A going-down principle for ample groupoids and the Baum-Connes conjecture (Q783227)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A going-down principle for ample groupoids and the Baum-Connes conjecture
scientific article

    Statements

    A going-down principle for ample groupoids and the Baum-Connes conjecture (English)
    0 references
    0 references
    11 August 2020
    0 references
    Let \(G\) be a locally compact group and let \(A\) be a \(C^*\)-algebra with a continuous action of \(G\). The Baum-Connes assembly map for \(G\) with coefficients in \(A\) is a group homomorphism whose target is the K-theory of the reduced crossed product for \(G\) acting on \(A\). Its domain \(\mathrm K^{\mathrm{top}}_*(G,A)\) is called the topological K-theory of \(G\) with coefficients in \(A\). The Baum-Connes assembly map is known to be an isomorphism for many groups and coefficients. This is the Baum-Connes conjecture. One technique to prove this are permanence properties. These say that the map is an isomorphism for some \(G\) and \(A\) if it is an isomorphism in some other situation. This article extends several key results needed to prove such permanence properties to the setting where \(G\) is an ample, Hausdorff groupoid. That is, \(G\) is a Hausdorff, étale groupoid and its object space is totally disconnected and locally compact. A first key result is an adjunction between restriction and induction in groupoid equivariant KK-theory. This is proven for induction from a subgroupoid \(H\subseteq G\) that is closed, open, and proper in the sense that the map \(H\to H^0 \times H^0\), \(g\mapsto (r(g),s(g))\), is proper. The reason why the object space of \(G\) is assumed totally disconnected here is that otherwise there may be too few such subgroupoids. Induction may still be defined for open subgroupoids in étale groupoids, and the adjunction between induction and restriction remains true in this generality. This is shown in the thesis of Valerio Proietti, using a different construction of the induction functor. Roughly speaking, the idea is to replace a fixed point algebra for the subgroupoid by a crossed product. The second key result in this article is an ample groupoid version of the Going-Down principle that was developed by Chabert, Echterhoff and Oyono-Oyono for groups. It says that an element \(x\) in \(\mathrm{KK}^G(A,B)\) induces an isomorphism \(\mathrm K^{\mathrm{top}}_*(G,A) \to\mathrm K^{\mathrm{top}}_*(G,B)\) if it induces an isomorphism \(\mathrm{KK}^H(C(H^0),A|_H) \to\mathrm{KK}^H(C(H^0),B|_H)\) for all compact open subgroupoids \(H\) of \(G\). In particular, this happens if \(x\) restricts to an invertible element in \(\mathrm{KK}^H(A|_H,B|_H)\) for all such \(H\). In the group case, the Going-Down Principle paved the way for the reinterpretation of the Baum-Connes assembly map through localisation of triangulated categories in [\textit{R. Meyer} and \textit{R. Nest}, Topology 45, No. 2, 209--259 (2006; Zbl 1092.19004)]. Roughly speaking, this says that the functor \(A\mapsto\mathrm K^{\mathrm{top}}_*(G,A)\) is the best approximation to the \(K\)-theory of the reduced crossed product that inverts all \(x\) in \(\mathrm{KK}^G(A,B)\) that become invertible in \(\mathrm{KK}^H(A|_H,B|_H)\) for all compact subgroups \(H\) in \(G\). It is very plausible that this localisation approach also generalises to groupoids. But this has not yet been shown in general. It certainly holds for those groupoids that admit a \(\gamma\)-element; this is a key tool to prove that the Baum-Connes assembly map is split injective. As an application of the Going-Down Principle, the Baum-Connes assembly map is shown to be split injective for those ample, Hausdorff groupoids that are strongly amenable at infinity. Strong amenability at infinity requires an amenable action of the groupoid on a space with a proper anchor map which, in addition, admits a continuous non-equivariant section. As for groups, this implies that the groupoid in question is exact, that is, its reduced crossed product functor is exact. Let \(G\) be a groupoid that is strongly amenable at infinity and let \(A\) be a \(G\)-\(C^*\)-algebra whose underlying field of \(C^*\)-algebras over \(G^0\) is continuous. Assume that for each \(x \in G^0\), the Baum-Connes assembly map is invertible for the isotropy group \(G_x\) of \(G\) at \(x\) with coefficients in \(A_x\). Then the Baum-Connes assembly map is also invertible for \(G\). This is a remarkable result that sometimes allows to reduce the validity of the Baum-Connes conjecture from groupoids to groups. Along the way, the article establishes various basic results about groupoid-equivariant KK-theory, such as results about the automatic equivariance of the Fredholm operator in an equivariant Kasparov cycle.
    0 references
    0 references
    Baum-Connes conjecture
    0 references
    étale groupoid
    0 references
    equivariant Kasparov theory
    0 references
    induction-restriction adjunction
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references

    Identifiers

    0 references
    0 references
    0 references
    0 references
    0 references
    0 references
    0 references