Weighted spectral large sieve inequalities for Hecke congruence subgroups of \(\mathrm{SL}(2,\mathbb{Z}[i])\) (Q351242)

The theory of Kloosterman sums and sums of Kloosterman sums has been the subject of very intensive study in analytic number theory. The summation formula of \textit{R. W. Bruggeman} [Invent. Math. 45, 1--18 (1978; Zbl 0351.10019)] and \textit{N. V. Kuznetsov} [Mat. Sb., N. Ser. 111(153), 334--383 (1980; Zbl 0427.10016)] revealed a deep connection between Kloosterman sums and spectral theory of automorphic forms. Results of Bruggeman and Kuznetsov were generalized by \textit{R. W. Bruggeman} and \textit{Y. Motohashi} [Funct. Approximatio, Comment. Math. 31, 23--92 (2003; Zbl 1068.11057)] from the case of the modular group \(\mathrm{PSL}(2, \mathbb{Z})\) to the group \(\mathrm{PSL}(2, \mathbb{Z}[i])\). The first formula of Bruggeman and Motohashi shows that a class of sums involving Fourier coefficients of modular forms may be expressed in terms of sums of appropriately defined Kloosterman sums. The second formula is a kind of inverse, showing that a wide class of sums of Kloosterman sums may be represented in terms of the spectral data of the group \(\mathrm{PSL}(2, \mathbb{Z}[i])\). Those results were further generalized by \textit{H. Lokvenc-Guleska} in her PhD thesis [Sum formula for \(\mathrm{SL}_2\) over imaginary quadratic number fields, University of Utrecht (2004)] to Hecke congruence subgroups \(\Gamma\) of the group \(\mathrm{SL}(2, \mathfrak{O}_F)\) where \(\mathfrak{O}_F\) denotes the ring of integers of imaginary quadratic number field \(F=\mathbb{Q}(\sqrt{D})\). In the paper under review several applications of the Lokvenc-Guleska summation formulas in the case when \(F=\mathbb{Q}(i)\) are derived. They include new bounds for sums involving Fourier coefficients of modular forms and new bounds for weighted sums of generalized Kloosterman sums. The main feature of these bounds is that they depend upon the best available lower bound for the first non-zero eigenvalue of the Laplacian on the space \(L^2(\Gamma \setminus \mathrm{SL}(2, \mathbb{C}) / \mathrm{SU}(2))\), where \(\Gamma\) is a Hecke congruence subgroup of \(\mathrm{SL}(2, \mathbb{Z}[i])\). The main result of the paper is an upper bound for the sum of certain weighted generalized Kloosterman sums, whose precise statement is too complicated to be stated here. It may be considered as a generalization of the author's results from [``Kloosterman sums and a mean value for Dirichlet polynomials'', J. Number Theory 53, No. 1, 179--210 (1995; Zbl 0837.11050)] and has a number of applications. For example, the obtained upper bound for the sum of generalized Kloosterman sums yields a new upper bound for the mean value \[ J(D;N)=\frac{1}{D^2} \sum_{-D\leq k\leq D} \int_{-D}^D \left | \sum_{0 < | n| ^2 \leq N} \alpha_n \lambda^k(n) | n | ^{-2it} \right |^2 | \zeta(1/2 +it, \lambda ^k)|^4 \,dt, \] where the summation runs over all \(n\in\mathbb{Z}[i]\), the coefficients \(\alpha_n\) are arbitrary complex numbers, \(\lambda^k\) denotes Hecke ``Größencharacter'' given by \(\lambda^k(n)=(n/| n|)^{4k}\), for \(0\neq n \in \mathbb{Z}[i]\) and \(\zeta(s,\lambda^k)\) is the corresponding Hecke zeta function. This represents a generalization of a Riemann zeta-function result of the author (loc. cit.) and of \textit{J.-M. Deshouillers} and \textit{H. Iwaniec} [Mathematika 29, 202--212 (1982; Zbl 0506.10032)]. The proof of the main results makes extensive use of generalization of Bruggeman-Motohashi summation formulae for \(\mathrm{PSL}(2, \mathbb{Z}[i])\), bounds of Kim and Shahidi for the first non-zero eigenvalues of the Laplacian and an ``unweighted'' spectral large sieve inequality. The paper is written very carefully and contains many technical results. We refer an interested reader to consult the full text of the paper.