Given a finite subgroup of $G$ sitting inside the Morava stabilizer group $S_n$, we can form the homotopy fixed point spectrum $E_n^{hG}$. There is a spectral sequence with $E_2^{s,t} = H^s(G;\pi_t(E_n)) \Rightarrow \pi_{t-s}(E_n^{hG})$.

For resolving further differentials in this spcetral sequence I have seen a way by claiming some elements in the $E_2$ page is actually the image of certain elements of $\pi_*(S)$ under Hurewicz map and then by the known multiplicative relation(for example $h_1^3 = h_0^2 h_2$) we shall know that some elements in this spectral sequence must die and try to figure out who kills it or who it kills.

Question: what is the general way, or is there any general way of proving this kind of claim? i.e. how to prove that something in HFPSS $E_2$ is actually corresponding to the image of an element from $\pi_*(S)$?

I only know some classical results that might be related to this: If we start with $Ext_{BP_*BP}$ and keep track of the element we are interested in inside the chromatic spectral sequence, we will end up with continuous group cohomology of $S_n$(after tensoring with $\mathbb{F}_{p^n}$ at some point) thus cohomology of finite subgroup of it. Then we can look at how elements $t_i\in BP_*BP$ works as functions from $S_n$ to $\mathbb{F}_{p^n}$ to tell the image of $t_i$ in group cohomology and finally detecting elements of $\pi_*(S)$.

The second question comes from a concrete example of this. By classical result, we know that the non-exist elements $\beta_{p^j/p^j}$ and their monomials for $p \geq 5$ can be detected in the way above, in group cohomology of a finite group of order $p$ inside $S_{p-1}$ with coefficient $\mathbb{F}_{p^{p-1}}$. And my second question is the following:

Question: I believe that if I want to rephrase the detection of $\beta_{p^j/p^j}$ in modern language, it will becomes things like "If $\beta_{p^j/p^j}$ survives to $\pi_*(S)$, then it can be detected in the $C_p$ homotopy fixed point of $E_{p-1} $ for $p \geq 5$". So is this true? If so, if I can draw the $E_2$ of this spectral sequence, could I actually identify elements in $H^*(C_p;\mathbb{F}_{p^{p-1}})$ (where the classical detection happens) inside this $E_2$? Or is there a suitably defined map from $E_2$ to $H^*(C_p;\mathbb{F}_{p^{p-1}})$ which maps image of $\beta_{p^j/p^j}$ to non-zero elements?

Any help or hint is really appreciated.


In Ravenel's paper on the Arf invariant, he shows (p. 439) that there is a composite of maps $$ \mathrm{Ext}_{BP_*BP}(BP_*,BP_*) \to H^*_c(\mathbb{S}_n,E_*) \to H^*(C_p,E_*/\frak{m}), $$ under which the images of $\alpha_1,\beta_1$ and $\beta_{p,p}$ are non-zero. You can use these to determine the Hurewicz image of classes under the composite $$ \mathrm{Ext}_{BP_*BP}(BP_*,BP_*) \to H^*_c(\mathbb{S}_n,E_*) \to H^*(G,E_*). $$

Since this is probably the argument you outline in the first paragraph, it is useful to point out another method, via the functoriality of the Greek letter construction; this is the argument used in Proposition 7 of this paper by Goerss, Henn, and Mahowald.

I'm afraid I don't fully understand the second part of the question; if a class survives in the $BP$-Adams spectral sequence, and is detected in the $C_p$-HFPSS spectral sequence under the map $$ \mathrm{Ext}_{BP_*BP}(BP_*,BP_*) \to H^*(C_p,E_*), $$ then it must also survive in the latter spectral sequence.

  • $\begingroup$ Thank you for the answer! There is one simple thing confusing me: In Ravenel's paper, he constructed such a map via mapping ANSS into $Ext_{BP_*BP}(BP_*,v_n^{-1}BP_*/I_n)$ and play with the Ravenel-Wilson change of ring. So what's the map $Ext_{BP_*BP}(BP_*,BP_*) \rightarrow H^*_c(\mathbb{S}_n,E_*)$ and why the classical detection map factors through it? $\endgroup$ – Mingcong Zeng Nov 3 '15 at 14:47
  • 1
    $\begingroup$ @MingcongZeng: This is Morava's change of rings theorem. Map $S$ to $L_{K(n)}S^0$, and look at the induced map. in $\mathrm{Ext}_{BP_*}$, and then use the change of rings theorem. $\endgroup$ – Drew Heard Nov 3 '15 at 16:33

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.