The Stacks project

Lemma 68.9.3. Let $S$ be a scheme. Let $X$ be a quasi-compact and quasi-separated algebraic space over $S$ such that $|X|$ has finitely many irreducible components.

  1. There exists a surjective integral morphism $Y \to X$ where $Y$ is a scheme such that $f$ is finite étale over a quasi-compact dense open $U \subset X$,

  2. given a surjective étale morphism $V \to X$ we may choose $Y \to X$ such that for every $y \in Y$ there is an open neighbourhood $W \subset Y$ such that $W \to X$ factors through $V$.

Proof. The proof is the (roughly) same as the proof of Lemma 68.9.2 with additional technical comments to obtain the dense quasi-compact open $U$ (and unfortunately changes in notation to keep track of $U$).

Part (1) is the special case of part (2) where $V = X$.

Proof of (2). Choose a surjective étale morphism $V' \to V$ where $V'$ is a scheme. It is clear that we may replace $V$ by $V'$ and hence we may assume $V$ is a scheme. Since $X$ is quasi-compact, there exist finitely many affine opens $V_ i \subset V$ such that $V' = \coprod V_ i \to X$ is surjective. After replacing $V$ by $V'$ again, we see that we may assume $V$ is affine. Since $X$ is quasi-separated, hence reasonable, there exists an integer $d$ bounding the degree of the geometric fibres of $V \to X$ (see Lemma 68.5.1).

By induction on $d \geq 1$ we will prove the following induction hypothesis $(H_ d)$:

  • for any quasi-compact and quasi-separated algebraic space $X$ with finitely many irreducible components, for any $m \geq 0$, for any quasi-compact and separated schemes $V_ j$, $j = 1, \ldots , m$, for any étale morphisms $\varphi _ j : V_ j \to X$, $j = 1, \ldots , m$ such that $d$ bounds the degree of the geometric fibres of $\varphi _ j : V_ j\to X$ and $\varphi = \coprod \varphi _ j : V = \coprod V_ j \to X$ is surjective, the statement of the lemma holds for $\varphi : V \to X$.

If $d = 1$, then each $\varphi _ j$ is an open immersion. Hence $X$ is a scheme and the result holds with $Y = V$. Assume $d > 1$, assume $(H_{d - 1})$ and let $m$, $\varphi : V_ j \to X$, $j = 1, \ldots , m$ be as in $(H_ d)$.

Let $\eta _1, \ldots , \eta _ n \in |X|$ be the generic points of the irreducible components of $|X|$. By Properties of Spaces, Proposition 66.13.3 there is an open subscheme $U \subset X$ with $\eta _1, \ldots , \eta _ n \in U$. By shrinking $U$ we may assume $U$ affine and by Morphisms, Lemma 29.51.1 we may assume each $\varphi _ j : V_ j \to X$ is finite étale over $U$. Of course, we see that $U$ is quasi-compact and dense in $X$ and that $\varphi _ j^{-1}(U)$ is dense in $V_ j$. In particular each $V_ j$ has finitely many irreducible components.

Fix $j \in \{ 1, \ldots , m\} $. As in Morphisms of Spaces, Lemma 67.52.2 we let $Y_ j$ be the normalization of $X$ in $V_ j$. We obtain a factorization

\[ \xymatrix{ V_ j \ar[rr] \ar[rd]_{\varphi _ j} & & Y_ j \ar[ld]^{\pi _ j} \\ & X } \]

with $\pi _ j$ integral and $V_ j \to Y_ j$ a quasi-compact open immersion. Since $Y_ j$ is the normalization of $X$ in $V_ j$, we see from Morphisms of Spaces, Lemmas 67.48.4 and 67.48.10 that $\varphi _ j^{-1}(U) \to \pi _ j^{-1}(U)$ is an isomorphism. Thus $\pi _ j$ is finite étale over $U$. Observe that $V_ j$ is scheme theoretically dense in $Y_ j$ because $Y_ j$ is the normalization of $X$ in $V_ j$ (follows from the characterization of relative normalization in Morphisms of Spaces, Lemma 67.48.5). Since $V_ j$ is quasi-compact we see that $|V_ j| \subset |Y_ j|$ is dense, see Morphisms of Spaces, Section 67.17 (and especially Morphisms of Spaces, Lemma 67.17.7). It follows that $|Y_ j|$ has finitely many irreducible components. Then $V_ j \times _ X Y_ j$ is a quasi-compact, separated scheme (being finite over $V_ j$) and

\[ V_ j \times _ X Y_ j = V_ j \amalg W_ j \]

Here the first summand is the image of $V_ j \to V_ j \times _ X Y_ j$ (which is closed by Morphisms of Spaces, Lemma 67.4.6 and open because it is étale as a morphism between algebraic spaces étale over $Y$) and the second summand is the (open and closed) complement.

The étale morphism $W_ j \to Y_ j$ has geometric fibres of cardinality $< d$. Namely, this is clear for geometric points of $V_ j \subset Y_ j$ by inspection. Since $|V_ j| \subset |Y_ j|$ is dense, it holds for all geometric points of $Y_ j$ by Lemma 68.8.1 (the degree of the fibres of a quasi-compact étale morphism does not go up under specialization). By $(H_{d - 1})$ applied to $V_ j \amalg W_ j \to Y_ j$ we find a surjective integral morphism $Y_ j' \to Y_ j$ with $Y_ j'$ a scheme, which Zariski locally factors through $V_ j \amalg W_ j$, and which is finite étale over a quasi-compact dense open $U_ j \subset Y_ j$. After shrinking $U$ we may and do assume that $\pi _ j^{-1}(U) \subset U_ j$ (we may and do choose the same $U$ for all $j$; some details omitted).

We claim that

\[ Y = \coprod \nolimits _{j = 1, \ldots , m} Y'_ j \longrightarrow X \]

is the solution to our problem. First, this morphism is integral as on each summand we have the composition $Y'_ j \to Y \to X$ of integral morphisms (Morphisms of Spaces, Lemma 67.45.4). Second, this morphism Zariski locally factors through $V = \coprod V_ j$ because we saw above that each $Y'_ j \to Y_ j$ factors Zariski locally through $V_ j \amalg W_ j = V_ j \times _ X Y_ j$. Finally, since both $Y'_ j \to Y_ j$ and $Y_ j \to X$ are finite étale over $U$, so is the composition. This finishes the proof. $\square$

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