MathOverflow will be down for maintenance for approximately 3 hours, starting Monday evening (06/24/2013) at approximately 9:00 PM Eastern time (UTC-4).

## When is a extension of $\mathbb{Z}$ by a free group a CAT(0) group?

The question has an easy answer, if one replaces free by free abelian: Then the resulting group is always solvable and a solvable subgroup of a CAT(0) group is virtually abelian. If the resulting was CAT(0), then the chosen automorphism $\varphi$ in $\mathbb{Z}^n\rtimes_\varphi \mathbb{Z}$ would have finite order - otherwise the group would not be virtually abelian.

Now one can ask the same question for the free group instead or the free abelian group. I would like to know for which automorphisms $\varphi$ of the free group $F_n$ the group $F_n\rtimes_\varphi \mathbb{Z}$ is CAT(0).

I only know, that $F_n \times \mathbb{Z}$ is CAT(0). I think that if the chosen automorphism has finite order, then the result should be CAT(0) (although I don't have a proof). And I do not know automorphism, that gives a non-CAT(0) group.

-
If the automorphism is finite order then it can be realised by a combinatorial automorphism of a graph. (This basically follows from Stallings' Ends Theorem.) The resulting mapping torus is CAT(0). – HW Feb 1 2011 at 15:49

You'll find some examples of CAT(0) free-by-cyclics in

and

Barnard and Brady, "Distorsion of surface groups in CAT(0) free-by-cyclic groups"

-
 For future reference, there is a free version of the first paper mentioned

An example of a free-by-cyclic group that is not CAT(0) was given by Gersten. It is constructed from the automorphism of $F_3\cong\langle a,b,c\rangle$ that sends

$a\mapsto a,~b\mapsto ba,~ c\mapsto ca^2~.$

The idea of the proof is to think about translation lengths and flats in any CAT(0) space on which it acts. As $\langle a,t\rangle\cong\mathbb{Z}^2$, it stabilises some flat. But $t$, $at$ and $a^2t$ are all conjugate, so have the same translation lengths. A little thought shows that this is impossible in a flat.

Note that, in many respects, (fg free)-by-cyclic groups are difficult to distinguish from CAT(0) groups. For instance, they have quadratic isoperimetric inequality.

As pointed out in this blog post, the question

Which free-by-cyclic groups are CAT(0)?

is Question 7.9 of Bridson's AIM article about `Problems concerning hyperbolic and CAT(0) groups'.

-
 Here, $t$ is the stable letter, of course. – HW Feb 1 2011 at 15:57

If you take the mapping torus of an automorphism of a surface with boundary, then it has a non-positively curved metric by a result of Leeb. Such automorphisms though will be sparse in the set of all automorphisms of a free group (except in rank 2, as noted in a paper of Brady).

-