I am interested in the existence of a surface $X$ over $\mathbb{C}$ with the following properties (or a reason for why one cannot exist):

  • $X$ is slc (and not-normal)
  • There is rational curve $C \subset X$ contained in the double locus (i.e. non-normal locus)
  • The cotangent $\Omega^1_Y$ is ample, where $Y \to X$ is the normalization

Note that this last condition implies that $Y$ has no rational or elliptic curves. In particular, since the pre-image of the double locus under the normalization is a 2-to-1 cover branched over the pinch points, Riemann-Hurwitz gives that $C$ contains at least 6 pinch points.

The only examples I know of non-normal surfaces with a rational curve in the double locus containing $> 2$ pinch points have non-positive Kodaira dimension.

  • $\begingroup$ What does "slc" mean? $\endgroup$ – Sasha Mar 28 '18 at 10:08
  • $\begingroup$ @Sasha slc stands for Semi Log Canonical singularities. It is a higher dimensional analogue of nodes for curves and it appears naturally in the study of moduli spaces of higher dimensional varieties (and pairs). More in detail is a non-normal version of log canonical singualirities, in which one requires the normalization to have lc singualirities plus other technical conditions. For a formal definition and properties see Chapter 3 of Kollár's book. $\endgroup$ – Srks Mar 28 '18 at 17:55
  • $\begingroup$ Here you can see an explicit list of slc singularities in dim 2. $\endgroup$ – Srks Mar 28 '18 at 17:55
  • $\begingroup$ Thanks. So, if you find a genus 2 curve $C \subset Y$ and glue together the orbits of the hyperelliptic involution of $C$ on $Y$, will the resulted surface be slc? $\endgroup$ – Sasha Mar 28 '18 at 19:19
  • $\begingroup$ @Sasha can you explain what you mean by "glue together the orbits of the involution of the curve inside the surface? $\endgroup$ – Srks Mar 29 '18 at 0:28

Here is a possible example building up on what I think was the idea of @Sasha.

Take $Y$ to be a non-isotrivial fibration of genus $g_1 \geq 3$ curves over a genus $g_2 \geq 2$ curve that contains a closed fiber $C'$ which is hyperelliptic. Such surfaces (Kodaira fibrations) have ample cotangent.

Consider the surface $X$ obtained by the pushout of $Y$ given by the quotient by the involution, i.e.

$\require{AMScd}$ \begin{CD} C' @>\sigma>> \mathbb{P}^1\\ @V V V @VV V\\ Y @>>> X \end{CD}

This is an slc surface $X$, and $Y$ has a finite map to $X$ that is birational and finite. Moreover one can show that $Y \to X$ is the normalization. By construction of the pushout the image of the hyperlliptic fiber $C'$ is a $\mathbb{P}^1$ inside the double locus of $X$.

Follow up question: Is this surface contained in the closure of the main component of the moduli space of surfaces of general type (i.e. is $X$ a smoothable surface in the sense of KSBA)?


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.