10
$\begingroup$

Let $X$ be a smooth complex projective variety. Let $M_{Higgs}(X, P)$ be the coarse moduli which universally corepresents the functor:

$$M^{\#}_{Higgs}(X, P): Sch/\mathbb{C}\longrightarrow \mathcal{Set}$$ which assigns to any scheme $S$ the set of isomorphism classes of semi-stable Higgs bundles $(E, \theta)$ over $X\times S$ with Hilbert poly $P$. Here $P=nP_0$, with $P_0$ being the Hilbert polynomial of $\mathcal{O}_X$ with respect to a fixed polarization $\mathcal{O}_X(1)$.

Then we have the proper Hitchin map $$H_1: M_{Higgs}(X, P) \rightarrow \bigoplus_{i=1}^nH^0(X, sym^i\Omega^1(X))$$

which maps $(E,\theta)$ to the characteristic polynomial of $\theta$. (Take the coefficients.)

Consider the Dolbeault Moduli space $M_{Dol}(X, P)$, which universally corepresents $$M^{\#}_{Dol}(X, P): Sch/\mathbb{C}\longrightarrow \mathcal{Set}$$ which assigns to any scheme $S$ the set of isomorphism classes of semi-stable Higgs bundles $(E, \theta)$ over $X\times S$ with Hilbert polynomial $P$ for which the Chern classes of $E$ vanish.

Since $M_{Dol}(X, P)$ is disjoint union of some of the connected components of $M_{Higgs}(X, P)$, we should also have the following proper Hitchin map $$H_2: M_{Dol}(X, P) \rightarrow \bigoplus_{i=1}^nH^0(X, sym^i\Omega^1(X))$$

Can any one give some interesting, even small, geometric applications of the Hitchin map $H_1$ or $H_2$? Any answers and references are appreciated.

$\endgroup$

2 Answers 2

10
$\begingroup$

You may know this already, but I'm not sure if you'll get many other answers. A nice result of Brunebarbe, Klingler and Totaro [Symmetric differentials and the fundamental group, Duke 2013] is that if the fundamental group of a smooth complex projective variety $X$ has a finite dimensional representation with infinite image, then $X$ carries a nonzero holomorphic symmetric differential form, i.e. $H^0(X, sym^i\Omega_X^1)\not=0$ for some $i>0$. As I recall, one step of their proof uses a special case, due to me, that the same conclusion holds if $\pi_1(X)$ has a nonrigid representation. The latter assumption amounts to saying that the "Betti" moduli space $$M_{Betti}(X,n)=Hom(\pi_1(X),GL_n(\mathbb{C}))//GL_n(\mathbb{C})$$ is an infinite set. The proof of the special case is easy, modulo Simpson's work. Since $M_{Betti}(X,n)$ is affine, and infinite, it must be noncompact in the classical topology. But $M_{Betti}(X,n)^{an}$ is homeomorphic to $M_{Dol}(X,n)^{an}$ (the moduli space of rank $n$ Higgs bundles with the usual conditions). Since the Hitchin map $H_2$ is proper, the non compactness of $M_{Dol}(X,n)^{an}$ forces $H_2$ to have infinite image, so $sym^i\Omega_X$ must have sections for some $i$ in the range $i=1,\ldots, n$.

$\endgroup$
1
  • $\begingroup$ Dear Prof. Arapura, I only saw your result in your paper. But not the one in [Symmetric differentials and the fundamental group, Duke 2013]. Thank you for you answer. $\endgroup$
    – Feng Hao
    Jul 15, 2017 at 15:09
2
$\begingroup$

If $X$ is a smooth projective curve with genus $g \geq 2$ and $M_{Higgs}(X,P)$ is smooth, the map: $$ H_1 : M_{Higgs}(X,P) \longrightarrow \bigoplus_{i} H^0(X, Sym^i \Omega_X)$$ gives $M_{Higgs}(X,P)$ a structure of complete algebraic integrable system. I think this result is due to Beauville. See : https://projecteuclid.org/download/pdf_1/euclid.acta/1485890603. As far as I know, it is quite difficult to provide examples of complete algebraic integrable systems.

EDIT : as Beauville comments below, the algebraic complete integrability of this moduli space was proved by Hitchin himself.

$\endgroup$
1
  • 2
    $\begingroup$ Actually this is due to Hitchin himself, and it was the main reason for him to considered this fibration. $\endgroup$
    – abx
    Jul 15, 2017 at 14:37

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.