4
$\begingroup$

I'm reading this paper and at page 67, he states that for any line bundle $L$ over a Rieman surface there is a connection $A$ whose curvature is $$ F(A)=(\deg L)\omega, $$ where $\omega$ is a positive form. Does anyone know how can I prove it without using algebraic geometry arguments?

$\endgroup$

2 Answers 2

5
$\begingroup$

Let me expand a bit Henri's (hi Henri!) answer, even if this is completely standard. In general, given a compact Kähler manifold $X$ of any dimension, given a holomorphic line bundle $L\to X$, and given any $(1,1)$-form $\eta$ representing $c_1(L)$, there exists a smooth hermitian metric $h_\eta$ on $L$ such that its Chern curvature satisfies $$ \frac i{2\pi}\Theta(L,h_\eta)=\eta. $$ The proof goes as Henri was saying, that is: fix any hermitian metric $h$ on $L$, take its Chern curvature form, then $\frac i{2\pi}\Theta(L,h)\in c_1(L)$. But then, $\frac i{2\pi}\Theta(L,h)-\eta$ is an exact form, and the $\partial\bar\partial$-lemma tells us that there exists a real function $f$ on $X$ such that $$ \frac i{2\pi}\Theta(L,h)-\eta=\frac i{2\pi}\partial\bar\partial f. $$ Now, make a conformal change of the metric $h$ using $f$ as follows: $h_\eta:=e^fh$. The Chern curvature of this new metric satisfies $$ \frac i{2\pi}\Theta(L,e^fh)=-\frac i{2\pi}\partial\bar\partial f+\frac i{2\pi}\Theta(L,h)=\eta. $$

To finish with, if your $X$ is a compact Riemann surface (you didn't mention explicitly compactness in your question, but I guess you just missed it? ) then $H^2(X,\mathbb Z)\simeq\mathbb Z$. Thus (I also also imagine that your Kähler class $[\omega]$ was chosen to be a positive generator of $H^2(X,\mathbb Z)$, otherwise what you say is trivially false...), $c_1(L)=\deg L$ and $\deg L\,\omega$ is in the right cohomology class (namely $c_1(L)$), so that in the preceding argument you can replace $\eta$ by $\deg L\,\omega$.

$\endgroup$
1
  • 1
    $\begingroup$ Thanks for the details. It helped me a lot! $\endgroup$ Jun 10, 2019 at 16:11
4
$\begingroup$

Let $h_0$ be any hermitian metric on $L$, with curvature $F(h_0)$.

By Hodge theory, there exists a function $f\in \mathcal C^{\infty}(X)$ such that $$\Delta_{\omega} f =\Lambda_{\omega} F(h_0) - \mathrm{deg}(L).$$

Define $h:=e^{f}h_0$, it satisfies $\Lambda_{\omega} F(h)=\mathrm{deg}(L)$.

$\endgroup$
2
  • $\begingroup$ Thanks for your answer Henri! Could you please give me more details? Which theorem from Hodge theory are you using? Can you recommend me a reference where I can find it? Sorry about this, but I'm not familiar with this subject yet. $\endgroup$ Jun 10, 2019 at 14:53
  • 1
    $\begingroup$ Essentially, the $\partial\bar\partial$-lemma. $\endgroup$
    – diverietti
    Jun 10, 2019 at 15:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.