# Topological rigidity of compact manifolds in dimension three

The Borel Conjecture asserts that homotopy equivalent aspherical closed manifolds are homeomorphic, which is still open in general. But, for three-dimensional manifolds, this conjecture holds (I read this in Bessieres-Besson-Boileau), whose proof depends on the geometrization theorem (Perelman).

Question: Does the relative version of the Borel conjecture also hold for compact 3-manifolds with boundary (by the geometrization)? The relative version: If there is a homotopy equivalence between two compact aspherical manifolds that is a homeomorphism between their boundaries, are those manifolds homeomorphic?

-

@unknown: The Poincar&eacute; Conjecture implies that a $3$--manifold is irreducibile if it is aspherical. Say your manifolds are $M$ and $N$. If they are homotopy equivalent, they have isomorphic fundamental groups. If this group does not contain a $\mathbb{Z} \oplus \mathbb{Z}$ subgroup, then $M$ and $N$ are hyperbolic, by Geometrization, and so, by Mostow Rigidity, they are homeomorphic. – Richard Kent Oct 7 '11 at 3:10
@unkown: Continued: If this group contains a $\mathbb{Z} \oplus \mathbb{Z}$ subgroup, then, by the (strong) Torus Theorem, $M$ and $N$ are either Haken or small Seifert fibered spaces. If both are Haken, then Waldhausen's theorem comes to the rescue. I think you do something like show that $\pi_1$ of a small SFS isn't an amalgamated product to rule out $M$ Haken and $N$ a small SFS. If both are small SFSs, then the euler class tells you they're homeomorphic. – Richard Kent Oct 7 '11 at 3:15
@unkown: Actually, to rule out $M$ Haken and $N$ small SFS, argue like this: $\pi_1$ has infinite cyclic center, and so $M$ is a SFS too (this is part of the proof of the Strong Torus Theorem). The center is a completely algebraic thing, and so the quotient of $\pi_1(M)$ by the center is the orbifold fundamental group of the underlying orbifold of the Seifert fibered structure on $M$. This must be isomorphic to the orbifold group for the underlying orbifold for $N$, and so $M$ is a small SFS. Euler class to the rescue. – Richard Kent Oct 7 '11 at 3:28