MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Suppose that $K$ is a $p$-adic field, that is a field of characteristic $0$ whose ring of integers is a complete discrete valuation ring $\mathcal O_K$ and with residue field $k$ (algebraic closed) of characteristic $p$.

Let $X\rightarrow \mathop{Spec}(\mathcal{O}_K)$ be a projective threefold with generic fibre $X_K$ a (smooth) $K3$-surface. If $p>5$ we can assume that $X$ is a semi-stable minimal model of $X_K$.

Is it true that the special fibre $X_0 = X \otimes{k}$ is a combinatorial $K3$-surface?

I know that it is true for the complex numbers, that is, if $X\rightarrow \Delta$ is a projective semi-stable degeneration of $K3$-surfaces over the open complex disk, then the central fibre $X_0$ is combinatorial (in the sense of Firedman-Kulikov). (1. Persson, U.; Pinkham, H. Degeneration of surfaces with trivial canonical bundle. Ann. of Math. (2) 1981, 113, 45–66.) and (Kulikov).

Moreover Friedman studied the converse situation over the complex numbers and proved that any combinatorial $K3$-surface over the complex numbers is the central fibre of a family of $K3$-surfaces. 1. Persson, U.; Pinkham, H. Degeneration of surfaces with trivial canonical bundle. Ann. of Math. (2) 1981, 113, 45–66. 2. Friedman, R. Global smoothings of varieties with normal crossings. The Annals of Mathematics 1983, 118, 75–114.

Latter Nakkajima prove it for combinatorial $K3$ surfaces that admit a log-strtucture family of $K3$-surface. 1. Nakkajima, Y. Liftings of simple normal crossing log $K3$ and log Enriques surfaces in mixed characteristics. J. Algebraic Geom. 2000, 9, 355–393.

My question is then to know if it is know that for a semi-stable $K3$ surface over a field $K$ of mixed characteristic with minimal semi-stable model, the generic fibre is combinatorial.

share|cite|improve this question
Some of your references seem to have ended up in the middle of sentences. Also, "desecrate" should be "discrete". – user5117 Nov 18 '13 at 17:10
ups… that is right! – Yoyontzin Nov 18 '13 at 17:12
you should have a look at Maulik's paper "Supersingular K3 surfaces for large primes" (, and in particular to Section 4 – Christian Liedtke Nov 19 '13 at 10:47
Yes, I have seen that paper. He works on characteristic $p$ and I am on mixed characteristic. Looking at Maulik's arguments on section 4, he claims that having a semi-stable minimal model, then indeed the special fibre is combinatorial, however there is something that I am missing: He said that this follows by Nakkajima, but for Nakkajima we need to have that the special fibre to be a simple normal crossing log K3-surface. Why is it true in Maulik's paper? – Yoyontzin Nov 20 '13 at 3:25
Is is true that if we have a semi-stable minimal model $X\rightarrow \mathop{spec}(R)$ with $R$ a complete discreat valuation ring, with smooth generic fibre a $K3$-surface, then the special fibre $X_0$ (with the induced log structure) is a simple normal crossing Log $K3$-surface? Even if we assume that the special fibre is reduced and that its components are geometrically irreducible why is it $d$-semistable and why $H^1(X_0,\mathcal O_{X_0}) = 0$ and $\Omega_{X_0}^2(log) = \mathcal{O}_{X_0}$? – Yoyontzin Nov 20 '13 at 3:38

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.