13
$\begingroup$

For compact manifolds, Hodge Theory tells us that (de Rham) cohomology is finite dimensional. What about non-compact manifolds? That is:

Can non-compact manifolds have infinite dimensional cohomology?

If the answer is yes, is there an example for which this is easy to see?

$\endgroup$
1
  • $\begingroup$ A necessary condition is that the manifold must not admit a finite good cover $\endgroup$ Commented Jul 10, 2020 at 10:57

2 Answers 2

12
$\begingroup$

The (non-compact) manifold $\mathbb R^2\setminus \mathbb Z^2$ should have infinite first homology and cohomology.

$\endgroup$
11
$\begingroup$

Sure. Here are two examples:

  1. Let $M$ be the real line with all integer points removed. Then $M$ is the disjoint union of countably many open intervals, and $H^0 (M, \mathbb{R})$ is a real vector space of uncountably infinite dimension. (To be precise, it is the dual of the real vector space of countably infinite dimension.)

  2. Let $M$ be the real plane with the set $\{ (n, 0) : n \in \mathbb{Z} \}$ removed. It's not too hard to see that $M$ is homotopy-equivalent to a chain of countably many circles joined side-by-side. It follows that $H^1 (M, \mathbb{R})$ is a real vector space of uncountably infinite dimension – we can concoct differential forms with a prescribed singularity at each hole independently, and their cohomology classes are distinct because they can be distinguished by integrating along a loop around the relevant singularities.

$\endgroup$
3
  • 1
    $\begingroup$ However the last De Rham cohomology vanishes for non compact manifold. $\endgroup$ Commented Feb 23, 2016 at 13:44
  • $\begingroup$ Interesting answer. Anyway I don't get why we have unconuntable dimension in 1. Shouldn't the zero cohomology group count the number of connected components? Does this hold only for the finite case ? $\endgroup$ Commented Jul 10, 2020 at 10:55
  • $\begingroup$ The 0th homology group counts the number of connected components. The 0th cohomology group counts the number of locally constant functions. $\endgroup$
    – Zhen Lin
    Commented Jul 10, 2020 at 13:48

You must log in to answer this question.

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