The Stacks project

Lemma 32.4.14. In Situation 32.4.5 if $S$ is separated, then for some $i_0 \in I$ the schemes $S_ i$ for $i \geq i_0$ are separated.

Proof. Choose a finite affine open covering $S_0 = U_{0, 1} \cup \ldots \cup U_{0, m}$. Set $U_{i, j} \subset S_ i$ and $U_ j \subset S$ equal to the inverse image of $U_{0, j}$. Note that $U_{i, j}$ and $U_ j$ are affine. As $S$ is separated the intersections $U_{j_1} \cap U_{j_2}$ are affine. Since $U_{j_1} \cap U_{j_2} = \mathop{\mathrm{lim}}\nolimits _{i \geq 0} U_{i, j_1} \cap U_{i, j_2}$ we see that $U_{i, j_1} \cap U_{i, j_2}$ is affine for large $i$ by Lemma 32.4.13. To show that $S_ i$ is separated for large $i$ it now suffices to show that

\[ \mathcal{O}_{S_ i}(U_{i, j_1}) \otimes _{\mathcal{O}_ S(S)} \mathcal{O}_{S_ i}(U_{i, j_2}) \longrightarrow \mathcal{O}_{S_ i}(U_{i, j_1} \cap U_{i, j_2}) \]

is surjective for large $i$ (Schemes, Lemma 26.21.7).

To get rid of the annoying indices, assume we have affine opens $U, V \subset S_0$ such that $U \cap V$ is affine too. Let $U_ i, V_ i \subset S_ i$, resp. $U, V \subset S$ be the inverse images. We have to show that $\mathcal{O}(U_ i) \otimes \mathcal{O}(V_ i) \to \mathcal{O}(U_ i \cap V_ i)$ is surjective for $i$ large enough and we know that $\mathcal{O}(U) \otimes \mathcal{O}(V) \to \mathcal{O}(U \cap V)$ is surjective. Note that $\mathcal{O}(U_0) \otimes \mathcal{O}(V_0) \to \mathcal{O}(U_0 \cap V_0)$ is of finite type, as the diagonal morphism $S_ i \to S_ i \times S_ i$ is an immersion (Schemes, Lemma 26.21.2) hence locally of finite type (Morphisms, Lemmas 29.15.2 and 29.15.5). Thus we can choose elements $f_{0, 1}, \ldots , f_{0, n} \in \mathcal{O}(U_0 \cap V_0)$ which generate $\mathcal{O}(U_0 \cap V_0)$ over $\mathcal{O}(U_0) \otimes \mathcal{O}(V_0)$. Observe that for $i \geq 0$ the diagram of schemes

\[ \xymatrix{ U_ i \cap V_ i \ar[r] \ar[d] & U_ i \ar[d] \\ U_0 \cap V_0 \ar[r] & U_0 } \]

is cartesian. Thus we see that the images $f_{i, 1}, \ldots , f_{i, n} \in \mathcal{O}(U_ i \cap V_ i)$ generate $\mathcal{O}(U_ i \cap V_ i)$ over $\mathcal{O}(U_ i) \otimes \mathcal{O}(V_0)$ and a fortiori over $\mathcal{O}(U_ i) \otimes \mathcal{O}(V_ i)$. By assumption the images $f_1, \ldots , f_ n \in \mathcal{O}(U \otimes V)$ are in the image of the map $\mathcal{O}(U) \otimes \mathcal{O}(V) \to \mathcal{O}(U \cap V)$. Since $\mathcal{O}(U) \otimes \mathcal{O}(V) = \mathop{\mathrm{colim}}\nolimits \mathcal{O}(U_ i) \otimes \mathcal{O}(V_ i)$ we see that they are in the image of the map at some finite level and the lemma is proved. $\square$

Comments (2)

Comment #4527 by awllower on

Two minor typos:

should be

And should be

Post a comment

Your email address will not be published. Required fields are marked.

In your comment you can use Markdown and LaTeX style mathematics (enclose it like $\pi$). A preview option is available if you wish to see how it works out (just click on the eye in the toolbar).

Unfortunately JavaScript is disabled in your browser, so the comment preview function will not work.

All contributions are licensed under the GNU Free Documentation License.

In order to prevent bots from posting comments, we would like you to prove that you are human. You can do this by filling in the name of the current tag in the following input field. As a reminder, this is tag 086Q. Beware of the difference between the letter 'O' and the digit '0'.