Let $K$ be a knot in the 3-sphere $S^3$.

Here we denote by $s(K)$ Rasmussen's s-invariant for $K$, and by $X_{K}(n)$ the 4-manifold obtained from the standard 4-ball $B^4$ by attaching a $2$-handle along $K$ with framing $n$.

My question is the following.

For two knots $K$ and $K'$ such that $X_{K}(0)$ and $X_{K'}(0)$ are diffeomorphic, is it true that $s(K)=s(K')$?

I am also interested in the same question for Ozsváth-Szabó's $\tau$-invariant.

Note that it is known that there exist knots $J$ and $J'$ such that $\partial X_{J}(0)$ and $\partial X_{J'}(0)$ are diffeomorphic, and $s(J) \neq s(J')$, see Yasui's paper Corks, exotic 4-manifolds and knot concordance.

  • $\begingroup$ I think that the question is true if $K'$ is the unknot. If $X_K(0)$ is diffeomorphic to $S^2\times D^2=X_U(0)$ where $U$ is the unknot, then $K$ is slice in homotopy 4-ball. By Kronheimer-Mrowka, $s(K)=0$. Similar argument works for $\tau$. $\endgroup$
    – user156937
    Aug 11, 2015 at 17:44
  • 1
    $\begingroup$ @user156937 Thank you very much! If one of them is the unknot, the other is also the unknot by Property R theorem by Gabai. If one of them is slice, then the other, say $K'$, is slice in a homotopy 4-ball, therefore $s(K')=0$.The problem is when $K$ and $K'$ are not slice. $\endgroup$ Aug 11, 2015 at 17:52

1 Answer 1


Lisa Piccirillo just posted a few days ago this paper on the arXiv. Corollary 1.3 asserts exactly that $s$ is not a 0-trace invariant of the knot.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.