Let $X$ be a compact metric space and consider the sheaf cohomology group $H^1(X, \mathbb{T})$. From a class in $H^1(X, \mathbb{T})$, I can get a principal $\mathbb{T}$-bundle over $X$ and from this, an associated complex line bundle. What is the relationship between classes in $H^1(X, \mathbb{T})$ and complex line bundles? Does one class give me the same line bundle up to isomorphism?
-
1$\begingroup$ What's $\mathbb{T}$? $\endgroup$– Donu ArapuraCommented Dec 7, 2020 at 22:15
-
2$\begingroup$ @Donu Arapura: presumably $U(1)$? $\endgroup$– QfwfqCommented Dec 7, 2020 at 23:08
-
4$\begingroup$ I assume $\mathbb T =U(1)$ the circle group. From line bundles to cohomology it's: take a trivializing cover $\{U_\alpha\}$ and transition functions $\{g_{\alpha\beta}\}$; the latter represent a Cech cohomology class. From principal bundles to line bundles it's associated bundle construction $L=P\times^\mathbb{T}\mathbb{C}$. From line bundles to principal bundles it's taking the (unitary) frame bundle $L\mapsto \mathrm{Fr}^{U(1)}(L)$. From cohomology to line bundles it's: take a Cech cocycle $(\{U_\alpha\},\{g_{\alpha \beta}\}$ representative and use it as transition functions. $\endgroup$– QfwfqCommented Dec 7, 2020 at 23:17
-
2$\begingroup$ Yes, by $\mathbb{T}$ I mean $U(1)$. $\endgroup$– Francine LaporteCommented Dec 8, 2020 at 0:21
-
1$\begingroup$ Are you seeing $\mathbb{T}$ with the discrete topology or are you imagining as some kind of sheaf of topological groups? $\endgroup$– Denis NardinCommented Dec 8, 2020 at 8:41
1 Answer
At least if $X$ has the homotopy type of a CW complex, there is a natural isomorphism between $H^1(X; \mathbb T)$ and the group of isomorphism classes of line bundles on $X$ under tensor product.
The usual way this is phrased is that the first Chern class defines an isomorphism from the group of line bundles to $H^2(X;\mathbb Z)$. For example, and for a proof, see Hatcher, "Vector bundles and $K$-theory," Prop. 3.10 (p. 86).
Now consider the short exact sequence of sheaves
$$0\to \mathbb Z\to\mathbb R\to\mathbb R/\mathbb Z\to 0,$$
where $\mathbb R$ carries the continuous topology (i.e., this is the sheaf of continuous real-valued functions on $X$). We have $\mathbb R/\mathbb Z\cong\mathbb T$. There is an induced long exact sequence in cohomology, but as Donu Arapura notes in an answer to a different MathOverflow question, $H^k(X;\mathbb R)$ vanishes for $k > 0$. Therefore the long exact sequence simplifies to
$$ 0 \to H^1(X; \mathbb T)\longrightarrow H^2(X; \mathbb Z)\to 0, $$
so $H^1(X;\mathbb T)$ is isomorphic to the group of line bundles. It takes a little more work to see that the isomorphism is the same as the map you described (associated line bundle to a principal $\mathbb T$-bundle), but that is also true.
Not all compact metric spaces have the homotopy type of CW complexes, as noted by Milnor (end of section 1). I unfortunately don't know what the answer to your question is for those spaces.