I am trying to visualize the $SU(3)$ group used in quantum field theory. I have a (reasonably) good understanding of $SU(2)$ as the double cover of $SO(3)$ and also that this is homeomorphic to $S^3$. I have also read from other questions here that $SU(3)$ is "something like" $S^5\times SU(2)$ (would that make it "something like" $S^5\times S^3$?) but I am not sure if the "something like" entails homeomorphism or if it is some more complicated bundle.

Is there a way to visualize $SU(3)$ as a real manifold (ignoring its group structure if necessary to focus on its properties as a manifold)?


1 Answer 1


The special unitary groups fit into a sequence of fiber bundles

$$SU(n-1) \to SU(n) \to S^{2n-1}$$

coming from their actions on the unit spheres of $\mathbb{C}^n$ equipped with the standard inner product. For $n = 2$ we get a fiber bundle

$$SU(2) \to SU(3) \to S^5$$

exhibiting $SU(3)$ as a (nontrivial) $S^3$ bundle over $S^5$. This bundle splits rationally in the sense that $SU(3)$ is rationally homotopy equivalent to $S^3 \times S^5$; in particular it has the same rational homology, cohomology, and rational homotopy groups as $S^3 \times S^5$, and so in particular its Betti polynomial is

$$\sum b_i(SU(3)) t^i = (1 + t^3)(1 + t^5).$$

See also this MO question.

  • $\begingroup$ Thanks very much. I can see that I have a ways to go in order to grasp the non-triviality of this $S^3$ bundle over $S^5$ (at least it appears that there is only one such non-trivial bundle, according to the response on the MO question) but this gives me somewhere to start! $\endgroup$ Nov 12, 2015 at 2:55
  • 1
    $\begingroup$ @MattDickau: If you'd like to visualize nontrivial sphere bundles, the Klein bottle $S^1 \widetilde \times S^1$ is a good start. There's also a nontrivial $S^2$-bundle over $S^1$ and a nontrivial $S^2$-bundle over $S^2$. $\endgroup$
    – user98602
    Nov 12, 2015 at 3:38
  • $\begingroup$ The Klein bottle I am familiar with - it is like a torus but the longitudinal circles "invert" when you traverse around the axis. From that I can imagine a non-trivial $S^2$ over $S^1$ - it would be similar but with the longitudinal circles replaced by 2-spheres.Traversing around the axis would return you to the antipode of where you started. Having trouble when the base space is not $S^1$ though. $\endgroup$ Nov 12, 2015 at 6:27

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.