3
$\begingroup$

Let $D$ be a normal crossing divisors on a projective smooth algebraic variety $X$ over a field $k$ of characteristic zero. Put $n=\dim X$ and denote by $D_i$ the irreducible components of $D$.

I'm trying to understand why the degree of the (n-1)-th Chern class of the restriction of the sheaf of logarithmic differential forms to a component $D_i$, that is

$\deg c_{n-1}(\Omega^1_X(\log D)_{| D_i})$

equals the Euler characteristic of $D_i - \cup_{j \neq i} D_j$.

Can anybody help me?

Thanks!

$\endgroup$
  • 1
    $\begingroup$ Maybe you could try to restrict the exact sequence $$0 \to \Omega^1_X \to \Omega^1_X(\log D) \to \bigoplus \mathcal{O}_{D_j} \to 0$$ to the divisor $D_i$. $\endgroup$ – Francesco Polizzi Jan 29 '13 at 15:15
  • $\begingroup$ I don't manage to get it from here. Could you please give more details? Thanks! $\endgroup$ – ncd Jan 29 '13 at 16:41
3
$\begingroup$

Consider the following commutative diagram:

$$\begin{aligned} &&&& 0\qquad \qquad&\to &0\qquad\qquad && \\ &&&&\downarrow\qquad\qquad && \downarrow\qquad\qquad \\ 0 &\to &\Omega_X(\log D)(-D_i)\quad &\to &\Omega_X(\log (D-D_i))\quad &\to &\Omega_{D_i}(\log (D-D_i)|_{D_i})\quad &\to &0\\ && \downarrow\,\simeq \qquad\qquad &&\downarrow\qquad\qquad && \downarrow\qquad\qquad \\ 0 &\to &\Omega_X(\log D)(-D_i)\quad &\to &\Omega_X(\log D)\qquad &\to &\Omega_X(\log D)|_{D_i}\quad &\to &0\\ &&&&\downarrow\qquad\qquad && \downarrow\qquad\qquad \\ &&&& \mathscr O_{D_i}\quad \qquad&\overset\simeq\to & \mathscr O_{D_i}\quad\qquad && \\ \end{aligned}$$

The rows and the middle column are standard short exact sequences for logarithmic differentials. The 9-lemma implies that then the last column is also exact. Then the usual formula for Chern classes of short exact sequences (of locally free sheaves!) shows that $$ c_{n-1}(\Omega_X(\log D)|_{D_i})=c_{n-1}(\Omega_{D_i}(\log (D-D_i)|_{D_i})). $$

This looks very close to what you probably need, but it seems to me that there might be a sign difference of $(-1)^{n-1}$. It is possible that I made a mistake somewhere, but in any case, this method should work.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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