# Free action of SL_2(F_p) on a sphere

Let $p>2$ be prime. Then for abstract reasons the special linear group $\text{SL}_2({\mathbb F}_p)$ possesses a free action on some sphere (one has to check that any abelian subgroup of $\text{SL}_2({\mathbb F}_p)$ is cyclic and that there's at most one element of order $2$).

Does somebody know a concrete example for such a free action for general $p$? (For $p=5$, for example, $\text{SL}_2({\mathbb F}_p)$is the binary icosahedral group which is a subgroup of ${\mathbb S}^3$ thus acting freely on it by multiplication; I'd like to know if there's one single action that can be written down for all $p$ simultaneously).

• Does acting on a sphere mean acting by orthogonal matrices, or is it just a topological action? The former is impossible, because the smallest nontrivial representation of SL_2(F_p) has dimension (p-1)/2. – David E Speyer Feb 1 '10 at 13:43
• Yes, I just mean a free topological action, not necessarily orthogonal. But concerning your remark on orthogonal actions: How do know that there cannot exist any orthogonal action from the fact on the dimension of the smallest nontrivial representation? – Hanno Becker Feb 1 '10 at 14:25
• Could the issue lie in your phrasing "some sphere"? I read this as meaning $\mathbb{S}^n$ for some $n$; maybe David was thinking $\mathbb{S}^2$? – Pete L. Clark Feb 1 '10 at 14:34
• Hanno, could you please explain a bit what you mean by "abstract reasons"? – Dmitri Panov Feb 2 '10 at 0:07
• I've added the gt tag. I don't know the answer to either question; all I can say is that there is no single sphere upon which all SL_2(F_p) act orthogonally. This follows from a classical result by Jordan on abelian normal subgroups of finite linear groups in characteristic zero. – algori Feb 2 '10 at 1:33

Apparently, a linear free action exists only for $p=5$ (if $p\ge 5$), see paper by C. Thomas "Almost linear actions by $SL_2(p)$ on $S^{2n-1}$". There is a weaker notion of an "almost linear" action, and it seems that constructing such actions is a fairly complicated business, using state-of-the-art differential geometry and topology; see arXiv:math/9911250. It seems that simple explicit actions for higher $p$ are not expected (also, of course, the sphere must be odd dimensional, since an orinentation preserving self-map of an even dimensional sphere has a fixed point by the Lefschetz theorem).