Group bundle duality (Q733363)

Suppose that \(G\) is a locally compact Hausdorff groupoid. We say that \(p:G\to X\) is a group bundle if \(X\) is the unit space of \(G\) and the range and source maps coincide with \(p\). Note that \(G_{x}:=p^{-1}(x)\) is a locally compact group for each \(x\in X\). If \(G\) has a continuous Haar system \(\beta=\{\beta^{x}\}_{x\in X}\), and if each \(G_{x}\) is abelian, then the groupoid \(C^{*}\)-algebra, \(C^{*}(G,\beta)\), is abelian, and its spectrum, \(\widehat G:=C^{*}(G,\beta)^{\wedge}\), is a locally compact Hausdorff space. As was observed in Section~3 of [\textit{P. Muhly, J. N. Renault} and \textit{D. P. Williams}, Trans. Am. Math. Soc. 348, No.~9, 3621--3641 (1996; Zbl 0859.46039)], \(\widehat G\) is itself and group bundle over \(X\) with fibres \(\widehat G_{x}\). Furthermore, if \(\widehat \beta_{x}\) is the usual dual Haar system on \(\widehat G_{x}\), then \(\widehat\beta:=\{\widehat\beta_{x}\}_{x\in X}\) is a continuous Haar system for \(\widehat G\). The main result of this paper is that \(\widehat {\widehat G}\) and \(G\) are isomorphic as topological groupoids. This is a generalization of the usual Pontryagin Duality Theorem for locally compact abelian groups. As the author makes clear, establishing a continuous bijection of \(G\) onto \(\widehat{\widehat G}\) is relatively routine. Showing that this map is open requires some work and occupies the majority of the proof of the main result.