4
$\begingroup$

There exists the following result in the literature: There exists a polarized $K3$ surface $(X, H)$ of genus $3$ and a smooth irreducible curve $C$ on $X$ satisfying $C^2 =4$, $C.H=6$ such that $\text{Pic}(X) \cong \mathbb Z[H] \oplus Z[C]$. The theorem follows from [https://arxiv.org/pdf/math/9805140.pdf] theorem $1.1(iv)$.

Now Let's consider $X$ to be a smooth quintic hypersurface in $\mathbb P^3$ with Picard number $2$. Then I think it can be shown that $\text{Pic}(X) \cong \mathbb Z[H] \oplus Z[H']$, where $H$ is the hyperplane class and $H'$ is some divisor. Now in order to locate the ample line bundles in $\text{Pic}(X)$ using the Nakai-Moishezon criterion, we must know the intersection numbers $H.H'$ and $H'^2$.

In this context my question is the following: Does there exist in the literature an analogous existence result as the first-mentioned theorem for smooth quintic hypersurface with Picard number $2$?

To be more precise: Does there exist a polarized smooth quintic hypersurface $(X, H)$ in $\mathbb P^3$ and a smooth irreducible curve $C$ for which the intersection numbers $C^2$ and $C.H$ are known and $\text{Pic}(X) \cong \mathbb Z[H] \oplus Z[C]$?

Can someone give me any reference which could be even remotely useful in the context of finding out such $(X,H)$ and $C$

Any help from anyone is welcome.

$\endgroup$

1 Answer 1

4
$\begingroup$

A natural source of surfaces in $\mathbb{P}^{3}$ with Picard number $> 1$ is given by linear determinantal surfaces, i.e. zero sets of square matrices of linear forms. In what follows, let $X \subseteq \mathbb{P}^{3}$ be a smooth linear determinantal surface of degree $d \geq 2$ (smoothness can arranged by taking the $d \times d$ matrix to be sufficiently general) and let $H \in {\rm Pic}(X)$ be the class of a hyperplane section.

It can be checked that $X$ contains a curve $C$ which is the degeneracy locus of a $d \times (d-1)$ matrix of linear forms. For degree/genus reasons, this $C$ cannot be a complete intersection of $X$ and another surface in $\mathbb{P}^{3}$.

There is a Noether-Lefschetz-type result which says that if $X$ is a generic linear determinantal surface of degree $d,$ the Picard group ${\rm Pic}(X)$ is isomorphic to $\mathbb{Z}H \oplus \mathbb{Z}C.$ Here is a reference for this result, as well as others which go further than determinantal surfaces:

Lopez, Angelo Felice; Noether-Lefschetz theory and the Picard group of projective surfaces, Mem. Amer. Math. Soc. 89 (1991), no. 438, x+100 pp.

EDIT: The aforementioned result on generic linear determinantal varieties is a corollary of the following result of Lopez; if $C \subseteq \mathbb{P}^{3}$ is a smooth connected curve whose homogeneous ideal is generated in degree at most $d-1,$ then the general surface of $S$ of degree $d$ containing $C$ is smooth and ${\rm Pic}(S) \cong \mathbb{Z}H \oplus \mathbb{Z}C.$ This holds in particular when $d \geq 2$ and $C$ is a line in $\mathbb{P}^{3}.$

$\endgroup$
8
  • 1
    $\begingroup$ thanks for the answer. Is the curve $C$ mentioned in the answer irreducible? Are the intersection numbers $C.C$ and $C.H$ immediate from the discussion? (apology in advance, if this is trivial to compute). Finally, Is theorem $(II)3.1$ mentioned there relevant in this context? $\endgroup$
    – User
    Commented May 26, 2020 at 11:49
  • 1
    $\begingroup$ also as the general surface of degree $d \geq4$ is not linear determinantal, but linear pfaffian if $d \leq 15$, so in this case are we talking about smooth quintic surfaces which is not general? $\endgroup$
    – User
    Commented May 26, 2020 at 14:26
  • 1
    $\begingroup$ The curve $C$ is cut out by the maximal minors of a $d \times (d-1)$-submatrix of the matrix of linear forms cutting out $X,$ so by general results on determinantal varieties we can ensure $C$ is smooth and connected, and therefore irreducible. The degree $C \cdot H$ and genus can of course be read off the Hilbert polynomial, which in turn can be read off the Eagon-Northcott resolution of the associated determinantal ideal. You can then determine $C \cdot C$ from the adjunction formula on $X.$ I can't access the reference at the moment, so I can't speak to the relevance of Theorem (II)3. $\endgroup$ Commented May 26, 2020 at 17:12
  • 1
    $\begingroup$ As to your second comment, you are correct that the smooth surfaces in under discussion are not general when $d \geq 4.$ $\endgroup$ Commented May 26, 2020 at 17:14
  • 1
    $\begingroup$ In this context I have one further question : Does it make sense to expect the existence of a smooth quintic hypersurface $X$ in $\mathbb P^3$ containing a line $L$ (even if this is rare) such that $\text{Pic}(X)$ is generated by the hyperplane section $H$ and the line $L$? $\endgroup$
    – User
    Commented Jun 28, 2021 at 7:00

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.