Anick spaces and Kac-Moody groups (Q1713061)

For an odd prime \(p\geq 3\) and an integer \(r\geq 1\), let \(P^{n+1}(p^r)=S^n\cup_{p^r}e^{n+1}\) denote the mod \(p^r\) Moore space and let \(T^{2n+1}(p^r)\) denote the Anick space constructed by \textit{D. Anick} [Differential algebras in topology. Wellesley, MA: A. K. Peters (1993; Zbl 0770.55001)] such that there is a homotopy fibration \[ S^{2n-1}\longrightarrow T^{2n+1}(p^r)\longrightarrow \Omega S^{2n+1} \] and that there is a coalgebra isomorphism \(H_*(T^{2n+1}(p^r);\mathbb{Z}/p)\cong E[u_{2n-1}]\otimes \mathbb{Z}/p [v_{2n}]\) with \(\beta^r (v_{2n})=u_{2n-1}\) (where \(\beta^r\) denotes the \(r\)-th Bockstein). Let \(C^{2n+1}(p^r)\) denote the homotopy cofibration of \[ p^{4n}(p^r)\mathop{\longrightarrow}\limits^{[\nu ,\mu]} P^{2n+1}(p^r) \longrightarrow C^{2n+1}(p^r), \] where \([\nu,\mu]\) denotes the mod \(p^r\) Whitehead product of the identity map \(\nu\) on \(P^{2n+1}(p^r)\) and the Bockstein map \(\mu\). Note that F. Cohen, J. Moore and J. Neisendorfer conjectured that the space \(\Omega T^{2n+1}(p^r)\) retracts off \(\Omega^2P^{2n+1}(p^r)\) and \textit{J. Neisendorfer} [Topology 38, No. 6, 1293--1311 (1999; Zbl 0935.55008)] proved this conjecture affirmatively for \(r\geq 2\). In this paper, firstly the authors consider this problem for \(r=1\) and they prove that \(\Omega T^{2n+1}(p^r)\) is a retract of \(\Omega^2C^{2n+1}(p^r)\) when \(p\geq 5\), \(r\geq 1\) and \(n\geq 2\). Next they consider the relation between Anick spaces and certain rank \(2\) Kac-Moody groups constructed by Kitchloo. Let \(k\in \{p,2p\}\) or let \(k\) be a divisor of \(p-1\) or \(p+1\). Then Kitchloo showed that there is a non-empty set \(\mathcal{V}_k\subset \mathbb{Z}_{\geq 1}\) with the property that if \(r\in \mathcal{V}_k\) there is a rank \(2\) Kac-Moody group such that there is an isomorphism \(H_*(K;\mathbb{Z}/p)\cong E[z_3,y_{2k-1}]\otimes \mathbb{Z}/p [x_{2k}]\) with \(\beta^r(x_{2k})=y_{2k-1}\) and that there is a homotopy fibration sequence \[ S^3\mathop{\longrightarrow}\limits^{i} K\mathop{\longrightarrow}\limits^{\delta} X \longrightarrow BS^3\mathop{\longrightarrow}\limits^{Bi}BK \] where \(i\) is an inclusion of groups and \(\delta\) is the projection. It is known that there is an isomorphism \(H_*(X;\mathbb{Z}/p)\cong H_*(T^{2n+1}(p^r);\mathbb{Z}/p)\) and he conjectured that there is a \(p\)-local fibration \(S^{2k-1}\to X\to \Omega S^{2k+1}\) that is equivalent to Anick's fibration for \(n=k\). His weaker conjecture is that there is a \(p\)-local homotopy equivalence \(X\simeq_pT^{2k+1}(p^r)\). The second purpose of the authors is to consider this weaker conjecture and they show that if \(1<k<p-1\) and \(r\in \mathcal{V}_k\) with \(p\geq 5\) there are a \(p\)-local homotopy equivalences \[ \Omega X\simeq_p \Omega T^{2k+1}(p^r) \quad \text{and}\quad \Omega K\simeq_p\Omega S^3\times \Omega T^{2k+1}(p^r). \] As an application they also obtain that the \(p\) primary homotopy exponent of \(K\) is \(p^r\).