The spectrum for commutative complex \(K\)-theory (Q1743524)

Let \(G\) be a topological group with identity element a nondegenerate basepoint. A simplicial model for the classifying space \(BG\) is the geometric realization of \(N_*G\), the nerve of \(G\) (viewed as a category with one object.) If one considers only those simplices which are ordered tuples of pairwise commuting elements of \(G\), one gets a subcomplex \(B_{\text{com}}G\) of \(BG\) and a natural inclusion map \(i : B_{\text{com}}G \rightarrow BG\). In some cases, \(B_{\text{com}}G\) carries an \(E_\infty\)-ring space structure, and represents thus a generalised cohomology theory called \textit{commutative \(K\)-theory}, see [\textit{A. Adem} et al., Algebr. Geom. Topol. 17, No. 2, 869--893 (2017; Zbl 1360.55003)]. The present work deals with the case \(G=U\) (and the associated case \(G=SU\)), whose associated theory is \textit{commutative complex \(K\)-theory}, determining in particular the homotopy type of \(B_{\text{com}}U\), its rational Hopf ring, and how it relates to previous work by \textit{A. Adem} and \textit{J. Gómez} [Algebr. Geom. Topol. 15, No. 1, 493--535 (2015; Zbl 1397.55010)] on the rational cohomology of \(B_{\text{com}}U\). Section 2 determines a spectrum for commutative complex \(K\)-theory. The main result here is Theorem 2.11, stating that there exists a commutative \(ku\)-algebra spectrum \(E\) such that \(\Omega^\infty E \simeq {\mathbb Z} \times B_{\text{com}}U /\mkern-6mu/ U\) and an equivalence of \(ku\)-algebras \(E \simeq ku \wedge BU(1)_+\), where \(ku\) is the connective \(K\)-theory spectrum and \(B_{\text{com}}U /\mkern-6mu/ U\) is the homotopy orbit space coming from the action of \(U\) on \(B_{\text{com}}U\) by conjugation on each simplex. These results are obtained by means of Carlsson and Lawson's deformation \(K\)-theory of free abelian groups (Section 2.1) and the associated Bott cofibre sequence of Lawson (from Corollary 2.10). Corollary 2.13 presents the homotopy groups of \(B_{\text{com}}U\). Section 3 deals with the determination of the homotopy types of \(B_{\text{com}}U\) and \(B_{\text{com}}SU\). This appears in Theorem 3.4, stating that \(B_{\text{com}}U \simeq BU \times E_{\text{com}}U\) and \(B_{\text{com}}SU \simeq BSU \times E_{\text{com}}U\) as \(E_\infty\)-\({\mathbb Z} \times BU\)-modules. Moreover, \(E_{\text{com}}U\), the homotopy fiber of the inclusion \(B_{\text{com}}G \rightarrow BG\), is homotopically equivalent to \(\prod_{n \geq 2} BU \langle 2n \rangle\) (here, \(BU \langle 2n \rangle\) denotes the \(2n-1\) connected cover of \(BU\).) This result is achieved using the fact that \(ku \wedge BU(1)_+\) splits as a wedge of suspensions of \(ku\), the main ingredient in the proof of Lemma 3.3. Section 4 calculates the Hopf ring structure of the rational homology of \({\mathbb Z} \times B_{\text{com}}U /\mkern-6mu/ U\): Theorem 4.4 states that it is generated by one class in degree \(0\), \([1] \otimes 1\), and two classes in degree \(2\), \([0] \otimes \zeta_{1,0}\) and \([0] \otimes \zeta_{0,1}\), where the \(\zeta_{a,b}\) are dual to the polynomial generators \(z_{a,b}\) of the rational cohomology algebra of \(B_{\text{com}}U\) previously determined by Adem and Gómez [loc. cit.]. Corollary 4.6 uses the Hopf ring description to relate the homotopy and the rational homology of \(B_{\text{com}}U\), obtaining a formula that features the Sterling numbers of the first kind and that connects to Adem and Gómez's calculations of the rational cohomology. Corollary 4.7 describes the splitting from Theorem 3.4 on rational cohomology, obtaining a formula that gives the \(n^{th}\) component of the Chern component in terms of the Stirling numbers of the second kind. Section 5 discusses some relationships between \(B_{\text{com}}U\) and \(B_{\text{com}}SU(2)\). The main result here is Proposition 5.2, declaring that their \(4^{th}\) homotopy groups are isomorphic.