖ𝔭r+n𝔭)|2+|χ1(ϖ𝔭-r-n𝔭)|2)if r0,0if r<0.

Proof.

For simplicity we write a=a(χ1). In particular, n=2a. Equations (A.3) and (A.4) remain true and cover la. So let us assume l=a. If there is a character μ such that a(μχ1)=a(μχ1)=0, we get the contribution

δr=n𝔭-2q-2-aζF𝔭(1)2+δr=n𝔭-1q𝔭-1-a|χ1(ϖ𝔭)+χ1(ϖ𝔭-1)|2
+δr>n𝔭q𝔭-t-aζF𝔭(1)2|-q𝔭-1ζF𝔭(1)-1(χ1(ϖ𝔭t+2)+χ2(ϖ𝔭t+2))+ζF𝔭(1)-2l=0tχ1(ϖ𝔭l)χ2(ϖ𝔭t-l)|2
tq𝔭-r2(|χ1(ϖ𝔭-r+n𝔭-3)|2+|χ1(ϖ𝔭r-n𝔭+3)|2).

The other exceptional contribution comes from a character μ satisfying a(μχj)a(μχi)=0. In this case we write a(μχj)=a-r0. By assumption we have r0<a so that for ra-1+r0 we have r0r2. This situation contributes

δr=n𝔭-a(μχj)-1ζF𝔭(1)2q𝔭-1-a|χi(ϖn𝔭-r-2)|2+δrn𝔭-a(μχj)q𝔭-a-t-a(μχj)|χi(ϖ𝔭-1)|2q-r2(|χi(ϖ-1)|2+|χi(ϖa)|2).

At last we deal with the contribution of generic μ. Together with the cases above we have

St,a=μ𝔛a,a(μχi)0,t=-a(μχ1)-a(μχ2)ζF𝔭(1)2q𝔭-l|χ1(ϖ𝔭a(μχ2)-a(μχ1))|2=Sg(r)+O(max(t,1)q𝔭-r2(|χ1(ϖ𝔭r+n𝔭)|2+|χ1(ϖ𝔭-r-n𝔭)|2)).

First, we observe that the generic characters only contribute when r<n-1. We write a(μχi)=a-ri for 0r1,r2<a and define

𝔛r1,r2={μ𝔛a:a(μχi)=a-ri}.

One has the trivial bound

𝔛r1,r2ζF𝔭(1)-1q𝔭a-max(r1,r2).

Therefore, by ordering the summation in Sg(r) accordingly, we obtain

Sg(r)=0r1,r2<a,r1+r2=r𝔛r1,r2ζF𝔭(1)2q𝔭-l|χ1(ϖ𝔭a(μχ2)-a(μχ1))|2
ζF𝔭(1)rq𝔭-r2(|χ1(ϖ𝔭r)|2+|χ1(ϖ𝔭-r)|2).

The stated inequality follows after putting everything together. ∎

The last three lemmata together imply an extension of [20, Proposition 2.10].

Proposition A.1.

Let πp=χ1χ2 such that ωπp(ϖp)=1 and gGL2(Fp) such that

t(g)=-max(2l(g),l(g)+m𝔭,n𝔭)+r(g).

Then we have:

  1. If Wπ𝔭(g)0, then r(g)0.

  2. If r(g)0, then

    (𝔬𝔭×|Wπ𝔭(a(v)g)|2d×v)12ζF𝔭(1)max(1,r)q𝔭-r4(|χ1(ϖ𝔭r+n𝔭)|+|χ1(ϖ𝔭-r-n𝔭)|).

Proof.

First we note that there is w𝔬𝔭, zZ(F𝔭), xF𝔭 and kK1,𝔭(n𝔭) such that

g=zn(x)gt(g),l(g),wk.

By the transformation behavior of Wπ𝔭 we get

|Wπ𝔭(a(v)g)|2=|Wπ𝔭(gt(g),l(g),wv-1)|2.

With this at hand the proposition follows from the previous lemmata. ∎

Acknowledgements

I would like to thank the referee for carefully reading the manuscript, which has led to a significant improvement of this work.

References

[1] E. Assing, On sup-norm bounds part I: Ramified Maaß newforms over number fields, preprint (2017), https://arxiv.org/abs/1710.00362. Search in Google Scholar

[2] E. Assing, On the size of p-adic Whittaker functions, Trans. Amer. Math. Soc. (2018), 10.1090/tran/7685. 10.1090/tran/7685Search in Google Scholar

[3] V. Blomer, G. Harcos, P. Maga and D. Milićević, The sup-norm problem for GL(2) over number fields, preprint (2016), https://arxiv.org/abs/1605.09360. 10.4171/JEMS/916Search in Google Scholar

[4] D. Bump, Automorphic Forms and Representations, Cambridge Stud. Adv. Math. 55, Cambridge University, Cambridge, 1997. 10.1017/CBO9780511609572Search in Google Scholar

[5] M. D. Coleman, A zero-free region for the Hecke L-functions, Mathematika 37 (1990), no. 2, 287–304. 10.1112/S0025579300013000Search in Google Scholar

[6] E. Fogels, Über die Ausnahmenullstelle der Heckeschen L-Funktionen, Acta Arith. 8 (1962/1963), 307–309. 10.4064/aa-8-3-307-309Search in Google Scholar

[7] S. Gelbart and H. Jacquet, Forms of GL(2) from the analytic point of view, Automorphic Forms, Representations and L-functions (Corvallis 1977), Proc. Sympos. Pure Math. 33 Part 1, American Mathematical Society, Providence (1979), 213–251. 10.1090/pspum/033.1/546600Search in Google Scholar

[8] S. S. Gelbart and E. M. Lapid, Lower bounds for L-functions at the edge of the critical strip, Amer. J. Math. 128 (2006), no. 3, 619–638. 10.1353/ajm.2006.0024Search in Google Scholar

[9] I. S. Gradshteyn and I. M. Ryzhik, Table of Integrals, Series, and Products, 7th ed., Elsevier/Academic Press, Amsterdam, 2007. Search in Google Scholar

[10] G. Harcos and P. Michel, The subconvexity problem for Rankin–Selberg L-functions and equidistribution of Heegner points. II, Invent. Math. 163 (2006), no. 3, 581–655. 10.1007/s00222-005-0468-6Search in Google Scholar

[11] J. G. Hinz, Eine Erweiterung des nullstellenfreien Bereiches der Heckeschen Zetafunktion und Primideale in Idealklassen, Acta Arith. 38 (1980/81), no. 3, 209–254. 10.4064/aa-38-3-209-254Search in Google Scholar

[12] B. Huang and Z. Xu, Sup-norm bounds for Eisenstein series, Forum Math. 29 (2017), no. 6, 1355–1369. 10.1515/forum-2015-0195Search in Google Scholar

[13] H. Iwaniec and E. Kowalski, Analytic Number Theory, Amer. Math. Soc. Colloq. Publ. 53, American Mathematical Society, Providence, 2004. 10.1090/coll/053Search in Google Scholar

[14] H. Iwaniec and P. Sarnak, L norms of eigenfunctions of arithmetic surfaces, Ann. of Math. (2) 141 (1995), no. 2, 301–320. 10.2307/2118522Search in Google Scholar

[15] J. Jung, Quantitative quantum ergodicity and the nodal domains of Hecke–Maass cusp forms, Comm. Math. Phys. 348 (2016), no. 2, 603–653. 10.1007/s00220-016-2694-8Search in Google Scholar

[16] W. Z. Luo and P. Sarnak, Quantum ergodicity of eigenfunctions on PSL2()2, Publ. Math. Inst. Hautes Études Sci. (1995), no. 81, 207–237. 10.1007/BF02699377Search in Google Scholar

[17] J. Neukirch, Algebraic Number Theory, Grundlehren Math. Wiss. 322, Springer, Berlin, 1999. 10.1007/978-3-662-03983-0Search in Google Scholar

[18] Y. N. Petridis and M. S. Risager, Averaging over Heegner points in the hyperbolic circle problem, Int. Math. Res. Not. IMRN 2018 (2018), no. 16, 4942–4968. 10.1093/imrn/rnx026Search in Google Scholar

[19] A. Saha, Large values of newforms on GL(2) with highly ramified central character, Int. Math. Res. Not. IMRN 2016 (2016), no. 13, 4103–4131. 10.1093/imrn/rnv259Search in Google Scholar

[20] A. Saha, Hybrid sup-norm bounds for Maass newforms of powerful level, Algebra Number Theory 11 (2017), no. 5, 1009–1045. 10.2140/ant.2017.11.1009Search in Google Scholar

[21] P. Sarnak, Letter to Morawetz, 2004. Search in Google Scholar

[22] R. Schmidt, Some remarks on local newforms for GL(2), J. Ramanujan Math. Soc. 17 (2002), no. 2, 115–147. Search in Google Scholar

[23] E. C. Titchmarsh, The Theory of the Riemann Zeta-Function, 2nd ed., The Clarendon Press, New York, 1986. Search in Google Scholar

[24] J. L. Truelsen, Quantum unique ergodicity of Eisenstein series on the Hilbert modular group over a totally real field, Forum Math. 23 (2011), no. 5, 891–931. 10.1515/form.2011.031Search in Google Scholar

[25] H. Wu, Burgess-like subconvex bounds for GL2×GL1, Geom. Funct. Anal. 24 (2014), no. 3, 968–1036. 10.1007/s00039-014-0277-4Search in Google Scholar

[26] M. P. Young, A note on the sup norm of Eisenstein series, Q. J. Math. 69 (2018), no. 4, 1151–1161. 10.1093/qmath/hay019Search in Google Scholar

[27] L. Zhang, Quantum Unique Ergodicity of Degenerate Eisenstein Series on GL(n), ProQuest LLC, Ann Arbor, 2017. Search in Google Scholar

Received: 2018-01-16
Revised: 2019-03-10
Published Online: 2019-05-07
Published in Print: 2019-07-01

© 2019 Walter de Gruyter GmbH, Berlin/Boston

Downloaded on 14.9.2025 from https://www.degruyterbrill.com/document/doi/10.1515/forum-2018-0014/html
Scroll to top button