A Cuntz-Pimsner model for the \(C^\ast \)-algebra of a graph of groups (Q1997232)

Motivated by the Bass-Serre theory for graphs of groups, the first author with Brownlowe, Pask, Spielberg and Thomas considered C*-algebras related to a graph \(\mathcal{G} = (\Gamma, G)\) of (discrete) groups [\textit{N. Brownlowe} et al., Adv. Math. 316, 114--186 (2017; Zbl 1382.46040)]. In this context (and for the paper under review), the graph \(\mathcal{G}\) is assumed to be locally finite and nonsingular. A reconstruction analogue is obtained therein, namely, that \(\mathrm{C}^*(\mathcal{G})\) is \(*\)-isomorphic to the stabilization of the full crossed product of the action of the fundamental group on the boundary of the Bass-Serre tree. As such, \(\mathrm{C}^*(\mathcal{G})\) attains a groupoid model that can help its investigation. In the paper under review, the authors consider alternative realizations of \(\mathrm{C}^*(\mathcal{G})\). Towards this end, they construct a C*-correspondence \(D\) as a direct sum of group C*-correspondences over the \(2\)-paths. Next they consider an amplified version \(E\) by conjugating the summands of \(D\) with the imprimitivity bimodule on the range vertex group C*-correspondences. Both \(D\) and \(E\) are shown to be non-degenerate and regular. Since \(D\) and \(E\) are strongly Morita equivalent in the sense of Muhly-Solel, their Cuntz-Pimsner algebras \(\mathcal{O}_D\) and \(\mathcal{O}_E\) are strongly Morita equivalent [\textit{P. S. Muhly} and \textit{B. Solel}, Proc. Lond. Math. Soc. (3) 81, No. 1, 113--168 (2000; Zbl 1036.46046)]. The strong Morita equivalence can be realized by the same imprimitivity bimodule, and for other relative Cuntz-Pimsner algebras [\textit{G. K. Eleftherakis} et al., Isr. J. Math. 222, No. 2, 949--972 (2017; Zbl 1388.46046)]. The authors first show that \(\mathrm{C}^*(\mathcal{G})\) is \(*\)-isomorphic to the Cuntz-Pimsner algebra \(\mathcal{O}_E\). On the other hand, they show that \(\mathcal{O}_D\) is an Exel-Pardo algebra on a new graph constructed from the edges and paths of \(\Gamma\), and a natural cocycle. In this way they can import results from the theory of these C*-constructs. As an immediate application, they produce a six-term short exact sequence for the $K$-theory of \(\mathrm{C}^*(\mathcal{G})\). This is studied further for the case of: (i) single-edge graphs of groups; (ii) graphs of trivial groups; and (iii) generalized Baumslag-Solitar graphs of groups (namely, when all of the vertex and edge groups are the group of integers). The authors close with a preliminary investigation of Poincaré duality for the case where \(\mathcal{G}\) is a loop of groups.