4

Question 1: How does one show that the $n$-th Hirzebruch surface $$\mathcal{H}_n:=V(u^nX-v^nY)\subset \mathbb{P}^2\times \mathbb{P}^1$$ ($[u:v]\in \mathbb{P}^1, [X:Y:Z]\in \mathbb{P}^2$) and the projective product $\mathbb{P}^1\times \mathbb{P}^1$ are birational? I am having trouble finding which open set of $\mathcal{H}_n$ to make isomorphic to $\mathbb{A}^1\times \mathbb{A}^1$.

Question 2: Also, I want to understand how $\mathcal{H}_n$ is isomorphic to the blow-up of $\mathbb{P}^2$ at some point $p$. I understand that the latter is a geometrically ruled surface and should be isomorphic as such, but can one construct an explicit isomorphism (regular map with inverse) between the two?

Question 3: In a distinguished open set, $\mathcal{H}_n$ is also a line bundle over the projective line. How does one show $\mathcal{H}_n=\mathbb{P}(\mathcal{O}_{\mathbb{P}^1}(n)\bigoplus \mathcal{O}_{\mathbb{P}^1})$?

Are there any useful references involving Hirzebruch surfaces that I can read to learn more?

ff90
  • 377
  • For the first question, I found it easier to just construct a birational map from $\mathcal{H}_m\to \mathbb{P}^1\times \mathbb{P}^1$ (consider the projection). I will add my results as I go along in the comments in case it might help anyone else. – ff90 Oct 04 '13 at 06:01
  • For the third question, I ultimately showed the result by interpreting the surface with sheaves. I would still be interested in seeing an explicit isomorphism for Question 2, or a more elementary proof for Question 3 :-) – ff90 Oct 04 '13 at 07:23
  • I don't have time for a full answer, but I'd guess Barth--Hulek--Peters--van de Ven has plenty of material on this. For Question 2, $H_n$ is not a blowup of $P^2$ at a point unless $n=1$, because the blowup of $P^2$ at a point is exactly $P(O \oplus O(1))$. (And of course the $H_n$ are not isomorphic for different $n$.) –  Oct 04 '13 at 07:43

1 Answers1

7

Question 2. Let $Y=\textrm{Bl}_P\mathbb P^2$, the blow-up at a point $P\in\mathbb P^2$. To convince you that $Y$ is a projective bundle of rank 1 over $\mathbb P^1$, you can project $Y$ onto the exceptional divisor $E\cong\mathbb P^1$ and see that the fibers are $\mathbb P^1$'s, because they are the strict transforms of lines through $P$. This is not quite what you want, though - it was just to make a remark a posteriori.

Let $\mathscr E=\mathcal O_{\mathbb P^1}\oplus \mathcal O_{\mathbb P^1}(1)$ and $X=\mathbb P(\mathscr E)$. We want to show that $X\cong Y$. We have $$Y=\{((x_0,x_1,x_2);(u,v))\in \mathbb P^2\times\mathbb P^1\,|\,x_1v=x_2u\}\subset \mathbb P^2\times\mathbb P^1.$$ Let us take the standard open covering $(U_0,U_1)$ of $\mathbb P^1$; if $p:\mathscr E\to \mathbb P^1$ comes with trivializations $\phi_i:p^{-1}(U_i)\to U_i\times \mathbb A^2$, and hence with transition maps $\phi_{ij}:U_{ij}\times\mathbb A^2\to U_{ij}\times\mathbb A^2$, then on $\mathbb P(\mathscr E)\to \mathbb P^1$ we have the same (but projectivized) transition functions $\phi_{ij}:U_{ij}\times\mathbb P^1\to U_{ij}\times\mathbb P^1$. In our case, we only have two open sets, so there is only one isomorphism which we have to use in order to get $X$ by gluing two copies of $U_0\times\mathbb P^1$ and $U_1\times \mathbb P^1$ together: for instance, by using the isomorphism $$\phi_{01}:U_{01}\times\mathbb P^1\to U_{01}\times\mathbb P^1$$ sending $(z,(u,v))\mapsto (1/z,(u,zv))$, we can construct open immersions \begin{align} U_0\times\mathbb P^1\to Y\,\,\,\,\,\,\,\,\,\,(z,(u,v))\mapsto ((u,zv,v),(z,1)) \notag\\ U_1\times\mathbb P^1\to Y\,\,\,\,\,\,\,\,\,\,(z,(u,v))\mapsto ((u,v,zv),(1,z)) \notag\\ \end{align} which glue to an isomorphism $\mathbb P(\mathscr E)\to Y$, by construction.

Brenin
  • 14,072