Let $k(n)$ be the $n$th connective Morava K-theory, with $k(n)_* = \mathbb F_p[v]$ where $|v| = 2p^n-2$. If $X$ is a space or a spectrum (assumed bounded below), one can compute $k(n)_*(X)$ using either the classical Adams spectral sequence or the even more classical Atiyah-Hirzebruch spectral sequence.

Both spectral sequences are spectral sequences of modules over $k(n)_*$. (In the ASS the bidegree of $v$ is $(1,2p^n-1)$.) Both spectral sequences start with $k(n)_* \otimes H_*(X;\mathbb F_p)$, at $E_1$ for the ASS, and $E_2$ for the AHSS. Both have first possible nontrivial differential given by the formula $ d(x) = vQ_n(x)$, where $Q_n$ is the $n$th Milnor primitive in the Steenrod algebra (acting on homology by going down in degree by $2p^n-1$).

So it seems that these must really be the same spectral sequence. Is this true? Is this a fact in the literature? (I am a tad bothered by the fact that the AHSS arises from an increasing filtration of $k(n) \wedge X$ while the ASS arises from a decreasing filtration of $k(n) \wedge X$.)

  • 11
    $\begingroup$ Hi Nick, I think this is true. The AHSS for $E_* X$ can either use a cellular filtration of X (starting at $E_2$) or the Postnikov filtration of E (starting at $E_1$), and "shearing" the spectral sequence accounts for the increasing/decreasing difference -- I think this is in Appendix B of Greenlees-May's "Generalized Tate cohomology". The result should then follow from the fact that the Postnikov tower for k(n) is an Adams tower, because mutliplication by $v_n$ becomes null after smashing with $H$. $\endgroup$ – Tyler Lawson May 6 '20 at 19:19
  • 1
    $\begingroup$ @TylerLawson: this is perfect! And that Appendix is well written and definitive. (And I am pretty sure Peter taught me this in grad school, though that was awhile ago!) $\endgroup$ – Nicholas Kuhn May 6 '20 at 19:59

Tyler's comment answers my question. A bit more detail: the Postnikov tower of $k(n)$ is an Adams resolution, because the `bottom class' map $k(n) \rightarrow H\mathbb F_p$ is onto in mod $p$ cohomology; indeed $H^*(k(n);\mathbb F_p) = A_p//E(Q_n)$.

The appendix by Greenlees and May has the details that two spectral sequences converging to $\pi_*(Y \wedge X)$, one coming from filtering $Y$ by its Postnikov tower and the other by filtering $X$ by its skeleta, agree.


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.