On the Bott periodicity, $J$-homomorphisms, transfer maps and $H_{\ast }{Q}_0{S}^{-n}$

Abstract

The Curtis conjecture predicts that the only spherical classes in $H_{\ast }(Q_0{S}^0;\mathbb{Z}/2)$ are the Hopf invariant one and the Kervaire invariant one elements. We consider Sullivan's decomposition $$Q_0S^0=J\times\cok J$$ where $J$ is the fibre of ${\psi }^q-1$ ($q=3$ at the prime $2$) and observe that the Curtis conjecture holds when we restrict to $J$. We then use the Bott periodicity and the $J$-homomorphism $SO\to Q_0{S}^0$ to define some generators in $H_{\ast }(Q_0{S}^0;\mathbb{Z}/p)$, when $p$ is any prime, and determine the type of subalgebras that they generate. For $p=2$ we determine spherical classes in $H_{\ast }({\Omega }_0^{k}J;\mathbb{Z}/2)$. We determine truncated subalgebras inside $H_{\ast }(Q_0{S}^{-k};\mathbb{Z}/2)$. Applying the machinery of the Eilenberg-Moore spectral sequence we define classes that are not in the image of by the $J$-homomorphism. We shall make some partial observations on the algebraic structure of $H_*(Ω^k_0\cok J;\Z/2)$. Finally, we shall make some comments on the problem in the case equivariant $J$-homomorphisms.

arXiv comment20 pages. Updated, with more details proofs