4
$\begingroup$

Let S be a singular complex algebraic surface. Let $\psi: \chi \rightarrow \Delta$ be a QQ-Gorenstein smoothing of S (central fibre is isomorphic to S). Let G be a finite group acting on the fibres of above deformation with no fixed points. What can we say about the new deformation family (where fibres are quotients of fibres of $\psi: \chi \rightarrow \Delta$). Is it QQ-Gorenstein smoothing of S/G? Also provide some reference related to this.

$\endgroup$
2
  • $\begingroup$ If I understand your question correctly you are essentially asking whether a fixed point free quotient of a $\mathbb Q$-Gorenstein variety is still $\mathbb Q$-Gorenstein. Right? $\endgroup$ Aug 15, 2013 at 19:47
  • $\begingroup$ @SándorKovács: Thats right. Is it true in general, or we need some extra conditions? $\endgroup$
    – SAG1
    Aug 16, 2013 at 19:27

1 Answer 1

2
$\begingroup$

I assume that by a "$\mathbb Q$-Gorenstein smoothing" you mean that $\chi$ is $\mathbb Q$-Gorenstein and probably $\Delta$ is the unit disk.

I also assume that by "$\mathbb Q$-Gorenstein" you probably mean something like

  1. $\chi$ is normal
  2. $K_\chi$ is $\mathbb Q$-Cartier,

but you're probably not requiring that $\chi$ be Cohen-Macaulay.

Anyway, I think that the answer is "yes" and the proof is very simple:

Let $\pi:\chi\to \chi'=\chi/G$ be the quotient map. Since $G$ is finite and acts without fixed points, $\pi$ is étale and hence $\chi'$ is locally analytically isomorphic to $\chi$. More precisely for all $x\in\chi'$, and for all $y\in\pi^{-1}(x)\subset \chi$, the analytical or étale local schemes $(\chi',x)$ and $(\chi, y)$ are isomorphic. Hence $\omega_{\chi'}^{[m]}$ (the reflexive hull of the $m^{\text{th}}$ tensor power) is locally free in the étale topology for some $m\in \mathbb N$ and hence the stalk $(\omega_{\chi'}^{[m]})_x$ has rank $1$. This is independent of the topology and so it follows that $\omega_{\chi'}^{[m]}$ is a line bundle and so $\chi'$ is $\mathbb Q$-Gorenstein.

You might want to replace "étale" with analytic if you prefer that language, but I think the proof works the same way.

$\endgroup$
1
  • $\begingroup$ Sandor Kovacs: Thats perfect. Thanks. $\endgroup$
    – SAG1
    Aug 28, 2013 at 11:51

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.