The Stacks project

96.18 When is a quotient stack algebraic?

In Groupoids in Spaces, Section 77.20 we have defined the quotient stack $[U/R]$ associated to a groupoid $(U, R, s, t, c)$ in algebraic spaces. Note that $[U/R]$ is a stack in groupoids whose diagonal is representable by algebraic spaces (see Bootstrap, Lemma 79.11.5 and Algebraic Stacks, Lemma 93.10.11) and such that there exists an algebraic space $U$ and a $1$-morphism $(\mathit{Sch}/U)_{fppf} \to [U/R]$ which is an “fppf surjection” in the sense that it induces a map on presheaves of isomorphism classes of objects which becomes surjective after sheafification. However, it is not the case that $[U/R]$ is an algebraic stack in general. This is not a contradiction with Theorem 96.16.1 as the $1$-morphism $(\mathit{Sch}/U)_{fppf} \to [U/R]$ may not be flat and locally of finite presentation.

The easiest way to make examples of non-algebraic quotient stacks is to look at quotients of the form $[S/G]$ where $S$ is a scheme and $G$ is a group scheme over $S$ acting trivially on $S$. Namely, we will see below (Lemma 96.18.3) that if $[S/G]$ is algebraic, then $G \to S$ has to be flat and locally of finite presentation. An explicit example can be found in Examples, Section 109.52.

Lemma 96.18.1. Let $S$ be a scheme and let $B$ be an algebraic space over $S$. Let $(U, R, s, t, c)$ be a groupoid in algebraic spaces over $B$. The quotient stack $[U/R]$ is an algebraic stack if and only if there exists a morphism of algebraic spaces $g : U' \to U$ such that

  1. the composition $U' \times _{g, U, t} R \to R \xrightarrow {s} U$ is a surjection of sheaves, and

  2. the morphisms $s', t' : R' \to U'$ are flat and locally of finite presentation where $(U', R', s', t', c')$ is the restriction of $(U, R, s, t, c)$ via $g$.

Proof. First, assume that $g : U' \to U$ satisfies (1) and (2). Property (1) implies that $[U'/R'] \to [U/R]$ is an equivalence, see Groupoids in Spaces, Lemma 77.25.2. By Theorem 96.17.2 the quotient stack $[U'/R']$ is an algebraic stack. Hence $[U/R]$ is an algebraic stack too, see Algebraic Stacks, Lemma 93.12.4.

Conversely, assume that $[U/R]$ is an algebraic stack. We may choose a scheme $W$ and a surjective smooth $1$-morphism

\[ f : (\mathit{Sch}/W)_{fppf} \longrightarrow [U/R]. \]

By the $2$-Yoneda lemma (Algebraic Stacks, Section 93.5) this corresponds to an object $\xi $ of $[U/R]$ over $W$. By the description of $[U/R]$ in Groupoids in Spaces, Lemma 77.24.1 we can find a surjective, flat, locally finitely presented morphism $b : U' \to W$ of schemes such that $\xi ' = b^*\xi $ corresponds to a morphism $g : U' \to U$. Note that the $1$-morphism

\[ f' : (\mathit{Sch}/U')_{fppf} \longrightarrow [U/R]. \]

corresponding to $\xi '$ is surjective, flat, and locally of finite presentation, see Algebraic Stacks, Lemma 93.10.5. Hence $(\mathit{Sch}/U')_{fppf} \times _{[U/R]} (\mathit{Sch}/U')_{fppf}$ which is represented by the algebraic space

\[ \mathit{Isom}_{[U/R]}(\text{pr}_0^*\xi ', \text{pr}_1^*\xi ') = (U' \times _ S U') \times _{(g \circ \text{pr}_0, g \circ \text{pr}_1), U \times _ S U} R = R' \]

(see Groupoids in Spaces, Lemma 77.22.1 for the first equality; the second is the definition of restriction) is flat and locally of finite presentation over $U'$ via both $s'$ and $t'$ (by base change, see Algebraic Stacks, Lemma 93.10.6). By this description of $R'$ and by Algebraic Stacks, Lemma 93.16.1 we obtain a canonical fully faithful $1$-morphism $[U'/R'] \to [U/R]$. This $1$-morphism is essentially surjective because $f'$ is flat, locally of finite presentation, and surjective (see Stacks, Lemma 8.4.8); another way to prove this is to use Algebraic Stacks, Remark 93.16.3. Finally, we can use Groupoids in Spaces, Lemma 77.25.2 to conclude that the composition $U' \times _{g, U, t} R \to R \xrightarrow {s} U$ is a surjection of sheaves. $\square$

Lemma 96.18.2. Let $S$ be a scheme and let $B$ be an algebraic space over $S$. Let $G$ be a group algebraic space over $B$. Let $X$ be an algebraic space over $B$ and let $a : G \times _ B X \to X$ be an action of $G$ on $X$ over $B$. The quotient stack $[X/G]$ is an algebraic stack if and only if there exists a morphism of algebraic spaces $\varphi : X' \to X$ such that

  1. $G \times _ B X' \to X$, $(g, x') \mapsto a(g, \varphi (x'))$ is a surjection of sheaves, and

  2. the two projections $X'' \to X'$ of the algebraic space $X''$ given by the rule

    \[ T \longmapsto \{ (x'_1, g, x'_2) \in (X' \times _ B G \times _ B X')(T) \mid \varphi (x'_1) = a(g, \varphi (x'_2))\} \]

    are flat and locally of finite presentation.

Proof. This lemma is a special case of Lemma 96.18.1. Namely, the quotient stack $[X/G]$ is by Groupoids in Spaces, Definition 77.20.1 equal to the quotient stack $[X/G \times _ B X]$ of the groupoid in algebraic spaces $(X, G \times _ B X, s, t, c)$ associated to the group action in Groupoids in Spaces, Lemma 77.15.1. There is one small observation that is needed to get condition (1). Namely, the morphism $s : G \times _ B X \to X$ is the second projection and the morphism $t : G \times _ B X \to X$ is the action morphism $a$. Hence the morphism $h : U' \times _{g, U, t} R \to R \xrightarrow {s} U$ from Lemma 96.18.1 corresponds to the morphism

\[ X' \times _{\varphi , X, a} (G \times _ B X) \xrightarrow {\text{pr}_1} X \]

in the current setting. However, because of the symmetry given by the inverse of $G$ this morphism is isomorphic to the morphism

\[ (G \times _ B X) \times _{\text{pr}_1, X, \varphi } X' \xrightarrow {a} X \]

of the statement of the lemma. Details omitted. $\square$

slogan

Lemma 96.18.3. Let $S$ be a scheme and let $B$ be an algebraic space over $S$. Let $G$ be a group algebraic space over $B$. Endow $B$ with the trivial action of $G$. Then the quotient stack $[B/G]$ is an algebraic stack if and only if $G$ is flat and locally of finite presentation over $B$.

Proof. If $G$ is flat and locally of finite presentation over $B$, then $[B/G]$ is an algebraic stack by Theorem 96.17.2.

Conversely, assume that $[B/G]$ is an algebraic stack. By Lemma 96.18.2 and because the action is trivial, we see there exists an algebraic space $B'$ and a morphism $B' \to B$ such that (1) $B' \to B$ is a surjection of sheaves and (2) the projections

\[ B' \times _ B G \times _ B B' \to B' \]

are flat and locally of finite presentation. Note that the base change $B' \times _ B G \times _ B B' \to G \times _ B B'$ of $B' \to B$ is a surjection of sheaves also. Thus it follows from Descent on Spaces, Lemma 73.8.1 that the projection $G \times _ B B' \to B'$ is flat and locally of finite presentation. By (1) we can find an fppf covering $\{ B_ i \to B\} $ such that $B_ i \to B$ factors through $B' \to B$. Hence $G \times _ B B_ i \to B_ i$ is flat and locally of finite presentation by base change. By Descent on Spaces, Lemmas 73.11.13 and 73.11.10 we conclude that $G \to B$ is flat and locally of finite presentation. $\square$

Later we will see that the quotient stack of a smooth $S$-space by a group algebraic space $G$ is smooth, even when $G$ is not smooth (Morphisms of Stacks, Lemma 100.33.7).


Comments (3)

Comment #7531 by Anonymous on

In the first paragraph it reads "This is not a contradiction with Theorem 96.16.1 as the 1-morphism is not representable by algebraic spaces in general, ..."

Maybe I misunderstood something, but I thought having diagonal representable by algebraic spaces (as claimed earlier in the same paragraph) implies that for any algebraic space , any morphism is representable by algebraic spaces? (Lemma 93.10.11(7)).

Comment #7534 by on

Indeed. The rest of the sentence does make sense.


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 06PI. Beware of the difference between the letter 'O' and the digit '0'.