In a handful of contexts people study Calabi-Yau threefolds formed by taking the fiber product of two rational elliptic surfaces. I can't find any detailed explanation of why such geometries are actually Calabi-Yau, so I think it's just a straightforward computation which I don't fully understand.

Let $\pi: S \to \mathbb{P}^{1}$ and $\pi' : S' \to \mathbb{P}^{1}$ be two rational elliptic surfaces, and define their fiber product

$$X = S \times_{\mathbb{P}^{1}} S'.$$

With some mild assumptions on $S$ and $S'$, $X$ should be a Calabi-Yau threefold, and I'm hoping someone can help me complete the proof of this. In other words, I want to see that $\omega_{X}=0$ or $K_{X}=0$ (however, note that in general, $X$ will certainly not be smooth).

I believe one should start by considering the obvious map induced by $\pi$ and $\pi'$

$$f: S \times S' \to \mathbb{P}^{1} \times \mathbb{P}^{1}.$$

We can then write $X$ as the pullback of the diagonal $\Delta \subset \mathbb{P}^{1} \times \mathbb{P}^{1}$,

$$X = f^{*} \Delta.$$

So we can realize $X$ as a hypersurface in $S \times S'$, so you should then be able to apply the adjunction formula:

$$\omega_{X} = \omega_{S \times S'}|_{X} \otimes \mathcal{N}_{X/S \times S'},$$

where $\mathcal{N}_{X/S \times S'}$ is the normal bundle of $X$ in $S \times S'$. However, I'm sort of stuck on how to proceed -- How can one explicitly handle both of the two factors in the above tensor product and show they somehow cancel to give 0?


2 Answers 2


The diagonal $\Delta $ is linearly equivalent to $\{p\}\times \mathbb{P}^1 +\mathbb{P}^1\times \{p\} $ for any $p$ in $\mathbb{P}^1$. Therefore $X$ is the zero locus in $S\times S'$ of a section of $L:=\pi^*\mathcal{O}(1) \boxtimes \pi'^*\mathcal{O}(1) $. On the other hand, standard theory of elliptic surfaces gives $\omega _S=\pi ^*\mathcal{O}(-1) $ and $\omega _{S'}=\pi' ^*\mathcal{O}(-1) $, therefore $\omega _{S\times S'}= L^{-1}$. Then the adjunction formula gives indeed $\omega_X\cong \mathcal{O}_{X}$.

  • $\begingroup$ Thanks a lot, that nearly clears everything up for me. In your expressions for $\omega_{S}$ and $\omega_{S'}$ presumably you're specializing to certain rational elliptics? Because I believe for general rational elliptic surfaces the fiber product will not be Calabi-Yau. $\endgroup$
    – Benighted
    Jan 31, 2019 at 17:15
  • 1
    $\begingroup$ I am only assuming that the elliptic fibrations have no multiple fibers and are relatively minimal (no $(−1)$-curve in any fiber). Then the "canonical bundle formula" gives $\omega _{S}\cong \mathcal{O}_{S}(-F)$, where $F$ is a fiber (see e.g. Barth et al., Corollary 12.3). $\endgroup$
    – abx
    Jan 31, 2019 at 17:49

$S\times_{\mathbb{P}^1}S'$ is a complete intersection in $\mathbb{P}^1\times \mathbb{P}^2\times \mathbb{P}^2$: it is given by two equations of degree $(1,3,0)$ and $(1,0,3)$. The canonical bundle is trivial by the adjunction formula.

  • $\begingroup$ I see, my argument only applies directly in the generic case. @abx gives an argument which applies to all relatively minimal rational elliptic surfaces with a section. $\endgroup$
    – Jim Bryan
    Jan 31, 2019 at 17:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

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