MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Consider six countable Abelian groups and six group homomorphims as in the following diagram

G → H → I
↑       ↓
L ← K ← J

Assume that the resulting sequence is exact at all six entries.

Question: Is there a (second countable) locally compact Hausdorff space X with a closed subspace A, such that the resulting six-term sequence in K-theory

K0(X,A) → K0(X) →  K0(A)
  ↑                 ↓
 K1(A)  ← K1(X) ← K1(X,A)

is isomorphic to the above one?

An answer in the finitely generated case would also be interesting.

share|cite|improve this question
First, it's equivalent to ask whether the sequence comes from a cofibration $X\xrightarrow{f}Y\xrightarrow{g}Z\xrightarrow{h}\Sigma X$ of finite spectra, which seems more natural. It is illuminating to consider the case where all groups are $\mathbb{Z}/4$ and all maps are multiplication by 2. I suspect that this cannot be done. Certainly, we cannot take $Y=X$ and $f=2.1_X$ because then $Z=S/2\wedge X$ and $K\wedge S/2$ has exponent $2$ (even though $S/2$ does not) so $K_0(Z)$ cannot be $\mathbb{Z}/4$. We could try taking $f$ to be twice an Adams self-map, but that does not work either. – Neil Strickland Feb 22 '11 at 11:30
@Neil Strickland: Yes, this is the simplest example in which problems arise. For, if the given algebraic sequence splits as a direct sum of sequences in which at least one map vanishes, I can accomplish the realisation. – Rasmus Bentmann Feb 22 '11 at 15:11
Can any 6-term sequence be realized as a 6-term sequence coming from an extension of $C^*$-algebras? – Andreas Thom Feb 22 '11 at 16:08
@Andreas: Yes (provided that it is exact, of course). This is in Alexander Bonkat's thesis. – Rasmus Bentmann Feb 22 '11 at 17:00
The $C^*$-algebra case should imply the topological case, at least for finitely generated groups. Using $C^*$-algebras we get a cofibration sequence of $K$-module spectra. Any $K$-module spectrum with finitely generated homotopy is equivalent to $K\wedge X$, where $X$ is a wedge of terms $S^0$, $S^0/n$, $S^1$ or $S^1/n$. In this context it is not hard to calculate $[X,Y]$ and $[K\wedge X,K\wedge Y]^K$ and I think that with a bit of bookkeeping we can make things work out. It would probably be easier to adapt Bonkat's proof, however. What is his approach? – Neil Strickland Feb 22 '11 at 17:38

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.