The Stacks project

Lemma 32.10.5. For a scheme $T$ denote $\mathcal{C}_ T$ the full subcategory of schemes $W$ over $T$ such that $W$ is quasi-compact and quasi-separated and such that the structure morphism $W \to T$ is locally of finite presentation. Let $S = \mathop{\mathrm{lim}}\nolimits S_ i$ be a directed limit of schemes with affine transition morphisms. Then there is an equivalence of categories

\[ \mathop{\mathrm{colim}}\nolimits \mathcal{C}_{S_ i} \longrightarrow \mathcal{C}_ S \]

given by the base change functors.

Proof. Fully faithfulness. Suppose we have $i \in I$ and objects $X_ i$, $Y_ i$ of $\mathcal{C}_{S_ i}$. Denote $X = X_ i \times _{S_ i} S$ and $Y = Y_ i \times _{S_ i} S$. Suppose given a morphism $f : X \to Y$ over $S$. We can choose a finite affine open covering $Y_ i = V_{i, 1} \cup \ldots \cup V_{i, m}$ such that $V_{i, j} \to Y_ i \to S_ i$ maps into an affine open $W_{i, j}$ of $S_ i$. Denote $Y = V_1 \cup \ldots \cup V_ m$ the induced affine open covering of $Y$. Since $f : X \to Y$ is quasi-compact (Schemes, Lemma 26.21.14) after increasing $i$ we may assume that there is a finite open covering $X_ i = U_{i, 1} \cup \ldots \cup U_{i, m}$ by quasi-compact opens such that the inverse image of $U_{i, j}$ in $Y$ is $f^{-1}(V_ j)$, see Lemma 32.4.11. By Lemma 32.10.1 applied to $f|_{f^{-1}(V_ j)}$ over $W_ j$ we may assume, after increasing $i$, that there is a morphism $f_{i, j} : V_{i, j} \to U_{i, j}$ over $S$ whose base change to $S$ is $f|_{f^{-1}(V_ j)}$. Increasing $i$ more we may assume $f_{i, j}$ and $f_{i, j'}$ agree on the quasi-compact open $U_{i, j} \cap U_{i, j'}$. Then we can glue these morphisms to get the desired morphism $f_ i : X_ i \to Y_ i$. This morphism is unique (up to increasing $i$) because this is true for the morphisms $f_{i, j}$.

To show that the functor is essentially surjective we argue in exactly the same way. Namely, suppose that $X$ is an object of $\mathcal{C}_ S$. Pick $i \in I$. We can choose a finite affine open covering $X = U_1 \cup \ldots \cup U_ m$ such that $U_ j \to X \to S \to S_ i$ factors through an affine open $W_{i, j} \subset S_ i$. Set $W_ j = W_{i, j} \times _{S_ i} S$. This is an affine open of $S$. By Lemma 32.10.1, after increasing $i$, we may assume there exist $U_{i, j} \to W_{i, j}$ of finite presentation whose base change to $W_ j$ is $U_ j$. After increasing $i$ we may assume there exist quasi-compact opens $U_{i, j, j'} \subset U_{i, j}$ whose base changes to $S$ are equal to $U_ j \cap U_{j'}$. Claim: after increasing $i$ we may assume the image of the morphism $U_{i, j, j'} \to U_{i, j} \to W_{i, j}$ ends up in $W_{i, j} \cap W_{i, j'}$. Namely, because the complement of $W_{i, j} \cap W_{i, j'}$ is closed in the affine scheme $W_{i, j}$ it is affine. Since $U_ j \cap U_{j'} = \mathop{\mathrm{lim}}\nolimits U_{i, j, j'}$ does map into $W_{i, j} \cap W_{i, j'}$ we can apply Lemma 32.4.9 to get the claim. Thus we can view both

\[ U_{i, j, j'} \quad \text{and}\quad U_{i, j', j} \]

as schemes over $W_{i, j'}$ whose base changes to $W_{j'}$ recover $U_ j \cap U_{j'}$. Hence after increasing $i$, using Lemma 32.10.1, we may assume there are isomorphisms $U_{i, j, j'} \to U_{i, j', j}$ over $W_{i, j'}$ and hence over $S_ i$. Increasing $i$ further (details omitted) we may assume these isomorphisms satisfy the cocycle condition mentioned in Schemes, Section 26.14. Applying Schemes, Lemma 26.14.1 we obtain an object $X_ i$ of $\mathcal{C}_{S_ i}$ whose base change to $S$ is isomorphic to $X$; we omit some of the verifications. $\square$

Comments (0)

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 0EY1. Beware of the difference between the letter 'O' and the digit '0'.