7
$\begingroup$

It is known that closed spherical and hyperbolic 3-manifolds are rigid. I.e., if two such manifolds are diffeomorphic, then they are isometric (moreover, I think, that every diffeomorphism is isotopic to an isometry). On the other hand, closed Euclidean 3-manifolds admit non-trivial Teichmüller spaces of Euclidean structures. What is known about more exotic Thurston's geometries, namely, Nil, $\widetilde{\rm SL}_2(\mathbb R)$ and Sol?

Let $M$ be a closed 3-manifold, $G$ be one of the mentioned Lie groups and $\mathcal T(M, G)$ be the space of the respective geometric structures on $M$ up to isotopy (say, endowed with the topology inherited from the respective space of representations of $\pi_1(M)$). Could it be more than one point? If yes, what is known about it?

$\endgroup$
3
  • 2
    $\begingroup$ Yes, typically not rigid, but there are some rigid exceptions among small Seifert manifolds. You can start by looking at Teichmuller spaces of base-orbifolds of Seifert manifolds. $\endgroup$ Feb 18 at 21:10
  • 1
    $\begingroup$ One issue comes from the non-uniqueness of metric on the simply connected cover $G/K$ itself. For instance, for $G=\mathrm{SOL}$ I don't know if there is a "better" choice (there is just a very weak condition of being invariant by $K$, but then $K$ is finite of order $8$). $\endgroup$
    – YCor
    Feb 18 at 21:13
  • 2
    $\begingroup$ A key point in recent work of Bamler and Kleiner on Smale conjecture for 3-manifold modelled on Nil is the proof (in lemma 2.4 in arxiv.org/abs/2108.03302) that any non-Haken closed 3-manifold modelled on Nil has a unique unit volume Nil-structure up to isometry. $\endgroup$ Feb 19 at 18:15

1 Answer 1

3
$\begingroup$

$\newcommand{\PSLt}{\widetilde{\mathrm{PSL}_2}}$Manifolds with $\PSLt$ geometry can have non-trivial moduli. For example, suppose that $S$ is a closed, connected, oriented surface with genus at least two. Let $M$ be a copy of $\mathrm{UT}S$, the unit tangent bundle of $S$. Then there is a (distinct) $\PSLt$ structure on $M$ coming from every (distinct) hyperbolic structure on $S$.

The argument is similar when $M$ is instead the unit tangent bundle over a hyperbolic orbifold $B$. The exception is when $B$ is a triangle orbifold: that is, a copy of the two-sphere with three cone points having angles $2\pi/p$, $2\pi/q$, and $2\pi/r$ so that $1/p + 1/q + 1/r > 1$. In this case $B$ has a unique hyperbolic structure (the double of the corresponding triangle); also $\mathrm{UT}B$ has a unique $\PSLt$ structure.

The story for $\mathbb{H}^2 \times \mathbb{R}$ geometry follows similar lines.

$\endgroup$
2
  • $\begingroup$ What do you call PSL geometry? Do you mean $\widetilde{\mathrm{PSL}_2}$? $\endgroup$
    – YCor
    Feb 19 at 18:59
  • $\begingroup$ Yes, I mean that - but that takes a long time to typeset... $\endgroup$
    – Sam Nead
    Feb 19 at 21:08

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.