Segal-Bargmann and Weyl transforms on compact Lie groups (Q1034729): Difference between revisions

Let \(K\) be a connected and compact Lie group and suppose that it is contained in its complexification \(\widetilde K= K_{\mathbb{C}}\). The authors introduce the Fock space associated with \(K\) and give a simple proof of the unitarity of the Segal-Bargmann transform proved by B. C. Hall for the group \(K\). First, let \(\Delta_K\) be the \(K\)-invariant Laplacian operator on \(K\). Let \(\rho_t: K\to \mathbb{C}\) be a heat kernel of \(\Delta_K\), i.e., a solution of the equation \(\partial\rho_t/\partial t= 1/2\Delta_K\rho_t\), \(\lim_{t\to 0} \rho_t=\delta_1\), where \(\delta_1\) is the delta distribution at the unit element 1 of \(K\). After recalling a result of \textit{B. C. Hall} [J. Funct. Anal. 122, No. 1, 103--151 (1994; Zbl 0838.22004)], the authors assert that \(\rho_t(g)\) can be holomorphically extended to \(\widetilde K\). Let \(\mu_t\) \((t> 0)\) be a heat kernel of the Laplacian \(\Delta_{\widetilde K}\) on \(K_{\mathbb{C}}\). By a theorem in the book by \textit{D. W. Robinson} [Elliptic operators and Lie groups, Oxford etc.: Clarendon Press (1991; Zbl 0747.47030)] we know that \(\mu_t\) is a positive function for each \(t> 0\). We define \(\nu_t: K\to\mathbb{C}\) by \[ \nu_t(g)= \int_K\mu_t(gk)\,dk. \] Now, we define the Fock space \(F(\widetilde K,\nu_t)\) \((t> 0)\) for the group \(K\) by the space of all holomorphic functions \(f:\widetilde K\to\mathbb{C}\) such that \[ \int_K|f(z)|^2\nu_t(z)\, dz< \infty, \] where \(dz\) denotes a Haar measure on \(\widetilde K\). The authors prove that the Fock space \(F(\widetilde K,\nu_t)\) has a reproducing kernel \(E_t\) given by \(E_t(g,h)= \rho_{2t}(g\overline{h^{-1}})\) for \(g,h\in\widetilde K\). In fact, they prove that \[ E_t(g,h)= \int_K\rho_t(k^{-1} g)\overline{\rho_t(k^{-1} h)}\,dk \] holds for \(g\), \(h\in\widetilde K\). Next, they prove that there exists a linear map \(C_t: L^2(K)\to F(\widetilde K,\nu_t)\) which is a unitary \(K\)-equivariant operator. The authors call \(C_t\) the Segal-Bargmann transformation. To define the Weyl transform we put \(\omega_t(g)= E_t(g,g)\nu_t(g)\) for \(g\in K\) and consider the space \(L^2(\widetilde K,\omega_t(g)\,dg)\). The authors define an operator \(W: L^2(\widetilde K,\omega_t dg)\to F(\widetilde K,\nu_t)\otimes \overline{F(\widetilde K,\nu_t)}\), which they call the Weyl transform, and then they prove the boundedness of this operator.