3
$\begingroup$

Suppose $G$ is a connected, compact Lie group and $S^1 \subset G$ is a central subgroup. Can I write $G$ as a quotient of a product group $$G=(S^1 \times H)/Z$$

where the $S^1$ factor maps onto the circle group $S^1 \subset G$, $Z \subset S^1 \times H$ is a finite subgroup which maps faithfully to the $S^1$ factor, and $H$ another compact Lie group?

$\endgroup$

1 Answer 1

1
$\begingroup$

First let me address if $\mathcal Z$ is compact abelian, but not necessarily connected. Let $\mathcal Z^0$ be the identity component. Then $\mathcal Z^0$ is a torus and $S^1 \subset \mathcal Z^0$ splits as a direct factor (its cocharacter is primitive). Further the component group of an abelian Lie group splits off, so $\mathcal Z \cong \mathcal Z^0 \times \mathcal Z/\mathcal Z^0$. Thus $\mathcal Z = S^1 \times \mathcal Z'$ for some $\mathcal Z'$.

Now let $G$ be as in the question. Let $\mathfrak g$ be the Lie algebra of $G$. By Levi decomposition $\mathfrak g = \mathfrak z \oplus \mathfrak k$ where $\mathfrak z$ is the center of $\mathfrak g$ and $\mathfrak k$ is (real) semisimple. Then $\mathfrak k$ is the Lie algebra of a simply connected compact group $\tilde K$. By Lie's third theorem $\mathfrak k \to \mathfrak g$ lifts to $\tilde K \to G$. If $\mathcal Z$ is the center of $G$, then the multiplication $\mathcal Z \times \tilde K \to G$ is a homomorphism of compact groups which is an isomorphism on tangent spaces, hence an isogeny. Let $K$ be the image of $\tilde K$. Write as above $\mathcal Z = S^1 \times \mathcal Z'$, and define $H = \mathcal Z'K \subset G$. Again $S^1 \times H \to G$ is an isogeny with kernel $Z \subset S^1 \cap H \subset \mathcal Z$.

$\endgroup$
2
  • 1
    $\begingroup$ Although it's a nice touch, you don't need to handle disconnectedness of the centre; a compact Lie group is already the almost-direct product of the simply connected cover of its derived group and the maximal central torus--no need for the full centre, by your own second-paragraph argument about tangent spaces! $\endgroup$
    – LSpice
    Commented Jul 28 at 16:47
  • 1
    $\begingroup$ @LSpice good point, thank you. The second paragraph goes through perfectly well with $\mathcal Z^0$ everywhere instead of $\mathcal Z$. $\endgroup$ Commented Jul 28 at 16:59

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .