Maass-Jacobi forms over complex quadratic fields (Q2372679)

The aim of this paper is to show that the Jacobi theta functions used in [\textit{H. Skogman}, Acta Arith. 96, No. 4, 333--350 (2001; Zbl 1077.11035)] to construct Jacobi forms over imaginary quadratic fields are Maass forms, i.e.\ are eigenfunctions for a suitable differential operator invariant under the Jacobi group. To do this the authors have to use methods of representation theory to compute the relevant differential operators and show that the functions satisfy the resulting PDE. They do this in reverse order, first writing down the operators and then justifying them later. For an imaginary quadratic field \(K\) the Jacobi group \(\Gamma^J(K)=\text{ SL}_2({\mathcal O}_K)\ltimes{\mathcal O}_K^2\) acts on \({\mathbb H}_{\mathcal Q}\times{\mathcal Q}\), where \({\mathcal Q}\) is the ring of quaternions and \({\mathbb H}_{\mathcal Q}\) is the quaternionic upper half-plane. This extends the action of \(\text{ SL}_2({\mathcal O}_K)\) on \({\mathbb H}_{\mathcal Q}\), and a choice of \(k\in{\mathbb C}\) and a vector \(m\in{\mathbb C}^l\) with \({}^tmm\in{\mathcal O}_K\) defines an action \(| _{k,m}\) (the slash action) of \(\Gamma^J(K)\) on functions \(f:{\mathbb H}_{\mathcal Q}\times{\mathcal Q} \to {\mathbb C}\). The authors write down a basis \(\{D^z_{k,m},D^\tau_{k,m}\}\) for the space of 2nd-order differential operators on \({\mathbb H}_{\mathcal Q}\times{\mathcal Q}\) invariant for \(| _{k,m}\), and define a \((k,m)\)-Maass-Jacobi form to be a \(C^\infty\) function \(f\) that is both invariant under \(| _{k,m}\) and an eigenfunction for \(D^z_{k,m}\) (and satisfies a boundedness condition). They define Jacobi-Eisentein series \(E_{k,m,s}(\tau,z)\) for \(s\in{\mathbb C}\) under the assumptions that \(k\in{\mathbb Z}\), \(m\) is real and \({}^tmm\in{\mathbb N}\). These are shown to converge for \(\Re\, s>7/2\) and to have a Fourier expansion, which is computed. The resulting functions are Maass-Jacobi forms. The expected meromorphic continuation and functional equation are shown for the case \(K={\mathbb Q}(i)\) by direct methods, but an appeal to general theory fails because \(\Gamma^J(K)\) is not reductive. In the rest of the paper, which is fully half of it, the authors compute \(D^z_{k,m}\) and \(D^\tau_{k,m}\). They first describe the general setting in terms of Lie groups. This part of the paper (Section~4) is independent of the rest, and this independence extends to the notation: \(K\) is used for a compact subgroup of a real Lie group \(G\) and \(k\) for an element of \(K\). Differential operators on a bundle over \(G/K\) appear as sections of another bundle, and the situation is analysed carefully. In Section~5 this is applied to the case in hand: \(G\) becomes the complex Jacobi group, considered as a real Lie group. The specific calculations are intricate and much detail is sensibly omitted: even checking that the slash action is indeed a group action is nontrivial.