Suppose G, a discrete group acting on the compact operators $\mathcal{K}$ by automorphism of C*-algebra $\mathcal{K}$. Can we view the crossed product as a groupoid C*-algebra of some groupoid?

This question occurred to me when I was thinking of possible formulation of Baum-Connes conjecture for projective representation of the group.

  • $\begingroup$ Since all automorphisms of $\mathcal K$ are inner, the crossed product ${\mathcal K} \rtimes G$ is $*$-isomorphic to ${\mathcal K} \otimes \mathrm{C}^*_\lambda(G)$. The latter is a groupid $\mathrm{C}^*$-algebra. $\endgroup$ – Narutaka OZAWA Jun 30 '14 at 21:47
  • $\begingroup$ @NarutakaOZAWA if you look at this paper arxiv.org/pdf/math/0609784v2.pdf (page 5, second paragraph from top), this action by $G$ comes from a 2-cocycle from 2nd cohomology group of $G$, and the crossed product $\mathcal{K} \rtimes G$ highly depends on that cocycle. So can you please clarify the 1st statement? $\endgroup$ – SiOn Jul 1 '14 at 20:32
  • $\begingroup$ You are right. It was isomorphic to ${\mathcal K}\otimes\mathrm{C}^*_\lambda(G,\bar{\omega})$. It isn't clear if it's a groupoid $\mathrm{C}^*$-algebra. $\endgroup$ – Narutaka OZAWA Jul 1 '14 at 22:03
  • $\begingroup$ yes, even not clear for finite groups(for trivial group its obvious only)! $\endgroup$ – SiOn Jul 1 '14 at 23:03
  • $\begingroup$ Depending on what you want to do, describing a C*-algebra as a twisted groupoid C"-algebra is often good enough. So @Ozawa's last comment could be what you need. $\endgroup$ – Ruy Oct 13 '16 at 21:48

Have you seen Packer-Raeburn Stabilization trick? ( Transactions of the AMS 334(2), p. 685-718, 1992). Specially theorem 3.4 therein stablishes that the algebra is morita equivalent to a Twisted Group C*- algebra, summarizing the answer by N. Ozawa and the correction above.

By considering the trick of viewing the algebra $K\rtimes G$ as a locally trivial bundle with a $G$-action over a contractible $G$- space $X$ ( For instance, the total space of the universal $G$-Bundle $EG$, but also the space $\underline{E}G$ works), you can see the $C*$ algebra in question is morita equivalent to the algebr a of sections of the Fell bundle over the transformation groupoid $X\rtimes G$. THis has ben documented in Kumjian's monograph http://www.theta.ro/jot/archive/1988-020-002/1988-020-002-002.pdf

Other references on this kind of tricks include http://jlms.oxfordjournals.org/content/50/1/170.abstract , and part 5 here: https://arxiv.org/abs/1501.06050


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.