Proof.
Write Y = \mathop{\mathrm{lim}}\nolimits _{i \in I} Y_ i as a limit of algebraic spaces over a directed set I with affine transition morphisms and with Y_ i Noetherian, see Limits of Spaces, Proposition 70.8.1. We will use the material from Limits of Spaces, Section 70.23.
Choose a presentation \mathcal{X} = [U/R]. Denote (U, R, s, t, c, e, i) the corresponding groupoid in algebraic spaces over Y. We may and do assume U is affine. Then U, R, R \times _{s, U, t} R are quasi-separated algebraic spaces of finite type over Y. We have two morpisms s, t : R \to U, three morphisms c : R \times _{s, U, t} R \to R, \text{pr}_1 : R \times _{s, U, t} R \to R, \text{pr}_2 : R \times _{s, U, t} R \to R, a morphism e : U \to R, and finally a morphism i : R \to R. These morphisms satisfy a list of axioms which are detailed in Groupoids, Section 39.13.
According to Limits of Spaces, Remark 70.23.5 we can find an i_0 \in I and inverse systems
(U_ i)_{i \geq i_0},
(R_ i)_{i \geq i_0},
(T_ i)_{i \geq i_0}
over (Y_ i)_{i \geq i_0} such that U = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} U_ i, R = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} R_ i, and R \times _{s, U, t} R = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} T_ i and such that there exist morphisms of systems
(s_ i)_{i \geq i_0} : (R_ i)_{i \geq i_0} \to (U_ i)_{i \geq i_0},
(t_ i)_{i \geq i_0} : (R_ i)_{i \geq i_0} \to (U_ i)_{i \geq i_0},
(c_ i)_{i \geq i_0} : (T_ i)_{i \geq i_0} \to (R_ i)_{i \geq i_0},
(p_ i)_{i \geq i_0} : (T_ i)_{i \geq i_0} \to (R_ i)_{i \geq i_0},
(q_ i)_{i \geq i_0} : (T_ i)_{i \geq i_0} \to (R_ i)_{i \geq i_0},
(e_ i)_{i \geq i_0} : (U_ i)_{i \geq i_0} \to (R_ i)_{i \geq i_0},
(i_ i)_{i \geq i_0} : (R_ i)_{i \geq i_0} \to (R_ i)_{i \geq i_0}
with s = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} s_ i, t = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} t_ i, c = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} c_ i, \text{pr}_1 = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} p_ i, \text{pr}_2 = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} q_ i, e = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} e_ i, and i = \mathop{\mathrm{lim}}\nolimits _{i \geq i_0} i_ i. By Limits of Spaces, Lemma 70.23.7 we see that we may assume that s_ i and t_ i are smooth (this may require increasing i_0). By Limits of Spaces, Lemma 70.23.6 we may assume that the maps R \to U \times _{U_ i, s_ i} R_ i given by s and R \to R_ i and R \to U \times _{U_ i, t_ i} R_ i given by t and R \to R_ i are isomorphisms for all i \geq i_0. By Limits of Spaces, Lemma 70.23.9 we see that we may assume that the diagrams
\xymatrix{ T_ i \ar[r]_{q_ i} \ar[d]_{p_ i} & R_ i \ar[d]^{t_ i} \\ R_ i \ar[r]^{s_ i} & U_ i }
are cartesian. The uniqueness of Limits of Spaces, Lemma 70.23.4 then guarantees that for a sufficiently large i the relations between the morphisms s, t, c, e, i mentioned above are satisfied by s_ i, t_ i, c_ i, e_ i, i_ i. Fix such an i.
It follows that (U_ i, R_ i, s_ i, t_ i, c_ i, e_ i, i_ i) is a smooth groupoid in algebraic spaces over Y_ i. Hence \mathcal{X}_ i = [U_ i/R_ i] is an algebraic stack (Algebraic Stacks, Theorem 94.17.3). The morphism of groupoids
(U, R, s, t, c, e, i) \to (U_ i, R_ i, s_ i, t_ i, c_ i, e_ i, i_ i)
over Y \to Y_ i determines a commutative diagram
\xymatrix{ \mathcal{X} \ar[d] \ar[r] & \mathcal{X}_ i \ar[d] \\ Y \ar[r] & Y_ i }
(Groupoids in Spaces, Lemma 78.21.1). We claim that the morphism \mathcal{X} \to Y \times _{Y_ i} \mathcal{X}_ i is a closed immersion. The claim finishes the proof because the algebraic stack \mathcal{X}_ i \to Y_ i is of finite presentation by construction. To prove the claim, note that the left diagram
\xymatrix{ U \ar[d] \ar[r] & U_ i \ar[d] \\ \mathcal{X} \ar[r] & \mathcal{X}_ i } \quad \quad \xymatrix{ U \ar[d] \ar[r] & Y \times _{Y_ i} U_ i \ar[d] \\ \mathcal{X} \ar[r] & Y \times _{Y_ i} \mathcal{X}_ i }
is cartesian by Groupoids in Spaces, Lemma 78.25.3 and the results mentioned above. Hence the right commutative diagram is cartesian too. Then the desired result follows from the fact that U \to Y \times _{Y_ i} U_ i is a closed immersion by construction of the inverse system (U_ i) in Limits of Spaces, Lemma 70.23.3, the fact that Y \times _{Y_ i} U_ i \to Y \times _{Y_ i} \mathcal{X}_ i is smooth and surjective, and Properties of Stacks, Lemma 100.9.4.
\square
There is a version for separated algebraic stacks.
Proof.
First we use exactly the same procedure as in the proof of Lemma 102.6.1 (and we borrow its notation) to construct the embedding \mathcal{X} \to \mathcal{X}' as a morphism \mathcal{X} \to \mathcal{X}' = Y \times _{Y_ i} \mathcal{X}_ i with \mathcal{X}_ i = [U_ i/R_ i]. Thus it is enough to show that \mathcal{X}_ i \to Y_ i is separated for sufficiently large i. In other words, it is enough to show that \mathcal{X}_ i \to \mathcal{X}_ i \times _{Y_ i} \mathcal{X}_ i is proper for i sufficiently large. Since the morphism U_ i \times _{Y_ i} U_ i \to \mathcal{X}_ i \times _{Y_ i} \mathcal{X}_ i is surjective and smooth and since R_ i = \mathcal{X}_ i \times _{\mathcal{X}_ i \times _{Y_ i} \mathcal{X}_ i} U_ i \times _{Y_ i} U_ i it is enough to show that the morphism (s_ i, t_ i) : R_ i \to U_ i \times _{Y_ i} U_ i is proper for i sufficiently large, see Properties of Stacks, Lemma 100.3.3. We prove this in the next paragraph.
Observe that U \times _ Y U \to Y is quasi-separated and of finite type. Hence we can use the construction of Limits of Spaces, Remark 70.23.5 to find an i_1 \in I and an inverse system (V_ i)_{i \geq i_1} with U \times _ Y U = \mathop{\mathrm{lim}}\nolimits _{i \geq i_1} V_ i. By Limits of Spaces, Lemma 70.23.9 for i sufficiently large the functoriality of the construction applied to the projections U \times _ Y U \to U gives closed immersions
V_ i \to U_ i \times _{Y_ i} U_ i
(There is a small mismatch here because in truth we should replace Y_ i by the scheme theoretic image of Y \to Y_ i, but clearly this does not change the fibre product.) On the other hand, by Limits of Spaces, Lemma 70.23.8 the functoriality applied to the proper morphism (s, t) : R \to U \times _ Y U (here we use that \mathcal{X} is separated) leads to morphisms R_ i \to V_ i which are proper for large enough i. Composing these morphisms we obtain a proper morphisms R_ i \to U_ i \times _{Y_ i} U_ i for all i large enough. The functoriality of the construction of Limits of Spaces, Remark 70.23.5 shows that this is the morphism is the same as (s_ i, t_ i) for large enough i and the proof is complete.
\square
Comments (0)