11
$\begingroup$

It's known that the Seifert–Weber space (obtained from a dodecahedron by gluing opposite faces with a 3/10 turn) is an example of a non-Haken 3-manifold. Since every closed 3-manifold is virtually Haken, I was wondering: is there was a known finite cover of the Seifert-Weber space that is Haken?

$\endgroup$
  • 1
    $\begingroup$ It's not clear to me immediately if there are any smaller covers than Hempel's 5-sheeted one. But in principle you could turn this into a fairly simple algorithmic problem by iterating over the covering spaces and applying Reidemeister-Schrier to the corresponding subgroups. $\endgroup$ – Ryan Budney Jan 18 '17 at 23:14
  • 2
    $\begingroup$ There are no covers of index $<5$, since the homology is $Z/5^3$ (hence any homomorphism to $S_4$ will be trivial). $\endgroup$ – Ian Agol Jan 18 '17 at 23:36
8
$\begingroup$

This is constructed (reasonably explicitly) in John Hempel's 1982 paper.

John Hempel, MR 664329 Orientation reversing involutions and the first Betti number for finite coverings of $3$-manifolds, Invent. Math. 67 (1982), no. 1, 133--142.

$\endgroup$
  • $\begingroup$ This is what I was looking for - thanks! $\endgroup$ – user Jan 18 '17 at 21:33
9
$\begingroup$

I just checked using SnapPea that there is a cover of the Seifert-Weber dodecahedral space of index 25 (a 5-fold cyclic cover of a 5-fold cyclic cover) which has positive first betti number hence is Haken.

The Seifert-Weber space is a 5-fold cyclic branched cover over the Whitehead link complement. There are two such 5-fold covers (up to homeomorphism), which one may compute using SnapPea (perform $(5,0)$ surgery on each cusp of the Whitehead link, then compute all 5-fold cyclic covers of this orbifold, giving four manifold covers, with two isometry types). One may then compute the 5-fold cyclic covers of these two manifolds. One of them (not Seifert-Weber) has a 5-fold cyclic cover with positive betti number, whereas the 5-fold cyclic covers of the Seifert-Weber space have trivial betti number. However, one of them will be a 5-fold cover of its sibling, and hence will have a 5-fold cyclic cover which has positive betti number.

There are many other ways that we know the Seifert-Weber space to have a finite cover with positive first betti number (Hempel's paper pointed out by Igor shows that there is a 5-fold irregular cover), but it is an interesting question whether given a manifold $M$ with $b_1(M,\mathbb{F}_p)\geq 4$, is there a $p$-cover which has positive first betti number (I asked this as question 5 in a survey paper). Since all 5-fold covers of the Seifert-Weber space have $b_1(*; \mathbb{F}_5)\geq 4$, then this computation shows that a single 5-fold cyclic cover works in some cases.

$\endgroup$

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.