Functional equations for quantum theta functions (Q1769584)

Quantum theta functions were introduced by \textit{Yuri I. Manin} [Prog. Theor. Phys., Suppl. 102, 219--228 (1990; Zbl 0796.14005)]. They are certain elements in the function rings of quantum tori. By definition, they satisfy a version of the classical functional equations involving shifts by multiplicative periods. This contribution addresses two interrelated questions: (a) What is the connection between quantum theta functions and theta vectors? (b) Does there exists a quantum analog of the classical functional equation for thetas (related to the action of the metaplectic group)? Briefly, (partial) answers the author gives here look as follows. (i) Schwarz's theta vectors are certain elements of projective modules over \(C^{\infty}\)- or \(C^{*}\)-rings of unitary quantum tori. When such a module is induced from the basic Heisenberg representation by a lattice embedding into a vector Heisenberg group, the respective theta vectors \(f_{T}\) are parametrized by the points \(T\) of the Siegel upper half space, and in different models of the basic representation take the form of a ``quadratic exponent'' \(e^{\pi i x ^{t}T x},\) a classical theta, or Fock vacuum state; (ii) The classical functional equation relates two thetas considered as sections of line bundles over two isomorphic complex tori (Fourier series). Bundles and sections are lifted to the universal covers which are then identified compatibly with period lattices. Consider a classical theta function of \(z \in {\mathbb C}^{N}\) \[ \theta(z, \Omega) = \sum_{n \in {\mathbb Z}^{N}} e^{\pi i n ^{t}\Omega n + 2\pi i n^{t}z}, \tag{0.1} \] where \(\Omega\) is a symmetric complex matrix with positive defined imaginary part. This function satisfies two sets of functional equations. Firstly, for all \(m \in {\mathbb Z}^{N}\) \[ \theta(z + m, \Omega) = \theta(z, \Omega), \] \[ \theta(z + \Omega m, \Omega) = e^{-\pi i m ^{t}\Omega m - 2\pi i m ^{t}z} \theta(z, \Omega). \] Secondly, \[ \theta(\Omega^{-1} z, -\Omega^{-1}) = (det(\Omega/i))^{1/2} e^{\pi i z ^{t}\Omega^{-1} z } \theta(z, \Omega).\tag{0.4} \] In fact, (0.4) is the most important special case of a more general modular functional equation related to the action of \(Sp(2N, {\mathbb Z}).\) In this paper, the author proposes an analog of the functional equation (0.4) corresponding this time to the change \(A \rightarrow A^{-1}\) of the quantization matrix. The noncommutative geometric context replacing the classical isomorphism of triples \(({\mathcal T}, {L}, \theta),\) consisting of a complex torus, a line bundle, and its section, involves now the (strong) Morita equivalence of the relevant quantum tori, compatible complex structures on these quantum tori, and theta vectors in the respective projective bimodule. Let \({K}\) (resp. \({Z}\)) be an Abelian group written additively (resp. multiplicatively). Morever, describe \({G}\) as a central extension of \({K}\) by \({Z}\): \[ 1 \rightarrow {Z} \rightarrow {G} \rightarrow {K} \rightarrow 1. \] Theorem 2.2: The completion of \({H}_{\infty}\) with respect to the norm \[ \| \Phi\| ^{2}: = \| _{D}\langle \Phi, \Phi \rangle\| = \| \sum_{h \in D} \langle \Phi, e_{D,\alpha}(h)\Psi \rangle e_{D,\alpha}(h)\| \;\text{(the norm in } C^{\infty}(D, \alpha)) \] is a finitely generated projective left \(C^{*}(D, \alpha)\)-module \(P.\) The scalar product has a natural extension to \(_{D}\langle , \rangle : P \times P \rightarrow C^{*}(D, \alpha).\) The properties : (i) Symmetry: \(_{D}\langle \Phi, \Psi \rangle^{*} = _{D}\langle \Psi, \Phi \rangle;\) (ii) Bilinearity: \( _{D}\langle a \Psi, \Phi \rangle = a _{D} \langle \Psi, \Phi \rangle\) for any \(a \in C^{\infty}(D, \alpha);\) (iii) Positivity: \(_{D}\langle \Phi, \Phi \rangle\) belongs to the cone of positive elements of \(C^{\infty}(D, \alpha).\) Moreover, if \(_{D}\langle \Phi, \Phi \rangle = 0\) then \(\Phi = 0;\) (iv) Density: The image of \(\langle , \rangle\) is dense in \(C^{\infty}(D, \alpha),\) hold for this extension as well. Theorem 3.2.1: (i) We have \[ _{D}\langle f_{T}, f_{T} \rangle = \frac{1}{\sqrt{2^{N}\det \operatorname{Im} T}} \sum_{h \in D} e^{-\frac{\pi}{2} H(\underline{h}, \underline{h})} e_{D,\alpha}(h). \] Moreover, \[ \Theta_{D}: = \sum_{h \in D} e^{-\frac{\pi}{2} H(\underline{h}, \underline{h})} e_{D,\alpha}(h) \] is a quantum theta function in the ring \(C^{\infty}(D, \alpha)\) satisfying the following functional equations: \[ \forall g \in D: C_{g} e_{D,\alpha}(g) x^{*}_{g}(\Theta_{D}) = \Theta_{D}, \] where \[ C_{g} = e^{-\frac{\pi}{2} H(\underline{g}, \underline{g})}, x^{*}_{g}(e_{D,\alpha}(h)) = e ^{-\pi H(\underline{g}, \underline{h})} e_{D,\alpha}(h). \tag{3.10} \] (ii) We have \[ \Theta_{D} {\mathbf 1} = \sum_{h \in D} e^{ -\pi H(\underline{h}, \underline{h}) - \pi H(\underline{x}, \underline{h})}, \] where \textbf{1} is the vacuum vector in the model \(\text{II}_{T}\) represented by the function identically equal to 1. Theorem 3.5.1: Let \(T(D,\alpha)\) be a quantum torus over \({\mathbb C},\) with unitary quantization form \(\alpha.\) Let \({L} : B \rightarrow {G}(D,\alpha)\) be an ample multiplier such that the left representative projection \(B \rightarrow D\) is an isomorphism which we will use to identify \(B\) with \(D.\) Assume moreover that one can define a Kähler structure on \({\mathbb R} \times D\) such that \[ \langle g, h \rangle = e^{-\pi H(\underline{g}, \underline{h})}, \alpha(g, h) = e^{\pi i \Im H(\underline{g}, \underline{h})} \] for all \(g, h \in D\) holds (\(\langle \,,\, \rangle\) being the structure form of \({L}).\) Then there exists a real space \({K}\) endowed with a bicharacter \(\psi,\) an compatible Kähler structure, and a lattice embedding \(D \subset {K}\) such that \(\psi\) induces \(\alpha\) on \(D,\) and an appropriate generator of \(\Gamma({L})\) is of the form \(\Theta_{D}\) as above. Theorem 3.6.1: (i) The scalar products \(_{D}\langle f_{T,a}, f_{T,b} \rangle\) are quantum theta functions belonging to the space \(\Gamma({L})\) where \({L}\) is the multiplier \[ D_{0} \rightarrow {G}(D,\alpha),\;g \mapsto [C_{g}; x_{g}, g], \] \(C_{g}, x_{g}\) being defined by (3.10), with \(H\) is lifted to \({\mathbb R}^{N} \times F \times \widehat{F}\) via projection. (ii) The scalar products \(_{D}\langle f_{T,a}, f_{T,b} \rangle\) form a basis of \(\Gamma({L}).\)