Quillen famously constructed a fibration sequence $BGL(\mathbb{F}_q)^+ \to BU \to BU$ to compute the algebraic K-groups of finite fields, where the second map is $\psi^\ell-1$ for $\ell$ a generator of $\mathbb{Z}_q^\times$. Does this lift to the level of the K-theory spectra?

I don't know much about delooping, but I would guess that if we had a fibration sequence $BGL(\mathbb{F}_q)^+ \to BU\times \mathbb{Z} \to BU\times \mathbb{Z}$ of $E_\infty$-spaces, this would then deloop to a fibration sequence of the associated connective $\Omega$-spectra. However, this seems out of reach, since Quillen's method proceeds via the Atiyah-Segal completion theorem and only produces a homotopy class of map.

  • 2
    $\begingroup$ As far as I know, such a fiber sequence of spectra (with $ku$ in place of $KU$ for obvious connectivity reasons), exists only when you $l$-complete where $l$ is a prime different from $p$. In fact I'm not even sure the Brauer lift is a map of H-spaces if you do not complete away from the characteristic $\endgroup$ Dec 6, 2018 at 22:57
  • $\begingroup$ is there a reference where this is all written down? $\endgroup$
    – xir
    Dec 7, 2018 at 1:36
  • 3
    $\begingroup$ First, as Denis said, you need connective K-theory. If you're fine with p-completion (where p is different from char(F_q)), then there is a fiber sequence L_K(1) K(F_q) -> KU_p -> KU_p. So you'd need to show that the p-completion of K(F_q) is the connective cover of its K(1)-localization. This seems to be true, just by inspection. I don't know if there is an integral version. $\endgroup$
    – skd
    Dec 7, 2018 at 1:43

1 Answer 1


The infinite loop space/spectrum level statements were written down in

May, Quinn, Ray, Tornehave: "$E_\infty$ Ring Spaces and $E_\infty$ Ring Spectra" (1977) http://www.math.uchicago.edu/~may/BOOKS/e_infty.pdf ,

see Theorem VIII.3.2 and its proof on pages 224-225. For a prime $p$ and a prime power $r=q$ generating the $p$-adic units (take $r=3$ for $p=2$), May writes $j^\delta_p$ for $K(\mathbb{F}_r)$ completed at $p$, viewed as a discrete model for the connective complex image-of-$J$ spectrum, and argues that this spectrum is equivalent to the homotopy fiber $F\psi^r$ of $\psi^r-1 \colon kU \to bu$, after $p$-completion. Here $kU = ku$ is the connective complex $K$-theory spectrum and $bu$ is its $1$-connected cover. See also Section 19 of the survey paper

May: "Infinite Loop Space Theory" (1977) http://www.math.uchicago.edu/~may/PAPERS/18.pdf ,

which may be easier to read.

  • $\begingroup$ thanks, this is very helpful! is there any way to see how the brauer lift fails to be good enough before completion? (e.g. fails to even be a map of h-spaces, as denis said?) $\endgroup$
    – xir
    Dec 30, 2018 at 0:41

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.