forked from stacks/stacks-project
-
Notifications
You must be signed in to change notification settings - Fork 0
/
limits.tex
4478 lines (4058 loc) · 169 KB
/
limits.tex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
\input{preamble}
% OK, start here.
%
\begin{document}
\title{Limits of Schemes}
\maketitle
\phantomsection
\label{section-phantom}
\tableofcontents
\section{Introduction}
\label{section-introduction}
\noindent
In this chapter we put material related to limits of schemes. We mostly
study limits of inverse systems over directed sets
(Categories, Definition \ref{categories-definition-directed-set})
with affine transition maps. We discuss absolute Noetherian
approximation. We characterize schemes locally of finite presentation
over a base as those whose associated functor of points is limit
preserving. As an application of absolute Noetherian approximation
we prove that the image of an affine under an integral morphism is affine.
Moreover, we prove some very general variants of Chow's lemma.
A basic reference is \cite{EGA}.
\section{Directed limits of schemes with affine transition maps}
\label{section-limits}
\noindent
In this section we construct the limit.
\begin{lemma}
\label{lemma-directed-inverse-system-affine-schemes-has-limit}
Let $I$ be a directed set. Let $(S_i, f_{ii'})$ be an inverse system of
schemes over $I$. If all the schemes $S_i$
are affine, then the limit $S = \lim_i S_i$ exists
in the category of schemes.
In fact $S$ is affine and $S = \Spec(\colim_i R_i)$
with $R_i = \Gamma(S_i, \mathcal{O})$.
\end{lemma}
\begin{proof}
Just define $S = \Spec(\colim_i R_i)$.
It follows from Schemes, Lemma \ref{schemes-lemma-morphism-into-affine}
that $S$ is the limit even in the category of locally ringed spaces.
\end{proof}
\begin{lemma}
\label{lemma-directed-inverse-system-has-limit}
Let $I$ be a directed set. Let $(S_i, f_{ii'})$ be an
inverse system of schemes over $I$. If all the morphisms
$f_{ii'} : S_i \to S_{i'}$ are affine, then the limit $S = \lim_i S_i$ exists
in the category of schemes. Moreover,
\begin{enumerate}
\item each of the morphisms $f_i : S \to S_i$ is affine,
\item for an element $0 \in I$ and any open subscheme $U_0 \subset S_0$
we have
$$
f_0^{-1}(U_0) = \lim_{i \geq 0} f_{i0}^{-1}(U_0)
$$
in the category of schemes.
\end{enumerate}
\end{lemma}
\begin{proof}
Choose an element $0 \in I$. Note that $I$ is nonempty as the limit is
directed. For every $i \geq 0$ consider the quasi-coherent sheaf of
$\mathcal{O}_{S_0}$-algebras $\mathcal{A}_i = f_{i0, *}\mathcal{O}_{S_i}$.
Recall that $S_i = \underline{\Spec}_{S_0}(\mathcal{A}_i)$,
see Morphisms, Lemma \ref{morphisms-lemma-characterize-affine}.
Set $\mathcal{A} = \colim_{i \geq 0} \mathcal{A}_i$.
This is a quasi-coherent sheaf of $\mathcal{O}_{S_0}$-algebras,
see Schemes, Section \ref{schemes-section-quasi-coherent}.
Set $S = \underline{\Spec}_{S_0}(\mathcal{A})$.
By Morphisms, Lemma \ref{morphisms-lemma-affine-equivalence-algebras}
we get for $i \geq 0$ morphisms $f_i : S \to S_i$ compatible with
the transition morphisms. Note that the morphisms $f_i$ are
affine by Morphisms, Lemma \ref{morphisms-lemma-affine-permanence} for example.
By Lemma \ref{lemma-directed-inverse-system-affine-schemes-has-limit} above
we see that for any affine open $U_0 \subset S_0$ the
inverse image $U = f_0^{-1}(U_0) \subset S$ is the limit of the
system of opens $U_i = f_{i0}^{-1}(U_0)$, $i \geq 0$ in the
category of schemes.
\medskip\noindent
Let $T$ be a scheme. Let $g_i : T \to S_i$ be a compatible system
of morphisms. To show that $S = \lim_i S_i$ we have
to prove there is a unique morphism $g : T \to S$ with
$g_i = f_i \circ g$ for all $i \in I$.
For every $t \in T$ there exists an affine open
$U_0 \subset S_0$ containing $g_0(t)$. Let $V \subset g_0^{-1}(U_0)$
be an affine open neighbourhood containing $t$.
By the remarks above we obtain a unique morphism
$g_V : V \to U = f_0^{-1}(U_0)$ such that $f_i \circ g_V = g_i|_{U_i}$
for all $i$. The open sets $V \subset T$ so constructed form
a basis for the topology of $T$. The morphisms $g_V$ glue to a morphism
$g : T \to S$ because of the uniqueness property. This gives the
desired morphism $g : T \to S$.
\medskip\noindent
The final statement is clear from the construction of the limit above.
\end{proof}
\begin{lemma}
\label{lemma-scheme-over-limit}
Let $I$ be a directed set.
Let $(S_i, f_{ii'})$ be an inverse system of schemes over $I$.
Assume all the morphisms $f_{ii'} : S_i \to S_{i'}$ are affine,
Let $S = \lim_i S_i$. Let $0 \in I$.
Suppose that $T$ is a scheme over $S_0$.
Then
$$
T \times_{S_0} S = \lim_{i \geq 0} T \times_{S_0} S_i
$$
\end{lemma}
\begin{proof}
The right hand side is a scheme by
Lemma \ref{lemma-directed-inverse-system-has-limit}.
The equality is formal, see
Categories, Lemma \ref{categories-lemma-colimits-commute}.
\end{proof}
\section{Infinite products}
\label{section-inifinite-products}
\noindent
Infinite products of schemes usually do not exist. For example in
Examples, Section \ref{examples-section-not-algebraic}
it is shown that an infinite product of
copies of $\mathbf{P}^1$ is not even an algebraic space.
\medskip\noindent
On the other hand, infinite products of affine schemes do exist
and are affine. Using Schemes, Lemma \ref{schemes-lemma-morphism-into-affine}
this corresponds to the fact that in the category of rings
we have infinite coproducts: if $I$ is a set and $R_i$ is a ring
for each $i$, then we can consider the ring
$$
R = \otimes R_i =
\colim_{\{i_1, \ldots, i_n\} \subset I}
R_{i_1} \otimes_\mathbf{Z} \ldots \otimes_\mathbf{Z} R_{i_n}
$$
Given another ring $A$ a map $R \to A$ is the same thing as a
collection of ring maps $R_i \to A$ for all $i \in I$ as
follows from the corresponding property of finite tensor products.
\begin{lemma}
\label{lemma-infinite-product}
Let $S$ be a scheme. Let $I$ be a set and for each $i \in I$
let $f_i : T_i \to S$ be an affine morphism. Then the
product $T = \prod T_i$ exists in the category of schemes
over $S$. In fact, we have
$$
T = \lim_{\{i_1, \ldots, i_n\} \subset I}
T_{i_1} \times_S \ldots \times_S T_{i_n}
$$
and the projection morphisms $T \to T_{i_1} \times_S \ldots \times_S T_{i_n}$
are affine.
\end{lemma}
\begin{proof}
Omitted. Hint: Argue as in the discussion preceding the lemma
and use Lemma \ref{lemma-directed-inverse-system-has-limit}
for existence of the limit.
\end{proof}
\begin{lemma}
\label{lemma-infinite-product-surjective}
Let $S$ be a scheme. Let $I$ be a set and for each $i \in I$
let $f_i : T_i \to S$ be a surjective affine morphism. Then the
product $T = \prod T_i$ in the category of schemes over $S$
(Lemma \ref{lemma-infinite-product})
maps surjectively to $S$.
\end{lemma}
\begin{proof}
Let $s \in S$. Choose $t_i \in T_i$ mapping to $s$.
Choose a huge field extension $K/\kappa(s)$ such that
$\kappa(s_i)$ embeds into $K$ for each $i$. Then we get
morphisms $\Spec(K) \to T_i$ with image $s_i$ agreeing
as morphisms to $S$. Whence a morphism $\Spec(K) \to T$
which proves there is a point of $T$ mapping to $s$.
\end{proof}
\begin{lemma}
\label{lemma-infinite-product-integral}
Let $S$ be a scheme. Let $I$ be a set and for each $i \in I$
let $f_i : T_i \to S$ be an integral morphism. Then the
product $T = \prod T_i$ in the category of schemes over $S$
(Lemma \ref{lemma-infinite-product})
is integral over $S$.
\end{lemma}
\begin{proof}
Omitted. Hint: On affine pieces this reduces to the following
algebra fact: if $A \to B_i$ is integral for all $i$, then
$A \to \otimes_A B_i$ is integral.
\end{proof}
\section{Descending properties}
\label{section-descent}
\noindent
First some basic lemmas describing the topology of a limit.
\begin{lemma}
\label{lemma-inverse-limit-sets}
Let $S = \lim S_i$ be the limit of a directed inverse system
of schemes with affine transition morphisms
(Lemma \ref{lemma-directed-inverse-system-has-limit}). Then
$S_{set} = \lim_i S_{i, set}$ where $S_{set}$
indicates the underlying set of the scheme $S$.
\end{lemma}
\begin{proof}
Pick $i \in I$. Take $U_i \subset S_i$ an affine open.
Denote $U_{i'} = f_{i'i}^{-1}(U_i)$ and $U = f_i^{-1}(U_i)$.
Here $f_{i'i} : S_{i'} \to S_i$ is the transtion morphism
and $f_i : S \to S_i$ is the projection.
By Lemma \ref{lemma-directed-inverse-system-has-limit}
we have $U = \lim_{i' \geq i} U_i$.
Suppose we can show that $U_{set} = \lim_{i' \geq i} U_{i', set}$. Then
the lemma follows by a simple argument using an affine covering of $S_i$.
Hence we may assume all $S_i$ and $S$ affine. This reduces us to the
algebra question considered in the next paragraph.
\medskip\noindent
Suppose given a system of rings $(A_i, \varphi_{ii'})$
over $I$. Set $A = \colim_i A_i$ with canonical maps $\varphi_i : A_i \to A$.
Then
$$
\Spec(A) = \lim_i \Spec(A_i)
$$
Namely, suppose that we are given primes $\mathfrak p_i \subset A_i$
such that $\mathfrak p_i = \varphi_{ii'}^{-1}(\mathfrak p_{i'})$
for all $i' \geq i$. Then we simply set
$$
\mathfrak p =
\{x \in A
\mid
\exists i, x_i \in \mathfrak p_i \text{ with }\varphi_i(x_i) = x\}
$$
It is clear that this is an ideal and has the property that
$\varphi_i^{-1}(\mathfrak p) = \mathfrak p_i$. Then it follows
easily that it is a prime ideal as well.
\end{proof}
\begin{lemma}
\label{lemma-inverse-limit-top}
Let $S = \lim S_i$ be the limit of a directed inverse system
of schemes with affine transition morphisms
(Lemma \ref{lemma-directed-inverse-system-has-limit}). Then
$S_{top} = \lim_i S_{i, top}$ where $S_{top}$
indicates the underlying topological space of the scheme $S$.
\end{lemma}
\begin{proof}
We will use the criterion of
Topology, Lemma \ref{topology-lemma-characterize-limit}.
We have seen that $S_{set} = \lim_i S_{i, set}$ in
Lemma \ref{lemma-inverse-limit-sets}.
The maps $f_i : S \to S_i$ are morphisms of schemes
hence continuous. Thus $f_i^{-1}(U_i)$ is open for each
open $U_i \subset S_i$. Finally,
let $s \in S$ and let $s \in V \subset S$ be an open
neighbourhood. Choose $0 \in I$ and choose an
affine open neighbourhood $U_0 \subset S_0$ of the image of $s$.
Then $f_0^{-1}(U_0) = \lim_{i \geq 0} f_{i0}^{-1}(U_0)$, see
Lemma \ref{lemma-directed-inverse-system-has-limit}.
Then $f_0^{-1}(U_0)$ and $f_{i0}^{-1}(U_0)$ are affine and
$$
\mathcal{O}_S(f_0^{-1}(U_0)) =
\colim_{i \geq 0} \mathcal{O}_{S_i}(f_{i0}^{-1}(U_0))
$$
either by the proof of
Lemma \ref{lemma-directed-inverse-system-has-limit}
or by
Lemma \ref{lemma-directed-inverse-system-affine-schemes-has-limit}.
Choose $a \in \mathcal{O}_S(f_0^{-1}(U_0))$ such that
$s \in D(a) \subset V$. This is possible because the principal
opens form a basis for the topology on the affine scheme $f_0^{-1}(U_0)$.
Then we can pick an $i \geq 0$ and
$a_i \in \mathcal{O}_{S_i}(f_{i0}^{-1}(U_0))$ mapping to $a$.
It follows that $D(a_i) \subset f_{i0}^{-1}(U_0) \subset S_i$
is an open subset whose inverse image in $S$ is $D(a)$.
This finishes the proof.
\end{proof}
\begin{lemma}
\label{lemma-limit-nonempty}
Let $S = \lim S_i$ be the limit of a directed inverse system
of schemes with affine transition morphisms
(Lemma \ref{lemma-directed-inverse-system-has-limit}).
If all the schemes $S_i$ are nonempty and quasi-compact,
then the limit $S = \lim_i S_i$ is nonempty.
\end{lemma}
\begin{proof}
Choose $0 \in I$. Note that $I$ is nonempty as the limit is directed.
Choose an affine open covering $S_0 = \bigcup_{j = 1, \ldots, m} U_j$.
Since $I$ is directed there exists a $j \in \{1, \ldots, m\}$
such that $f_{i0}^{-1}(U_j) \not = \emptyset$ for all
$i \geq 0$. Hence $\lim_{i \geq 0} f_{i0}^{-1}(U_j)$ is not
empty since a directed colimit of nonzero rings is nonzero
(because $1 \not = 0$). As $\lim_{i \geq 0} f_{i0}^{-1}(U_j)$
is an open subscheme of the limit we win.
\end{proof}
\begin{lemma}
\label{lemma-inverse-limit-irreducibles}
Let $S = \lim S_i$ be the limit of a directed inverse system
of schemes with affine transition morphisms
(Lemma \ref{lemma-directed-inverse-system-has-limit}).
Let $s \in S$ with images $s_i \in S_i$.
Then $\overline{\{s\}} = \lim_i \overline{\{s_i\}}$ as sets
and as schemes if endowed with the reduced induced scheme structure.
\end{lemma}
\begin{proof}
Choose $0 \in I$ and an affine open covering
$U_0 = \bigcup_{j \in J} U_{0, j}$.
For $i \geq 0$ let $U_{i, j} = f_{i, 0}^{-1}(U_{0, j})$
and set $U_j = f_0^{-1}(U_{0, j})$.
Here $f_{i'i} : S_{i'} \to S_i$ is the transtion morphism
and $f_i : S \to S_i$ is the projection.
For $j \in J$ the following are equivalent:
(a) $x \in U_j$, (b) $x_0 \in U_{0, j}$,
(c) $x_i \in U_{i, j}$ for all $i \geq 0$.
Let $J' \subset J$ be the set of indices for which (a), (b), (c) are true.
Then $\overline{\{s\}} = \bigcup_{j \in J'} (\overline{\{s\}} \cap U_j)$
and simiarly for $\overline{\{s_i\}}$ for $i \geq 0$.
Note that $\overline{\{s\}} \cap U_j$ is the closure of the
set $\{s\}$ in the topological space $U_j$. Similarly for
$\overline{\{s_i\}} \cap U_{i, j}$ for $i \geq 0$.
Hence it suffices to prove the lemma in the case
$S$ and $S_i$ affine for all $i$. This reduces us to the
algebra question considered in the next paragraph.
\medskip\noindent
Suppose given a system of rings $(A_i, \varphi_{ii'})$
over $I$. Set $A = \colim_i A_i$ with canonical maps
$\varphi_i : A_i \to A$. Let $\mathfrak p \subset A$ be a
prime and set $\mathfrak p_i = \varphi_i^{-1}(\mathfrak p)$.
Then
$$
V(\mathfrak p) = \lim_i V(\mathfrak p_i)
$$
This follows from Lemma \ref{lemma-inverse-limit-sets}
because $A/\mathfrak p = \colim A_i/\mathfrak p_i$.
This equality of rings also shows the final statement
about reduced induced scheme structures holds true.
\end{proof}
\noindent
In the rest of this section we work in the following situation.
\begin{situation}
\label{situation-descent}
Let $S = \lim_{i \in I} S_i$ be the limit of a directed system of schemes
with affine transition morphisms $f_{i'i} : S_{i'} \to S_i$
(Lemma \ref{lemma-directed-inverse-system-has-limit}).
We assume that $S_i$ is quasi-compact and quasi-separated for all $i \in I$.
We denote $f_i : S \to S_i$ the projection.
We also choose an element $0 \in I$.
\end{situation}
\noindent
The type of result we are looking for is the following:
If we have an object over $S$, then for some $i$ there is a
similar object over $S_i$.
\begin{lemma}
\label{lemma-topology-limit}
In Situation \ref{situation-descent}.
\begin{enumerate}
\item We have $S_{set} = \lim_i S_{i, set}$ where $S_{set}$
indicates the underlying set of the scheme $S$.
\item We have $S_{top} = \lim_i S_{i, top}$ where $S_{top}$
indicates the underlying topological space of the scheme $S$.
\item If $s, s' \in S$ and $s'$ is not a specialization of $s$
then for some $i \in I$ the image $s'_i \in S_i$ of $s'$ is not
a specialization of the image $s_i \in S_i$ of $s$.
\item Add more easy facts on topology of $S$ here.
(Requirement: whatever is added should be easy in the affine case.)
\end{enumerate}
\end{lemma}
\begin{proof}
Part (1) is a special case of Lemma \ref{lemma-inverse-limit-sets}.
\medskip\noindent
Part (2) is a special case of Lemma \ref{lemma-inverse-limit-top}.
\medskip\noindent
Part (3) is a special case of Lemma \ref{lemma-inverse-limit-irreducibles}.
\end{proof}
\begin{lemma}
\label{lemma-descend-section}
In Situation \ref{situation-descent}.
Suppose that $\mathcal{F}_0$ is a quasi-coherent sheaf on $S_0$.
Set $\mathcal{F}_i = f_{i0}^*\mathcal{F}_0$ for $i \geq 0$ and set
$\mathcal{F} = f_0^*\mathcal{F}_0$.
Then
$$
\Gamma(S, \mathcal{F}) = \colim_{i \geq 0} \Gamma(S_i, \mathcal{F}_i)
$$
\end{lemma}
\begin{proof}
Write $\mathcal{A}_j = f_{i0, *} \mathcal{O}_{S_i}$.
This is a quasi-coherent sheaf of $\mathcal{O}_{S_0}$-algebras
(see Morphisms, Lemma \ref{morphisms-lemma-affine-equivalence-algebras})
and $S_i$ is the relative spectrum of $\mathcal{A}_i$ over $S_0$.
In the proof of Lemma \ref{lemma-directed-inverse-system-has-limit}
we constructed $S$ as the relative spectrum of
$\mathcal{A} = \colim_{i \geq 0} \mathcal{A}_i$
over $S_0$. Set
$$
\mathcal{M}_i = \mathcal{F}_0 \otimes_{\mathcal{O}_{S_0}} \mathcal{A}_i
$$
and
$$
\mathcal{M} = \mathcal{F}_0 \otimes_{\mathcal{O}_{S_0}} \mathcal{A}.
$$
Then we have $f_{i0, *} \mathcal{F}_i = \mathcal{M}_i$
and $f_{0, *}\mathcal{F} = \mathcal{M}$. Since $\mathcal{A}$
is the colimit of the sheaves $\mathcal{A}_i$ and since tensor
product commutes with directed colimits, we conclude that
$\mathcal{M} = \colim_{i \geq 0} \mathcal{M}_i$.
Since $S_0$ is quasi-compact and quasi-separated we see that
\begin{eqnarray*}
\Gamma(S, \mathcal{F})
& = &
\Gamma(S_0, \mathcal{M}) \\
& = &
\Gamma(S_0, \colim_{i \geq 0} \mathcal{M}_i) \\
& = &
\colim_{i \geq 0} \Gamma(S_0, \mathcal{M}_i) \\
& = &
\colim_{i \geq 0} \Gamma(S_i, \mathcal{F}_i)
\end{eqnarray*}
see Sheaves, Lemma \ref{sheaves-lemma-directed-colimits-sections} and
Topology, Lemma \ref{topology-lemma-topology-quasi-separated-scheme}
for the middle equality.
\end{proof}
\begin{lemma}
\label{lemma-limit-closed-nonempty}
In Situation \ref{situation-descent}.
Suppose for each $i$ we are given a nonempty closed subset
$Z_i \subset S_i$ with $f_{ii'}(Z_i) \subset Z_{i'}$.
Then there exists a point $s \in S$ with $f_i(s) \in Z_i$ for
all $i$.
\end{lemma}
\begin{proof}
Let $Z_i \subset S_i$ also denote the reduced closed subscheme
associated to $Z_i$, see Schemes,
Definition \ref{schemes-definition-reduced-induced-scheme}.
A closed immersion is affine, and a composition of affine
morphisms is affine (see
Morphisms, Lemmas \ref{morphisms-lemma-closed-immersion-affine}
and \ref{morphisms-lemma-composition-affine}), and hence $Z_i \to S_{i'}$ is
affine when $i \geq i'$. We conclude that the morphism
$f_{ii'} : Z_i \to Z_{i'}$ is affine by
Morphisms, Lemma \ref{morphisms-lemma-affine-permanence}.
Each of the schemes $Z_i$ is quasi-compact as a closed
subscheme of a quasi-compact scheme. Hence we may apply
Lemma \ref{lemma-limit-nonempty} to see that
$Z = \lim_i Z_i$ is nonempty. Since there is a
canonical morphism $Z \to S$ we win.
\end{proof}
\begin{lemma}
\label{lemma-limit-fibre-product-empty}
In Situation \ref{situation-descent}.
Suppose we are given an $i$ and a morphism $T \to S_i$ such that
\begin{enumerate}
\item $T \times_{S_i} S = \emptyset$, and
\item $T$ is quasi-compact.
\end{enumerate}
Then $T \times_{S_i} S_{i'} = \emptyset$ for all sufficiently large $i'$.
\end{lemma}
\begin{proof}
By Lemma \ref{lemma-scheme-over-limit}
we see that $T \times_{S_i} S = \lim_{i' \geq i} T \times_{S_i} S_{i'}$.
Hence the result follows from
Lemma \ref{lemma-limit-nonempty}.
\end{proof}
\begin{lemma}
\label{lemma-limit-contained-in-constructible}
In Situation \ref{situation-descent}.
Suppose we are given an $i$ and a locally constructible subset
$E \subset S_i$ such that $f_i(S) \subset E$.
Then $f_{ii'}(S_{i'}) \subset E$ for all sufficiently large $i'$.
\end{lemma}
\begin{proof}
Writing $S_i$ as a finite union of open affine subschemes reduces
the question to the case that $S_i$ is affine and $E$ is constructible, see
Lemma \ref{lemma-directed-inverse-system-has-limit}
and
Properties, Lemma \ref{properties-lemma-locally-constructible}.
In this case the complement $S_i \setminus E$ is constructible too.
Hence there exists an affine scheme $T$ and a morphism $T \to S_i$
whose image is $S_i \setminus E$, see
Algebra, Lemma \ref{algebra-lemma-constructible-is-image}.
By
Lemma \ref{lemma-limit-fibre-product-empty}
we see that $T \times_{S_i} S_{i'}$ is empty for all sufficiently large
$i'$, and hence $f_{ii'}(S_{i'}) \subset E$ for all sufficiently large $i'$.
\end{proof}
\begin{lemma}
\label{lemma-descend-opens}
In Situation \ref{situation-descent} we have the following:
\begin{enumerate}
\item Given any quasi-compact open $V \subset S = \lim_i S_i$
there exists an $i \in I$ and a quasi-compact open $V_i \subset S_i$
such that $f_i^{-1}(V_i) = V$.
\item Given $V_i \subset S_i$ and $V_{i'} \subset S_{i'}$
quasi-compact opens such that $f_i^{-1}(V_i) = f_{i'}^{-1}(V_{i'})$
there exists an index $i'' \geq i, i'$ such that
$f_{i''i}^{-1}(V_i) = f_{i''i'}^{-1}(V_{i'})$.
\item If $V_{1, i}, \ldots, V_{n, i} \subset S_i$ are quasi-compact
opens and $S = f_i^{-1}(V_{1, i}) \cup \ldots \cup f_i^{-1}(V_{n, i})$
then $S_{i'} = f_{i'i}^{-1}(V_{1, i}) \cup \ldots \cup f_{i'i}^{-1}(V_{n, i})$
for some $i' \geq i$.
\end{enumerate}
\end{lemma}
\begin{proof}
Choose $i_0 \in I$. Note that $I$ is nonempty as the limit is directed.
For convenience we write $S_0 = S_{i_0}$ and $i_0 = 0$.
Choose an affine open covering $S_0 = U_{1, 0} \cup \ldots \cup U_{m, 0}$.
Denote $U_{j, i} \subset S_i$ the inverse image of $U_{j, 0}$
under the transition morphism for $i \geq 0$.
Denote $U_j$ the inverse image of $U_{j, 0}$ in $S$.
Note that $U_j = \lim_i U_{j, i}$ is a limit of affine
schemes.
\medskip\noindent
We first prove the uniqueness statement: Let
$V_i \subset S_i$ and $V_{i'} \subset S_{i'}$
quasi-compact opens such that $f_i^{-1}(V_i) = f_{i'}^{-1}(V_{i'})$.
It suffices to show that $f_{i''i}^{-1}(V_i \cap U_{j, i''})$ and
$f_{i''i'}^{-1}(V_{i'} \cap U_{j, i''})$ become equal
for $i''$ large enough. Hence we reduce to the case
of a limit of affine schemes. In this case write
$S = \Spec(R)$ and $S_i = \Spec(R_i)$ for all $i \in I$.
We may write $V_i = S_i \setminus V(h_1, \ldots, h_m)$
and $V_{i'} = S_{i'} \setminus V(g_1, \ldots, g_n)$.
The assumption means that the ideals
$\sum g_jR$ and $\sum h_jR$ have the same radical
in $R$. This means that $g_j^N = \sum a_{jj'}h_{j'}$ and
$h_j^N = \sum b_{jj'} g_{j'}$ for some $N \gg 0$ and $a_{jj'}$
and $b_{jj'}$ in $R$.
Since $R = \colim_i R_i$ we can chose an index
$i'' \geq i$ such that the equations
$g_j^N = \sum a_{jj'}h_{j'}$ and
$h_j^N = \sum b_{jj'} g_{j'}$ hold in $R_{i''}$ for some
$a_{jj'}$ and $b_{jj'}$ in $R_{i''}$. This implies that
the ideals $\sum g_jR_{i''}$ and $\sum h_jR_{i''}$ have the same radical
in $R_{i''}$ as desired.
\medskip\noindent
We prove existence: If $S_0$ is affine, then $S_i = \Spec(R_i)$ for all
$i \geq 0$ and $S = \Spec(R)$ with $R = \colim R_i$. Then
$V = S \setminus V(g_1, \ldots, g_n)$ for some $g_1, \ldots, g_n \in R$.
Choose any $i$ large enough so that each of the $g_j$ comes from an
element $g_{j, i} \in R_i$ and take
$V_i = S_i \setminus V(g_{1, i}, \ldots, g_{n, i})$.
If $S_0$ is general, then the opens $V \cap U_j$
are quasi-compact because $S$ is quasi-separated. Hence by the
affine case we see that for each $j = 1, \ldots, m$
there exists an $i_j \in I$ and a quasi-compact open
$V_{i_j} \subset U_{j, i_j}$ whose inverse image in $U_j$
is $V \cap U_j$. Set $i = \max(i_1, \ldots, i_m)$
and let $V_i = \bigcup f_{ii_j}^{-1}(V_{i_j})$.
\medskip\noindent
The statement on coverings follows from the uniqueness statement
for the opens $V_{1, i} \cup \ldots \cup V_{n, i}$ and $S_i$ of $S_i$.
\end{proof}
\begin{lemma}
\label{lemma-limit-quasi-affine}
In Situation \ref{situation-descent} if $S$ is quasi-affine, then
for some $i_0 \in I$ the schemes $S_i$ for $i \geq i_0$ are quasi-affine.
\end{lemma}
\begin{proof}
Choose $i_0 \in I$. Note that $I$ is nonempty as the limit is directed.
For convenience we write $S_0 = S_{i_0}$ and $i_0 = 0$.
Let $s \in S$. We may choose an affine open
$U_0 \subset S_0$ containing $f_0(s)$. Since $S$ is quasi-affine
we may choose an element $a \in \Gamma(S, \mathcal{O}_S)$ such
that $s \in D(a) \subset f_0^{-1}(U_0)$, and such that
$D(a)$ is affine. By Lemma \ref{lemma-descend-section}
there exists an $i \geq 0$ such that $a$
comes from an element $a_i \in \Gamma(S_i, \mathcal{O}_{S_i})$.
For any index $j \geq i$ we denote $a_j$
the image of $a_i$ in the global sections of the
structure sheaf of $S_j$.
Consider the opens $D(a_j) \subset S_j$
and $U_j = f_{j0}^{-1}(U_0)$. Note that
$U_j$ is affine and $D(a_j)$ is a quasi-compact open of $S_j$,
see Properties, Lemma \ref{properties-lemma-affine-cap-s-open}
for example. Hence we may apply Lemma \ref{lemma-descend-opens} to the opens
$U_j$ and $U_j \cup D(a_j)$ to conclude that
$D(a_j) \subset U_j$ for some $j \geq i$.
For such an index $j$ we see that $D(a_j) \subset S_j$ is an affine open
(because $D(a_j)$ is a standard affine open of the affine open $U_j$)
containing the image $f_j(s)$.
\medskip\noindent
We conclude that for every $s \in S$ there exist
an index $i \in I$, and a global section
$a \in \Gamma(S_i, \mathcal{O}_{S_i})$
such that $D(a) \subset S_i$ is an affine open
containing $f_i(s)$. Because $S$ is quasi-compact we
may choose a single index $i \in I$ and global sections
$a_1, \ldots, a_m \in \Gamma(S_i, \mathcal{O}_{S_i})$
such that each $D(a_j) \subset S_i$ is affine open
and such that $f_i : S \to S_i$ has image contained
in the union $W_i = \bigcup_{j = 1, \ldots, m} D(a_j)$.
For $i' \geq i$ set $W_{i'} = f_{i'i}^{-1}(W_i)$.
Since $f_i^{-1}(W_i)$ is all of $S$ we see
(by Lemma \ref{lemma-descend-opens} again)
that for a suitable $i' \geq i$ we
have $S_{i'} = W_{i'}$. Thus we may replace $i$ by
$i'$ and assume that $S_i = \bigcup_{j = 1, \ldots, m} D(a_j)$.
This implies that $\mathcal{O}_{S_i}$ is an ample invertible
sheaf on $S_i$ (see Properties, Definition \ref{properties-definition-ample})
and hence that $S_i$ is quasi-affine, see
Properties, Lemma \ref{properties-lemma-quasi-affine-O-ample}.
Hence we win.
\end{proof}
\begin{lemma}
\label{lemma-limit-affine}
In Situation \ref{situation-descent} if $S$ is affine,
then for some $i_0 \in I$ the schemes $S_i$ for $i \geq i_0$
are affine.
\end{lemma}
\begin{proof}
By Lemma \ref{lemma-limit-quasi-affine} we may assume that $S_0$ is
quasi-affine for some $0 \in I$. Set $R_0 = \Gamma(S_0, \mathcal{O}_{S_0})$.
Then $S_0$ is a quasi-compact open of $T_0 = \Spec(R_0)$. Denote
$j_0 : S_0 \to T_0$ the corresponding quasi-compact open immersion.
For $i \geq 0$ set $\mathcal{A}_i = f_{0i, *}\mathcal{O}_{S_i}$.
Since $f_{0i}$ is affine we see that
$S_i = \underline{\Spec}_{S_0}(\mathcal{A}_i)$.
Set $T_i = \underline{\Spec}_{T_0}(j_{0, *}\mathcal{A}_i)$.
Then $T_i \to T_0$ is affine, hence $T_i$ is affine. Thus
$T_i$ is the spectrum of
$$
R_i = \Gamma(T_0, j_{0, *}\mathcal{A}_i) = \Gamma(S_0, \mathcal{A}_i) =
\Gamma(S_i, \mathcal{O}_{S_i}).
$$
Write $S = \Spec(R)$. We have $R = \colim_i R_i$
by Lemma \ref{lemma-descend-section}.
Hence also $S = \lim_i T_i$. As formation of the relative spectrum commutes
with base change, the inverse image
of the open $S_0 \subset T_0$ in $T_i$ is $S_i$.
Let $Z_0 = T_0 \setminus S_0$ and let $Z_i \subset T_i$
be the inverse image of $Z_0$. As $S_i = T_i \setminus Z_i$, it suffices
to show that $Z_i$ is empty for some $i$. Assume $Z_i$ is nonempty for all
$i$ to get a contradiction. By Lemma \ref{lemma-limit-closed-nonempty}
there exists a point $s$ of $S = \lim T_i$ which maps to a point of $Z_i$
for every $i$. But $S = \lim_i S_i$, and hence we arrive at a contradiction
by Lemma \ref{lemma-topology-limit}.
\end{proof}
\begin{lemma}
\label{lemma-limit-separated}
In Situation \ref{situation-descent} if $S$ is separated,
then for some $i_0 \in I$ the schemes $S_i$ for $i \geq i_0$
are separated.
\end{lemma}
\begin{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
$U_{j_1} \cap U_{j_2} = \lim_{i \geq 0} U_{i, j_1} \cap U_{i, j_2}$
we see that $U_{i, j_1} \cap U_{i, j_2}$ is affine for large $i$
by Lemma \ref{lemma-limit-affine}. To show that $S_i$ is separated
for large $i$ it now suffices to show that
$$
\mathcal{O}_{S_i}(V_{i, j_1})
\otimes_{\mathcal{O}_S(S)}
\mathcal{O}_{S_i}(V_{i, j_2})
\longrightarrow
\mathcal{O}_{S_i}(V_{i, j_1} \cap V_{i, j_2})
$$
is surjective for large $i$
(Schemes, Lemma \ref{schemes-lemma-characterize-separated}).
\medskip\noindent
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 \ref{schemes-lemma-diagonal-immersion})
hence locally of finite type
(Morphisms, Lemmas \ref{morphisms-lemma-locally-finite-type-characterize} and
\ref{morphisms-lemma-immersion-locally-finite-type}).
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
$\mathcal{O}(U) \otimes \mathcal{O}(V) =
\colim \mathcal{O}(U_i) \otimes \mathcal{O}(V_i)$
we see that they are in the image of the map at some finite level
and the lemma is proved.
\end{proof}
\begin{lemma}
\label{lemma-limit-ample}
In Situation \ref{situation-descent} let $\mathcal{L}_0$ be an invertible
sheaf of modules on $S_0$. If the pullback $\mathcal{L}$ to $S$ is ample,
then for some $i \in I$ the pullback $\mathcal{L}_i$ to $S_i$ is ample.
\end{lemma}
\begin{proof}
The assumption means there are finitely many sections
$s_1, \ldots, s_m \in \Gamma(S, \mathcal{L})$ such that
$S_{s_j}$ is affine and such that $S = \bigcup S_{s_j}$, see
Properties, Definition \ref{properties-definition-ample}.
By Lemma \ref{lemma-descend-section} we can find an $i \in I$
and sections $s_{i, j} \in \Gamma(S_i, \mathcal{L}_i)$ mapping to $s_j$.
By Lemma \ref{lemma-limit-affine} we may, after increasing $i$, assume
that $(S_i)_{s_{i, j}}$ is affine for $j = 1, \ldots, m$.
By Lemma \ref{lemma-descend-opens} we may, after increasing $i$ a
last time, assume that $S_i = \bigcup (S_i)_{s_{i, j}}$.
Then $\mathcal{L}_i$ is ample by definition.
\end{proof}
\begin{lemma}
\label{lemma-finite-type-eventually-closed}
Let $S$ be a scheme. Let $X = \lim X_i$ be a directed limit of
schemes over $S$ with affine transition morphisms. Let $Y \to X$
be a morphism of schemes over $S$.
\begin{enumerate}
\item If $Y \to X$ is a closed immersion, $X_i$ quasi-compact, and
$Y$ locally of finite type over $S$, then $Y \to X_i$ is a closed
immersion for $i$ large enough.
\item If $Y \to X$ is an immersion, $X_i$ quasi-separated, $Y \to S$ locally
of finite type, and $Y$ quasi-compact, then $Y \to X_i$ is an
immersion for $i$ large enough.
\end{enumerate}
\end{lemma}
\begin{proof}
Proof of (1). Choose $0 \in I$ and a finite affine open covering
$X_0 = U_{0, 1} \cup \ldots \cup U_{0, m}$ with the property that
$U_{0, j}$ maps into an affine open $W_j \subset S$.
Let $V_j \subset Y$, resp.\ $U_{i, j} \subset X_i$, $i \geq 0$,
resp. $U_j \subset X$ be the inverse image of $U_{0, j}$. It suffices
to prove that $V_j \to U_{i, j}$ is a closed immersion for $i$
sufficiently large and we know that $V_j \to U_j$ is a closed immersion.
Thus we reduce to the following algebra fact: If $A = \colim A_i$ is a
directed colimit of $R$-algebras, $A \to B$ is a surjection of $R$-algebras,
and $B$ is a finitely generated $R$-algebra, then
$A_i \to B$ is surjective for $i$ sufficiently large.
\medskip\noindent
Proof of (2). Choose $0 \in I$. Choose a quasi-compact open
$X'_0 \subset X_0$ such that $Y \to X_0$ factors through $X'_0$.
After replacing $X_i$ by the inverse image of $X'_0$ for $i \geq 0$
we may assume all $X_i'$ are quasi-compact and quasi-separated.
Let $U \subset X$ be a quasi-compact open such that $Y \to X$ factors
through a closed immersion $Y \to U$ ($U$ exists as $Y$ is quasi-compact). By
Lemma \ref{lemma-descend-opens}
we may assume that $U = \lim U_i$ with $U_i \subset X_i$ quasi-compact
open. By part (1) we see that $Y \to U_i$ is a closed immersion for some
$i$. Thus (2) holds.
\end{proof}
\begin{lemma}
\label{lemma-eventually-separated}
Let $S$ be a scheme. Let $X = \lim X_i$ be a directed
limit of schemes over $S$ with affine transition morphisms.
Assume
\begin{enumerate}
\item $S$ quasi-separated,
\item $X_i$ quasi-compact and quasi-separated,
\item $X \to S$ separated.
\end{enumerate}
Then $X_i \to S$ is separated for all $i$ large enough.
\end{lemma}
\begin{proof}
Let $0 \in I$. Note that $I$ is nonempty as the limit is directed.
As $X_0$ is quasi-compact we can find finitely many
affine opens $U_1, \ldots, U_n \subset S$ such that
$X_0 \to S$ maps into $U_1 \cup \ldots \cup U_n$.
Denote $h_i : X_i \to S$ the structure morphism.
It suffices to check that for some $i \geq 0$ the morphisms
$h_i^{-1}(U_j) \to U_j$ are separated for $j = 1, \ldots, n$.
Since $S$ is quasi-separated the morphisms $U_j \to S$ are quasi-compact.
Hence $h_i^{-1}(U_j)$ is quasi-compact and quasi-separated.
In this way we reduce to the case $S$ affine. In this case we
have to show that $X_i$ is separated and we know that $X$ is separated.
Thus the lemma follows from Lemma \ref{lemma-limit-separated}.
\end{proof}
\begin{lemma}
\label{lemma-eventually-affine}
Let $S$ be a scheme. Let $X = \lim X_i$ be a directed limit of schemes
over $S$ with affine transition morphisms. Assume
\begin{enumerate}
\item $S$ quasi-compact and quasi-separated,
\item $X_i$ quasi-compact and quasi-separated,
\item $X \to S$ affine.
\end{enumerate}
Then $X_i \to S$ is affine for $i$ large enough.
\end{lemma}
\begin{proof}
Choose a finite affine open covering $S = \bigcup_{j = 1, \ldots, n} V_j$.
Denote $f : X \to S$ and $f_i : X_i \to S$ the structure morphisms.
For each $j$ the scheme $f^{-1}(V_j) = \lim_i f_i^{-1}(V_j)$
is affine (as a finite morphism is affine by definition). Hence by
Lemma \ref{lemma-limit-affine} there exists an $i \in I$ such that
each $f_i^{-1}(V_j)$ is affine. In other words, $f_i : X_i \to S$ is
affine for $i$ large enough, see
Morphisms, Lemma \ref{morphisms-lemma-characterize-affine}.
\end{proof}
\begin{lemma}
\label{lemma-eventually-finite}
Let $S$ be a scheme. Let $X = \lim X_i$ be a directed limit of schemes
over $S$ with affine transition morphisms. Assume
\begin{enumerate}
\item $S$ quasi-compact and quasi-separated,
\item $X_i$ quasi-compact and quasi-separated,
\item the transition morphisms $X_{i'} \to X_i$ are finite,
\item $X_i \to S$ locally of finite type
\item $X \to S$ integral.
\end{enumerate}
Then $X_i \to S$ is finite for $i$ large enough.
\end{lemma}
\begin{proof}
By Lemma \ref{lemma-eventually-affine}
we may assume $X_i \to S$ is affine for all $i$.
Choose a finite affine open covering $S = \bigcup_{j = 1, \ldots, n} V_j$.
Denote $f : X \to S$ and $f_i : X_i \to S$ the structure morphisms.
It suffices to show that there exists an $i$ such that
$f_i^{-1}(V_j)$ is finite over $V_j$ for $j = 1, \ldots, m$
(Morphisms, Lemma \ref{morphisms-lemma-finite-local}).
Namely, for $i' \geq i$ the composition $X_{i'} \to X_i \to S$
will be finite as a composition of finite morphisms
(Morphisms, Lemma \ref{morphisms-lemma-composition-finite}).
This reduces us to the affine case: Let $R$ be a ring and
$A = \colim A_i$ with $R \to A$ integral and $A_i \to A_{i'}$
finite for all $i \leq i'$. Moreover $R \to A_i$ is of finite type
for all $i$. Goal: Show that $A_i$ is finite over $R$ for some $i$.
To prove this choose an $i \in I$ and pick generators
$x_1, \ldots, x_m \in A_i$ of $A_i$ as an $R$-algebra.
Since $A$ is integral over $R$ we can find monic polynomials
$P_j \in R[T]$ such that $P_j(x_j) = 0$ in $A$.
Thus there exists an $i' \geq i$ such that $P_j(x_j) = 0$ in $A_{i'}$
for $j = 1, \ldots, m$. Then the image $A'_i$ of $A_i$ in $A_{i'}$
is finite over $R$ by
Algebra, Lemma \ref{algebra-lemma-characterize-finite-in-terms-of-integral}.
Since $A'_i \subset A_{i'}$ is finite too we conclude
that $A_{i'}$ is finite over $R$ by
Algebra, Lemma \ref{algebra-lemma-finite-transitive}.
\end{proof}
\begin{lemma}
\label{lemma-eventually-closed-immersion}
Let $S$ be a scheme. Let $X = \lim X_i$ be a directed limit of schemes
over $S$ with affine transition morphisms. Assume
\begin{enumerate}
\item $S$ quasi-compact and quasi-separated,
\item $X_i$ quasi-compact and quasi-separated,
\item the transition morphisms $X_{i'} \to X_i$ are closed immersions,
\item $X_i \to S$ locally of finite type
\item $X \to S$ a closed immersion.
\end{enumerate}
Then $X_i \to S$ is a closed immersion for $i$ large enough.
\end{lemma}
\begin{proof}
By Lemma \ref{lemma-eventually-affine}
we may assume $X_i \to S$ is affine for all $i$.
Choose a finite affine open covering $S = \bigcup_{j = 1, \ldots, n} V_j$.
Denote $f : X \to S$ and $f_i : X_i \to S$ the structure morphisms.
It suffices to show that there exists an $i$ such that
$f_i^{-1}(V_j)$ is a closed subscheme of $V_j$ for $j = 1, \ldots, m$
(Morphisms, Lemma \ref{morphisms-lemma-closed-immersion}).
This reduces us to the affine case: Let $R$ be a ring and
$A = \colim A_i$ with $R \to A$ surjective and $A_i \to A_{i'}$
surjective for all $i \leq i'$. Moreover $R \to A_i$ is of finite type
for all $i$. Goal: Show that $R \to A_i$ is surjective for some $i$.
To prove this choose an $i \in I$ and pick generators
$x_1, \ldots, x_m \in A_i$ of $A_i$ as an $R$-algebra.
Since $R \to A$ is surjective we can find
$r_j \in R$ such that $r_j$ maps to $x_j$ in $A$.
Thus there exists an $i' \geq i$ such that $r_j$ maps to the image
of $x_j$ in $A_{i'}$ for $j = 1, \ldots, m$. Since $A_i \to A_{i'}$
is surjective this implies that $R \to A_{i'}$ is surjective.
\end{proof}
\section{Absolute Noetherian Approximation}
\label{section-approximation}
\noindent
A nice reference for this section is Appendix C of the article
by Thomason and Trobaugh \cite{TT}.
See Categories, Section \ref{categories-section-posets-limits}
for our conventions regarding directed systems.
We will use the existence result and properties of the limit
from Section \ref{section-limits} without further mention.
\begin{lemma}
\label{lemma-quasi-affine-finite-type-over-Z}
Let $W$ be a quasi-affine scheme of finite type over
$\mathbf{Z}$. Suppose $W \to \Spec(R)$ is an
open immersion into an affine scheme. There exists a
finite type $\mathbf{Z}$-algebra $A \subset R$
which induces an open immersion $W \to \Spec(A)$.
Moreover, $R$ is the directed colimit of such subalgebras.
\end{lemma}
\begin{proof}
Choose an affine open covering $W = \bigcup_{i = 1, \ldots, n} W_i$
such that each $W_i$ is a standard affine open in $\Spec(R)$.
In other words, if we write $W_i = \Spec(R_i)$
then $R_i = R_{f_i}$ for some $f_i \in R$.
Choose finitely many $x_{ij} \in R_i$ which generate
$R_i$ over $\mathbf{Z}$.
Pick an $N \gg 0$ such that each $f_i^Nx_{ij}$ comes from an
element of $R$, say $y_{ij} \in R$.
Set $A$ equal to the $\mathbf{Z}$-algebra generated by
the $f_i$ and the $y_{ij}$ and (optionally) finitely many
additional elements of $R$. Then $A$ works. Details omitted.
\end{proof}