Question. Is there a $3$-dimensional integer homology sphere whose universal cover is $\mathbb{R}^3$?

I believe the answer is in the positive and I am looking for (precise) references. If not in dimension $3$, I would be happy with higher dimensional examples.

This question is related to my post: Linking topological spheres and also to the post Have examples of non-simple connected higher-dimensions integer homology sphere? (This link was provided by Ian Agol).

  • 4
    $\begingroup$ Yes: an integral homology 3-sphere that admits a hyperbolic metric. These are fairly easy to construct, but I'm on a train right now, so I can't post a precise reference to one of the constructions (this is why this is a comment rather than an answer). $\endgroup$ – Andy Putman Jan 24 '19 at 19:54
  • $\begingroup$ Related question in higher dimensional case: mathoverflow.net/q/247502/1345 $\endgroup$ – Ian Agol Feb 1 '19 at 18:59

In a sense, most $3$-manifolds have universal cover $R^3$. In particular, this is the case for hyperbolic $3$-manifolds. And there do exist integer homology spheres which are hyperbolic. Two explicit examples I found by googling: Auckly: Surgery numbers of 3-manifolds: a hyperbolic example and Hom, Lidman: Surgery obstructions and hyperbolic integer homology spheres. It should be possible to get more examples from SnapPy.

  • 4
    $\begingroup$ If you tell me your name (you can e-mail me) I will put you in the acknowledgements since I will use this result in a paper. $\endgroup$ – Piotr Hajlasz Jan 25 '19 at 1:57

In fact, this is true for any closed connected 3-manifold with aspherical fundamental group. More generally, the universal cover of a closed 3-manifold is $S^3$ punctured at $0, 1, 2$ points or a tame Cantor set.

Aspherical homology spheres abound, for example $1/n$ surgery on a knot for $n$ sufficiently large. Other examples are Brieskorn spheres $\Sigma(p,q,r)$ with $p,q,r$ pairwise relatively prime and $1/p+1/q+1/r \leq 1$.

  • 2
    $\begingroup$ Your first link seems to take me to a dropbox page with no associated pdf. $\endgroup$ – Mike Miller Jan 24 '19 at 21:18
  • $\begingroup$ @MikeMiller there should be a gif, it’s a slide from a talk I gave. Are you logged in to dropbox? $\endgroup$ – Ian Agol Jan 24 '19 at 21:57
  • $\begingroup$ I get the notice: "The folder ‘/Public/public_html/cover’ doesn’t exist." It seems that your link doesn't include any user data so I am not sure it should link to one of your files. (I am mostly just curious about the argument!) $\endgroup$ – Mike Miller Jan 24 '19 at 22:57
  • 3
    $\begingroup$ Sorry for the many comments. I found a link to the slide here following through your webpage. $\endgroup$ – Mike Miller Jan 25 '19 at 0:34

There are several good answers but I thought I'd chuck in one more. Prior to all of the important results on geometrization mentioned above, Waldhausen had shown that any Haken manifold (any 2-sphere bounds a ball and the manifold contains an incompressible surface) has universal cover $\mathbb{R}^3$. Simple examples of Haken homology spheres are obtained by gluing two non-trivial knot complements in such a way that meridian and longitude are interchanged.

Piotr also asked about higher dimensional examples. These are harder to come by. In dimension $4$ there is "Some examples of aspherical 4-manifolds that are homology 4-spheres" by Ratcliffe and Tschantz, Topology Volume 44, Issue 2, March 2005, Pages 341-350. I don't know about dimensions higher than that.


Here are two references which also might be of help:

The introduction of

Saveliev, Nikolai, Invariants for Homology 3-spheres, Encyclopaedia of Mathematical Sciences, 140. Springer (2002).

has a description and construction of Seifert Fibered Homology Spheres with infinite fundamental group, $\S$1.1.4 in the 2002 edition. (I have yet to read the 2013 edition.) Also, Peter Scott's paper:

Scott, Peter, The geometries of 3-manifolds, Bull. Lond. Math. Soc. 15, 401-487 (1983). ZBL0561.57001.

could then serve as a reference for the fact that the universal cover of a Seifert fibered homology sphere with infinite fundamental group has $R^3$ as a universal cover. Here I am relying on the complete classification of orientable $S^2 \times \mathbb{R}$ manifolds (see p. 457-459 of Scott's paper).


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.