5
$\begingroup$

The question is the following:

Let $K\subset S^{3}$ be a knot, consider the double branched cover $Y$ of $S^{3}$ over $K$. We know $Y$ has a unique spin structure $\mathfrak{s}$, now the question is:

When do we have $\widehat{HF}(Y,\mathfrak{s})\cong \mathbb{Z}$? If yes, what do we know about the d-invariant?

I know that this is true for quasi-alternating knot, however, I don't know any other examples...

$\endgroup$
1
  • $\begingroup$ It definitely holds whenever $K$ is (Khovanov-reduced, I believe) homologically thin, and when $K$ is a 2-bridge knot. The first case follows from the spectral sequence defined by Ozsvath-Szabo, the second from the classical fact that double covers branched over 2-bridge knots are lens spaces. $\endgroup$ May 22, 2014 at 6:39

3 Answers 3

5
$\begingroup$

The paper of Manolescu and Owens "A concordance invariant from the floer homology of double branched covers" seems to answer your question. They compute many examples of the Froyshov/d-invariant in their paper, using software of Saso Strle. It's available on the arXiv. In that case it's at least reasonably-computable.

$\endgroup$
3
$\begingroup$

John Baldwin said something about both questions in this paper. He considers closures of 3-braids and classifies which of these have an $L$-space branched double cover (Theorem 4.1). He then computes the correction term for the spin structure (Theorem 5.1); he also seems to say that there are no new results in this last computation (see the discussion before the statement of the theorem).

$\endgroup$
3
$\begingroup$

Let assume $Y$ is a closed oriented rational homology sphere.

The set of spin$^c$ structure of $Y$ is in one to one correspondence with $H_1(Y,\mathbb{Z})$. So if $H_1(Y,\mathbb{Z}) \neq 0$ then there are more than one spin$^c$ structure.

The question then becomes: When does $\widehat{HF}(Y,\mathfrak{s}) \cong \mathbb{Z}$ for each spin$^c$ structure $\mathfrak{s}$?

Three-manifolds with such property are called $L$-spaces. Double branched cover of quasi-alternating knots (in $S^3$) are L-spaces. I think your question is a work in progress. For some other examples see Steven Sivek comment below.

Note also that the collection of quasi-alternating knots contains the collection of alternating knots, and 2-bridge knots are examples of alternating knots.

In the special case $H_1(Y,\mathbb{Z}) = 0$, in which situation there is a unique spin$^c$ structure, it is conjectured that $S^3$ and the Poincar\'e homology sphere ($\Sigma=S^3/I$ where $I$ is the full icosahedral group) are the only such 3-manifold. We know that $d(S^3)=0$ and that $d(\Sigma)=\pm 2$ depending on the orientation you chose. For the general case I think knowing exactly the $d$-invariant is still an open problem except when $Y$ is a lens space in which case there is a formula for each $d$ invariant.

$\endgroup$
1
  • $\begingroup$ There are actually lots of non-QA knots whose branched double covers are L-spaces. See section 6.1 of arxiv.org/pdf/1205.5261.pdf for some discussion and explicit examples, including the $P(p_1,\dots,p_n,-q) $ pretzel knots where $p_i,q>0$ and $q = \min(p_1,\dots,p_n)$. $\endgroup$ Jun 13, 2014 at 3:08

Your Answer

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

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