A bivariant theory for the Cuntz semigroup (Q2421528)

From MaRDI portal
scientific article
Language Label Description Also known as
English
A bivariant theory for the Cuntz semigroup
scientific article

    Statements

    A bivariant theory for the Cuntz semigroup (English)
    0 references
    0 references
    0 references
    0 references
    17 June 2019
    0 references
    The authors introduce a bivariant version of the Cuntz semigroup, establish some basic properties of this object, provide computations of it in a number of examples, and finally present an application to the classification problem of unital and stably finite \(C^\ast\)-algebras. The Cuntz semigroup was introduced in [\textit{J. Cuntz}, Math. Ann. 233, 145--153 (1978; Zbl 0354.46043)]. Its bivariant version, denoted by \(W(A,B)\), is introduced here in the setting of local \(C^\ast\)-algebras in the sense of \textit{R. Antoine} et al. [Tensor products and regularity properties of Cuntz semigroups. Providence, RI: American Mathematical Society (AMS) (2018; Zbl 1414.46035)], as equivalence classes of c.p.c oder zero maps from \(A\) to \(M_\infty(B)\). More precisely, the equivalence relation is obtained as the antisymmetrization of the following pre-order among c.p.c. order zero maps from \(A\) to \(B\): \[ \phi \lesssim \psi \iff \exists \{b_n\}_{n\in\mathbb{N}}\subset B:\quad \forall a\in A\quad \lVert b^\ast_n \psi(a)b_n - \phi(a)\rVert\to 0. \] The structure theory of c.p.c. order zero maps is developed in [\textit{W. Winter} and \textit{J. Zacharias}, Münster J. Math. 2, No. 1, 311--324 (2009; Zbl 1190.46042)] and extended to include local \(C^\ast\)-algebras in the second section of this paper. The authors go on to prove that \(W(\cdot,\cdot)\) is additive in both variables, contravariant in the first one, covariant in the second one, stable in the first one, and matrix-stable in the second one (e.g., we can replace \(B\) with \(M_\infty(B)\) up to isomorphism). Moreover, as a functor, the bivariant Cuntz semigroup is well defined on the target category \(W\), defined in [Antoine et al., loc. cit.] (roughly speaking, an object in this category is a partially ordered monoid equipped with an auxiliary relation). In this setting, the functor is continuous w.r.t. abitrary inductive limits in the second variable, whenever the first variable is an elementary \(C^\ast\)-algebra. Examples are given, showing that more general continuity properties fail to hold. These issues open up a possibility of studying a Cuntz analogue of the notion of \(K\!K\)-semiprojectivity [\textit{M. Dadarlat}, Adv. Math. 222, No. 5, 1850--1881 (2009; Zbl 1190.46040)]. Leaving the world of local \(C^\ast\)-algebras, the authors also define a stabilized version of the semigroup as \(W\!W(A,B)=W(A\otimes \mathbb{K},A\otimes \mathbb{K})\). They raise the interesting question of whether \(W\!W(A,B)\) belongs to the richer category of semigroups \(\mathrm{Cu}\) defined in [\textit{K. T. Coward} et al., J. Reine Angew. Math. 623, 161--193 (2008; Zbl 1161.46029)]; an answer to this question would clarify the continuity properties of \(W\!W\) as a functor. The invariant \(W\!W\) admits a module picture which resembles the standard description of \(K\!K\)-theory. The functor \(W\) can also be seen as a refinement of bivariant \(K\)-theory: the subsemigroup given by a \(\ast\)-homomorphism from the first algebra to the matrix-stabilization of the second corresponds to an unstable version of \(K\!K\)-theory. In addition, in analogy with the Kasparov product, the authors introduce a product for the bivariant Cuntz semigroup (essentially given by composition) which makes \(W(A,A)\) a semiring, yields a notion of Cuntz equivalence between \(C^\ast\)-algebras, and a map \(W(A,B)\to \mathrm{Hom}(W(A),W(B))\). An in-depth study of this map could lead to a notion of Cuntz analogue of the UCT class. Passing to computations, the authors prove: \begin{itemize} \item[(1)] When \(A\) is exact and \(B\) is a Kirchberg \(C^\ast\)-algebra, \(W\!W(A,B)\) is isomorphic to the two-sided ideal lattice of \(A\). \item[(2)] When \(D\) is a strongly self-absorbing \(C^\ast\)-algebra [\textit{A. S. Toms} and \textit{W. Winter}, Trans. Am. Math. Soc. 359, No. 8, 3999--4029 (2007; Zbl 1120.46046)], the following isomorphism holds: \(W(A\otimes D,B\otimes D)\cong W(A,B\otimes D)\). An analogous result holds for \(W\!W\). \item[(3)] Cuntz homology (i.e., the functor \(A\mapsto WW(A,\mathbb{C})\)) is a complete invariant of compact metrizable spaces. \end{itemize} A strictly invertible element in \(W\!W(A,B)\) is the class of a c.p.c. order zero map from \(A\) to \(B\), admitting an inverse induced by a map from \(B\) to \(A\). In the last section, the authors prove that, if \(A\) and \(B\) are unital and stably finite \(C^\ast\)-algebras, they are isomorphic if and only if there is a strictly invertible element in \(W\!W(A,B)\). Stably finite \(C^\ast\)-algebras include AF- and AI-algebras. This result can be regarded as an analogue of the celebrated Kirchberg-Phillips classification theorem [\textit{E. Kirchberg} and \textit{C. N. Phillips}, J. Reine Angew. Math. 525, 17--53 (2000; Zbl 0973.46048); J. Reine Angew. Math. 525, 55--94 (2000; Zbl 0973.46047)].
    0 references
    Cuntz semigroup
    0 references
    bivariant $K$-theory
    0 references
    classification of ${C}^{\ast}$-algebras
    0 references
    local \(C^\ast\)-algebras
    0 references

    Identifiers

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