1
$\begingroup$

Given a Segre product $\mathbb P^m \times \mathbb P^n$, or more generally $\mathbb P^{m_1}\times\cdots\times\mathbb P^{m_n}$, is there a characterization in terms of $m$ and $n$, or the $m_i$, for the Segre product to be Gorenstein?

Similar question for $\nu_d(\mathbb P^n)$

Added: I am interested in the case when when the corresponding graded ring of Segre/Veronese variety is Gorenstein i.e. it has a self dual free resolution or equivalently its Hilbert series $f(t)$ is symmetric so that $f(1/t)=(-1)^n t^sf(t) $.

$\endgroup$
2
  • 3
    $\begingroup$ Do you really mean Gorenstein (in the intrinsic sense) or arithmetically Gorenstein? All smooth varieties are Gorenstein. $\endgroup$ Apr 3, 2016 at 9:05
  • 3
    $\begingroup$ The Veronese varieties will almost never be arithmetically Gorenstein. The affine cone over $\nu_d(\mathbb{P}^n)$ is the quotient of $\mathbb{A}^{n+1}$ by the diagonal action of the group of roots of unity, $\mu_{d}$. The determinant of that action is the character $(\bullet)^{n+1}:\mu_d\to \mu_d$. Thus, this action factors through the special linear group if and only if $d$ divides $n+1$. On the positive side, these are always arithmetically Cohen-Macaulay. That follows from the Hochster-Roberts theorem (among others). $\endgroup$ Apr 3, 2016 at 10:16

1 Answer 1

9
$\begingroup$

I assume that you mean arithmetically Gorenstein, i.e. the cone over the variety is Gorenstein. Then:

$\bullet$ The Veronese variety $\nu_d(\mathbb{P}^n)\subset \mathbb{P}^N$ is always arithmetically Cohen-Macaulay; the extra condition you need is $\omega _{\mathbb{P}^n}\cong \nu_d^*\mathcal{O}_{\mathbb{P}^N}(\ell)$ for some $\ell$, which means that $n+1$ is a multiple of $d$.

$\bullet$ Similarly the Segre variety $s(\mathbb{P}^{m_1}\times \ldots \times \mathbb{P}^{m_k})\subset \mathbb{P}^N$ is ACM; it is AG if and only if $\omega_{\mathbb{P}^{m_1}\times \ldots \times \mathbb{P}^{m_k}} \cong s^*\mathcal{O}_{\mathbb{P}^N}(\ell)$ for some $\ell$, that is, if and only if $m_1=\ldots =m_k\,$.


Edit: Here are some precisions at the request of the OP. By e.g. Migliore's book Introduction to Liaison Theory and Deficiency Modules, pp. 9-11, a smooth projective variety $V\subset \mathbb{P}^N$ is arithmetically Gorenstein if and only if:

1) $H^i(V, \mathcal{O}_V(k))=0$ for $0<i<\dim(V)$ and all $k$;

2) The canonical line bundle $\omega_V$ is induced by $\mathcal{O}_{\mathbb{P}^N}(k)$ for some integer $k$.

Condition 1) always holds for a product of projective spaces. When $V=\mathbb{P}^n$ embedded by the Veronese embedding $\nu_d$, one has $\omega _V=\mathcal{O}_{\mathbb{P}^n}(-n-1)$ and $\nu_d^*\mathcal{O}_{\mathbb{P}^N}(k)=\mathcal{O}_{\mathbb{P}^n}(kd)$, hence the condition $d\mid n+1$. If $V=\mathbb{P}^{m_1}\times \ldots \times \mathbb{P}^{m_k}$ embedded by the Segre embedding $s$, $\omega _V= \mathcal{O}_{\mathbb{P}^{m_1}}(-m_1-1)\boxtimes\ldots \boxtimes \mathcal{O}_{\mathbb{P}^{m_k}}(-m_k-1)$ and $s^*\mathcal{O}_{\mathbb{P}^N}(k)=\mathcal{O}_{\mathbb{P}^{m_1}}(k)\boxtimes\ldots \boxtimes \mathcal{O}_{\mathbb{P}^{m_k}}(k)$, hence the condition $m_1=\ldots =m_k$.

$\endgroup$
5
  • $\begingroup$ I have edited my question further and can you point me to some reference please $\endgroup$
    – MIQ
    Apr 4, 2016 at 5:34
  • 1
    $\begingroup$ 1) In case you are not sure: saying that the graded ring you mention is Gorenstein (or Cohen-Macaulay) is the same as saying that the variety is arithmetically Gorenstein (or Cohen-Macaulay). 2) One possible reference is chapter 4 of Migliore's book Introduction to Liaison Theory and Deficiency Modules. There is also a paper by Hartshorne: Geometry of arithmetically Gorenstein curves in $\mathbb{P}^4$. $\endgroup$
    – abx
    Apr 4, 2016 at 7:07
  • $\begingroup$ Thanks but I meant a reference for Veronese and Segre to be Arithmetically Gorenstein when satisfying the above conditions, if one needs to cite the reference in some piece of thesis or article... $\endgroup$
    – MIQ
    Apr 8, 2016 at 14:00
  • 1
    $\begingroup$ I do not know a direct reference, but the result is obvious in view of the characterization of arithmetically Gorenstein by the canonical line bundle being $\mathcal{O}(\ell)$. $\endgroup$
    – abx
    Apr 8, 2016 at 15:31
  • 1
    $\begingroup$ This is too long for a comment, so I am editing my answer. $\endgroup$
    – abx
    Apr 9, 2016 at 11:47

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.