A Primer of Commutative Algebra

James S. Milne
May 2, 2013, v3.00
Abstract
These notes collect the basic results in commutative algebra used in the rest of my
notes and books.
Contents
1 Rings and algebras . . . . . . . . . . . . . . . . . . . . . . 3
2 Ideals . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3 Noetherian rings . . . . . . . . . . . . . . . . . . . . . . . 9
4 Unique factorization . . . . . . . . . . . . . . . . . . . . . . 14
5 Rings of fractions . . . . . . . . . . . . . . . . . . . . . . . 17
6 Integrality . . . . . . . . . . . . . . . . . . . . . . . . . 23
7 Artinian rings . . . . . . . . . . . . . . . . . . . . . . . . 32
8 Direct and inverse limits . . . . . . . . . . . . . . . . . . . . 34
9 Tensor Products . . . . . . . . . . . . . . . . . . . . . . . 37
10 Flatness . . . . . . . . . . . . . . . . . . . . . . . . . . 41
11 Finitely generated projective modules . . . . . . . . . . . . . . . 46
12 Zariskis lemma and the Hilbert Nullstellensatz . . . . . . . . . . . . 53
13 The spectrum of a ring . . . . . . . . . . . . . . . . . . . . . 57
14 Jacobson rings and max spectra. . . . . . . . . . . . . . . . . . 62
15 Quasi-nite algebras and Zariskis main theorem. . . . . . . . . . . . 66
16 Dimension theory for nitely generated k-algebras . . . . . . . . . . . 73
17 Primary decompositions . . . . . . . . . . . . . . . . . . . . 76
18 Dedekind domains . . . . . . . . . . . . . . . . . . . . . . 80
19 Dimension theory for noetherian rings . . . . . . . . . . . . . . . 85
20 Regular local rings . . . . . . . . . . . . . . . . . . . . . . 89
21 Completions. . . . . . . . . . . . . . . . . . . . . . . . . 91
References . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
c _2009, 2010, 2011, 2012, 2013 J.S. Milne. Single paper copies for noncommercial personal use may be
made without explicit permission from the copyright holder. Available at www.jmilne.org/math/.
1
CONTENTS 2
Notations and conventions
Our convention is that rings have identity elements,
1
and homomorphisms of rings respect
the identity elements. A unit of a ring is an element admitting an inverse. The units of a ring
form a group, which we denote by
2

. Throughout ring means commutative ring.

Following Bourbaki, we let N ={0. 1. 2. . . .]. For a eld k, k
al
denotes an algebraic closure
of k.
X Y X is a subset of Y (not necessarily proper).
X
def
=Y X is dened to be Y , or equals Y by denition.
X ~Y X is isomorphic to Y .
X .Y X and Y are canonically isomorphic
(or there is a given or unique isomorphism).
Prerequisites
A knowledge of the algebra usually taught in advanced undergraduate or rst-year graduate
courses.
References
A reference to monnnn is to question nnnn on mathoverow.net.
Acknowledgements
I thank the following for providing corrections and comments for earlier versions of these
notes: Florian Herzig, Chun Yin Hui, Keenan Kidwell, Leon Lampret, Andrew McLennan,
Shu Otsuka, Bhupendra Nath Tiwari.
1
An element e of a ring is an identity element if ea =a =ae for all elements a of the ring. It is usually
denoted 1

or just 1. Some authors call this a unit element, but then an element can be a unit without being a
unit element. Worse, a unit need not be the unit.
2
This notation differs from that of Bourbaki, who writes

for the multiplicative monoid {0] and

+
for the group of units. We shall rarely need the former, and
+
is overused.
1 RINGS AND ALGEBRAS 3
1 Rings and algebras
A ring is an integral domain if it is not the zero ring and if ab =0 in the ring implies that
a =0 or b =0.
Let be a ring. A subring of is a subset that contains 1

and is closed under addition,

multiplication, and the formation of negatives. An -algebra is a ring T together with a
homomorphism i
B
: T. A homomorphism of -algebras T C is a homomorphism
of rings : T C such that (i
B
(a)) =i
C
(a) for all a .
Elements .
1
. . . . . .
n
of an -algebra T are said to generate it if every element of T can
be expressed as a polynomial in the .
i
with coefcients in i
B
(), i.e., if the homomorphism
of -algebras X
1
. . . . . X
n
| T acting as i
B
on and sending X
i
to .
i
is surjective.
When T and .
1
. . . . . .
n
T, we let .
1
. . . . . .
n
| denote the -subalgebra of T
generated by the .
i
.
A ring homomorphism T is of nite type, and T is a nitely generated -algebra,
if T is generated by a nite set of elements as an -algebra, i.e, if T is a quotient of a
polynomial ring X
1
. . . . . X
n
|. An -algebra T is nitely presented if it is the quotient of
a polynomial ring kX
1
. . . . . X
n
| by a nitely generated ideal.
A ring homomorphism T is nite, and T is a nite
3
-algebra, if T is nitely
generated as an -module. If T and T C are nite ring homomorphisms, then so
also is their composite C.
Let k be a eld, and let be a k-algebra. When 1

=0, the map k is injective,

and we can identify k with its image, i.e., we can regard k as a subring of . When 1

=0,
the ring is the zero ring {0].
Let X| be the ring of polynomials in the symbol X with coefcients in . If is an
integral domain, then deg(g) =deg( ) deg(g), and so X| is also an integral domain;
moreover, X|

.
Let be an integral domain and an algebra over a eld k. If is nite over k (more
generally, if every element of is algebraic over k), then is a eld. To see this, let a be a
nonzero element of . Because is an integral domain, the k-linear map . a.:
is injective, and hence is surjective if is nite, which shows that a has an inverse. More
generally, if a is algebraic over k, then ka| is nite over k, and hence contains an inverse of
a; again is a eld.
Products and idempotents
An element e of a ring is idempotent if e
2
=e. For example, 0 and 1 are both idempotents
they are called the trivial idempotents. Idempotents e
1
. . . . . e
n
are orthogonal if e
i
e
}
=0
for i = . Every sum of orthogonal idempotents is again idempotent. A set {e
1
. . . . . e
n
] of
orthogonal idempotents is complete if e
1
e
n
=1. Every set of orthogonal idempotents
{e
1
. . . . . e
n
] can be made into a complete set of orthogonal idempotents by adding the
idempotent e =1(e
1
e
n
).
If =
1

n
(direct product of rings), then the elements
e
i
=(0. . . . .
i
1. . . . . 0). 1 _i _n.
3
This is Bourbakis terminology (AC V 1, 1). Finite homomorphisms of rings correspond to nite maps of
varieties and schemes. Some other authors say module-nite.
2 IDEALS 4
form a complete set of orthogonal idempotents in . Conversely, if {e
1
. . . . . e
n
] is a com-
plete set of orthogonal idempotents in , then e
i
becomes a ring
4
with the addition and
multiplication induced by that of , and .e
1
e
n
.
2 Ideals
Let be a ring. An ideal a in is a subset such that
a is a subgroup of regarded as a group under addition;
a a, r == ra a.
The ideal generated by a subset S of is the intersection of all ideals a containing S it
is easy to verify that this is in fact an ideal, and that it consists of all nite sums of the form

r
i
s
i
with r
i
, s
i
S. The ideal generated by the empty set is the zero ideal {0]. When
S ={a. b. . . .], we write (a. b. . . .) for the ideal it generates.
An ideal is principal if it is generated by a single element. Such an ideal (a) is proper if
and only if a is not a unit. Thus a ring is a eld if and only if 1

=0 and contains no
nonzero proper ideals.
Let a and b be ideals in . The set {ab [ a a. b b] is an ideal, denoted ab. The
ideal generated by {ab [ a a. b b] is denoted by ab. Clearly ab consists of all nite
sums

a
i
b
i
with a
i
a and b
i
b, and if a = (a
1
. . . . . a
n
) and b = (b
1
. . . . . b
n
), then
ab =(a
1
b
1
. . . . . a
i
b
}
. . . . . a
n
b
n
). Note that ab a =a and ab b =b, and so
ab ab. (1)
The kernel of a homomorphism T is an ideal in . Conversely, for every ideal a in
a ring , the set of cosets of a in forms a ring ,a, and a a a is a homomorphism
: ,a whose kernel is a. There is a one-to-one correspondence
{ideals of containing a]
b|-(b)

1
(b)-b
{ideals of ,a]. (2)
For an ideal b of ,
-1
(b) =ab.
The ideals of T are all of the form a b with a and b ideals in and T. To see
this, note that if c is an ideal in T and (a. b) c, then (a. 0) = (1. 0)(a. b) c and
(0. b) =(0. 1)(a. b) c. Therefore, c =ab with
a ={a [ (a. 0) c]. b ={b [ (0. b) c].
An ideal p in is prime if p = and ab p =a p or b p. Thus p is prime if and
only if the quotient ring ,p is nonzero and has the property that
ab =0 == a =0 or b =0.
i.e., ,p is an integral domain. In particular, the zero ideal is prime if and only if the ring is
an integral domain. Note that if p is prime and a
1
a
n
p, then at least one of the a
i
p
(because either a
1
p or a
2
a
n
p; if the latter, then either a
2
p or a
3
a
n
p; etc.).
When p is prime, we write k(p) for the eld of fractions of ,p.
4
But e
i
is not a subring of if n =1 because its identity element is e
i
=1

. However, the map a

ae
i
: e
i
realizes e
i
as a quotient of .
2 IDEALS 5
An ideal m in is maximal if it is a maximal element of the set of proper ideals in .
Therefore an ideal m is maximal if and only if the quotient ring ,m is nonzero and has no
proper nonzero ideals (by (2)), and so is a eld. Note that
m maximal == m prime.
A multiplicative subset of a ring is a subset S with the property:
1 S. a. b S == ab S.
For example, the following are multiplicative subsets:
the multiplicative subset {1. . . . . .
i
. . . .] generated by an element of ;
the complement of a prime ideal (or of a union of prime ideals);
1a
def
={1a [ a a] for any ideal a of .
PROPOSITION 2.1. Let S be a subset of a ring , and let a be an ideal disjoint from S.
The set of ideals in containing a and disjoint from S contains maximal elements (i.e., an
element not properly contained in any other ideal in the set). If S is multiplicative, then
every such maximal element is prime.
PROOF. The set of ideals containing a and disjoint from S is nonempty (it contains a).
If is noetherian (see 3 below), automatically contains maximal elements. Otherwise,
we apply Zorns lemma. Let b
1
b
2
be a chain of ideals in , and let b =
_
b
i
.
Then b , because otherwise some element of S lies in b, and hence in some b
i
, which
contradicts the denition of . Therefore b is an upper bound for the chain. As every chain
in has an upper bound, Zorns lemma implies that has a maximal element.
Now assume that S is a multiplicative subset of , and let c be maximal in . Let
bb
t
c. If b is not in c, then c (b) properly contains c, and so it is not in . Therefore
there S contains an element in c (b), say,
=c ab. c c. a .
Similarly, if b
t
is not in c, then S contains an element

t
=c
t
a
t
b. c
t
c. a
t
.
Now

t
=cc
t
abc
t
a
t
b
t
c aa
t
bb
t
c.
which contradicts

t
S.
Therefore, at least one of b or b
t
is in c, which is therefore prime.
2
COROLLARY 2.2. Every proper ideal in a ring is contained in a maximal ideal.
PROOF. Apply the proposition with S ={1].
2
The radical rad(a) of an ideal a is
{ [
i
a, some r N, r >0].
2 IDEALS 6
An ideal a is said to be radical if it equals its radical. Thus a is radical if and only if the
quotient ring ,a is reduced, i.e., without nonzero nilpotent elements (elements some power
of which is zero). Since integral domains are reduced, prime ideals (a fortiori maximal
ideals) are radical. The radical of (0) consists of the nilpotent elements of it is called
the nilradical of .
If b -b
t
under the one-to-one correspondence (2) between ideals of and ideals of
,a, then ,b .(,a),b
t
, and so b is prime (resp. maximal, radical) if and only if b
t
is
prime (resp. maximal, radical).
PROPOSITION 2.3. Let a be an ideal in a ring .
(a) The radical of a is an ideal.
(b) rad(rad(a)) =rad(a).
PROOF. (a) If rad(a), then clearly a rad(a) for all a . Suppose that a. b rad(a),
with say a
i
a and b
x
a. When we expand (a b)
ix
using the binomial theorem, we
nd that every term has a factor a
i
or b
x
, and so lies in a.
(b) If a
i
rad(a), then a
ix
=(a
i
)
x
a for some s >0, and so a rad(a).
2
Note that (b) of the proposition shows that rad(a) is radical. In fact, it is the smallest
radical ideal containing a.
If a and b are radical, then ab is radical, but ab need not be: consider, for example,
a =(X
2
Y ) and b =(X
2
Y ); they are both prime ideals in kX. Y | (by 4.10 below), but
ab =(X
2
. Y ), which contains X
2
but not X.
PROPOSITION 2.4. The radical of an ideal is equal to the intersection of the prime ideals
containing it. In particular, the nilradical of a ring is equal to the intersection of the prime
ideals of .
PROOF. If a =, then the set of prime ideals containing it is empty, and so the intersection
is . Thus we may suppose that a is a proper ideal of . Then rad(a)
_
pa
p because
prime ideals are radical and rad(a) is the smallest radical ideal containing a.
For the reverse inclusion, let rad(a). According to Proposition 2.1, there exists a
prime ideal containing a and disjoint from the multiplicative subset {1. . . . .]. Therefore

_
pa
p.
2
DEFINITION 2.5. The Jacobson radical J of a ring is the intersection of the maximal ideals
of the ring:
J() =
_
{m[ m maximal in ].
A ring is local if it has exactly one maximal ideal. For such a ring, the Jacobson
radical is m.
PROPOSITION 2.6. An element c of is in the Jacobson radical of if and only if 1ac
is a unit for all a .
PROOF. We prove the contrapositive: there exists a maximal ideal m such that c m if and
only if there exists an a such that 1ac is not a unit.
=: As 1ac is not a unit, it lies in some maximal ideal m of (by 2.2). Then c m,
because otherwise 1 =(1ac) ac m.
=: Suppose that c is not in the maximal ideal m. Then m(c) =, and so 1 =mac
for some m m and a . Now 1ac m, and so it is not a unit.
2
2 IDEALS 7
PROPOSITION 2.7 (PRIME AVOIDANCE). Let p
1
. . . . . p
i
, r _1, be ideals in with p
2
. . . . . p
i
prime. If an ideal a is not contained in any of the p
i
, then it is not contained in their union.
PROOF. When r = 1, there is nothing to prove, and so we may assume that r > 1 and
(inductively) that the statement is true for r 1. Then a is not contained in the union of the
ideals p
1
. . . . . p
i-1
. p
i1
. . . . . p
i
, and so there exists an a
i
a
_
}yi
p
}
. If some a
i
does
not lie in p
i
, then that a
i
a
_
1_i_i
p
i
, and the proof is complete. Thus suppose that
every a
i
p
i
, and consider
a =a
1
a
i-1
a
i
a.
I claim that a belongs to no p
i
. Because p
i
is prime and none of the elements a
1
. . . . . a
i-1
lies in p
i
, their product does not lie in p
i
; as a
i
p
i
, we see that a p
i
. Next consider a
prime p
i
with i _ r 1. In this case a
1
a
i-1
p
i
because the product involves a
i
, but
a
i
p
i
, and so again a p
i
. This completes the proof.
2
ASIDE 2.8. In general, the condition in (2.7) that the ideals p
2
. . . . . p
i
be prime is necessary: the
ideal (.. ,) in F
2
.. ,| is the union of three smaller nonprime ideals. However, when contains an
innite eld, the condition can be dropped (see mo108594, Mohan).
Extension and contraction of ideals
Let : T be a homomorphism of rings.
NOTATION 2.9. For an ideal b of T,
-1
(b) is an ideal in , called the contraction of b to
, which is often denoted b
c
. For an ideal a of , the ideal in T generated by (a) is called
the extension of a to T, and is often denoted a
e
. When is surjective, (a) is already an
ideal, and when is a subring of T, b
c
=b.
2.10. There are the following equalities (a. a
t
ideals in ; b. b
t
ideals in T):
(aa
t
)
e
=a
e
a
te
. (aa
t
)
e
=a
e
a
te
. (bb
t
)
c
=b
c
b
tc
. rad(b)
c
=rad(b
c
).
2.11. Obviously (i) a a
ec
and (ii) b
ce
b (a an ideal of ; b an ideal of T). On applying
e to (i), we nd that a
e
a
ece
, and (ii) with b replaced by a
e
shows that a
ece
a
e
; therefore
a
e
=a
ece
. Similarly, b
cec
=b
c
. It follows that extension and contraction dene inverse
bijections between the set of contracted ideals in and the set of extended ideals in T:
{b
c
[ b an ideal in T]
a|-a
e

b
c
-b
{a
e
T [ a an ideal in ]
Note that, for every ideal b in T, the map ,b
c
T,b is injective, and so b
c
is prime
(resp. radical) if b is prime (resp. radical).
The Chinese remainder theorem
Recall the classical form of the theorem: let J
1
. .... J
n
be integers, relatively prime in pairs;
then for any integers .
1
. .... .
n
, the congruences
. .
i
mod J
i
have a simultaneous solution . Z; moreover, if . is one solution, then the other solutions
are the integers of the form . mJ with m Z and J =

J
i
.
2 IDEALS 8
We want to translate this in terms of ideals. Integers m and n are relatively prime if and
only if (m. n) =Z, i.e., if and only if (m) (n) =Z. This suggests dening ideals a and b
in a ring to be relatively prime (or coprime) if ab =.
If m
1
. .... m
k
are integers, then
_
(m
i
) =(m) where m is the least common multiple of
the m
i
. Thus
_
(m
i
) (

m
i
), which equals

(m
i
). If the m
i
are relatively prime in pairs,
then m=

m
i
, and so we have
_
(m
i
) =

(m
i
). Note that in general,
a
1
a
2
a
n
a
1
a
2
... a
n
.
but the two ideals need not be equal.
These remarks suggest the following statement.
THEOREM 2.12 (CHINESE REMAINDER THEOREM). Let a
1
. . . . . a
n
be ideals in a ring .
If a
i
is relatively prime to a
}
whenever i = , then the map
a (. . . . aa
i
. . . .): ,a
1
,a
n
(3)
is surjective with kernel

a
i
(so

a
i
=
_
a
i
).
PROOF. Suppose rst that n =2. As a
1
a
2
=, there exist a
i
a
i
such that a
1
a
2
=1.
Then a
1
.
2
a
2
.
1
maps to (.
1
moda
1
. .
2
moda
2
), which shows that (3) is surjective.
For each i , there exist elements a
i
a
1
and b
i
a
i
such that
a
i
b
i
=1, all i _2.
The product

i_2
(a
i
b
i
) =1, and lies in a
1

i_2
a
i
, and so
a
1

i_2
a
i
=.
We can now apply the theorem in the case n =2 to obtain an element ,
1
of such that
,
1
1 mod a
1
. ,
1
0 mod

i_2
a
i
.
These conditions imply
,
1
1 mod a
1
. ,
1
0 mod a
}
, all >1.
Similarly, there exist elements ,
2
. .... ,
n
such that
,
i
1 mod a
i
. ,
i
0 mod a
}
for =i.
The element . =

.
i
,
i
maps to (.
1
moda
1
. . . . . .
n
moda
n
), which shows that (3) is
surjective.
The kernel of the map is
_
a
i
, and so it remains to prove that
_
a
i
=

a
i
. Obviously

a
i

_
a
i
. Suppose rst that n =2, and let a
1
a
2
=1, as before. For c a
1
a
2
, we
have
c =a
1
c a
2
c a
1
a
2
which proves that a
1
a
2
=a
1
a
2
. We complete the proof by induction. This allows us to
assume that

i_2
a
i
=
_
i_2
a
i
. We showed above that a
1
and

i_2
a
i
are relatively prime,
and so
a
1

i_2
a
i
_
=a
1

i_2
a
i
_
by the n =2 case. Now a
1

_
i_2
a
i
_
=

i_1
a
i
and a
1

_
i_2
a
i
_
=a
1

__
i_2
a
i
_
=
_
i_1
a
i
, which completes the proof.
2
3 NOETHERIAN RINGS 9
3 Noetherian rings
PROPOSITION 3.1. The following three conditions on a ring are equivalent:
(a) every ideal in is nitely generated;
(b) every ascending chain of ideals a
1
a
2
eventually becomes constant, i.e., for
some m, a
n
=a
n1
= .
(c) every nonempty set of ideals in has a maximal element.
PROOF. (a) =(b): If a
1
a
2
is an ascending chain, then a =
_
a
i
is an ideal, and
hence has a nite set {a
1
. . . . . a
n
] of generators. For some m, all the a
i
belong a
n
, and then
a
n
=a
n1
= =a.
(b) =(c): Let be a nonempty set of ideals in . If has no maximal element, then
the axiom of dependent choice
5
shows that there exists a strictly ascending sequence of
ideals in , which contradicts (b).
(c) =(a): Let a be an ideal, and let be the set of nitely generated ideals contained
in a. Then is nonempty because it contains the zero ideal, and so it contains a maximal
element c =(a
1
. . . . . a
i
). If c =a, then there exists an element a ac, and (a
1
. . . . . a
i
. a)
will be a nitely generated ideal in a properly containing c. This contradicts the denition of
c.
2
A ring is noetherian if it satises the equivalent conditions of the proposition. For
example, elds and principal ideal domains are noetherian. On applying (c) to the set of all
proper ideals containing a xed proper ideal, we see that every proper ideal in a noetherian
ring is contained in a maximal ideal. We saw in (3.6) that this is, in fact, true for every ring,
but the proof for non-noetherian rings requires Zorns lemma.
A quotient ,a of a noetherian ring is noetherian, because the ideals in ,a are all of
the form b,a with b an ideal in , and every set of generators for b generates b,a.
PROPOSITION 3.2. Let be a ring. The following conditions on an -module M are
equivalent:
(a) every submodule of M is nitely generated (in particular, M is nitely generated);
(b) every ascending chain of submodules M
1
M
2
eventually becomes constant.
(c) every nonempty set of submodules of M has a maximal element.
PROOF. Essentially the same as that of (3.1).
2
An -module M is noetherian if it satises the equivalent conditions of the proposition.
Let

denote regarded as a left -module. Then the submodules of

are exactly the
ideals in , and so

is noetherian (as an -module) if and only if is noetherian (as a
ring).
PROPOSITION 3.3. Let
0 M
t

M

M
tt
0
be an exact sequence of -modules.
5
This says: Let 1 be a binary relation on a nonempty set X, and suppose that, for each a in X, there exists
a b such that a1b; then there exists a sequence (a
n
)
nN
of elements of X such that a
n
1a
n1
for all n. It is
strictly stronger than the axiom of countable choice but weaker than the axiom of choice. See the Wikipedia
(axiom of dependent choice).
3 NOETHERIAN RINGS 10
(a) If N 1 are submodules of M such that (M
t
) N =(M
t
) 1 and (N) =
(1), then N =1.
(b) If M
t
and M
tt
are nitely generated, so also is M.
(c) M is noetherian if and only if M
t
and M
tt
are both noetherian.
PROOF. (a) Let . 1. The second condition implies that there exists a , N such
that (,) = (.). Now (. ,) = 0, and so . , M
t
1 = M
t
N. Thus . =
(. ,) , N.
(b) Let S
t
be a nite set of generators for M, and let S
tt
be a nite subset of M such
that S
tt
generates M
tt
. The submodule N of M generated by S
t
LS
tt
is such that
M
t
N =M
t
and N =M
tt
. Therefore (a) shows that N =M.
(c) =: An ascending chain of submodules of M
t
or of M
tt
gives rise to an ascending
chain in M, and therefore becomes constant.
=: Consider an ascending chain of submodules of M. As M
tt
is Noetherian, the image
of the chain in M
tt
becomes constant, and as M
t
is Noetherian, the intersection of the chain
with M
t
becomes constant. Now the (a) shows that the chain itself becomes constant.
2
For example, a direct sum
M =M
1
M
2
of -modules is noetherian if and only if M
1
and M
2
are both noetherian.
PROPOSITION 3.4. Every nitely generated module over a noetherian ring is noetherian.
PROOF. Let M be a module over a noetherian ring . If M is generated by a single element,
then M ~,a for some ideal a in , and the statement is obvious. We argue by induction on
the minimum number n of generators of M. Clearly M contains a submodule N generated
by n1 elements such that the quotient M,N is generated by a single element, and so the
statement follows from (3.3).
2
PROPOSITION 3.5. Every nitely generated module M over a noetherian ring contains a
nite chain of submodules M M
i
M
1
0 such that each quotient M
i
,M
i-1
is
isomorphic to ,p
i
for some prime ideal p
i
.
PROOF. The annihilator of an element . of M is
ann(.)
def
={a [ a. =0].
It is an ideal in , which is proper if . = 0. I claim that every ideal a that is maximal
among the annihilators of nonzero elements of is prime. Let a =ann(.), and let ab a,
so that ab. =0. Then a (a) a ann(b.). If b a, then b. =0, and so a = ann(b.)
by maximality, which implies that a a.
We now prove the proposition. Note that, for every . M, the submodule . of M is
isomorphic to ,ann(.). If M is nonzero, then there exists a nonzero . such that ann(.)
is maximal, and so M contains a submodule M
1
=. isomorphic to ,p
1
with p
1
prime.
Similarly, M,M
1
contains a submodule M
2
,M
1
isomorphic ,p
2
for some prime ideal p
2
,
and so on. The chain 0 M
1
M
2
terminates because M is noetherian (by 3.4).
2
ASIDE 3.6. The proofs of (2.1) and (3.5) are two of many in commutative algebra in which an ideal,
maximal with respect to some property, is shown to be prime. For a general examination of this
phenomenon, see Lam and Reyes, J. Algebra 319 (2008), no. 7, 30063027.
3 NOETHERIAN RINGS 11
THEOREM 3.7 (HILBERT BASIS THEOREM). Every nitely generated algebra over a noethe-
rian ring is noetherian.
PROOF. Let be noetherian. Since every nitely generated -algebra is a quotient of a
polynomial algebra, it sufces to prove the theorem for X
1
. . . . . X
n
|. Note that
X
1
. . . . . X
n
| =X
1
. . . . . X
n-1
|X
n
|. (4)
This simply says that every polynomial in n symbols X
1
. . . . . X
n
can be expressed uniquely
as a polynomial in X
n
with coefcients in kX
1
. . . . . X
n-1
|,
(X
1
. . . . . X
n
) =a
0
(X
1
. . . . . X
n-1
)X
i
n
a
i
(X
1
. . . . . X
n-1
).
Thus an induction argument shows that it sufces to prove the theorem for X|.
Recall that for a polynomial
(X) =c
0
X
i
c
1
X
i-1
c
i
. c
i
. c
0
=0.
c
0
is the leading coefcient of .
Let a be an ideal in X|, and let a(i ) be the set of elements of that occur as the
leading coefcient of a polynomial in a of degree i (we also include 0). Then a(i ) is
obviously an ideal in , and a(i 1) a(i ) because, if cX
i-1
a, then so also does
X(cX
i-1
).
Let b be an ideal of X| contained in a. Then b(i ) a(i ), and if equality holds for all
i , then b =a. Indeed, let be a polynomial of degree i in a. Because b(i ) =a(i ), there
exists a g b such that deg( g) < deg . On repeating this argument with g, we
eventually nd that b.
As is noetherian, the sequence of ideals
a(1) a(2) a(i )
eventually becomes constant, say, a(J) =a(J 1) =. . . (and then a(J) contains the lead-
ing coefcients of all polynomials in a). For each i _ J, choose a nite generating set
{c
i1
. c
i2
. . . .] for a(i ), and for each (i. ), choose a polynomial
i}
a of degree i with lead-
ing coefcient c
i}
. The ideal b generated by the
i}
is contained in a and has the property
that b(i ) =a(i ) for all i . Therefore b =a, and a is nitely generated.
2
COROLLARY 3.8. When 1 is noetherian, every nitely generated 1-algebra is nitely
presented.
PROOF. Obvious.
2
NAKAYAMAS LEMMA 3.9. Let be a ring, let a be an ideal in , and let M be an -
module. Assume that a is contained in all maximal ideals of and that M is nitely
generated.
(a) If M =aM, then M =0.
(b) If N is a submodule of M such that M =N aM, then M =N.
3 NOETHERIAN RINGS 12
PROOF. (a) Suppose that M =0. Choose a minimal set of generators {e
1
. . . . . e
n
] for M,
n _1, and write
e
1
=a
1
e
1
a
n
e
n
, a
i
a.
Then
(1a
1
)e
1
=a
2
e
2
a
n
e
n
and, as 1a
1
lies in no maximal ideal, it is a unit. Therefore e
2
. . . . . e
n
generate M, which
contradicts the minimality of the original set.
(b) The hypothesis implies that M,N =a(M,N), and so M,N =0.
2
Recall (2.5) that the Jacobson radical J of is the intersection of the maximal ideals of
, and so the condition on a is that a J. In particular, the lemma holds with a =J; for
example, when is a local ring, it holds with a the maximal ideal in .
COROLLARY 3.10. Let be a local ring with maximal ideal m and residue eld k
def
=,m,
and let M be a nitely generated module over . The action of on M,mM factors through
k, and elements a
1
. . . . . a
n
of M generate it as an -module if and only if the elements
a
1
mM. . . . . a
n
mM
span M,mM as k-vector space.
PROOF. If a
1
. . . . . a
n
generate M, then it is obvious that their images generate the vector
space M,mM. Conversely, suppose that a
1
mM. . . . . a
n
mM span M,mM, and let N
be the submodule of M generated by a
1
. . . . . a
n
. The composite N M M,mM is
surjective, and so M =N mM. Now Nakayamas lemma shows that M =N.
2
COROLLARY 3.11. Let be a noetherian local ring with maximal ideal m. Elements
a
1
. . . . . a
n
of m generate m as an ideal if and only if a
1
m
2
. . . . . a
n
m
2
span m,m
2
as a
vector space over ,m. In particular, the minimum number of generators for the maximal
ideal is equal to the dimension of the vector space m,m
2
.
PROOF. Because is noetherian, m is nitely generated, and we can apply the preceding
corollary with M =m.
2
EXAMPLE 3.12. Nakayamas lemma may fail if M is not nitely generated. For example,
let Z
(])
={
n
n
[ does not divide n] and consider the Z
(])
-module Q. Then Z
(])
is a local
ring with maximal ideal () (see 5 below) and Q=Q but Q=0.
DEFINITION 3.13. Let be a noetherian ring.
(a) The height ht(p) of a prime ideal p in is the greatest length J of a chain of distinct
prime ideals
p =p
d
p
d-1
p
0
. (5)
(b) The (Krull) dimension of is sup{ht(p) [ p . p prime].
Thus, the Krull dimension of a ring is the supremum of the lengths of chains of prime
ideals in (the length of a chain is the number of gaps, so the length of (5) is J). For
example, the integral domains of dimension 0 are the elds. The height of a nonzero prime
ideal in a principal ideal domain is 1, and so such a ring has Krull dimension 1 (provided it
3 NOETHERIAN RINGS 13
is not a eld). It is sometimes convenient to dene the Krull dimension of the zero ring to be
1.
We shall see in 19 that the height of every prime ideal in a noetherian ring is nite.
However, the Krull dimension of the ring may be innite, because it may contain a sequence
of prime ideals whose heights tend to innity (Krull 1938).
6
LEMMA 3.14. In a noetherian ring, every set of generators for an ideal contains a nite
generating set.
PROOF. Let S be a set of generators for a, and let a
t
be maximal among the ideals generated
by nite subsets of S. Then a
t
contains every element of S (otherwise it wouldnt be
maximal), and so equals a.
2
THEOREM 3.15 (KRULL INTERSECTION THEOREM). Let a be an ideal in a noetherian
ring . If a is contained in all maximal ideals of , then
_
n_1
a
n
={0].
PROOF. We shall show that, for every ideal a in a noetherian ring,
_
n_1
a
n
=a
_
n_1
a
n
. (6)
When a is contained in all maximal ideals of , Nakayamas lemma then shows that
_
n_1
a
n
is zero.
Let a
1
. . . . . a
i
generate a. Then a
n
consists of nite sums

i
1
i
r
=n
c
i
1
i
r
a
i
1
1
a
i
r
i
. c
i
1
i
r
.
In other words, a
n
consists of the elements of of the form g(a
1
. . . . . a
i
) for some homoge-
neous polynomial g(X
1
. . . . . X
i
) X
1
. . . . . X
i
| of degree n.
Let S
n
denote the set of homogeneous polynomials of degree msuch that (a
1
. . . . . a
i
)
_
n_1
a
n
, and let c be the ideal in X
1
. . . . . X
i
| generated by
_
n
S
n
. Because X
1
. . . . . X
i
|
is noetherian, c is nitely generated, and so c is generated by a nite set {
1
. . . . .
x
] of
elements of
_
n
S
n
. Let J
i
=deg
i
, and let J =maxJ
i
.
Let b
_
n_1
a
n
; then b a
d1
, and so b = (a
1
. . . . . a
i
) for some homogeneous
polynomial of degree J 1. By denition, S
d1
c, and so
=g
1

1
g
x

x
for some g
i
X
1
. . . . . X
n
|. As and the
i
are homogeneous, we can omit from each g
i
all terms not of degree deg deg
i
, since these terms cancel out. In other words, we may
choose the g
i
to be homogeneous of degree deg deg
i
=J 1J
i
>0. In particular,
the constant term of g
i
is zero, and so g
i
(a
1
. . . . . a
i
) a. Now
b =(a
1
. . . . . a
i
) =

i
g
i
(a
1
. . . . . a
i
)
i
(a
1
. . . . . a
i
) a
_
n
a
n
.
which completes the proof of (6).
2
6
In Nagata 1962, p.203, there is the following example. Let N = 1
0
L1
1
L. . . be a partition of N into
nite sets with strictly increasing cardinality. Let = kX
0
. X
1
. . . .| be the polynomial ring in a countably
innite number of symbols, and let p
i
be the prime ideal in generated by the X
}
s with in 1
i
. Let S be the
multiplicative set
_
p
i
. Then S
-1
is noetherian and regular, and the prime ideal S
-1
p
i
has height [1
i
[.
4 UNIQUE FACTORIZATION 14
The equality (6) can also be proved using primary decompositions see (17.15).
PROPOSITION 3.16. In a noetherian ring, every ideal contains a power of its radical; in
particular, some power of the nilradical of the ring is zero.
PROOF. Let a
1
. . . . . a
n
generate rad(a). For each i , some power of a
i
, say a
i
i
i
, lies in a.
Then every term of the expansion of
(c
1
a
1
c
n
a
n
)
i
1
i
n
. c
i
.
has a factor of the form a
i
i
i
for some i , and so lies in a.
2
ASIDE 3.17. In a noetherian ring, every ideal is nitely generated, but there is little that one can say
in general about the number of generators required. For example, in kX| every ideal is generated by
a single element, but in kX. Y | the ideal (X. Y )
n
requires at least n1 generators.
ASIDE 3.18. The following example shows that the Krull intersection theorem fails for nonnoethe-
rian rings. Let be the ring of germs of C
o
functions at 0 on the real line. Then is a local
ring with maximal ideal m equal to the set of germs zero at 0, and
_
n_1
m
n
consists of the germs
whose derivatives at zero are all zero. It therefore contains e
-1{x
2
. [Every germ of a function at
0 is represented by a function on an open neighbourhood U of 0; two pairs (. U) and (
t
. U
t
)
represent the same germ if and only if and
t
agree on some neighbourhood of 0 in U U
t
.]
4 Unique factorization
Let be an integral domain. An element a of is said to be irreducible if it is neither zero
nor a unit and admits only trivial factorizations, i.e.,
a =bc == b or c is a unit.
The element a is said to be prime if it is neither zero nor a unit and (a) is a prime ideal, i.e.,
a[bc == a[b or a[c.
An integral domain is called a unique factorization domain if every nonzero nonunit
a in can be written as a nite product of irreducible elements in exactly one way up to
units and the order of the factors. In more detail, the uniqueness means that if
a =

iJ
a
i
=

}J
b
}
with each a
i
and b
}
irreducible, then there exists a bijection i (i ): 1 J such that
b
}(i)
=a
i
unit for each i . Every principal ideal domain is a unique factorization domain
(proved in most algebra courses).
PROPOSITION 4.1. Let be an integral domain, and let a be an element of that is neither
zero nor a unit. If a is prime, then a is irreducible, and the converse holds when is a
unique factorization domain.
4 UNIQUE FACTORIZATION 15
PROOF. Assume that a is prime. If a = bc, then a divides bc and so a divides b or c.
Suppose the rst, and write b =aq. Now a =bc =aqc, which implies that qc =1 because
is an integral domain, and so c is a unit. We have shown that a is irreducible.
For the converse, assume that a is irreducible and that is a unique factorization domain.
If a[bc, then bc = aq for some q . On writing each of b, c, and q as a product of
irreducible elements, and using the uniqueness of factorizations, we see that a differs from
one of the irreducible factors of b or c by a unit. Therefore a divides b or c.
2
PROPOSITION 4.2. Let be an integral domain in which every nonzero nonunit element is
a nite product of irreducible elements. If every irreducible element of is prime, then is
a unique factorization domain.
PROOF. Suppose that
a
1
a
n
=b
1
b
n
(7)
with the a
i
and b
i
irreducible elements in . As a
1
is prime, it divides one of the b
i
, which
we may suppose to be b
1
, say b
1
=a
1
u. As b
1
is irreducible, u is a unit. On cancelling a
1
from both sides of (7), we obtain the equality
a
2
a
n
=(ub
2
)b
3
b
n
.
Continuing in this fashion, we nd that the two factorizations are the same up to units and
the order of the factors.
2
PROPOSITION 4.3. Let be an integral domain in which every ascending chain of principal
ideals becomes constant (e.g., a noetherian integral domain). Then every every nonzero
nonunit element in is a nite product of irreducible elements.
PROOF. The hypothesis implies that every nonempty set of principal ideals has a maximal
element (cf. the proof of 3.1). Assume that has nonfactorable elements, and let (a) be
maximal among the ideals generated by such elements. Then a is not itself irreducible, and
so a =bc with neither b nor c units. Now (b) and (c) both properly contain (a), and so b
and c are both factorable, which contradicts the nonfactorability of a.
2
PROPOSITION 4.4. Let be a unique factorization domain with eld of fractions J. If an
element of X| factors into the product of two nonconstant polynomials in JX|, then it
factors into the product of two nonconstant polynomials in X|.
In other words, if is not the product of two nonconstant polynomials in X|, then it
is irreducible in JX|.
PROOF. Let =gh in JX|. For suitable c. J , the polynomials g
1
=cg and h
1
=Jh
have coefcients in , and so we have a factorization
cJ =g
1
h
1
in X|.
If an irreducible element of divides cJ, then, looking modulo (), we see that
0 =g
1
h
1
in (,())X|.
4 UNIQUE FACTORIZATION 16
According to Proposition 4.1, the ideal () is prime, and so (,())X| is an integral
domain. Therefore, divides all the coefcients of at least one of the polynomials g
1
. h
1
,
say g
1
, so that g
1
=g
2
for some g
2
X|. Thus, we have a factorization
(cJ,) =g
2
h
1
in X|.
Continuing in this fashion, we can remove all the irreducible factors of cJ, and so obtain a
factorization of in X|.
2
The proof shows that every factorization = gh in JX| of an element of X|
gives a factorization =(cg)(c
-1
h) in X| for a suitable c J.
Let be a unique factorization domain. A nonzero polynomial
=a
0
a
1
X a
n
X
n
in X| is said to be primitive if the coefcients a
i
have no common factor other than units.
Every polynomial in JX| can be written =c( )
1
with c( ) J and
1
primitive.
The element c( ), which is well-dened up to multiplication by a unit, is called the content
of . Note that X| if and only if c( ) .
PROPOSITION 4.5. The product of two primitive polynomials is primitive.
PROOF. Let
=a
0
a
1
X a
n
X
n
g =b
0
b
1
X b
n
X
n
.
be primitive polynomials, and let be an irreducible element of . Let a
i
0
be the rst
coefcient of not divisible by and b
}
0
the rst coefcient of g not divisible by . Then
all the terms in

i}=i
0
}
0
a
i
b
}
are divisible by , except a
i
0
b
}
0
, which is not divisible
by . Therefore, doesnt divide the (i
0

0
)th-coefcient of g. We have shown that
no irreducible element of divides all the coefcients of g, which must therefore be
primitive.
2
Each of the last two propositions is referred to as Gausss lemma (Gauss proved them
with =Z).
PROPOSITION 4.6. Let be a unique factorization domain with eld of fractions J. For
polynomials . g JX|, c(g) =c( ) c(g); hence every factor in X| of a primitive
polynomial is primitive.
PROOF. Let =c( )
1
and g =c(g)g
1
with
1
and g
1
primitive. Then
g =c( )c(g)
1
g
1
with
1
g
1
primitive, and so c(g) =c( )c(g).
2
COROLLARY 4.7. The irreducible elements in X| are the irreducible elements a of
and the nonconstant primitive polynomials such that is irreducible in JX|.
PROOF. Obvious from (4.4) and (4.6).
2
5 RINGS OF FRACTIONS 17
THEOREM 4.8. If is a unique factorization domain, then so also is X|.
PROOF. Let X|, and write = c( )
1
. Then c( ) is a product of irreducible
elements in . If
1
is not irreducible, then it can be written as a product of two polynomials
of lower degree, which are necessarily primitive (4.6). Continuing in this fashion, we nd
that
1
is a product of irreducible primitive polynomials, and hence that is a product of
irreducible elements in X|.
According to Proposition 4.2, in order prove that X| is a unique factorization domain,
it remains to show that each irreducible element of X| is prime.
Let a be an irreducible element of . If a divides the product gh of g. h X|, then it
divides c(gh) =c(g)c(h). As a is prime, it divides c(g) or c(h), and hence also g or h.
Let be a nonconstant primitive polynomial in X| such that is irreducible in JX|.
If divides the product gh of g. h X|, then it divides g or h in JX|. Suppose the
rst, and write q =g with q JX|. Then c(q) =c( )c(q) =c(q) =c(g) , and
so q X|. Therefore divides g in X|.
2
Let k be a eld. A monomial in X
1
. . . . . X
n
is an expression of the form
X
o
1
1
X
o
n
n
. a
}
N.
The total degree of the monomial is

a
i
. The degree, deg( ), of a nonzero polyno-
mial (X
1
. . . . . X
n
) is the largest total degree of a monomial occurring in with nonzero
coefcient. Since
deg(g) =deg( ) deg(g).
kX
1
. . . . . X
n
| is an integral domain and kX
1
. . . . . X
n
|

=k

. Therefore, an element of
kX
1
. . . . . X
n
| is irreducible if it is nonconstant and =gh == g or h is constant.
THEOREM 4.9. The ring kX
1
. . . . . X
n
| is a unique factorization domain.
PROOF. This is trivially true when n =0, and an induction argument using (4), p.11, proves
it for all n.
2
COROLLARY 4.10. A nonzero proper principal ideal ( ) in kX
1
. . . . . X
n
| is prime if and
only is irreducible.
PROOF. Special case of (4.1).
2
5 Rings of fractions
Recall that a multiplicative subset of a ring is a nonempty subset closed under the formation
of nite products. In particular, it contains 1 (the empty product).
Let S be a multiplicative subset of a ring . Dene an equivalence relation on S by
(a. s) ~(b. t ) u(at bs) =0 for some u S.
Write
o
x
for the equivalence class containing (a. s), and dene addition and multiplication of
equivalence classes according to the rules:
o
x

b
t
=
otbx
xt
.
o
x
b
t
=
ob
xt
.
5 RINGS OF FRACTIONS 18
It is easily checked these do not depend on the choices of representatives for the equivalence
classes, and that we obtain in this way a ring
S
-1
={
o
x
[ a . s S]
and a ring homomorphism a
o
1
:
i
S
S
-1
whose kernel is
{a [ sa =0 for some s S].
If S contains no zero-divisors, for example, if is an integral domain and 0 S, then . At
the opposite extreme, if 0 S, then S
-1
is the zero ring.
A homomorphism T factors through
i
S
S
-1
if and only if the image of S in
T consists of units. More formally:
PROPOSITION 5.1. The pair (S
-1
. i
S
) has the following universal property:
every element of S maps to a unit in S
-1
, and
any other ring homomorphism : T with this
property factors uniquely through i
S
S
-1

T.
i
S

3
PROOF. Let : T be such a homomorphism, and let : S
-1
T be a homomorphism
such that i
S
=. Then
x
1
o
x
=
o
1
== (
x
1
)(
o
x
) =(
o
1
) == (s)(
o
x
) =(a).
and so
(
o
x
) =(a)(s)
-1
. (8)
This shows that there can be at most one such that i
S
=. We dene by the formula
(8). Then
o
x
=
b
t
== u(at bs) =0 some u S
(u)B

== (a)(t ) (b)(s) =0.

which shows that is well-dened, and it is easy to check that it is a homomorphism.
2
As usual, this universal property determines the pair (S
-1
. i
S
) uniquely up to a unique
isomorphism.
7
When is an integral domain and S ={0], the ring S
-1
is the eld of fractions
J of . In this case, for any other multiplicative subset T of not containing 0, the ring
T
-1
can be identied with the subring of J consisting of the fractions
o
t
with a and
t T .
7
Recall the proof: let (
1
. i
1
) and (
2
. i
2
) have the universal property in the proposition; because every
element of S maps to a unit in
2
, there exists a unique homomorphism :
1

2
such that i
1
= i
2
(universal property of
1
. i
1
); similarly, there exists a unique homomorphism
t
:
2

1
such that
t
i
2
=i
1
;
now

t
i
1
=
t
i
2
=i
1
=id

1
i
1
.
and so
t
= id

1
(universal property of
1
. i
1
); similarly,
t
= id

2
, and so and
t
are inverse
isomorphisms (and they are uniquely determined by the conditions i
1
=i
2
and
t
i
2
=i
1
).
5 RINGS OF FRACTIONS 19
EXAMPLE 5.2. Let h . Then S
h
={1. h. h
2
. . . .] is a multiplicative subset of , and we
let
h
=S
-1
h
. Thus every element of
h
can be written in the form a,h
n
, a , and
o
h
m
=
b
h
n
h
1
(ah
n
bh
n
) =0. some N.
If h is nilpotent, then
h
=0, and if is an integral domain with eld of fractions J and
h =0, then
h
is the subring of J of elements that can be written in the form a,h
n
, a ,
m N.
PROPOSITION 5.3. For every ring and h , the map

a
i
X
i

o
i
h
i
denes an
isomorphism
X|,(1hX)
h
.
PROOF. If h =0, both rings are zero, and so we may assume h =0. In the ring
.|
def
=X|,(1hX).
1 =h., and so h is a unit. Let : T be a homomorphism of rings such that (h) is a
unit in T. The homomorphism

i
a
i
X
i

i
(a
i
)(h)
-i
: X| T
factors through .| because 1hX 1(h)(h)
-1
=0, and this is the unique extension
of to .|. Therefore .| has the same universal property as
h
, and so the two are
(uniquely) isomorphic by an -algebra isomorphism that makes h
-1
correspond to ..
2
Let S be a multiplicative subset of a ring , and let S
-1
be the corresponding ring of
fractions. For every ideal a in , the ideal generated by the image of a in S
-1
is
S
-1
a ={
o
x
[ a a. s S].
If a contains an element of S, then S
-1
a contains 1, and so is the whole ring. Thus some of
the ideal structure of is lost in the passage to S
-1
, but, as the next proposition shows,
some is retained.
PROPOSITION 5.4. Let S be a multiplicative subset of the ring , and consider extension
a a
e
= S
-1
a and contraction a a
c
= {a [
o
1
a] of ideals with respect to the
homomorphism i
S
: S
-1
. Then
a
ce
=a for all ideals of S
-1

a
ec
=a if a is a prime ideal of disjoint from S.
Moreover, the map p p
e
is a bijection from the set of prime ideals of disjoint from S
onto the set of all prime ideals of S
-1
; the inverse map is p p
c
.
PROOF. Let a be an ideal in S
-1
. Certainly a
ce
a. For the reverse inclusion, let b a.
We can write b =
o
x
with a , s S. Then
o
1
=s(
o
x
) a, and so a a
c
. Thus b =
o
x
a
ce
,
and so a a
ce
.
Let p be a prime ideal of disjoint from S. Clearly p
ec
p. For the reverse inclusion,
let a p
ec
so that
o
1
=
o
0
x
for some a
t
p, s S. Then t (as a
t
) =0 for some t S, and
so ast p. Because st p and p is prime, this implies that a p, and so p
ec
p.
5 RINGS OF FRACTIONS 20
Let p be a prime ideal of disjoint from S, and let

S be the image of S in ,p.
Then (S
-1
),p
e
.

S
-1
(,p) because S
-1
,p
e
has the correct universal property, and

S
-1
(,p) is an integral domain because ,p is an integral domain and

S doesnt contain
0. Therefore p
e
is prime. From (2.11) we know that p
c
is prime if p is, and so p p
e
and
p p
c
are inverse bijections on the two sets.
2
COROLLARY 5.5. If is noetherian, then so also is S
-1
for any multiplicative set S.
PROOF. As b
c
is nitely generated, so also is (b
c
)
e
=b.
2
PROPOSITION 5.6. Let : T be a ring homomorphism. A prime ideal p of is the
contraction of a prime ideal in T if and only if p =p
ec
.
PROOF. If p =q
c
, then p
ec
=q
cec
2.11
= q
c
=p. Conversely, suppose that p =p
ec
, and let
S = p. Let s S; if (s) p
e
, then s p
ec
= p, contradicting the denition of S.
Therefore (S) is a multiplicative subset of T disjoint from p
e
, and so there exists a prime
ideal q in T containing p
e
and disjoint from (S) (apply 2.1). Now
-1
(q) contains p and
is disjoint from S, and so it equals p.
2
EXAMPLE 5.7. Let p be a prime ideal in . Then S
p
=p is a multiplicative subset of
, and we let
p
=S
-1
p
. Thus each element of
p
can be written in the form
o
c
, c p, and
o
c
=
b
d
s(aJ bc) =0, some s p.
According to (5.4), the prime ideals of
p
correspond to the prime ideals of disjoint from
p, i.e., contained in p. Therefore,
p
is a local ring with maximal ideal m=p
e
={
o
x
[
a p. s p].
PROPOSITION 5.8. Let m be a maximal ideal of a ring , and let n =m
m
be the maximal
ideal of
m
. For all n, the map
am
n
an
n
: ,m
n

m
,n
n
is an isomorphism. Moreover, it induces isomorphisms
m
i
,m
n
n
i
,n
n
for all pairs (r. n) with r _n.
PROOF. The second statement follows from the rst, because of the exact commutative
diagram (r <n):
0 m
i
,m
n
,m
n
,m
i
0
0 n
i
,n
n

m
,n
n

m
,n
i
0.
: :
We consider extension and contraction with respect to a
o
1
:
m
. In order to
show that the map ,m
n

m
,n
n
is injective, we have to show that (m
n
)
ec
= m
n
. If
a (m
n
)
ec
, then
o
1
=
b
x
with b m
n
and s S. Then s
t
sa m
n
for some s
t
S, and so
5 RINGS OF FRACTIONS 21
s
t
sa =0 in ,m
n
. The only maximal ideal containing m
n
is m, and so the only maximal
ideal in ,m
n
is m,m
n
. As s
t
s is not in m,m
n
, it must be a unit in ,m
n
, and so a =0 in
,m
n
, i.e., a m
n
. We have shown that (m
n
)
ec
m
n
, and the reverse inclusion is always
true.
We now prove that ,m
n

m
,n
n
is surjective. Let
o
x

m
, a , s m. The
only maximal ideal of containing m
n
is m, and so no maximal ideal contains both s and
m
n
. Therefore (s) m
n
=, and so sb q =1 for some b and q m
n
. Hence
s(ba) =a(1q). (9)
Because s is invertible in
m
,n
n
,
o
x
is the unique element of this ring such that s
o
x
=a. But
(9) shows that the image of ba in
m
also has this property and therefore equals
o
x
.
2
PROPOSITION 5.9. In a noetherian ring, only 0 lies in all powers of all maximal ideals.
PROOF. Let a be an element of a noetherian ring . If a =0, then its annihilator {b [ ba =0]
is a proper ideal in , and so it is contained in some maximal ideal m. Then
o
1
is nonzero in

m
, and so
o
1
(m
m
)
n
for some n (by the Krull intersection theorem 3.15), which implies
that a m
n
(by 5.8).
2
Modules of fractions
Let S be a multiplicative subset of the ring , and let M be an -module. Dene an
equivalence relation on MS by
(m. s) ~(n. t ) u(t msn) =0 for some u S.
Write
n
x
for the equivalence class containing (m. s), and dene addition and scalar multipli-
cation by the rules:
n
x

n
t
=
ntnx
xt
.
o
x
n
t
=
on
xt
. m. n M. s. t S. a .
It is easily checked these do not depend on the choices of representatives for the equivalence
classes, and that we obtain in this way an S
-1
-module
S
-1
M ={
n
x
[ m M. s S]
and a homomorphism m
n
1
: M
i
S
S
-1
M of -modules whose kernel is
{a M [ sa =0 for some s S].
A homomorphism M N of -modules factors through M S
-1
M if and only if
every element of S acts invertibly on N. More formally:
PROPOSITION 5.10. The pair (S
-1
M. i
S
) has the following universal property:
every element of S acts invertibly on S
-1
M, and
any homomorphism M N of -modules such
that every element of S acts invertibly on N factors
uniquely through i
S
M S
-1
M
N.
i
S
3
5 RINGS OF FRACTIONS 22
PROOF. Similar to that of Proposition 5.1.
2
In particular, for any homomorphism : M N of -modules, there is a unique
homomorphism S
-1
: S
-1
M S
-1
N such that S
-1
i
S
=i
S
:
M S
-1
M
N S
-1
N.
i
S
S
1

i
S
In this way, M S
-1
M becomes a functor from -modules to S
-1
-modules.
PROPOSITION 5.11. The functor M S
-1
M is exact. In other words, if the sequence of
-modules
M
t

M

M
tt
is exact, then so also is the sequence of S
-1
-modules
S
-1
M
t
S
1

S
-1
M
S
1

S
-1
M
tt
.
PROOF. Because =0, we have 0 =S
-1
() =S
-1
S
-1
. Therefore Im(S
-1
)
Ker(S
-1
). For the reverse inclusion, let
n
x
Ker(S
-1
) where m M and s S. Then
(n)
x
=0 and so, for some t S, we have t(m) =0. Then (t m) =0, and so t m=(m
t
)
for some m
t
M
t
. Now
n
x
=
tn
tx
=
(n
0
)
tx
Im(S
-1
).
2
PROPOSITION 5.12. Let M be an -module. The canonical map
M

{M
m
[ m a maximal ideal in ]
is injective.
PROOF. Let m M map to zero in all M
m
. The annilator a ={a [ am =0] of m is
an ideal in . Because m maps to zero M
m
, there exists an s m such that sm =0.
Therefore a is not contained in m. Since this is true for all maximal ideals m, a = (by 2.2),
and so it contains 1. Now m=1m=0.
2
COROLLARY 5.13. The -module M =0 if M
m
=0 for all maximal ideals m.
PROOF. Immediate consequence of the lemma.
2
PROPOSITION 5.14. A sequence
M
t

M

M
tt
(10)
is exact if and only if
M
t
m

m
M
m

m
M
tt
m
(11)
is exact for all maximal ideals m.
6 INTEGRALITY 23
PROOF. The necessity is a special case of (5.11). For the sufciency, let N =Ker(),Im().
Because the functor M M
m
is exact,
N
m
=Ker(
m
),Im(
m
).
If (11) is exact for all m, then N
m
=0 for all m, and so N =0 (by 5.13). But this means
that (10) is exact.
2
COROLLARY 5.15. A homomorphism M N of -modules is injective (resp. surjective)
if and only if M
m
N
m
is injective (resp. surjective) for all maximal ideals m.
PROOF. Apply the proposition to 0 M N (resp. M N 0).
2
EXAMPLE 5.16. Let M be an -module. For h , let M
h
= S
-1
h
M where S
h
=
{1. h. h
2
. . . .]. Then every element of M
h
can be written in the form
n
h
r
, m M, r N, and
n
h
r
=
n
0
h
r
0
if and only if h
1
(h
i
0
mh
i
m
t
) =0 for some N N.
EXERCISE 5.17. A multiplicative subset S of a ring is said to be saturated if
ab S == a and b S.
(a) Show that the saturated multiplicative subsets of are exactly the subsets S such that
S is a union of prime ideals.
(b) Let S be a multiplicative subset of , and let

S be the set of a such that ab S
for some b . Show that

S is a saturated multiplicative subset of (hence it is the
smallest such subset containing S), and that

S is the union of the prime ideals
of not meeting S. Show that for any -module M, the canonical homomorphism
S
-1
M

S
-1
M is bijective. In particular, S
-1
.

S
-1
. (Cf. Bourbaki AC, II 2,
Exercises 1,2.)
6 Integrality
Let be a subring of a ring T. An element of T is said to be integral over if it is a root
of a monic
8
polynomial with coefcients in , i.e., if it satises an equation

n
a
1

n-1
a
n
=0. a
i
.
If every element of T is integral over , then T is said to be integral over .
In the next proof, we shall need to apply Cramers rule. As usually stated in linear
algebra courses, this says that, if .
1
. . . . . .
n
is a solution to the system of linear equations
n

}=1
c
i}
.
}
=J
i
. i =1. . . . . m.
then
.
}
=
det(C
}
)
det(C)
. (12)
8
A polynomial is monic if its leading coefcient is 1, i.e., (X) =X
n
terms of degree less than n.
6 INTEGRALITY 24
where C =(c
i}
) and
C
}
=
_
_
_
c
11
c
1,}-1
J
1
c
1,}1
c
1n
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
c
n1
c
n,}-1
J
n
c
n,}1
c
nn
_
_
_
.
When one rewrites (12) in the form
det(C) .
}
=det(C
}
),
this statement becomes true over any ring (whether or not det(C) is a unit). The proof is
elementary expand out the right hand side of
det C
}
=det
_
_
_
c
11
. . . c
1}-1

c
1}
.
}
c
1}1
. . . c
1n
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
c
n1
. . . c
n}-1

c
n}
.
}
c
n}1
. . . c
nn
_
_
_
using standard properties of determinants.
PROPOSITION 6.1. Let be a subring of a ring T. An element of T is integral over if
and only if there exists a faithful
9
|-submodule M of T that is nitely generated as an
-module.
PROOF. =: Suppose that

n
a
1

n-1
a
n
=0. a
i
.
Then the -submodule M of T generated by 1, , ...,
n-1
has the property that M M,
and it is faithful because it contains 1.
=: Let M be an -module in T with a nite set {e
1
. . . . . e
n
] of generators such that
M M and M is faithful as an |-module. Then, for each i ,
e
i
=

a
i}
e
}
, some a
i}
.
We can rewrite this system of equations as
( a
11
)e
1
a
12
e
2
a
13
e
3
=0
a
21
e
1
( a
22
)e
2
a
23
e
3
=0
=0.
Let C be the matrix of coefcients on the left-hand side. Then Cramers formula tells us that
det(C) e
i
=0 for all i . As M is faithful and the e
i
generate M, this implies that det(C) =0.
On expanding out the determinant, we obtain an equation

n
c
1

n-1
c
2

n-2
c
n
=0. c
i
.
2
PROPOSITION 6.2. An -algebra T is nite if it is generated as an -algebra by a nite
number of elements, each of which is integral over .
9
An -module M is faithful if aM =0, a , implies a =0.
6 INTEGRALITY 25
PROOF. Suppose that T =
1
. . . . .
n
| and that

n
i
i
a
i1

n
i
-1
i
a
i n
i
=0. a
i}
. i =1. . . . . m.
Any monomial in the
i
s divisible by some
n
i
i
is equal (in T) to a linear combination of
monomials of lower degree. Therefore, T is generated as an -module by the monomials

i
1
1

i
m
n
, 1 _r
i
<n
i
.
2
COROLLARY 6.3. An -algebra T is nite if and only if it is nitely generated and integral
over .
PROOF. =: Immediate consequence of the proposition.
=: As an -module, T is faithful (because a 1
B
=a), and so (6.1) shows that every
element of T is integral over . As T is nitely generated as an -module, it is certainly
nitely generated as an -algebra.
2
The proof shows that, if an -algebra T is generated by a nite number of elements each
of which is integral over , then it is nitely generated as an -module.
THEOREM 6.4. Let be a subring of a ring T. The elements of T integral over form a
subring of T.
PROOF. Let and be two elements of T integral over . As just noted, . | is nitely
generated as an -module. It is stable under multiplication by and and it is faithful
as an |-module and as an |-module (because it contains 1

). Therefore (6.1)
shows that and are integral over .
2
DEFINITION 6.5. Let be a subring of the ring T. The integral closure of in T is the
subring of T consisting of the elements integral over .
PROPOSITION 6.6. Let be an integral domain with eld of fractions J, and let 1 be a
eld containing J. If 1 is algebraic over J, then there exists a J such that J is
integral over .
PROOF. By assumption, satises an equation

n
a
1

n-1
a
n
=0. a
i
J.
Let J be a common denominator for the a
i
, so that Ja
i
for all i , and multiply through
the equation by J
n
:
J
n

n
a
1
J
n

n-1
a
n
J
n
=0.
We can rewrite this as
(J)
n
a
1
J(J)
n-1
a
n
J
n
=0.
As a
1
J. . . . . a
n
J
n
, this shows that J is integral over .
2
COROLLARY 6.7. Let be an integral domain and let 1 be an algebraic extension of the
eld of fractions of . Then 1 is the eld of fractions of the integral closure of in 1.
6 INTEGRALITY 26
PROOF. In fact, the proposition shows that every element of 1 is a quotient ,J with
integral over and J .
2
DEFINITION 6.8. An integral domain is is said to be integrally closed or normal if it is
equal to its integral closure in its eld of fractions J, i.e., if
J. integral over == .
PROPOSITION 6.9. Every unique factorization domain is integrally closed.
PROOF. An element of the eld of fractions of not in can be written a,b with a. b
and b divisible by some irreducible element not dividing a. If a,b is integral over , then
it satises an equation
(a,b)
n
a
1
(a,b)
n-1
a
n
=0. a
i
.
On multiplying through by b
n
, we obtain the equation
a
n
a
1
a
n-1
b a
n
b
n
=0.
The element then divides every term on the left except a
n
, and hence must divide a
n
.
Since it doesnt divide a, this is a contradiction.
2
Let J 1 be elds, and let 1 be algebraic over J. The minimum polynomial
of over J is the unique element of smallest degree in the set of monic polynomials in
JX| having as a root. If is the minimum polynomial of , then the homomorphism
X : JX| J| denes an isomorphism JX|,( ) J|, i.e., J.| . J|,
. -.
PROPOSITION 6.10. Let be a normal integral domain, and let 1 be a nite extension of
the eld of fractions J of . An element of 1 is integral over if and only if its minimum
polynomial over J has coefcients in .
PROOF. Let be integral over , so that

n
a
1

n-1
a
n
=0. some a
i
. m> 0.
Let
t
be a conjugate of , i.e., a root of the minimum polynomial (X) of over J in
some eld containing 1. Then there is an J-isomorphism (see above)
o: J| J
t
|. o() =
t
On applying o to the above equation we obtain the equation

tn
a
1

tn-1
a
n
=0.
which shows that
t
is integral over . As the coefcients of are polynomials in the
conjugates of , it follows from (6.4) that the coefcients of (X) are integral over . They
lie in J, and is integrally closed, and so they lie in . This proves the only if part of the
statement, and the if part is obvious.
2
6 INTEGRALITY 27
COROLLARY 6.11. Let be a normal integral domain with eld of fractions J, and let
(X) be a monic polynomial in X|. Then every monic factor of (X) in JX| has
coefcients in .
PROOF. It sufces to prove this for an irreducible monic factor g of in JX|. Let be
a root of g in some extension eld of J. Then g is the minimum polynomial of , which,
being also a root of , is integral over . Therefore g has coefcients in .
2
PROPOSITION 6.12. Let T be rings, and let
t
be the integral closure of in T. For
any multiplicative subset S of , S
-1

t
is the integral closure of S
-1
in S
-1
T.
PROOF. Let b,s S
-1

t
with b
t
and s S. Then
b
n
a
1
b
n-1
a
n
=0
for some a
i
, and so
_
b
s
_
n

a
1
s
_
b
s
_
n-1

a
n
s
n
=0.
Therefore b,s is integral over S
-1
. This shows that S
-1

t
is contained in the integral
closure of S
-1
T.
For the converse, let b,s be integral over S
-1
with b T and s S. Then
_
b
s
_
n

a
1
s
1
_
b
s
_
n-1

a
n
s
n
=0.
for some a
i
and s
i
S. On multiplying this equation by s
n
s
1
s
n
, we nd that
s
1
s
n
b
t
, and therefore that b,s =s
1
s
n
b,ss
1
s
n
S
-1

t
.
2
COROLLARY 6.13. Let T be rings, and let S be a multiplicative subset of . If is
integrally closed in T, then S
-1
is integrally closed in S
-1
T.
PROOF. Special case of the proposition in which
t
=.
2
PROPOSITION 6.14. The following conditions on an integral domain are equivalent:
(a) is an integral domain;
(b)
p
is integrally closed for all prime ideals p;
(c)
m
is integrally closed for all maximal ideals m.
PROOF. The implication (a)=(b) follows from (6.13), and (b)=(c) is obvious. For (c)=(a),
let
t
be the integral closure of in its eld of fractions J. Then (
t
)
m
is the integral
closure of
m
in J (by 6.12). If (c) holds, then
n
(
t
)
m
is surjective for all maximal
ideals m, which implies that
t
is surjective (by 5.15), and so is integrally closed.
2
PROPOSITION 6.15. If is a normal integral domain, so also is the polynomial ring X|.
PROOF. Omitted for the present.
2
6 INTEGRALITY 28
The going-up theorem
PROPOSITION 6.16. Let T be integral domains, with T integral over . Then T is a
eld if and only if is a eld.
PROOF. Suppose that is a eld, and let b be a nonzero element of T. Then
b
n
a
1
b
n-1
a
n
=0
for some a
i
, and we may suppose that n is the minimum degree of such a relation. Then,
as T is an integral domain, a
n
=0, and the equation
b (b
n-1
a
1
b
n-2
a
n-1
)a
-1
n
=1
shows that b has an inverse in T.
Conversely, suppose that T is a eld, and let a be a nonzero element of . Then a has
an inverse a
-1
in T, and
a
-n
a
1
a
-(n-1)
a
n
=0
for some a
i
. On multiplying through by a
n-1
, we nd that
a
-1
a
1
a
2
a a
n
a
n-1
=0,
and so
a
-1
=(a
1
a
2
a a
n
a
n-1
) .
2
COROLLARY 6.17. Let T be rings with T integral over . Let q be a prime ideal of
T, and let p =q. Then q is maximal if and only if p is maximal.
PROOF. Apply the proposition to ,p T,q.
2
COROLLARY 6.18 (INCOMPARABILITY). Let T be rings with T integral over , and
let q q
t
be prime ideals of T. If q =q
t
, then q =q
t
.
PROOF. Let p = q = q
t
. Then
p
T
p
, and T
p
is integral over
p
. The ideals
qT
p
q
t
T
p
are both prime ideals of T
p
lying over p
p
, which is maximal, and so qT
p
=q
t
T
p
(by 6.17). Now
q
5.4
= (qT
p
)
c
=
_
q
t
T
p
_
c 5.4
= q
t
.
2
THEOREM 6.19. Let T be rings with T integral over , and let p be a prime ideal of
. Then there exists a prime ideal q of T such that p =q.
PROOF. We have
p
T
p
, and T
p
is integral over
p
. Let n be a maximal ideal in T
p
(which exists by 2.2). Then n
p
is maximal (6.17). But p
p
is the unique maximal ideal
of
p
, and so n
p
=p
p
. Let q be the inverse image of n in T. Then q is the inverse
image of p
p
in , because the diagram
T T
p

p
commutes. But the inverse image of p
p
in is p (as p
ec
= p; see 5.4). Therefore
q =p.
2
6 INTEGRALITY 29
COROLLARY 6.20. Let T be rings with T integral over . Let p p
t
be prime ideals
of , and let q be a prime ideal of T such that q =p. Then there exists a prime ideal q
t
of T containing q and such that q
t
=p
t
.
PROOF. We have ,p T,q, and T,q is integral over ,p. According to the theorem,
there exists a prime ideal q
tt
in T,q such that q
tt
(,p) =p
t
,p. The inverse image q
t
of
q
tt
in T has the required properties.
2
COROLLARY 6.21. Let T be rings with T integral over , and let p
1
p
n
be
prime ideals in . Let
q
1
q
n
(m<n) (13)
be prime ideals in T such that q
i
=p
i
for all i _ m. Then (13) can be extended to a
chain of prime ideals
q
1
q
n
such that q
i
=p
i
for all i _n.
PROOF. Immediate consequence of Corollary 6.20.
2
Theorem 6.19 and its corollaries are referred to as the going-up theorem (of Cohen and
Seidenberg).
The going-down theorem
Before proving the going-down theorem, we need to extend some of the denitions and
results from earlier in this section.
Let T be rings, and let a be an ideal of . An element b of T is said to be integral
over a if it satises an equation
b
n
a
1
b
n-1
a
n
=0 (14)
with the a
i
a. The set of elements of T integral over a is called the integral closure of a in
T. The proof of Proposition 6.1 shows that b T is integral over a if there exists a faithful
b|-submodule M of T such that bM aM and M is nitely generated as an -module.
Note that if b
n
is integral over a, so also is b (the equation (14) for b
n
can be read as a
similar equation for b).
LEMMA 6.22. Let
t
be the integral closure of in T. Then the integral closure of a in T
is the radical of a
t
.
PROOF. Let b T be integral over a. From (14) we see that b
t
and that b
n
a
t
, and
so b is in the radical of a
t
.
Conversely, let b be in the radical of a
t
, so that
b
n
=

i
a
i
.
i
. some m>0, a
i
a. .
i

t
.
As each .
i
is integral over , M
def
=.
1
. . . . . .
n
| is a nite -algebra (see 6.2). As b
n
M
M, we see that b
n
is integral over a, which implies that b is integral over a.
2
6 INTEGRALITY 30
In particular, the integral closure of a in T is an ideal in
t
, and so it is closed under the
formation of sums and (nonempty) products.
PROPOSITION 6.23. Let be a normal integral domain, and let 1 extension of the eld
of fractions J of . If an element of 1 is integral over an ideal a in , then its minimum
polynomial over J has coefcients in the radical of a.
PROOF. Let be integral over a, so that

n
a
1

n-1
a
n
=0
for some n > 0 and a
i
a. As in the proof of (6.10), the conjugates of satisfy the same
equation as , and so are also integral over a. The coefcients of the minimum polynomial
of over J are polynomials without constant term in its conjugates, and so they are also
integral over a. As these coefcients lie in J, they lie in the integral closure of a in J, which
is the radical of a (by 6.22).
2
THEOREM 6.24. Let T be integral domains with normal and T integral over . Let
p
t
p be prime ideals in , and let q be a prime ideal in T such that q =p. Then there
exists a prime ideal q
t
q in T such that q
t
=p
t
.
PROOF. The prime ideals of T contained in q are the contractions of prime ideals in T
q
(see
5.4), and so we have show to that p
t
is the contraction of a prime ideal of T
q
, or, equivalently
(see 5.6), that

_
p
t
T
q
_
=p
t
.
Let b p
t
T
q
. Then b =,,s with , p
t
T and s T q. By (6.22), , is integral over
p
t
, and so (by 6.23) the minimum equation
,
n
a
1
,
n-1
a
n
=0 (15)
of , over the eld of fractions J of has coefcients a
i
p
t
.
Suppose that b p
t
T
q
. Then b
-1
J, and so, on replacing , with bs in (15) and
dividing through by b
n
, we obtain the minimum equation for s over J:
s
n
(a
1
,b)s
n-1
(a
n
,b
n
) =0. (16)
But b is integral over , and so (by 6.10), each coefcient a
i
,b
i
. Suppose that
b p
t
. The coefcients a
i
,b
i
p
t
, and so (16) shows that s
n
p
t
T pT q, and so s q,
which contradicts its denition. Hence b p
t
, and so p
t
T
q
=p
t
as required.
2
COROLLARY 6.25. Let T be integral domains with normal and T integral over .
Let p
1
p
n
be prime ideals in T, and let
q
1
q
n
(m<n) (17)
be prime ideals in T such that q
i
=p
i
for all i . Then (17) can be extended to a chain of
prime ideals
q
1
q
n
such that q
i
=p
i
for all i.
PROOF. Immediate consequence of the theorem.
2
Theorem 6.24 and its corollary are referred to as the going-down theorem (of Cohen
and Seidenberg).
6 INTEGRALITY 31
The Noether normalization theorem
THEOREM 6.26 (NOETHER NORMALIZATION THEOREM). Every nitely generated alge-
bra over a eld k contains a polynomial algebra 1 such that is a nite 1-algebra.
In other words, there exist elements ,
1
. . . . . ,
i
of such that is a nitely generated
k,
1
. . . . . ,
i
|-module and ,
1
. . . . . ,
i
are algebraically independent
10
over k.
PROOF. We use induction on the minimum number n of generators of as a k-algebra. If
n =0, there is nothing to prove, and so we may suppose that n _1 and that the statement is
true for k-algebras generated by n1 (or fewer) elements.
Let =k.
1
. . . . . .
n
|. If the .
i
are algebraically independent, then there is nothing to
prove, and so we may suppose that there exists a nonconstant polynomial (T
1
. . . . . T
n
) such
that (.
1
. . . . . .
n
) =0. Some T
i
occurs in , say T
1
, and we can write
=c
0
T
1
1
c
1
T
1-1
1
c
1
. c
i
kT
2
. . . . . T
n
|. c
0
=0.
If c
0
k, then the equation
0 =(.
1
. . . . . .
n
) =c
0
.
1
1
c
1
(.
2
. . . . . .
n
).
1-1
1
c
1
(.
2
. . . . . .
n
)
shows that .
1
is integral over k.
2
. . . . . .
n
|. By induction, there exist algebraically indepen-
dent elements ,
1
. . . . . ,
i
such that k.
2
. . . . . .
n
| is nite over k,
1
. . . . . ,
i
|. It follows that
is nite over k,
1
. . . . . ,
i
| (a composite of nite ring homorphisms is nite).
If c
0
k, then we choose different generators for . Fix an integer m>0, and let
.
t
1
=.
1
. .
t
2
=.
2
.
n
2
1
. . . . . .
t
n
=.
n
.
n
n
1
.
Then
k.
t
1
. . . . . .
t
n
| =k.
1
. . . . . .
n
| =
because each .
t
i
k.
1
. . . . . .
n
| and, conversely, each
.
i
k.
1
. .
t
2
. . . . . .
t
n
| =k.
t
1
. . . . . .
t
n
|.
When we let
g(T
1
. . . . . T
n
) =(T
1
. T
2
T
n
2
1
. . . . . T
n
T
n
n
1
) kT
1
. . . . . T
n
|.
then
g(.
t
1
. . . . . .
t
n
) =(.
t
1
. .
t
2
.
tn
2
1
. . . . . .
t
n
.
tn
n
1
) =0.
I claim that, if m is chosen sufciently large, then
g(T
1
. . . . . T
n
) =c
t
0
T
1
1
c
t
1
T
1-1
1
c
t
1
.
with c
t
i
kT
2
. . . . . T
n
| and c
t
0
k

, and so the previous argument applies.

10
Recall that this means that the homomorphism of k-algebras kX
1
. . . . . X
i
| k,
1
. . . . . ,
i
| sending X
i
to
,
i
is an isomorphism, or, equivalently, that
1(,
1
. . . . . ,
i
) =0. 1 kX
1
. . . . . X
i
| == 1 =0.
7 ARTINIAN RINGS 32
To prove the claim, let
(T
1
. . . . . T
n
) =

c
}
1
}
n
T
}
1
1
T
}
n
n
.
Choose m so large that the numbers

1
m
2

2
m
n

n
. (18)
are distinct when (
1
. . . . .
n
) runs over the n-tuples with c
}
1
,...,}
n
=0. Then
(T
1
. T
2
T
n
2
1
. . . . . T
n
T
n
n
1
) =cT
1
1
c
1
T
1-1
1

with c k {0] and N equal to the largest value of (18).
2
REMARK 6.27. When k is innite, there is a simpler proof of a somewhat stronger result:
let =k.
1
. . . . . .
n
|; then there exist algebraically independent elements
1
. . . . .
i
that
are linear combinations of the .
i
such that is nite over k
1
. . . . .
i
| (see 8.13 of my
algebraic geometry notes).
EXERCISE 6.28. A ring is said to be normal if
p
is a normal integral domain for all
prime ideals p in . Show that a noetherian ring is normal if and only if it is a nite product
of normal integral domains.
7 Artinian rings
A ring is artinian if every descending chain of ideals a
1
a
2
in eventually
becomes constant; equivalently, if every nonempty set of ideals has a minimal element.
Similarly, a module M over a ring is artinian if every descending chain of submodules
N
1
N
2
in M eventually becomes constant.
PROPOSITION 7.1. An artinian ring has Krull dimension zero; in other words, every prime
ideal is maximal.
PROOF. Let p be a prime ideal of an artinian ring , and let
t
=,p. Then
t
is an artinian
integral domain. For any nonzero element a of
t
, the chain (a) (a
2
) eventually
becomes constant, and so a
n
=a
n1
b for some b
t
and n _ 1. We can cancel a
n
to
obtain 1 =ab. Thus a is a unit,
t
is a eld, and p is maximal.
2
COROLLARY 7.2. In an artinian ring, the nilradical and the Jacobson radical coincide.
PROOF. The rst is the intersection of the prime ideals (2.4), and the second is the intersec-
tion of the maximal ideals (2.5).
2
PROPOSITION 7.3. An artinian ring has only nitely many maximal ideals.
PROOF. Let m
1
. . . m
n
be minimal among nite intersections of maximal ideals in an
artinian ring, and let m be another maximal ideal in the ring. If m is not equal to one of the
m
i
, then, for each i , there exists an a
i
m
i
m. Now a
1
a
n
lies in m
1
. . . m
n
but not
in m (because m is prime), contradicting the minimality of m
1
. . . m
n
.
2
7 ARTINIAN RINGS 33
PROPOSITION 7.4. In an artinian ring, some power of the nilradical is zero.
PROOF. Let N be the nilradical of the artinian ring . The chain NN
2
eventually
becomes constant, and so N
n
=N
n1
= for some n _1. Suppose that N
n
=0. Then
there exist ideals a such that a N
n
=0, for example N, and we may suppose that a has been
chosen to be minimal among such ideals. There exists an a a such that a N
n
=0, and
so a =(a) (by minimality). Now (aN
n
)N
n
=aN
2n
=aN
n
=0 and aN
n
(a), and so
aN
n
=(a) (by minimality again). Hence a =a. for some . N
n
. Now a =a. =a.
2
=
=a0 =0 because . N. This contradicts the denition of a, and so N
n
=0.
2
LEMMA 7.5. Let be a ring in which some nite product of maximal ideals is zero. Then
is artinian if and only if it is noetherian.
PROOF. Suppose that m
1
m
n
=0 with the m
i
maximal ideals (not necessarily distinct),
and consider
m
1
m
1
m
i-1
m
1
m
i
m
1
m
n
=0.
The action of on the quotient M
i
def
=m
1
m
i-1
,m
1
m
i
factors through the eld ,m
i
,
and the subspaces of the vector space M
i
are in one-to-one correspondence with the ideals
of contained between m
1
m
i-1
and m
1
m
i
. If is either artinian or noetherian, then
M
i
satises a chain condition on subspaces and so it is nite-dimensional as a vector space
and both artinian and noetherian as an -module. Now repeated applications of Proposition
3.3 (resp. its analogue for artinian modules) show that if is artinian (resp. noetherian),
then it is noetherian (resp. artinian) as an -module, and hence as a ring.
2
THEOREM 7.6. A ring is artinian if and only if it is noetherian of dimension zero.
PROOF. =: Let be an artinian ring. After (7.1), it remains to show that is noetherian,
but according to (7.2), (7.3), and (7.4), some nite product of maximal ideals is zero, and so
this follows from the lemma.
=: Let be a noetherian ring of dimension zero. The zero ideal admits a primary
decomposition (17.11), and so has only nitely many minimal prime ideals, which are all
maximal because dim =0. Hence N is a nite intersection of maximal ideals (2.4), and
since some power of N is zero (3.16), we again have that some nite product of maximal
ideals is zero, and so can apply the lemma.
2
THEOREM 7.7. Every artinian ring is (uniquely) a product of local artinian rings.
PROOF. Let be artinian, and let m
1
. . . . . m
i
be the distinct maximal ideals in . We saw
in the proof of (7.6) that some product m
n
1
1
m
n
r
i
=0. For i = , the ideal m
n
i
i
m
n
j
}
is
not contained in any maximal ideal, and so equals . Now the Chinese remainder theorem
2.12 shows that
.,m
n
1
1
,m
n
r
i
,
and each ring ,m
n
i
i
is obviously local.
2
PROPOSITION 7.8. Let be a local artinian ring with maximal ideal m. If m is principal,
so also is every ideal in ; in fact, if m=(t ), then every ideal is of the form (t
i
) for some
r _0.
8 DIRECT AND INVERSE LIMITS 34
PROOF. Because m is the Jacobson radical of , some power of m is zero (by 7.4); in
particular, (0) =(t
i
) for some r. Let a be a nonzero ideal in . There exists an integer r _0
such that a m
i
but a ,m
i1
. Therefore there exists an element a of a such that a =ct
i
for some c but a (t
i1
). The second condition implies that c m, and so it is a unit;
therefore a =(a).
2
EXAMPLE 7.9. The ring =kX
1
. X
2
. X
3
. . . .|,(X
1
. X
2
2
. X
3
3
. . . .) has only a single prime
ideal, namely, (.
1
. .
2
. .
3
. . . .), and so has dimension zero. However, it is not noetherian
(hence not artinian).
ASIDE 7.10. Every nitely generated module over a principal Artin ring is a direct sum of cyclic
modules (see mo22722).
8 Direct and inverse limits
Direct limits
DEFINITION 8.1. A partial ordering _ on a set 1 is said to be directed, and the pair (1. _)
is called a directed set, if for all i. 1 there exists a k 1 such that i. _k.
DEFINITION 8.2. Let (1. _) be a directed set, and let be a ring.
A direct system of -modules indexed by (1. _)
is a family (M
i
)
iJ
of -modules together with a
family (
i
}
: M
i
M
}
)
i_}
of -linear maps such
that
i
i
=id

i
and
}
k

i
}
=
i
k
all i _ _k.
M
k
M
i
M
}

i
k

i
j

j
k
An -module M together with a family (
i
: M
i

M)
iJ
of -linear maps satisfying
i
=
}

i
}
all i _ is said to be a direct limit of the sys-
tem ((M
i
). (
i
}
)) if it has the following universal
property: for any other -module N and fam-
ily (
i
: M
i
N) of -linear maps such that

i
=
}

i
}
all i _ , there exists a unique mor-
phism : M N such that
i
=
i
for all
i .
M
M
i
M
}
N

i
j

As usual, the universal property determines the direct limit (if it exists) uniquely up to a
unique isomorphism. We denote it lim

(M
i
.
}
i
), or just lim

M
i
.
CRITERION
An -module M together with -linear maps
i
: M
i
M such that
i
=
}

i
}
for all
i _ is the direct limit of a system (M
i
.
}
i
) if and only if
(a) M =
_
iJ

i
(M
i
), and
(b) m
i
M
i
maps to zero in M if and only if it maps to zero in M
}
for some _i .
8 DIRECT AND INVERSE LIMITS 35
CONSTRUCTION
Let
M =

iJ
M
i
,M
t
where M
t
is the -submodule generated by the elements
m
i

i
}
(m
i
) all i < , m
i
M
i
.
Let
i
(m
i
) =m
i
M
t
. Then certainly
i
=
}

i
}
for all i _ . For every -module N
and -linear maps
}
: M
}
N, there is a unique map

iJ
M
i
N.
namely,

m
i

i
(m
i
), sending m
i
to
i
(m
i
), and this map factors through M and is
the unique -linear map with the required properties.
Direct limits of -algebras, etc., are dened similarly.
AN EXAMPLE
PROPOSITION 8.3. For every multiplicative subset S of a ring , S
-1
.lim

h
, where
h runs over the elements of S (partially ordered by division).
PROOF. When h[h
t
, say, h
t
=hg, there is a unique homomorphism
h

h
0 respecting
the maps
h
and
h
0 , namely,
o
h

o
h
0
, and so the rings
h
form a direct system
indexed by the set S. When h S, the homomorphism S
-1
extends uniquely to a
homomorphism
o
h

o
h
:
h
S
-1
(see 5.1), and these homomorphisms are compatible
with the maps in the direct system. Now apply the criterion p. 34 to see that S
-1
is the
direct limit of the
h
.
2
EXACTNESS
PROPOSITION 8.4. The direct limit of a system of exact sequences of modules is exact.
This means the following: suppose that (M
i
.
i
}
), (N
i
.
i
}
), and (1
i
. ;
i
}
) are direct
systems with repect to the directed set 1, and let
0 (M
i
.
i
}
)
(o
i
)
(N
i
.
i
}
)
(b
i
)
(1
i
. ;
i
}
) 0
be a sequence of maps of direct systems; if the sequences
0 M
i
o
i
N
i
b
i
1
i
0
are exact for all i , then the direct limit sequence
0 lim

M
i
lim

o
i
lim

N
i
lim

b
i
lim

1
i
0
is exact.
PROOF. Let (n
i
) lim

N
i
. If (b
i
(n
i
)) =0, then there exists an i
0
such that b
i
(n
i
) =0 for
all i _ i
0
. Let m
i
=0 unless i _ i
0
, in which case we let m
i
be the unique element such
that a
i
(m
i
) =n
i
. Then (m
i
) maps to (n
i
). This proves exactness at lim

N
i
, and the proof of
exactness at the other terms is obvious.
2
8 DIRECT AND INVERSE LIMITS 36
Inverse limits
Inverse limits are the same as direct limits except that the directions of the arrows is reversed.
Thus, formally, the theory of inverse limits is the same as that of inverse limits. However, in
concrete categories, they behave very differently. For example, the inverse limit of a system
of exact sequences of modules need not be exact.
We shall consider inverse limits only in the case that the indexing set if N with its usual
ordering. Thus, an inverse system of -modules is nothing more than a sequence of modules
and -homomorphisms
M
0

0
M
1

n1
M
n

n
.
Ahomomorphism(M
n
.
n
) (N
n
.
n
) of inverse systems is a sequence of -homomorphisms
;
n
: M
n
N
n
such that
n
;
n1
=;
n

n
for all n N.
Given an inverse system (M
n
.
n
) of -modules, we dene lim

M
n
and lim

1
M
n
to be
the kernel and cokernel of the map
(m
n
)
nN
(m
n

n
(m
n1
)):

M
n

M
n
.
PROPOSITION 8.5. For any inverse system (M
n
.
n
) and -module N,
Hom(lim

M
n
. N) .lim

Hom(M
n
. N).
PROOF. Obvious.
2
PROPOSITION 8.6. For any inverse system of exact sequence
0 (M
n
.
n
) (N
n
.
n
) (1
n
. ;
n
) 0.
there is an exact sequence
0 lim

M
n
lim

N
n
lim

1
n
lim

1
M
n
lim

1
N
n
lim

1
1
n
0.
PROOF. The sequence
0

M
n

N
n

1
n
0
is exact, and so this follows from the snake lemma.
2
COROLLARY 8.7. If the maps
n
: M
n1
M
n
are all surjective, then the sequence
0 lim

M
n
lim

N
n
lim

1
n
0
is exact.
PROOF. The hypothesis implies that lim

1
M
n
=0 (axiom of determined choice).
2
ASIDE 8.8. Direct (resp. inverse) limits are also called inductive (resp. projective) limits or colimits
(resp. limits).
9 TENSOR PRODUCTS 37
9 Tensor Products
Tensor products of modules
Let be a ring, and let M, N, and 1 be -modules. A map : MN 1 of -modules
is said to be -bilinear if
(. .
t
. ,) =(.. ,) (.
t
. ,). .. .
t
M. , N
(.. , ,
t
) =(.. ,) (.. ,
t
). . M. ,. ,
t
N
(a.. ,) =a(.. ,). a . . M. , N
(.. a,) =a(.. ,). a . . M. , N.
i.e., if is -linear in each variable.
MN T
T
t
.

0
3 linear
An -module T together with an -bilinear map
: MN T
is called the tensor product of M and N over if it has the
following universal property: every -bilinear map

t
: MN T
t
factors uniquely through .
As usual, the universal property determines the tensor product uniquely up to a unique
isomorphism. We write it M

N. Note that
Hom
-bilinear
(MN. T ) .Hom
-linear
(M

N. T ).
CONSTRUCTION
Let M and N be -modules, and let
(1)
be the free -module with basis MN. Thus
each element
(1)
can be expressed uniquely as a nite sum

a
i
(.
i
. ,
i
). a
i
. .
i
M. ,
i
N.
Let 1 be the submodule of
(1)
generated by the following elements
(. .
t
. ,) (.. ,) (.
t
. ,). .. .
t
M. , N
(.. , ,
t
) (.. ,) (.. ,
t
). . M. ,. ,
t
N
(a.. ,) a(.. ,). a . . M. , N
(.. a,) a(.. ,). a . . M. , N.
and dene
M

N =
(1)
,1.
Write . , for the class of (.. ,) in M

N. Then
(.. ,) . ,: MN M

N
9 TENSOR PRODUCTS 38
is -bilinear we have imposed the fewest relations necessary to ensure this. Every element
of M

N can be written as a nite sum

11

a
i
(.
i
,
i
). a
i
. .
i
M. ,
i
N.
and all relations among these symbols are generated by the following relations
(. .
t
) , =. , .
t
,
. (, ,
t
) =. , . ,
t
a(. ,) =(a.) , =. a,.
The pair (M

N. (.. ,) . ,) has the correct universal property because any bilinear

map
t
: M N T
t
denes an -linear map
(1)
T
t
, which factors through

(1)
,1, and gives a commutative triangle.
EXTENSION OF SCALARS
Let be a commutative ring and let T be an -algebra (not necessarily commutative) such
that the image of T lies in the centre of T. Then M T

M is a functor from
left -modules to left T-modules. Let M be an -module and N a T-module; an -linear
map : M N denes a T-linear map : T

M N such that b mb (m), and

- is an isomorphism:
Hom
-linear
(M. N) .Hom
B-linear
(T

M. N). (19)
If (e

)
J
is a family of generators (resp. basis) for M as an -module, then (1e

)
J
is
a family of generators (resp. basis) for T

M as a T-module.
BEHAVIOUR WITH RESPECT TO DIRECT LIMITS
PROPOSITION 9.1. Direct limits commute with tensor products:
lim

iJ
M
i

lim

}J
N
}
. lim

(i,})JJ
M
i

N
}
.
PROOF. Using the universal properties of direct limits and tensor products, one sees easily
that lim

(M
i

N
}
) has the universal property to be the tensor product of lim

M
i
and
lim

N
}
.
2
Tensor products of algebras
Let k be a ring, and let and T be k-algebras. Ak-algebra C together with homomorphisms
i : C and : T C is called the tensor product of and T if it has the following
universal property:
11
An element of the tensor product of two vector spaces is not necessarily a tensor product of two vectors,
but sometimes a sum of such. This might be considered a mathematical shenanigan but if you start with the
state vectors of two quantum systems it exactly corresponds to the notorious notion of entanglement which so
displeased Einstein. Georges Elencwajg on mathoverow.net.
9 TENSOR PRODUCTS 39
for every pair of homomorphisms (of k-algebras)
: 1 and g: T 1, there exists a unique
homomorphism (. g): C 1 such that (. g)
i = and (. g) =,
C T
1
i
}
(
3 ((, )
If it exists, the tensor product, is uniquely determined up to a unique isomorphism by this
property. We write it
k
T. Note that the universal property says that
Hom(
k
T. 1) .Hom(. 1) Hom(T. 1) (20)
(k-algebra homomorphisms).
CONSTRUCTION
Regard and T as k-modules, and form the tensor product
k
T. There is a multiplication
map
k
T
k
T
k
T for which
(ab)(a
t
b
t
) =aa
t
bb
t
. all a. a
t
. b. b
t
T.
This makes
k
T into a ring, and the homomorphism
c c(11) =c 1 =1c
makes it into a k-algebra. The maps
a a1:
k
T and b 1b: T
k
T
are homomorphisms, and they make
k
T into the tensor product of and T in the above
sense.
EXAMPLE 9.2. The algebra , together with the maps
k
id
A
,
is k
k
(because it has the correct universal property). In terms of the constructive
denition of tensor products, the map c a ca: k
k
is an isomorphism.
EXAMPLE 9.3. The ring kX
1
. . . . . X
n
. X
n1
. . . . . X
nn
|, together with the obvious inclu-
sions
kX
1
. . . . . X
n
| kX
1
. . . . . X
nn
| kX
n1
. . . . . X
nn
|
is the tensor product of the k-algebras kX
1
. . . . . X
n
| and kX
n1
. . . . . X
nn
|. To verify
this we only have to check that, for every k-algebra 1, the map
Hom(kX
1
. . . . . X
nn
|. 1) Hom(kX
1
. . . .|. 1) Hom(kX
n1
. . . .|. 1)
induced by the inclusions is a bijection. But this map can be identied with the bijection
1
nn
1
n
1
n
.
In terms of the constructive denition of tensor products, the map
kX
1
. . . . . X
n
|
k
kX
n1
. . . . . X
nn
| kX
1
. . . . . X
nn
|
sending g to g is an isomorphism.
9 TENSOR PRODUCTS 40
REMARK 9.4. (a) Let k k
t
be a homomorphism of rings. Then
k
t

k
kX
1
. . . . . X
n
| .k
t
1X
1
. . . . . 1X
n
| .k
t
X
1
. . . . . X
n
|.
If =kX
1
. . . . . X
n
|,(g
1
. . . . . g
n
), then
k
t

k
.k
t
X
1
. . . . . X
n
|,(g
1
. . . . . g
n
).
(b) If and T are algebras of k-valued functions on sets S and T respectively, then the
denition
( g)(.. ,) =(.)g(,). , g T, . S, , T.
realizes
k
T as an algebra of k-valued functions on S T .
The tensor algebra of a module
Let M be a module over a ring . For each _0, set
T
i
M =M

M (r factors),
so that T
0
M = and T
1
M =M, and dene
TM =

i_0
T
i
M.
This can be made into a noncommutative -algebra, called the tensor algebra of M, by
requiring that the multiplication map
T
i
MT
x
M T
ix
M
send (m
1
m
i
. m
i1
m
ix
) to m
1
m
ix
.
M TM
1
-linear 3-algebra
The pair (TM. M TM) has the following universal prop-
erty: every -linear map from M to an -algebra 1 (not neces-
sarily commutative) extends uniquely to an -algebra homomor-
phism TM 1.
If M is a free -module with basis .
1
. . . . . .
n
, then TM is
the (noncommutative) polynomial ring over in the noncommut-
ing symbols .
i
(because this -algebra has the same universal
property as TM).
The symmetric algebra of a module
The symmetric algebra Sym(M) of an -module M is the quotient of TM by the ideal
generated by all elements of T
2
M of the form
mnnm. m. n M.
It is a graded algebra Sym(M) =

i_0
Sym
i
(M) with Sym
i
(M) equal to the quotient of
M
i
by the -submodule generated by all elements of the form
m
1
m
i
m
c(1)
m
c(i)
. m
i
M. o T
i
(symmetric group).
10 FLATNESS 41
M Sym(M)
1
-linear
3-algebra
The pair (Sym(M). M Sym(M)) has the following
universal property: every -linear map M 1 from M
to a commutative -algebra 1 extends uniquely to an
-algebra homomorphism Sym(M) 1 (because it ex-
tends to an -algebra homomorphism TM 1, which
factors through Sym(M) because 1 is commutative).
If M is a free -module with basis .
1
. . . . . .
n
, then
Sym(M) is the polynomial ring over in the (commuting) symbols .
i
(because this -
algebra has the same universal property as TM).
10 Flatness
Let M be an -module. If the sequence of -modules
0 N
t
N N
tt
0 (21)
is exact, then the sequence
M

N
t
M

N M

N
tt
0
is exact, but M

N
t
M

N need not be injective. For example, when we tensor the

exact sequence of Z-modules
0 Z
n
Z Z,mZ 0
with Z,mZ, we get the sequence
Z,mZ
x|-nx=0
Z,mZ
x|-x
Z,mZ 0.
Moreover, M

N may be zero even when neither M nor N is nonzero. For example,

Z,2Z
Z
Z,3Z =0
because it is killed by both 2 and 3.
12
In fact, M

M may be zero without M being zero,

for example,
13
Q,Z
Z
Q,Z =0.
DEFINITION 10.1. An -module M is at if
N
t
N injective == M

N
t
M

N injective.
It is faithfully at if, in addition,
M

N =0 == N =0.
A homomorphism of rings T is said to be at (resp. faithfully at) when T is at
(resp. faithfully at) as an -module.
12
It was once customary in certain circles to require a ring to have an identity element 1 =0 (see, for example,
Northcott 1953, p.3). However, without the zero ring, tensor products dont always exist. Bourbakis rst
example of a ring is the zero ring.
13
Let .. , Q,Z; then n. =0 for some n Z, and , =n,
t
for some ,
t
Q,Z; now
. , =. n,
t
=n. ,
t
=0,
t
=0.
10 FLATNESS 42
Thus, an -module M is at if and only if M

is an exact functor, i.e.,

0 M

N
t
M

N M

N
tt
0 (22)
is exact whenever (21) is exact. An -algebra T is said to be at if T is at as an -module.
EXAMPLE 10.2. The functor M takes direct sums to direct sums, and therefore split-
exact sequences to split-exact sequences. Therefore, all vector spaces over a eld are at,
and nonzero vector spaces are faithfully at. In fact, every module over a product of elds
(even an innite product) is at.
EXAMPLE 10.3. Quotient maps ,a are rarely at. If is a product, =
1

2
,
then the quotient map
1
is obviously at. When is noetherian, all at quotient
maps are of this form.
14
PROPOSITION 10.4. Let T be a faithfully at homomorphism of rings. A sequence
of -modules
0 N
t
N N
tt
0 (23)
is exact if
0 T

N
t
T

N T

N
tt
0 (24)
is exact.
PROOF. Let N
0
be the kernel of N
t
N. Because T is at, T

N
0
is the kernel of
T

N
t
T

N, which is zero by assumption; because T is faithfully at, this

implies that N
0
=0. We have proved the exactness at N
t
, and the proof of the exactness
elsewhere is similar.
2
REMARK 10.5. There is a converse to the proposition: suppose that
(23) is exact =(24) is exact;
then T is faithfully at. The implication = shows that T is at. Now let N be
an -module, and consider the sequence
0 0 N 0 0.
If T

N =0, then this sequence becomes exact when tensored with T, and so is itself
exact, which implies that N =0. This shows that T is faithfully at.
COROLLARY 10.6. Let T be faithfully at. An -module M is at (resp. faithfully
at) if T

M is at (resp. faithfully at) as a T-module.

14
The set V(a) is closed in spec() (by denition of the topology on spec()). If ,a is at, then V(a) is
also open. Therefore =
1

2
and a is of the form b
2
with b an ideal in
1
such that V(b) =spec(
1
).
On tensoring
0 b
2

1
,b 0
with
1
,b we get an exact sequence
0 b,b
2

1
,b
id

1
,b 0.
Therefore b = b
2
, but b is contained in all prime ideals of
1
, and so this implies that b = 0 (Nakayamas
lemma, 3.9).
10 FLATNESS 43
PROOF. Assume that M
B
def
= T

N is at, and let N

t
N be an injective map of -
modules. We have that
T

(M

N
t
M

N) .M
B

B
(N
t
B
N
B
),
and the map at right is injective because T is at and M
B
is at. Now (10.4) shows
that M

N
t
M

N is injective. Thus M is at.

Assume that M
B
is faithfully at, and let N be an -module. If M

N =0, then
M
B

B
N
B
is zero because it is isomorphic to (M

N)
B
. Now N
B
=0 because M
B
is
faithfully at, and so N =0 because T is faithfully at.
2
PROPOSITION 10.7. Let i : T be a faithfully at homomorphism. For every -module
M, the sequence
0 M
d
0
T

M
d
1
T

M (25)
with
_
J
0
(m) = 1m.
J
1
(b m) = 1b mb 1m
is exact.
PROOF. Assume rst that there exists an -linear section to T, i.e., an -linear map
: T such that i =id

, and dene
k
0
: T

M M. k
0
(b m) =(b)m
k
1
: T

M T

M. k
1
(b b
t
m) =(b)b
t
m.
Then k
0
J
0
=id

, which shows that J

0
is injective. Moreover,
k
1
J
1
J
0
k
0
=id
B
A

which shows that, if J

1
(.) =0, then . =J
0
(k
0
(.)), as required.
We now consider the general case. Because T is faithfully at, it sufces to prove
that the sequence (25) becomes exact after tensoring in T. But the sequence obtained from
(25) by tensoring with T is isomorphic to the sequence (25) for the homomorphism of rings
b 1b: T T

T and the T-module T

M, because, for example,

T

(T

M) .(T

T)
B
(T

M).
Now T T

T has an T-linear section, namely, (b b

t
) =bb
t
, and so we can apply
the rst part.
2
COROLLARY 10.8. If T is faithfully at, then it is injective with image the set of
elements on which the maps
_
b 1b
b b 1
: T T

T
agree.
PROOF. This is the special case M = of the Proposition.
2
10 FLATNESS 44
PROPOSITION 10.9. Let
t
be a homomorphism of rings. If T is at (or
faithfully at), then so also is
t
T

t
.
PROOF. For any
t
-module M,
(T

t
)

0 M .T

(
t

0 M) .T

M.
from which the statement follows.
2
PROPOSITION 10.10. For every multiplicative subset S of a ring and -module M,
S
-1

M .S
-1
M.
The homomorphism a
o
1
: S
-1
is at.
PROOF. To give an S
-1
-module is the same as giving an -module on which the elements
of S act invertibly. Therefore S
-1

M and S
-1
M satisfy the same universal property
(see 9, especially (19)), which proves the rst statement. As M S
-1
M is exact (5.11),
so also is M S
-1

M, which proves the second statement.

2
PROPOSITION 10.11. A homomorphism of rings : T is at if

1
(n)
T
n
is at
for all maximal ideals n in T.
PROOF. Let N
t
N be an injective homomorphism of -modules, and let n be a maximal
ideal of T. Then p =
-1
(n) is a prime ideal in , and
p

(N
t
N) is injective (10.10).
Therefore, the map
T
n

(N
t
N) .T
n

p
(
p

(N
t
N))
is injective, and so the kernel M of T

(N
t
N) has the property that M
n
= 0. Let
. M, and let a ={b T [ b. =0]. For each maximal ideal n of T, . maps to zero in M
n
,
and so a contains an element not in n. Hence a =T, and so . =0.
2
PROPOSITION 10.12. The following conditions on a at homomorphism : T are
equivalent:
(a) is faithfully at;
(b) for every maximal ideal m of , the ideal (m)T =T;
(c) every maximal ideal m of is of the form
-1
(n) for some maximal ideal n of T.
PROOF. (a) =(b): Let m be a maximal ideal of , and let M =,m; then
T

M .T,(m)T.
As T

M =0, we see that (m)T =T.

(b) = (c): If (m)T = T, then (m) is contained in a maximal ideal n of T. Now

-1
(n) is a proper ideal in containing m, and hence equals m.
(c) =(a): Let M be a nonzero -module. Let . be a nonzero element of M, and let
a =ann(.)
def
={a [ a. =0]. Then a is an ideal in , and M
t
def
=. .,a. Moreover,
T

M
t
.T,(a) T and, because T is at, T

M
t
is a submodule of T

M.
Because a is proper, it is contained in a maximal ideal m of , and therefore
(a) (m) n
for some maximal ideal n of . Hence (a) T n =T, and so T

M T

M
t
=0.
2
10 FLATNESS 45
In more geometric terms, the proposition says that a homomorphism : T is
faithfully at if it is at and the map spmT spm is surjective.
THEOREM 10.13 (GENERIC FLATNESS). Let an integral domain with eld of fractions
J, and let T be a nitely generated -algebra contained in J

T. Then for some nonzero

elements a of and b of T, the homomorphism
o
T
b
is faithfully at.
PROOF. As J

T is a nitely generated J-algebra, the Noether normalization theorem

(6.26) shows that there exist elements .
1
. . . . . .
n
of J

T such that J.
1
. . . . . .
n
| is a
polynomial ring over J and J

T is a nite J.
1
. . . . . .
n
|-algebra. After multiplying
each .
i
by an element of , we may suppose that it lies in T. Let b
1
. . . . . b
n
generate T as an
-algebra. Each b
i
satises a monic polynomial equation with coefcients in J.
1
. . . . . .
n
|.
Let a be a common denominator for the coefcients of these polynomials. Then each b
i
is integral over
o
. As the b
i
generate T
o
as an
o
-algebra, this shows that T
o
is a nite

o
.
1
. . . . . .
n
|-algebra (by 6.2). Therefore, after replacing with
o
and T with T
o
, we
may suppose that T is a nite .
1
. . . . . .
n
|-algebra.
T J

T 1
x
1
,...,x
m
j
T
.
1
. . . . . .
n
| J.
1
. . . . . .
n
| 1
def
=J(.
1
. . . . . .
n
)
J.
injective
nite nite nite
Let 1 =J(.
1
. . . . . .
n
) be the eld of fractions of .
1
. . . . . .
n
|, and let b
1
. . . . . b
i
be
elements of T that form a basis for 1
x
1
,...,x
m
j
T as an 1-vector space. Each element
of T can be expressed as a linear combination of the b
i
with coefcients in 1. Let q be
a common denominator for the coefcients arising from a set of generators for T as an
.
1
. . . . . .
n
|-module. Then b
1
. . . . . b
i
generate T
q
as an .
1
. . . . . .
n
|
q
-module. In other
words, the map
(c
1
. . . . . c
i
)

c
i
b
i
: .
1
. . . . . .
n
|
i
q
T
q
(26)
is surjective. This map becomes an isomorphism when tensored with 1 over .
1
. . . . . .
n
|
q
,
which implies that each element of its kernel is killed by a nonzero element of .
1
. . . . . .
n
|
q
and so is zero (because .
1
. . . . . .
n
|
q
is an integral domain). Hence the map (26) is an
isomorphism, and so T
q
is free of nite rank over .
1
. . . . . .
n
|
q
. Let a be some nonzero
coefcient of the polynomial q, and consider the maps

o
.
1
. . . . . .
n
|
o
.
1
. . . . . .
n
|
q
T
oq
.
The rst and third arrows realize their targets as nonzero free modules over their sources,
and so are faithfully at. The middle arrow is at by (10.10). Let m be a maximal ideal in

o
. Then m
o
.
1
. . . . . .
n
| does not contain the polynomial q because the coefcient a of
q is invertible in
o
. Hence m
o
.
1
. . . . . .
n
|
q
is a proper ideal of
o
.
1
. . . . . .
n
|
q
, and so
the map
o

o
.
1
. . . . . .
n
|
q
is faithfully at (apply 10.12). This completes the proof.
2
REMARK 10.14. The theorem holds for every nitely generated T-algebra, i.e., without
the requirement that T J

T. To see this, note that J

T is the ring of fractions

11 FINITELY GENERATED PROJECTIVE MODULES 46
of T with respect to the multiplicative subset {0] (see 10.10), and so the kernel of
T J

T is the ideal
n ={b T [ ab =0 for some nonzero a ].
This is nitely generated (Hilbert basis theorem 3.7), and so there exists a nonzero c
such that cb =0 for all b n. I claim that the homomorphism T
c
J

c
T
c
is injective.
If
b
c
r
lies in its kernel, then
o
c
s
b
c
r
=0 in T
c
for some nonzero
o
c
s

c
, and so c
1
ab =0
in T for some N; therefore b n, and so cb =0, which implies that
b
c
r
=0 already in T
c
.
Therefore, after replacing , T, and M with
c
, T
c
, and M
c
, we may suppose that the map
T J

T is injective. On identifying T with its image, we arrive at the situation of the

theorem.
EXERCISE 10.15. Let (
i
.
i
}
) be a direct system of rings, and let (M
i
.
i
}
) be a direct
system of abelian groups with the same indexing set. Suppose that each M
i
has the structure
of an
i
-module, and that the diagrams

i
M
i
M
i

}
M
}
M
}

i
j

i
j

i
j
commute for all i _ . Let =lim

i
and M =lim

M
i
.
(a) Show that M has a unique structure of an -module for which the diagrams

i
M
i
M
i
M M

i
commute for all i .
(b) Show that M is at as an -module if each M
i
is at as an
i
-module.
(Bourbaki AC, I, 2, Prop. 9.)
11 Finitely generated projective modules
In many situations, the correct generalization of nite-dimensional vector space is not
nitely generated module but nitely generated projective module. From a different
perspective, they are the algebraists analogue of the differential geometers vector bundle.
Throughout this section, is a commutative ring.
Projective modules
DEFINITION 11.1. An -module 1 is projective if, for each surjective -linear map
: M N and -linear map g: 1 N, there exists an -linear map h: 1 M (not
11 FINITELY GENERATED PROJECTIVE MODULES 47
necessarily unique) such that h =g:
1
M N 0.
(

3h
In other words, 1 is projective if every map from 1 onto a quotient of a module M lifts to a
map to M. Equivalently, 1 is projective if the functor M Hom
-lin
(1. M) is exact.
As
Hom(

i
1
i
. M) .

i
Hom(1
i
. M)
we see that a direct sum of -modules is projective if and only if each direct summand
is projective. As itself is projective, this shows that every free -module is projective
and every direct summand of a free module is projective. Conversely, let 1 be a projective
module, and write it as a quotient of a free module,
J
(
1 0:
because 1 is projective, there exists an -linear map h: 1 J such that h =id
1
; then
J ~Im(h) Ker( ) ~1 Ker( ).
and so 1 is a direct summand of J. We conclude: the projective -modules are exactly the
direct summands of free -modules.
Finitely presented modules
DEFINITION 11.2. An -module M is nitely presented if there exists an exact sequence

n
M 0, some m. n N.
A nite family (e
i
)
iJ
of generators for an -module M denes a homomorphism
(a
i
)

iJ
a
i
e
i
:
J
M. The elements of the kernel of this homomorphism are called
the relations between the generators. Thus, M is nitely presented if it admits a nite family
of generators whose module of relations is nitely generated. Obviously
nitely presented =nitely generated,
and the converse is true when is noetherian (by 3.4).
PROPOSITION 11.3. If M is nitely presented, then the kernel of every surjective homo-
morphism
n
M, m N, is nitely generated.
In other words, if M is nitely presented, then the module of relations for every nite
generating set is nitely generated.
11 FINITELY GENERATED PROJECTIVE MODULES 48
PROOF. We are given that there exists a surjective homomorphism
n
M with nitely
generated kernel N, and we wish to show that the kernel N
t
of
n
M is nitely generated.
Consider the diagram:
0 N
n
M 0
0 N
t

n
M 0
id
M
(

The map g exists because
n
is projective, and it induces the map . From the diagram, we
get an exact sequence
N
(
N
t

n
,g
n
0,
either from the snake lemma or by a direct diagram chase. As N and
n
,g
n
are both
nitely generated, so also is N
t
(by 3.3(b)).
2
If M is nitely generated and projective, then the kernel of
n
M is a direct summand
(hence quotient) of
n
, and so is nitely generated. Therefore M is nitely presented.
Finitely generated projective modules
According to the above discussion, the nitely generated projective modules are exactly the
direct summands of free -modules of nite rank.
THEOREM 11.4. The following conditions on an -module are equivalent:
(a) M is nitely generated and projective;
(b) M is nitely presented and M
m
is a free
m
-module for all maximal ideals m of ;
(c) there exists a nite family (
i
)
iJ
of elements of generating the ideal and such
that, for all i 1, the
(
i
-module M
(
i
is free of nite rank;
(d) M is nitely presented and at.
Moreover, when is an integral domain and M is nitely presented, they are equivalent to:
(e) dim
k(p)
(M

k(p)) is the same for all prime ideals p of (here k(p) denotes the
eld of fractions of ,p).
PROOF. (a)=(d). As tensor products commute with direct sums, every free module is at
and every direct summand of a at module is at. Therefore, every projective module M is
at, and we saw above that such a module is nitely presented if it is nitely generated.
(b)=(c). Let m be a maximal ideal of , and let .
1
. . . . . .
i
be elements of M
whose images in M
m
form a basis for M
m
over
m
. The kernel N
t
and cokernel N of the
homomorphism
:
i
M. g(a
1
. . . . . a
i
) =

a
i
.
i
.
are both nitely generated, and N
t
m
=0 =N
m
. Therefore, there exists
15
an m such
that N
t
(
=0 =N
(
. Now becomes an isomorphism when tensored with
(
.
The set T of elements arising in this way is contained in no maximal ideal, and so
generates the ideal . Therefore, 1 =

iJ
a
i

i
for certain a
i
and
i
T .
15
To say that S
-1
N =0 means that, for each . N, there exists an s
x
S such that s
x
. =0. If .
1
. . . . . .
n
generate N, then s
def
=s
x
1
s
x
n
lies in S and has the property that sN =0. Therefore, N
x
=0.
11 FINITELY GENERATED PROJECTIVE MODULES 49
(c)=(d). Let T =

iJ

(
i
. Then T is faithfully at over , and T

M =

M
(
i
,
which is clearly a at T-module. It follows that M is a at -module (apply 10.6).
(c)=(e). This is obvious.
(e)=(c). Fix a prime ideal p of . For some p, there exist elements .
1
. . . . . .
i
of
M
(
whose images in M

k(p) form a basis. Then the map

:
i
(
M
(
. (a
1
. . . . . a
i
) =

a
i
.
i
.
denes a surjection
i
p
M
p
(Nakayamas lemma; note that k(p) .
p
,p
p
). Because
the cokernel of is nitely generated, the map itself will be surjective once has been
replaced by a multiple. For any prime ideal q of
(
, the map k(q)
i
M

k(q) dened
by is surjective, and hence is an isomorphism because dim(M

k(q)) = r. Thus
Ker() q
i
(
for every q, which implies that it is zero as
(
is reduced. Therefore M
(
is
free. As in the proof of (b), a nite set of such s will generate .
2
To prove the remaining implications, (d)=(a).(b) we shall need the following lemma.
LEMMA 11.5. Let
0 N J M 0 (27)
be an exact sequence of -modules with N a submodule of J.
(a) lf M and J are at over , then N aJ =aN (inside J) for all ideals a of .
(b) Assume that J is free with basis (,
i
)
iJ
and that M is at. If the element n =

iJ
a
i
,
i
of J lies in N, then there exist n
i
N such that n =

iJ
a
i
n
i
.
(c) Assume that M is at and J is free. For every nite set {n
1
. . . . . n
i
] of elements of
N, there exists an -linear map : J N with (n
}
) =n
}
. =1. . . . , r.
PROOF. (a) Consider
aN aJ aM
0 N aJ aJ aM
: :
The rst row is obtained from (27) by tensoring with a, and the second row is a subsequence
of (27). Both rows are exact. On tensoring a with J we get a map aJ J, which is
injective because J is at. Therefore aJ aJ is an isomorphism. Similarly, aM
aM is an isomorphism. From the diagram we get a surjective map aN N aJ, and
so the image of aN in aJ is N aJ. But this image is aN.
(b) Let a be the ideal generated by the a
i
. Then n N aJ =aN, and so there are
n
i
N such that n =

a
i
n
i
.
(c) We use induction on r. Assume rst that r =1, and write
n
1
=

iJ
0
a
i
,
i
where (,
i
)
iJ
is a basis for J and 1
0
is a nite subset of 1. Then
n
1
=

iJ
0
a
i
n
t
i
11 FINITELY GENERATED PROJECTIVE MODULES 50
for some n
t
i
N (by (b)), and may be taken to be the map such that (,
i
) =n
t
i
for i 1
0
and (,
i
) =0 otherwise. Now suppose that r >1, and that there are maps
1
.
2
: J N
such that
1
(n
1
) =n
1
and

2
(n
i

1
(n
i
)) =n
i

1
(n
i
). i =2. . . . r.
Then
: J N. =
1

1
has the required property.
2
We now complete the proof of the theorem.
(d)=(a). Because M is nitely presented, there is an exact sequence
0 N J M 0
in which J is free and N and J are both nitely generated. Because M is at, (c) of the
lemma shows that this sequence splits, and so M is projective.
(d)=(b). We may suppose that itself is local, with maximal ideal m. Let .
1
. . . . . .
i

M be such that their images in M,mM form a basis for this over the eld ,m. Then the
.
i
generate M (by Nakayamas lemma), and so there exists an exact
0 N J

M 0
in which J is free with basis {,
1
. . . . . ,
i
] and g(,
i
) =.
i
. According to (a) of the lemma,
mN =N (mJ), which equals N because N mJ . Therefore N is zero by Nakayamas
lemma.
EXAMPLE 11.6. (a) When regarded as a Z-module, Q is at but not projective (it is not
nitely generated, much less nitely presented, and so this doesnt contradict the theorem).
(b) Let 1 be a product of copies of F
2
indexed by N, and let a be the ideal in 1 consisting
of the elements (a
n
)
nN
such that a
n
is nonzero for only nitely many values of n (so a is a
direct sum of copies of F
2
indexed by N). The 1-module 1,1 is nitely generated and at,
but not projective (it is not nitely presented, and so this doesnt contradict the theorem).
ASIDE 11.7. An -module M is nitely generated and projective if and only if Hom(M. ) com-
mutes with arbitrary set-indexed direct sums (check; cf Keller 1998, 6.3).
ASIDE 11.8. Nonfree projective nitely generated modules are common: for example, the ideals
in a Dedekind domain are projective and nitely generated, but they are free only if principal. The
situation with modules that are not nitely generated is quite different: if is a noetherian ring with
no nontrivial idempotents, then every nonnitely generated projective -module is free (Bass, Hyman.
Big projective modules are free. Illinois J. Math. 7 1963, 2431, Corollary 4.5). The condition on the
idempotents is necessary because, for a ring T, the module
(J)
T
(J)
is not free when the sets
1 and J have different cardinalities.
Duals
The dual Hom
-lin
(M. ) of an -module M is denoted M

.
11 FINITELY GENERATED PROJECTIVE MODULES 51
PROPOSITION 11.9. For any -modules M, S, T with M nitely generated and projective,
the canonical maps
Hom
-lin
(S. T

M) Hom
-lin
(S

. T ) (28)
T

M Hom
-lin
(M

. T ) (29)
M

(MT )

(30)
M M

(31)
are isomorphisms.
PROOF. The canonical map (28) sends : S T

M to the map
t
: S

T
such that
t
(s g) =(T g)((s)). It becomes the canonical isomorphism
Hom
-lin
(S. T
n
) Hom
-lin
(S
n
. T )
when M =
n
. It follows that (28) is an isomorphism whenever M is a direct summand of
a nitely generated free module, i.e., whenever M is nitely generated and projective.
The canonical map (29) sends t m to the map (m)t . It is the special case of
(28) in which S =.
The canonical map (30) sends g M

to the map mt (m)g(t ): M

T , and the canonical map (31) sends m to the map (m): M

. Again, it is
obviously an isomorphism if one of M or T is free of nite rank, and hence also if one is a
direct summand of such a module.
2
We let ev: M

M denote the evaluation map m(m).

LEMMA 11.10. Let M and N be modules over commutative ring , and let e: N

M
be an -linear map. There exists at most one -linear map : M

N such that the

composites
M

MN M
e
M
N
1
N MN
e1
N
(32)
are the identity maps on M and N respectively. When such a map exists,
T

N .Hom
-lin
(M. T ) (33)
for all -modules T . In particular,
(N. e) .(M

. ev). (34)
PROOF. From e we get an -linear map
T e: T

M T.
which allows us to dene an -linear map
.
x
: T

N Hom
-lin
(M. T ) (35)
by setting

x
(m) =(T e)(. m). . T

N, m M.
11 FINITELY GENERATED PROJECTIVE MODULES 52
An -linear map : M T denes a map N: M

N T

N, and so a map
: M

N denes an -linear map

( N)((1)): Hom
-lin
(M. T ) T

N. (36)
When the rst (resp. the second) composite in (32) is the identity, then (36) is a right
(resp. a left) inverse to (35).
16
Therefore, when a map exists with the required properties,
the map (35) dened by e is an isomorphism. In particular, e denes an isomorphism
.
x
: M

N Hom
-lin
(M. M).
which sends (a) to the endomorphism . a. of M. This proves that is unique.
To get (34), take T =M in (33).
2
PROPOSITION 11.11. An -module M is nitely generated and projective if and only if
there exists an -linear map : MM

such that
(Mev) ( M) =id

and
(M

) (evM

) =id

_ .
PROOF. ==: Suppose rst that M is free with nite basis (e
i
)
iJ
, and let (e
t
i
)
iJ
be the
dual basis of M

. The linear map : MM

, 1 e
i
e
t
i
, satises the conditions.
Let (
i
)
iJ
be as in (11.4c). Then is dened for each module M
(
i
, and the uniqueness
assertion in Lemma 11.10 implies that the s for the different M
(
i
s patch together to give a
for M.
==: On taking T =M in (33), we see that M

M .End
-lin
(M). If

iJ

i
m
i
corresponds to id

, so that

iJ

i
(m)m
i
=m for all m M, then
M
n|-((
i
(n))

J
(o
i
)|-

o
i
n
i
M
is a factorization of id

. Therefore M is a direct summand of a free module of nite rank.

2
ASIDE 11.12. A module M over a ring is said to be reexive if the canonical map M M

is
an isomorphism. We have seen that for nitely generated modules projective implies reexive,
but the converse is false. In fact, for a nite generated module M over an integrally closed noetherian
integral domain , the following are equivalent (Bourbaki AC, VII 4, 2):
16
Assume satises the condition in the statement of the lemma.
Let . T

N; by denition, (
x
N)((1)) =(T e N)(. (1)). On tensoring the second sequence
in (32) with T , we obtain maps
T

N .T

T1
T

N
Te1
T

N
whose composite is the identity map on T

N. As . =. 1 maps to . (1) under T N , this shows

that (
x
N)((1)) =..
Let Hom
-lin
(M. T ), and consider the commutative diagram
T

M T
M M

M M.
T e
e
( 1 (
For m M, the two images of (1) m in T are (m) and
(( 1)((1))
(m), and so =
(( 1)((1))
.
12 ZARISKIS LEMMA AND THE HILBERT NULLSTELLENSATZ 53
(a) M is reexive;
(b) M is torsion-free and equals the intersection of its localizations at the prime ideals of of
height 1;
(c) M is the dual of a nitely generated module.
For noetherian rings of global dimension _ 2, for example, for regular local rings of Krull
dimension _2, every nitely generated reexive module is projective: for every nitely generated
module M over a noetherian ring , there exists an exact sequence

n
M 0
with m. n N; on taking duals and forming the cokernel, we get an exact sequence
0 M

n
N 0:
if has global dimension _2, then M

is projective, and if M is reexive, then M .(M

.
ASIDE 11.13. For a nitely generated torsion-free module M over an integrally closed noetherian
integral domain , there exists a free submodule 1 of M such that M,1 is isomorphic an ideal a in
(Bourbaki AC, VII, 4, Thm 6). When is Dedekind, every ideal is projective, and so M .1a.
In particular, M is projective. Therefore, the nitely generated projective modules over a Dedekind
domain are exactly the nitely generated torsion-free modules.
SUMMARY 11.14. Here is a summary of the assumptions under which the canonical mor-
phisms of -modules below are isomorphisms. If 1 is nitely generated projective:
1
:
1

A module 1 is nitely generated projective if and only if the following canonical map is an
isomorphism
1

1
:
End(1).
If 1 or 1
t
is nitely generated projective:
1

1
t
:
Hom(1. 1
t
).
If both 1 and 1
t
or both 1 and M or both 1
t
and M
t
are nitely generated projective
Hom(1. M) Hom(1
t
. M
t
)
:
Hom(1 1
t
. MM
t
).
In particular, for 1 or 1
t
nitely generated projective
1

1
t
:
(1 1
t
)

.
(Georges Elencwajg on mathoverow.net).
12 Zariskis lemma and the Hilbert Nullstellensatz
Zariskis lemma
In proving Zariskis lemma, we shall need to use that the ring kX| contains innitely many
distinct monic irreducible polynomials. When k is innite, this is obvious, because the
polynomials Xa, a k, are distinct and irreducible. When k is nite, we can adapt Euclids
argument: if
1
. . . . .
i
are monic irreducible polynomials in kX|, then
1

i
1 is
divisible by a monic irreducible polynomial distinct from
1
. . . . .
i
.
12 ZARISKIS LEMMA AND THE HILBERT NULLSTELLENSATZ 54
THEOREM 12.1 (ZARISKIS LEMMA). Let k 1 be elds. If 1 is nitely generated as a
k-algebra, then it is algebraic over k (hence nite over k, and it equals k if k is algebraically
closed).
PROOF. We shall prove this by induction on r, the smallest number of elements required to
generate 1 as a k-algebra. The case r =0 being trivial, we may suppose that
1 =k.
1
. . . . . .
i
| with r _1.
If 1 is not algebraic over k, then at least one .
i
, say .
1
, is not algebraic over k. Then, k.
1
|
is a polynomial ring in one symbol over k, and its eld of fractions k(.
1
) is a subeld of
1. Clearly 1 is generated as a k(.
1
)-algebra by .
2
. . . . . .
i
, and so the induction hypothesis
implies that .
2
. . . . . .
i
are algebraic over k(.
1
). According to Proposition 6.6, there exists a
c k.
1
| such that c.
2
. . . . . c.
i
are integral over k.
1
|. Let 1. For a sufciently large
N, c
1
k.
1
. c.
2
. . . . . c.
i
|, and so c
1
is integral over k.
1
| by 6.4. When we apply this
statement to an element of k(.
1
), it shows that c
1
k.
1
| because k.
1
| is integrally
closed. Therefore, k(.
1
) =
_
1
c
-1
k.
1
|, but this is absurd, because k.
1
| (.kX|) has
innitely many distinct monic irreducible polynomials that can occur as denominators of
elements of k(.
1
).
2
COROLLARY 12.2. Let be a nitely generated k-algebra. Every maximal ideal in is
the kernel of a homomorphism from into a nite eld extension of k.
PROOF. Indeed, ,m itself is a nite eld extension of k.
2
Alternative proof of Zariskis lemma
The following is a simplication of Swans simplication of a proof of Munshi see Swan.
LEMMA 12.3. For an integral domain , there does not exist an X| such that X|
(
is a eld.
PROOF. Suppose, on the contrary, that X|
(
is a eld. Then , and we can write
( 1)
-1
=g,
n
with g X| and n _1. Then
( 1)g =
n
=(1( 1))
n
=1( 1)h
with h X|, and so ( 1)(g h) =1. Hence 1 is a unit in X|, which is absurd
(it has degree _1).
2
LEMMA 12.4. Consider rings T. If T is integral over , then T

. In
particular, if T is a eld, then so also is .
PROOF. Let a be an element of that becomes a unit in T, say, ab =1 with b T. There
exist a
1
. . . . . a
n
such that b
n
a
1
b
n-1
a
n
=0. On multiplying through by a
n-1
,
we nd that b =a
1
a
n
a
n-1
, and so a

.
2
PROPOSITION 12.5. Let be an integral domain, and suppose that there exists a maximal
ideal m in X
1
. . . . . X
n
| such that m=(0). Then there exists a nonzero element a in
such that
o
is a eld and X
1
. . . . . X
n
|,m is a nite extension of
o
.
12 ZARISKIS LEMMA AND THE HILBERT NULLSTELLENSATZ 55
PROOF. Note that the condition m=(0) implies that (hence also
o
) is a subring of
the eld 1 =X
1
. . . . . X
n
|,m, and so the statement makes sense.
We argue by induction on n. When n =0, the hypothesis is that (0) is a maximal ideal in
; hence is a eld, and the statement is trivial. Therefore, suppose that n _1, and regard
X
1
. . . . . X
n
| as a polynomial ring in n1 symbols over X
i
|. Then mX
i
| =(0)
because otherwise the induction hypothesis would contradict Lemma 12.3. Let a
i
X
n
i
i

be a nonzero element of mX
i
|. The image .
i
of X
i
in 1 satises the equation
a
i
.
n
i
=0.
and so 1 is integral over its subring
o
1
o
n
. By Lemma 12.4,
o
1
o
n
is a eld, and 1 is
nite over it because it is integral (algebraic) and nitely generated.
2
We now prove Zariskis lemma. Write 1 =kX
1
. . . . . X
n
|,m. According to the proposi-
tion, 1 is a nite extension of k
o
for some nonzero a k, but because k is a eld k
o
=k.
The Nullstellensatz
Recall that k
al
denotes an algebraic closure of the eld k.
THEOREM 12.6 (NULLSTELLENSATZ). Every proper ideal a in kX
1
. . . . . X
n
| has a zero
in (k
al
)
n
, i.e., there exists a point (a
1
. . . . . a
n
) (k
al
)
n
such that (a
1
. . . . . a
n
) =0 for all
a.
PROOF. We have to show that there exists a k-algebra homomorphism kX
1
. . . . . X
n
| k
al
containing a in its kernel. Let m be a maximal ideal containing a. Then kX
1
. . . . . X
n
|,m
is a eld, which is nite over k by Zariskis lemma, and so there exists a k-algebra
homomorphism kX
1
. . . . . X
n
|,m k
al
. The composite of this with the quotient map
kX
1
. . . . . X
n
| kX
1
. . . . . X
n
|,m contains a in its kernel.
2
COROLLARY 12.7. When k is algebraically closed, the maximal ideals in kX
1
. . . . . X
n
|
are exactly the ideals (X
1
a
1
. . . . . X
n
a
n
), (a
1
. . . . . a
n
) k
n
.
PROOF. Clearly, kX
1
. . . . . X
n
|,(X
1
a
1
. . . . . X
n
a
n
) . k, and so (X
1
a
1
. . . . . X
n

a
n
) is maximal. Conversely, because k is algebraically closed, a maximal ideal m of
kX
1
. . . . . X
n
| has a zero (a
1
. . . . . a
n
) in k
n
. Let kX
1
. . . . . X
n
|; when we write as a
polynomial in X
1
a
1
. . . . . X
n
a
n
, its constant term is (a
1
. . . . . a
n
). Therefore
m == (X
1
a
1
. . . . . X
n
a
n
),
and so m=(X
1
a
1
. . . . . X
n
a
n
).
2
THEOREM 12.8 (STRONG NULLSTELLENSATZ). For an ideal a in kX
1
. . . . . X
n
|, let 7(a)
be the set of zeros of a in (k
al
)
n
. If a polynomial h kX
1
. . . . . X
n
| is zero on 7(a), then
some power of h lies in a.
PROOF. We may assume h = 0. Let g
1
. . . . . g
n
generate a, and consider the system of
m1 equations in n1 variables, X
1
. . . . . X
n
. Y.
_
g
i
(X
1
. . . . . X
n
) = 0. i =1. . . . . m
1Yh(X
1
. . . . . X
n
) = 0.
12 ZARISKIS LEMMA AND THE HILBERT NULLSTELLENSATZ 56
If (a
1
. . . . . a
n
. b) satises the rst m equations, then (a
1
. . . . . a
n
) 7(a); consequently,
h(a
1
. . . . . a
n
) =0, and (a
1
. . . . . a
n
. b) doesnt satisfy the last equation. Therefore, the equa-
tions are inconsistent, and so, according to the Nullstellensatz (12.6), the ideal
(g
1
. . . . . g
n
. 1Yh) =kX
1
. . . . . X
n
. Y |.
and so there exist
i
kX
1
. . . . . X
n
. Y | such that
1 =
n

i=1

i
g
i

n1
(1Yh). (37)
On applying the homomorphism
_
X
i
X
i
Y h
-1
: kX
1
. . . . . X
n
. Y | k(X
1
. . . . . X
n
)
to (37), we obtain the identity
1 =

i
(X
1
. . . . . X
n
. h
-1
) g
i
(X
1
. . . . . X
n
) (38)
in k(X
1
. . . . . X
n
). Clearly

i
(X
1
. . . . . X
n
. h
-1
) =
polynomial in X
1
. . . . . X
n
h
1
i
for some N
i
. Let N be the largest of the N
i
. On multiplying (38) by h
1
we obtain an
identity
h
1
=

i
(polynomial in X
1
. . . . . X
n
) g
i
(X
1
. . . . . X
n
).
which shows that h
1
a.
2
PROPOSITION 12.9. The radical of an ideal a in a nitely generated k-algebra is equal
to the intersection of the maximal ideals containing it: rad(a) =
_
ma
m. In particular, if
is reduced, then
_
m maximal
m=0.
PROOF. Because of the correspondence (2), p. 4, it sufces to prove this for =kX
1
. . . . . X
n
|.
The inclusion rad(a)
_
ma
m holds in any ring (because maximal ideals are radical
and rad(a) is the smallest radical ideal containing a). Let a be an ideal in kX
1
. . . . . X
n
|, and
let h
_
ma
m. For any (a
1
. . . . . a
n
) 7(a), the evaluation map
(a
1
. . . . . a
n
): kX
1
. . . . . X
n
| k
al
has image a subring of k
al
which is algebraic over k, and hence is a eld (see 1). Therefore,
the kernel of the map is a maximal ideal, which contains a, and therefore also contains h.
This shows that h(a
1
. . . . . a
n
) = 0, and we conclude from the strong Nullstellensatz that
h rad(a).
2
13 THE SPECTRUM OF A RING 57
13 The spectrum of a ring
Denition
Let be a ring, and let V be the set of prime ideals in . For an ideal a in , let
V(a) ={p V [ p a].
PROPOSITION 13.1. There are the following relations:
(a) a b == V(a) V(b):
(b) V(0) =V ; V() =0:
(c) V(ab) =V(ab) =V(a) LV(b):
(d) V(

iJ
a
i
) =
_
iJ
V(a
i
) for every family of ideals (a
i
)
iJ
.
PROOF. The rst two statements are obvious. For (c), note that
ab ab a. b == V(ab) V(ab) V(a) LV(b).
For the reverse inclusions, observe that if p V(a) LV(b), then there exist an ap and
a g bp; but then g abp, and so p V(ab). For (d) recall that, by denition,

a
i
consists of all nite sums of the form

i
,
i
a
i
. Thus (d) is obvious.
2
Statements (b), (c), and (d) show that the sets V(a) satisfy the axioms to be the closed
subsets for a topology on V : both the whole space and the empty set are closed; a nite
union of closed sets is closed; an arbitrary intersection of closed sets is closed. This topology
is called the Zariski topology on V . We let spec() denote the set of prime ideals in
endowed with its Zariski topology.
For h , let
D(h) ={p V [ h p].
Then D(h) is open in V , being the complement of V((h)). If S is a set of generators for an
ideal a, then
V V(a) =
_
hS
D(h).
and so the sets D(h) form a base for the topology on V . Note that
D(h
1
h
n
) =D(h
1
) D(h
n
).
For every element h of , spec(
h
) .D(h) (see 5.4), and for every ideal a in , spec(),a .
V(a) (isomorphisms of topological spaces).
The ideals in a nite product of rings =
1

n
are all of the form a
1
a
n
with a
i
an ideal in
i
(cf. p.8). The prime (resp. maximal) ideals are those of the form

1

i-1
a
i

i1

n
with a
i
prime (resp. maximal). It follows that spec() =

i
spec(
i
) (disjoint union of
open subsets).
13 THE SPECTRUM OF A RING 58
Idempotents and connected components
Let be a ring. In 1, we saw that complete sets of orthogonal idempotents in correspond
to decompositions of into a nite product of rings. We now see that they also correspond
to decompositions of spec into a nite disjoint union of open subsets.
LEMMA 13.2. The space spec is disconnected if and only if contains a nontrivial
idempotent.
PROOF. Let e be a nontrivial nilpotent, and let =1 e. For a prime ideal p, the map
,p must send exactly one of e or to a nonzero element. This shows that spec is a
disjoint union of the sets
17
D(e) and D( ), each of which is open. If D(e) =spec, then
e would be a unit (2.2), and hence can be cancelled from ee =e to give e =1. Therefore
D(e) =spec, and similarly, D( ) =spec.
Conversely, suppose that spec is disconnected, say, the disjoint union of two nonempty
closed subsets V(a) and V(b). Because the union is disjoint, no prime ideal contains both
a and b, and so a b =. Thus a b =1 for some a a and b b. As ab a b, all
prime ideals contain ab, which is therefore nilpotent (2.4), say (ab)
n
=0. Any prime ideal
containing a
n
contains a; similarly, any prime ideal containing b
n
contains b; thus no prime
ideal contains both a
n
and b
n
, which shows that (a
n
. b
n
) =. Therefore, 1 =ra
n
sb
n
for some r. s . Now
(ra
n
)(sb
n
) =rs(ab)
n
=0.
(ra
n
)
2
=(ra
n
)(1sb
n
) =ra
n
,
(sb
n
)
2
=sb
n
ra
n
sb
n
=1.
and so {ra
n
. sb
n
] is a complete set of orthogonal idempotents. Clearly V(a) V(ra
n
)
and V(b) V(sb
n
). As V(ra
n
) V(sb
n
) =0, we see that V(a) =V(ra
n
) and V(b) =
V(sb
n
), and so each of ra
n
and sb
n
is a nontrivial idempotent.
2
PROPOSITION 13.3. Let {e
1
. . . . . e
n
] be a complete set of orthogonal idempotents in .
Then
spec =D(e
1
) L. . . LD(e
n
)
is a decomposition of spec into a disjoint union of open subsets. Moreover, every such
decomposition arises in this way.
PROOF. Let p be a prime ideal in . Because ,p is an integral domain, exactly one of the
e
i
s maps to 1 in ,p and the remainder map to zero. This proves that spec is the disjoint
union of the sets D(e
i
).
Now consider a decomposition
spec =U
1
L. . . LU
n
each U
i
open. We use induction on n to show that it arises from a complete set of orthogonal
idempotents. When n =1, there is nothing to prove, and when n _2, we write
spec =U
1
L(U
2
L. . . LU
n
).
17
The set D(e) consists of the prime ideals of not containing e, and V(a) consists of all prime ideals
containing a.
13 THE SPECTRUM OF A RING 59
The proof of the lemma shows that there exist orthogonal idempotents e
1
, e
t
1
such that
e
1
e
t
1
=1 and
U
1
=D(e
1
)
U
2
L. . . LU
n
=D(e
t
1
) =spece
t
1
.
By induction, there exist orthogonal idempotents e
2
. . . . . e
n
in e
t
1
such that e
2
e
n
=
e
t
1
and U
i
= D(e
i
) for i = 2. . . . . n. Now {e
1
. . . . . e
n
] is a complete set of orthogonal
idempotents in such that U
i
=D(e
i
) for all i .
2
The topological space spec()
We study more closely the Zariski topology on spec(). For each subset S of , let V(S)
denote the set of prime ideals containing S, and for each subset W of spec(), let 1(W)
denote the intersection of the prime ideals in W:
S . V(S) ={p spec() [ S p].
W spec(). 1(W) =
_
pW
p.
Thus V(S) is a closed subset of spec() and 1(W) is a radical ideal in . If V(a) W, then
a 1(W), and so V(a) V1(W). Therefore V1(W) is the closure of W (smallest closed
subset of spec() containing W); in particular, V1(W) =W if W is closed.
PROPOSITION 13.4. Let V be a closed subset of spec().
(a) There is an order-inverting one-to-one correspondence W -1(W) between the closed
subsets of spec() and the radical ideals in .
(b) The closed points of V are exactly the maximal ideals in V .
(c) If is noetherien, then every ascending chain of open subsets U
1
U
2
of V
eventually becomes constant; equivalently, every descending chain of closed subsets
of V eventually becomes constant.
(d) If is noetherian, every open covering of V has a nite subcovering.
PROOF. (a) and (b) are obvious.
(c) We prove the second statement. A sequence V
1
V
2
of closed subsets of V
gives rise to a sequence of ideals 1(V
1
) 1(V
2
) . . ., which eventually becomes constant.
If 1(V
n
) =1(V
n1
), then V1(V
n
) =V1(V
n1
), i.e., V
n
=V
n1
.
(d) Let V =
_
iJ
U
i
with each U
i
open. Choose an i
0
1; if U
i
0
= V , then there
exists an i
1
1 such that U
i
0
U
i
0
LU
i
1
. If U
i
0
LU
i
1
=V , then there exists an i
2
1 etc..
Because of (c), this process must eventually stop.
2
A topological space V having the property (b) is said to be noetherian. This condition is
equivalent to the following: every nonempty set of closed subsets of V has a minimal element.
A topological space V having property (c) is said to be quasi-compact (by Bourbaki at least;
others call it compact, but Bourbaki requires a compact space to be Hausdorff). The proof
of (d) shows that every noetherian space is quasi-compact. Since an open subspace of a
noetherian space is again noetherian, it will also be quasi-compact.
18
18
In fact, spec() is always quasi-compact. To see this, let (U
i
)
iJ
be an open covering of spec(). On
covering each U
i
with basic open subsets, we get a covering (D(h
}
))
}J
of spec() by basic open subsets.
Because spec() =
_
}
D(h
}
), the ideal generated by the h
}
is , and so 1 =a
1
h
}
1
a
n
h
}
m
for some
a
1
. . . . . a
n
. Therefore spec() =
_
1_I_n
D(h
}
l
), and it follows that spec() is covered by nitely many
of the sets U
i
.
13 THE SPECTRUM OF A RING 60
DEFINITION 13.5. A nonempty topological space is said to be irreducible if it is not the
union of two proper closed subsets. Equivalent conditions: any two nonempty open subsets
have a nonempty intersection; every nonempty open subset is dense.
If an irreducible space W is a nite union of closed subsets, W =W
1
L. . . LW
i
, then
W =W
1
or W
2
L. . . LW
i
; if the latter, then W =W
2
or W
3
L. . . LW
i
, etc.. Continuing in
this fashion, we nd that W =W
i
for some i .
The notion of irreducibility is not useful for Hausdorff topological spaces, because the
only irreducible Hausdorff spaces are those consisting of a single point two points would
have disjoint open neighbourhoods.
PROPOSITION 13.6. A closed subset W of specm() is irreducible if and only if it is
irreducible. In particular, the spectrum of a ring is irreducible if and only if the nilradical
of is prime.
PROOF. =: Let W be an irreducible closed subset of spec(), and suppose that g 1(W).
Then g lies in each m in W, and so either m or g m; hence W V( ) LV(g), and
so
W =(W V( )) L(W V(g)).
As W is irreducible, one of these sets, say W V( ), must equal W. But then 1(W).
We have shown that 1(W) is prime.
=: Assume 1(W) is prime, and suppose that W =V(a) LV(b) with a and b radical
ideals we have to show that W equals V(a) or V(b). Recall that V(a) LV(b) =V(ab)
(see 13.1c) and that ab is radical; hence 1(W) =ab (by 14.2). If W =V(a), then there
exists an a1(W). For all g b,
g ab =1(W).
Because 1(W) is prime, this implies that b 1(W); therefore W V(b).
2
Thus, in the spectrum of a ring, there are one-to-one correspondences
radical ideals - closed subsets
prime ideals - irreducible closed subsets
maximal ideals - one-point sets.
EXAMPLE 13.7. Let kX
1
. . . . . X
n
|. According to Theorem 4.9, kX
1
. . . . . X
n
| is a
unique factorization domain, and so ( ) is a prime ideal if and only if is irreducible (4.1).
Thus
V( ) is irreducible is irreducible.
On the other hand, suppose that factors as
=

n
i
i
.
i
distinct irreducible polynomials.
Then
( ) =
_
(
n
i
i
). (
n
i
i
) distinct ideals,
rad(( )) =
_
(
i
). (
i
) distinct prime ideals,
V( ) =
_
V(
i
). V(
i
) distinct irreducible algebraic sets.
13 THE SPECTRUM OF A RING 61
PROPOSITION 13.8. Let V be a noetherian topological space. Then V is a nite union of
irreducible closed subsets, V =V
1
L. . . LV
n
. If the decomposition is irredundant in the
sense that there are no inclusions among the V
i
, then the V
i
are uniquely determined up to
order.
PROOF. Suppose that V can not be written as a nite union of irreducible closed subsets.
Then, because V is noetherian, there will be a closed subset W of V that is minimal
among those that cannot be written in this way. But W itself cannot be irreducible, and so
W =W
1
LW
2
, with each W
i
a proper closed subset of W. Because W is minimal, both W
1
and W
2
can be expressed as nite unions of irreducible closed subsets, but then so can W.
We have arrived at a contradiction.
Suppose that
V =V
1
L. . . LV
n
=W
1
L. . . LW
n
are two irredundant decompositions. Then V
i
=
_
}
(V
i
W
}
), and so, because V
i
is
irreducible, V
i
=V
i
W
}
for some . Consequently, there exists a function : {1. . . . . m]
{1. . . . . n] such that V
i
W
((i)
for each i . Similarly, there is a function g: {1. . . . . n]
{1. . . . . m] such that W
}
V
(})
for each . Since V
i
W
((i)
V
((i)
, we must have
g(i ) = i and V
i
= W
((i)
; similarly g = id. Thus and g are bijections, and the
decompositions differ only in the numbering of the sets.
2
The V
i
given uniquely by the proposition are called the irreducible components of V .
They are the maximal closed irreducible subsets of V . In Example 13.7, the V(
i
) are the
irreducible components of V( ).
COROLLARY 13.9. A radical ideal a in a noetherian ring is a nite intersection of prime
ideals, a =p
1
. . . p
n
; if there are no inclusions among the p
i
, then the p
i
are uniquely
determined up to order.
PROOF. Write V(a) as a union of its irreducible components, V(a) =
_
V
i
, and take p
i
=
1(V
i
).
2
In particular, a noetherian ring has only nitely many minimal prime ideals, and their
intersection is the radical of the ring.
COROLLARY 13.10. A noetherian topological space has only nitely many connected
components (each of which is open).
PROOF. Each connected component is closed, hence noetherian, and so is a nite union of
its irreducible components. Each of these is an irreducible component of the whole space,
and so there can be only nitely many.
2
REMARK 13.11. (a) An irreducible topological space is connected, but a connected topo-
logical space need not be irreducible. For example, 7(X
1
X
2
) is the union of the coordinate
axes in k
2
, which is connected but not irreducible. A closed subset V of spec() is not
connected if and only if there exist ideals a and b such that ab =1(V ) and ab =.
(b) A Hausdorff space is noetherian if and only if it is nite, in which case its irreducible
components are the one-point sets.
(c) In a noetherian ring, every proper ideal a has a decomposition into primary ideals:
a =
_
q
i
(see 17). For radical ideals, this becomes a simpler decomposition into prime
ideals, as in the corollary. For an ideal ( ) in kX
1
. . . . . X
n
| with =

n
i
i
, it is the
decomposition ( ) =
_
(
n
i
i
) noted in Example 13.7.
14 JACOBSON RINGS AND MAX SPECTRA 62
14 Jacobson rings and max spectra
DEFINITION 14.1. A ring is Jacobson if every prime ideal in is an intersection of
maximal ideals.
A eld is Jacobson. The ring Z is Jacobson because every nonzero prime ideal is
maximal and (0) =
_
]=2,3,5,...
(). A principal ideal domain (more generally, a Dedekind
domain) is Jacobson if it has innitely many maximal ideals.
19
A local ring is Jacobson
if and only if its maximal ideal is its only prime ideal. Proposition 12.9 shows that every
nitely generated algebra over a eld is Jacobson.
PROPOSITION 14.2. The radical of an ideal in a Jacobson ring is equal to the intersection of
the maximal ideals containing it. (Therefore, the radical ideals are precisely the intersections
of maximal ideals.)
PROOF. Proposition 2.4 says that the radical of an ideal is an intersection of prime ideals,
and so this follows from the denition of a Jacobson ring.
2
In a Jacobson ring , there are natural one-to-one correspondences between
the decompositions of spm() into a nite disjoint union of open subspaces,
the decompositions of into a nite direct products of rings, and
the complete sets of orthogonal idempotents in .
ASIDE 14.3. Any ring of nite type over a Jacobson ring is a Jacobson ring (EGA IV 10.4.6).
Moreover, if T is of nite type over and is Jacobson, then the map T denes a continuous
map specm(T) specm().
ASIDE 14.4. The spectrum spec() of a ring is the set of prime ideals in endowed with the
topology for which the closed subsets are those of the form
V(a) ={p [ p a]. a an ideal in .
Thus specm() is the subspace of spec() consisting of the closed points. When is Jacobson,
the map U U specm() is a bijection from the set of open subsets of spec() onto the set of
open subsets of specm(); therefore specm() and spec() have the same topologies only the
underlying sets differ.
ASIDE 14.5. Let k =R or C. Let X be a set and let be a k-algebra of k-valued functions on X.
In analysis, X is called the spectrum of if, for each k-algebra homomorphism : k, there
exists a unique . X such that ( ) =(.) for all , and every . arises from a (cf. Cartier
2007, 3.3.1, footnote).
Let be a nitely generated algebra over an arbitrary algebraically closed eld k, and let
X =specm(). An element of denes a k-valued function
m modm
on X. When is reduced, Proposition 12.9 shows that this realizes as a ring of k-valued functions
on X. Moreover, because (40) is an isomorphism in this case, for each k-algebra homomorphism
: k, there exists a unique . X such that ( ) =(.) for all . In particular, when
k =C and is reduced, specm() is the spectrum of in the sense of analysis.
19
In a principal ideal domain, a nonzero element a factors as a =u
i
1
1

i
s
x
with u a unit and the
i
prime.
The only prime divisors of a are
1
. . . . .
x
, and so a is contained in only nitely many prime ideals. Similarly,
in a Dedekind domain, a nonzero ideal a factors as a =p
i
1
1
p
i
s
x
with the p
i
prime ideals (cf. 18.7 below), and
p
1
. . . . . p
i
are the only prime ideals containing a. On taking a =(a), we see that again a is contained in only
nitely many prime ideals.
14 JACOBSON RINGS AND MAX SPECTRA 63
The max spectrum of a nitely generated k-algebra
Let k be a eld, and let be a nitely generated k-algebra. For any maximal ideal m of ,
the eld k(m)
def
=,m is a nitely generated k-algebra, and so k(m) is nite over k (Zariskis
lemma, 12.1). In particular, it equals k(m) =k when k is algebraically closed.
Now x an algebraic closure k
al
. The image of any k-algebra homomorphism k
al
is a subring of k
al
which is an integral domain algebraic over k and therefore a eld (see
1). Hence the kernel of the homomorphism is a maximal ideal in . In this way, we get a
surjective map
Hom
k-alg
(. k
al
) specm(). (39)
Two homomorphisms k
al
with the same kernel m factor as
k(m) k
al
.
and so differ by an automorphism
20
of k
al
. Therefore, the bres of (39) are exactly the orbits
of Gal(k
al
,k). When k is perfect, each extension k(m),k is separable, and so each orbit
has k(m): k| elements, and when k is algebraically closed, the map (39) is a bijection.
Set =kX
1
. . . . . X
n
|,a. Then to give a homomorphism k
al
is the same as giving
an n-tuple (a
1
. . . . . a
n
) of elements of k
al
(the images of the X
i
) such that (a
1
. . . . . a
n
) =0
for all a, i.e., an element of the zero-set 7(a) of a. The homomorphism corresponding
to (a
1
. . . . . a
n
) maps k(m) isomorphically onto the subeld of k
al
generated by the a
i
s.
Therefore, we have a canonical surjection
7(a) specm() (40)
whose bres are the orbits of Gal(k
al
,k). When the eld k is perfect, each orbit has
ka
1
. . . . . a
n
| : k|-elements, and when k is algebraically closed, 7(a) .specm().
Maps of max spectra
Let : T be a homomorphism of nitely generated k-algebras (k a eld). Because T is
nitely generated over k, its quotient T,m by any maximal ideal m is a nite eld extension
of k (Zariskis lemma, 12.1). Therefore the image of in T,m is an integral domain nite
over k, and hence is a eld (see 1). Since this image is isomorphic to ,
-1
(m), this
shows that the ideal
-1
(m) is maximal in . Therefore denes a map

+
: specm(T) specm(). m
-1
(m).
which is continuous because (
+
)
-1
(D( )) = D(( )). In this way, specm becomes a
functor from nitely generated k-algebras to topological spaces.
THEOREM 14.6. Let : T be a homomorphism of nitely generated k-algebras. Let
U be a nonempty open subset of specm(T), and let
+
(U)
-
be the closure of its image in
specm(). Then
+
(U) contains a nonempty open subset of each irreducible component of

+
(U)
-
.
20
Let and g be two k-homomorphisms from a nite eld extension k
t
of k into k
al
. We consider the set
of pairs (1. ) in which is a k-homomorphism from a subeld 1 of k
al
containing (k
t
) into k
al
such that
=g. The set is nonempty, and Zorns lemma can be applied to show that it has a maximal element (1
t
.
t
).
For such an element 1
t
will be algebraically closed, and hence equal to k
al
.
14 JACOBSON RINGS AND MAX SPECTRA 64
PROOF. Let W =specm(T) and V =specm(), so that
+
is a continuous map W V .
We rst prove the theorem in the case that is an injective homomorphism of integral
domains. For some b =0, D(b) U. According to Proposition 14.7 below, there exists
a nonzero element a such that every homomorphim : k
al
such that (a) = 0
extends to a homomorphism : T k
al
such that (b) =0. Let m D(a), and choose to
be a homomorphism k
al
with kernel m. The kernel of is a maximal ideal n D(b)
such that
-1
(n) =m, and so D(a)
+
(D(b)).
We now prove the general case. If W
1
. . . . . W
i
are the irreducible components of W,
then
+
(W)
-
is a union of the sets
+
(W
i
)
-
, and any irreducible component C of
+
(U)
-
is contained in one of
+
(W
i
)
-
, say
+
(W
1
)
-
. Let q =1(W
1
) and let p =
-1
(q). Because
W
1
is irreducible, they are both prime ideals. The homomorphism : T induces an
injective homomorphism : ,p T,q, and
+
can be identied with the restriction of
+
to W
1
. From the rst case, we know that
+
(U W
1
) contains a nonempty open subset of
C, which implies that
+
(U) does also.
2
In the next two statements, and T are arbitrary commutative rings they need not be
k-algebras.
PROPOSITION 14.7. Let T be integral domains with T nitely generated as an algebra
over , and let b be a nonzero element of T. Then there exists an element a = 0 in
with the following property: every homomorphism : from into an algebraically
closed eld such that (a) =0 can be extended to a homomorphism : T such that
(b) =0.
We rst need a lemma.
LEMMA 14.8. Let T be integral domains, and assume T =t | =T |,a. Let c
be the ideal of leading coefcients of the polynomials in a. Then every homomorphism
: from into an algebraically closed eld such that (c) =0 can be extended
to a homomorphism of T into .
PROOF. If a =0, then c =0, and every extends. Thus we may assume a =0. Let be a
homomorphism such that (c) =0. Then there exist polynomials a
n
T
n
a
0
in a such that (a
n
) =0, and we choose one, denoted , of minimum degree. Because
T =0, the polynomial is nonconstant.
Extend to a homomorphism T | T |, again denoted , by sending T to T , and
consider the subset (a) of T |.
FIRST CASE: (a) DOES NOT CONTAIN A NONZERO CONSTANT. If the -subspace
of T | spanned by (a) contained 1, then so also would (a),
21
contrary to hypothesis.
Because
T

c
i
(g
i
) =

c
i
(g
i
T ). c
i
. g
i
a.
this -subspace an ideal, which we have shown to be proper, and so it has a zero c in .
The composite of the homomorphisms
T |

T | . T T c.
factors through T |,a =T and extends .
21
Use that, if a system of linear equation with coefcients in a eld k has a solution in some larger eld, then
it has a solution in k.
14 JACOBSON RINGS AND MAX SPECTRA 65
SECOND CASE: (a) CONTAINS A NONZERO CONSTANT. This means that a contains a
polynomial
g(T ) =b
n
T
n
b
0
such that (b
0
) =0. (b
1
) =(b
2
) = =0.
On dividing (T ) into g(T ) we obtain an equation
a
d
n
g(T ) =q(T )(T ) r(T ). J N. q. r T |. degr <m.
When we apply , this becomes
(a
n
)
d
(b
0
) =(q)( ) (r).
Because ( ) has degree m>0, we must have (q) =0, and so (r) is a nonzero constant.
After replacing g(T ) with r(T ), we may suppose that n <m. If m=1, such a g(T ) cant
exist, and so we may suppose that m>1 and (by induction) that the lemma holds for smaller
values of m.
For h(T ) = c
i
T
i
c
i-1
T
i-1
c
0
, let h
t
(T ) = c
i
c
0
T
i
. Then the -
module generated by the polynomials T
x
h
t
(T ), s _0, h a, is an ideal a
t
in T |. Moreover,
a
t
contains a nonzero constant if and only if a contains a nonzero polynomial cT
i
, which
implies t =0 and =T (since T is an integral domain).
When a
t
does not contain a nonzero constant, we set T
t
=T |,a
t
=t
t
|. Then a
t
contains the polynomial g
t
= b
n
b
0
T
n
, and (b
0
)= 0. Because degg
t
< m, the
induction hypothesis implies that extends to a homomorphism T
t
. Therefore, there
exists a c such that, for all h(T ) =c
i
T
i
c
i-1
T
i-1
c
0
a,
h
t
(c) =(c
i
) (c
i-1
)c c
0
c
i
=0.
On taking h = g, we see that c = 0, and on taking h = , we obtain the contradiction
(a
n
) =0.
2
PROOF (OF 14.7). Suppose that we know the proposition in the case that T is generated by
a single element, and write T =t
1
. . . . . t
n
|. Then there exists an element b
n-1
such that any
homomorphism : t
1
. . . . . t
n-1
| such that (b
n-1
) =0 extends to a homomorphism
: T such that (b) =0. Continuing in this fashion (with b
n-1
for b), we eventually
obtain an element a with the required property.
Thus we may assume T = t |. Let a be the kernel of the homomorphism T t ,
T | t |.
Case (i). The ideal a =(0). Write
b =(t ) =a
0
t
n
a
1
t
n-1
a
n
. a
i
.
and take a =a
0
. If : is such that (a
0
) =0, then there exists a c such that
(c) =0, and we can take to be the homomorphism

J
i
t
i

(J
i
)c
i
.
Case (ii). The ideal a =(0). Let (T ) =a
n
T
n
a
0
, a
n
=0, be an element of
a of minimum degree. Let h(T ) T | represent b. Since b = 0, h a. Because is
irreducible over the eld of fractions of , it and h are coprime over that eld. In other
words, there exist u. T | and a nonzero c such that
uh =c.
It follows now that ca
n
satises our requirements, for if (ca
n
) =0, then can be extended
to : T by the lemma, and (u(t ) b) =(c) =0, and so (b) =0.
2
15 QUASI-FINITE ALGEBRAS AND ZARISKIS MAIN THEOREM. 66
REMARK 14.9. In case (ii) of the last proof, both b and b
-1
are algebraic over , and so
there exist equations
a
0
b
n
a
n
=0. a
i
. a
0
=0:
a
t
0
b
-n
a
t
n =0. a
t
i
. a
t
0
=0.
One can show that a =a
0
a
t
0
has the property required by the proposition (cf. AM 5.23,
p.66).
ASIDE 14.10. In general, the map X| does not induce a map spm(X|) spm(). Con-
sider for example a discrete valuation ring with maximal ideal () (e.g., Z
(])
with maximal ideal
()). The ideal (X 1) is maximal, because X|,(X 1) is the eld of fractions of (by 5.3),
but (X 1) =(0), which is not maximal.
15 Quasi-nite algebras and Zariskis main theorem.
In this section we prove a fundamental theorem of Zariski.
22
Throughout, k is a eld and
is a commutative ring.
Quasi-nite algebras
PROPOSITION 15.1. Let T be a nite generated k-algebra. A prime ideal q of T is an
isolated point of spec(T) if and only if T
q
is nite over k.
PROOF. To say that q is an isolated point of spec(T) means that there exists an T q
such that spec(T
(
) = {q]. Now T
(
is noetherian with only one prime ideal, namely,
m
def
=qT
(
, and so it is artinian (7.6). The quotient T
(
,m is a eld which is nitely generated
as a k-algebra, and hence is nite over k (Zariskis lemma 12.1). Because T
(
is artinian,
T
(
mm
2

can be rened to a nite ltration whose quotients are one-dimensional vector spaces over
T
(
,m. Therefore T
(
is a nite k-algebra. As q, we have T
q
=(T
(
)
q
, which equals
T
(
because T
(
is local. Therefore T
q
is also a nite k-algebra.
For the converse, suppose that T
q
is nite over k, and consider the exact seqence
0 M T T
q
N 0 (41)
of T-modules. When we apply the functor S
-1
q
to (41), it remains exact (5.11), but the
middle arrow becomes an isomorphism, and so M
q
=0 =N
q
. Because T is noetherian,
the T-module M is nitely generated, with generators e
1
. . . . . e
n
say. As M
q
=0, there
exists, for each i , an
i
T q such that
i
e
i
=0. Now
t
def
=
1
. . .
n
has the property
that
t
M =0, and so M
(
0 =0.
Because T
q
is a nite k-algebra, N is nitely generated as a k-module, and therefore
also as a T-module. As for M, there exists an
tt
T q such that M
(
00 = 0. Now

def
=
t

tt
T q has the property that M
(
=0 =N
(
. When we apply the functor S
-1
(
to (41), we obtain an isomorphism T
(
.T
q
, and so spec(T
(
) =spec(T
q
) ={q], which
shows that q is an isolated point.
2
22
Our exposition of the proof follows those in Raynaud 1970 and in Hochsters course notes from Winter,
2010.
15 QUASI-FINITE ALGEBRAS AND ZARISKIS MAIN THEOREM. 67
PROPOSITION 15.2. Let T be a nitely generated k-algebra. The space spec(T) is discrete
if and only if T is a nite k-algebra.
PROOF. If T is nite over k, then it is artinian and so (7.7)
T =

{T
m
[ m maximal] (nite product),
and
spec(T) =
_
m
spec(T
m
) =
_
m
{m] (disjoint union of open subsets).
Therefore each point is isolated in spec(T).
Conversely, if spec(T) is discrete then it is a nite disjoint union,
spec(T) =
_
1_i_n
spec(T
(
i
).
i
T.
with spec(T
(
i
) ={q
i
]. Hence T =

1_i_n
T
(
i
(by 13.3) with T
(
i
=T
q
i
. According to
Proposition (15.1), each k-algebra T
q
i
is nite over k, and so T is nite over k.
2
DEFINITION 15.3. Let T be a nitely generated -algebra.
(a) Let q be a prime ideal of T, and let p =q
c
. The ring T is said to be quasi-nite over
at q if T
q
,pT
q
is a nite k(p)-algebra.
(b) The ring T is said to be quasi-nite over if it is quasi-nite over at all the prime
ideals of T.
PROPOSITION 15.4. Let T be a nitely generated -algebra. Let q be a prime ideal of T,
and let p =q
c
. Then T is quasi-nite over at q if and only if q is an isolated point of
spec(T

k(p)).
PROOF. As
T
q
,pT
q
.(T,pT)
q{p
.(T

k(p))
q{p
.
this is an immediate consequence of (15.1) applied to the k(p)-algebra T

k(p).
2
The prime ideals of T,pT correspond to the prime ideals of T whose contraction to
contains p, and the prime ideals of T

k(p) correspond to the prime ideals of T whose

contraction to is p. To say that T is quasi-nite over at q means that q is both maximal
and minimal among the prime ideals lying over p (i.e., that each point of spec(T

k(p))
is closed).
PROPOSITION 15.5. A nitely generated -algebra T is quasi-nite over if and only if,
for all prime ideals p of , T

k(p) is nite over k(p).

PROOF. Immediate consequence of (15.2).
2
EXAMPLE 15.6. Let C be a nitely generated -algebra. If C is nite over , then
C

k(p) is nite over k(p) for all prime ideals p of , and so C is quasi-nite over . In
particular, spec(C) is discrete, and so if T is a nitely generated C-algebra such that the
map spec(T) spec(C) is an open immersion, then T is also quasi-nite over . Zariskis
main theorem says that all quasi-nite -algebras arise in this way.
The next two lemmas will be used in the proof of Zariskis main theorem.
15 QUASI-FINITE ALGEBRAS AND ZARISKIS MAIN THEOREM. 68
LEMMA 15.7. Let C T be homomorphisms of rings such that the composite
T is of nite type, and let q be a prime ideal of T. If T is quasi-nite over at q, then
it is quasi-nite over C at q.
PROOF. Let p

and p
C
be the inverse images of q in and C respectively. Then spec(T
C
k(p
C
)) is subspace of spec(T

k(p

)), and so if q is an isolated point in the second space,

then it is an isolated point in the rst space.
2
LEMMA 15.8. Let C T be rings. Let q be a prime ideal of T, and let r =qC and
p =q.
(a) If q is minimal among the primes lying over p and there exists a u C q such that
C
u
=T
u
, then r is minimal among the primes lying over p.
(b) If T is integral over a nitely generated -subalgebra T
0
and q is maximal among the
prime ideals lying over p, then r is maximal among the prime ideals lying over p.
(c) Assume that T is integral over a nitely generated -subalgebra T
0
, and that there
exists a u C q such that C
u
= T
u
. If T is quasi-nite over at q, then C is
quasi-nite over at r.
PROOF. (a) If r
t
is a prime ideal of C lying over p and strictly contained in r, then by
extending r
t
to C
u
=T
u
and then contracting the result to T, we obtain a prime ideal q
t
of
T lying over p and strictly contained in q.
(b) We may replace , C, and T with their localizations at p, and so assume that is
local with maximal ideal p. Then
,p C,r T,q
and we also have
,p T
0
,r
t
T,r
where r
t
=qT
0
. As q is maximal among the prime ideals lying over p, T,q is a eld. As
T,q is integral over T
0
,r
t
, the latter is also a eld (see 6.16), and it is nitely generated as
an ,p-algebra. Zariskis lemma (12.1) now shows that T
0
,r
t
is a nite algebraic extension
of ,p, and so T,q is an algebraic extension of ,p. It follows that C,r is a eld, and so r
is maximal among the prime ideals in C over p.
(c) Combine (a) and (b) (with the remark following (15.3)).
2
ASIDE 15.9. Geometrically, to say that T is quasi-nite means that the map SpecT Spec
has nite bres. The condition that T be nite is much stronger: it not only requires that
SpecT Spec have nite bres but also that it be universally closed. See, for example, my notes
on algebraic geometry.
Statement of Zariskis main theorem
THEOREM 15.10. Let T be a nitely generated -algebra, and let
t
be the integral closure
of in T. Then T is quasi-nite over at a prime ideal q if and only if
t
(
.T
(
for some

t
q.
The sufciency is obvious; the proof of the necessity will occupy the rest of this section.
First, we list some consequences.
15 QUASI-FINITE ALGEBRAS AND ZARISKIS MAIN THEOREM. 69
COROLLARY 15.11. Let T be a nitely generated -algebra. The set of prime ideals of T
at which T is quasi-nite over is open in spec(T).
PROOF. Let q be a prime ideal of T such that T is quasi-nite over at q. The theorem
shows that there exists an
t
q such that
t
(
. T
(
. Write
t
as the union of the
nitely generated -subalgebras
i
of
t
containing :

t
=
_
i

i
.
Because
t
is integral over , each
i
is nite over (see 6.3). We have
T
(
.
t
(
=
_
i

i(
.
Because T
(
is a nitely generated -algebra, T
(
=
i(
for all sufciently large
i
. As the

i
are nite over , T
(
is quasi-nite over , and spec(T
(
) is an open neighbourhood of q
consisting of quasi-nite points.
2
COROLLARY 15.12. Let T be a nitely generated -algebra, quasi-nite over , and let

t
be the integral closure of in T. Then
(a) the map SpecT Spec
t
is an open immersion, and
(b) there exists an -subalgebra
tt
of
t
, nite over , such that SpecT Spec
tt
is
an open immersion.
PROOF. (a) Because T is quasi-nite over at every point of spec(T), the theorem implies
that there exist
i

t
such that the open sets spec(T
(
i
) cover spec(T) and
t
(
i
.T
(
i
for
all i . As spec(T) quasi-compact, nitely many sets spec(T
(
i
) sufce to cover spec(T), and
it follows that spec(T) spec(
t
) is an open immersion.
(b) We have seen that spec(T) =
_
1_i_n
spec(T
(
i
) for certain
i

t
such that
t
(
i
.
T
(
i
. The argument in the proof of (15.11) shows that there exists an -subalgebra
tt
of
t
,
nite over , which contains
1
. . . . .
n
and is such that T
(
i
.
tt
(
i
for all i . Now the map
spec(T) spec(
tt
) is an open immersion.
2
Theorem 15.10, its corollary 15.12, and various global versions of these statements are
referred to as Zariskis main theorem.
A variant of Zariskis main theorem
PROPOSITION 15.13. Let C T be rings such that integrally closed in T, C is
nitely generated over , and T is nite over C. If T is quasi-nite over at a prime ideal
q, then T
p
=
p
with p =q.
PROOF THAT 15.13 IMPLIES 15.10
Let ,
t
, and T be as in the Theorem 15.10. We apply the proposition to
t
T =T
Lemma 15.7 shows that the ring T is quasi-nite over
t
at q. The proposition shows that
T
p
0 =
t
p
0
with p
t
=q
t
. Let b
1
. . . . . b
n
generate T as an
t
-algebra, and let b
t
i
denote
the image of b
i
in T
p
0 =
t
p
0
. Then b
t
i
=a
i
, for some a
i

t
and
t
p
t
. The b
t
i
are
in the image of the map
t
(
T
(
, which is therefore surjective. But
t
(
T
(
is injective
because T, and so the map is an isomorphism. This completes the proof of the theorem.
15 QUASI-FINITE ALGEBRAS AND ZARISKIS MAIN THEOREM. 70
Proof of Proposition 15.10
We proceed by proving four special cases of Proposition 15.10.
LEMMA 15.14. Let .| =T be rings such that is integrally closed in T. If T is
quasi-nite over at a prime ideal q, then T
p
=
p
with p =q.
PROOF. The hypotheses remain true when we invert the elements of S p to obtain
p

p
.| =T
p
. Thus, we may suppose that is local with maximal ideal p, and we have to
prove that T =. As is integrally closed in T and T =.|, it sufces to show that . is
integral over .
Let k =,p and consider the k-algebra
k .|
def
=.|

k =T

k(p).
By assumption, q is an isolated point in spec(k .|). Consequently, . is algebraic over k,
because otherwise k .| would be a polynomial ring over k, and its spectrum would have no
isolated points. Therefore there exists a polynomial J X| with nonconstant image in
kX| such that J(.) p.|. Now J J(.) is a polynomial in X| that vanishes on .
and has at least one coefcient not in p. Choose such a polynomial H of minimum degree
m, and write it
H(X) =a
n
X
n
a
0
.
The equation a
n-1
H(.) =0 can be written
(a
n
.)
n
a
n-1
(a
n
.)
n-1
a
0
a
n-1
n
=0.
It shows that a
n
. is integral over , and so lies in . Now the polynomial
(a
n
. a
n-1
)X
n-1
a
0
lies in X| and vanishes on .. As it has degree < m, all of its coefcients must lie in
p. In particular, a
n
. a
n-1
p. If a
n
is a unit, then . is integral over , as required.
Otherwise, a
n
p and a
n-1
is a unit (because otherwise all coefcients of H lie in p);
hence a
n-1
pT, which is contradiction because pT q.
2
LEMMA 15.15. Let T be an integral domain containing a polynomial ring X| and integral
over it. Then T is not quasi-nite over at any prime ideal q.
PROOF. Let q be a prime ideal of T, and let p =q. If T is quasi-nite over at q, then
q is both maximal and minimal among the prime ideals lying over p. We shall assume that q
is maximal and prove that it cant then be minimal.
Suppose rst that is integrally closed, and let r =q X|. If r were not maximal
among the prime ideals of X| lying over p, then the going-up theorem (6.20) would imply
that q is not either. Therefore r is maximal among the prime over p, and it follows that its
image r in k(p)X| is maximal. In particular, r =0, and so r strictly contains the prime ideal
pX| in X|. As is integrally closed, X| is also (6.15), and the going down theorem
(6.24) shows that q strictly contains a prime ideal lying over pX|. Therefore, q is not
minimal among the prime ideals lying over p.
In the general case, we let T
t
denote the integral closure of T in its eld of fractions.
Then T
t
contains the integral closure
t
of , and is integral over
t
T |. Let q
t
be a prime
15 QUASI-FINITE ALGEBRAS AND ZARISKIS MAIN THEOREM. 71
ideal of T
t
lying over q (which exists by 6.19), and let p
t
=q
t

t
. As q is maximal among
the primes lying over p, q
t
is maximal among those lying over p
t
(apply 6.18 to T T
t
).
But, according to the preceding paragraph, q
t
is not minimal, which implies that q is not
minimal (apply 6.18 again).
2
LEMMA 15.16. Let .| T be rings such that T is integral over .| and is
integrally closed in T. If there exists a monic polynomial J X| such that J(.)T .|,
then .| =T.
PROOF. Let b T be arbitrary. By assumption J(.)b .|, and so J(.)b =G(.) for
some polynomial G in X|. As J is monic, we can divide J into G to get
G =QJ 1. deg1 <degJ. Q. 1 X|.
Now
J(.)b =G(.) =Q(.)J(.) 1(.).
For c =b Q(.).
J(.)c =1(.). (42)
To show that b .|, it sufces to show that c , and for this it sufces to show that c is
integral over .
Let
t
be the image of in T
c
. As deg1 <degJ, the equality (42) shows that .,1, as
an element of T
c
, is integral over the subring
t
c
. As T is integral over .|, this implies that
T
c
is integral over
t
c
. In particular, c,1 is integral over
t
c
, and so it satises an equation
whose coefcients we can assume to have a common denominator c

:
(c,1)
n

a
1
c

(c,1)
n-1

a
n
c

=0. a
i
,
(equality in T
c
). Therefore
c
n
a
1
c
n-1
a
n
is an element of T whose image in T
c
is zero, and so is killed by a power of c. This shows
that c is integral over , as required.
2
Let T be a nite -algebra. The conductor of T in is
f(T,) ={a [ aT ].
This is an ideal of both and T. In fact, it is the largest ideal in that is also an ideal
in T, because every element a of such an ideal has the property that aT . For any
multiplicative subset S of ,
f(S
-1
T,S
-1
) =S
-1
f(T,). (43)
LEMMA 15.17. Let .| T be rings such that T is nite over .| and is integrally
closed in T. If T is quasi-nite over at a prime ideal q, then T
p
=
p
with p =q.
15 QUASI-FINITE ALGEBRAS AND ZARISKIS MAIN THEOREM. 72
PROOF. Let f =f(T,.|), so
f ={ .| [ T .|].
We rst consider the case that f , q. Let r = q .|. For any u f q, we have
.|
u
=T
u
, and so Lemma 15.8 shows that .| is quasi-nite over at r.
23
Now Lemma
15.14 shows that .|
p
=
p
. But T is nite over .|, and therefore T
p
is nite over
.|
p
=
p
. As is integrally closed in T,
p
is integrally closed in T
p
, and therefore

p
=T
p
, as required.
It remains to consider the case f q. We choose a prime ideal n q of T minimal
among those containing f. Let t denote the image of . in the ring T,n, and let m=n.
Now
,m(,m)t | T,n,
and T,n is integral over (,m)t |. As T is quasi-nite over at q, the quotient T,n is
quasi-nite over ,m at q,n. Now Lemma 15.15 implies that t is algebraic over ,m. We
shall complete the proof by obtaining a contradiction, which will show that this case doesnt
occur.
After making an extension of scalars
m
, we may assume that is a local ring
with maximal ideal m. Let n
t
=n.|. Because t is algebraic over ,m, the integral
domain .|,n
t
is a nite ,m-algebra, and hence a eld (see 1). Therefore, n
t
is maximal
in .|, and it follows from (6.17) that n is maximal in T. Thus T,n is a eld.
Because t is algebraic over ,m, there exists a monic polynomial J in X| such
that J(.) n. But n is minimal among the prime ideals of T containing f, and so nT
n
is
minimal among the prime ideals of T
n
containing f
n
. In fact, nT
n
is the only prime ideal
containing f
n
, and so nT
n
is the radical of f
n
. Therefore, there exists an integer r >0 such
that (J(.))
i
f
n
, and a , T n such that ,J(.)
i
f.
We therefore have ,J(.)
i
T .|. On applying Lemma 15.16 with .| T
t
,
T
t
= .|,T|. and J
t
= J
i
, we deduce that T
t
= .| and therefore that ,T .|.
Hence , f n, which contradicts the denition of ,.
2
PROOF OF PROPOSITION 15.10
We use induction on the number n of generators of the -algebra C. If n =0, then T is
integral over , and so T =. Assume that n > 0 and that the proposition has been proved
when C is generated by n1 elements.
Write C = .
1
. . . . . .
n
|, and let
t
be the integral closure of .
1
. . . . . .
n-1
| in T.
Then

t
.
n
| T.
and T is nite over
t
.
n
|. The ring T is nite over
t
.
n
| and it is quasi-nite over at q,
and so T is quasi-nite over
t
at q (by 15.7). From Lemma 15.17 we deduce that
t
p
0
=T
p
0
with p
t
=
t
q.
As
t
is integral over .
1
. . . . . .
n-1
|, it is a union of its nite subalgebras,

t
=
_
i

t
i
.
t
i
nite over .
1
. . . . . .
n-1
|.
Let p
t
i
=q
t
i
=p
t

t
i
. As T is nitely generated over .
1
. . . . . .
n-1
|, the canonical
homomorphism
(
t
i
)
p
0
i
T
p
0
i
23
Here we follow Hochster. Raynaud simply states that .| is quasi-nite over at r.
16 DIMENSION THEORY FOR FINITELY GENERATED k-ALGEBRAS 73
is an isomorphism for all sufciently large i . For such an i , we have a fortiori that
(
t
i
)
p
0
i
.T
q
.
and so
t
i
is quasi-nite over at p
t
i
. On applying the induction hypothesis to . .
1
. . . . . .
n-1
|,
and
t
i
, we deduce that

p
.(
t
i
)
p
.(
t
i
)
p
0
i
.
and consequently that
p
.T
p
. This completes the proof of Proposition 15.13 and hence of
Theorem 15.10.
16 Dimension theory for nitely generated k-algebras
Throughout this section, is both a nitely generated algebra over eld k and an integral
domain. We dene the transcendence degree of over k, tr deg
k
, to be the transcendence
degree over k of the eld of fractions of (see 8 of my notes Fields and Galois Theory).
Thus has transcendence degree J if it contains an algebraically independent set of J
elements, but no larger set (ibid. 8.12).
PROPOSITION 16.1. For any linear forms
1
. . . . .
n
in X
1
. . . . . X
n
, the quotient ring
kX
1
. . . . . X
n
|,(
1
. . . . .
n
)
is an integral domain of transcendence degree equal to the dimension of the subspace of k
n
dened by the equations

i
=0. i =1. . . . . m.
PROOF. This follows from the more precise statement:
Let c be an ideal in kX
1
. . . . . X
n
| generated by linearly independent linear
forms
1
. . . . .
i
, and let X
i
1
. . . . . X
i
nr
be such that
{
1
. . . . .
i
. X
i
1
. . . . . X
i
nr
]
is a basis for the linear forms in X
1
. . . . . X
n
. Then
kX
1
. . . . . X
n
|,c .kX
i
1
. . . . . X
i
nr
|.
This is obvious if the forms
i
are X
1
. . . . . X
i
. In the general case, because {X
1
. . . . . X
n
]
and {
1
. . . . .
i
. X
i
1
. . . . . X
i
nr
] are both bases for the linear forms, each element of one set
can be expressed as a linear combination of the elements of the other. Therefore,
kX
1
. . . . . X
n
| =k
1
. . . . .
i
. X
i
1
. . . . . X
i
nr
|.
and so
kX
1
. . . . . X
n
|,c =k
1
. . . . .
i
. X
i
1
. . . . . X
i
nr
|,c
.kX
i
1
. . . . . X
i
nr
|.
2
16 DIMENSION THEORY FOR FINITELY GENERATED k-ALGEBRAS 74
PROPOSITION 16.2. For any irreducible polynomial in kX
1
. . . . . X
n
|, the quotient ring
kX
1
. . . . . X
n
|,( ) has transcendence degree n1.
PROOF. Let
k.
1
. . . . . .
n
| =kX
1
. . . . . X
n
|,( ). .
i
=X
i
( ).
and let k(.
1
. . . . . .
n
) be the eld of fractions of k.
1
. . . . . .
n
|. Since is not zero, some X
i
,
say, X
n
, occurs in it. Then X
n
occurs in every nonzero multiple of , and so no nonzero
polynomial in X
1
. . . . . X
n-1
belongs to ( ). This means that .
1
. . . . . .
n-1
are algebraically
independent. On the other hand, .
n
is algebraic over k(.
1
. . . . . .
n-1
), and so {.
1
. . . . . .
n-1
]
is a transcendence basis for k(.
1
. . . . . .
n
) over k.
2
PROPOSITION 16.3. For every nonzero prime ideal p in a k-algebra ,
tr deg
k
(,p) <tr deg
k
().
PROOF. We may suppose that
=kX
1
. . . . . X
n
|,a =k.
1
. . . . . .
n
|.
For , let

denote the image of in ,p, so that ,p = k .
1
. . . . . .
n
|. Let J =
tr deg
k
,p, and number the X
i
so that .
1
. . . . . .
d
are algebraically independent (for a proof
that this is possible, see 8.9 of my notes Fields and Galois Theory). I shall show that, for
any nonzero p, the J 1 elements .
1
. . . . . .
d
. are algebraically independent, which
shows that tr deg
k
_J 1.
Suppose otherwise. Then there is a nontrivial algebraic relation, which we can write
a
0
(.
1
. . . . . .
d
)
n
a
1
(.
1
. . . . . .
d
)
n-1
a
n
(.
1
. . . . . .
d
) =0.
with a
i
kX
1
. . . . . X
d
| and a
0
= 0. Because is an integral domain, we can cancel a
power of if necessary to make a
n
(.
1
. . . . . .
d
) nonzero. On applying the homomorphism
,p to the above equality, we nd that
a
n
( .
1
. . . . . .
d
) =0.
which contradicts the algebraic independence of .
1
. . . . . .
d
.
2
PROPOSITION 16.4. Let be a unique factorization domain. If p is a prime ideal in such
that tr deg
k
,p =tr deg
k
1, then p =( ) for some .
PROOF. The ideal p is nonzero because otherwise and ,p would have the same tran-
scendence degree. Therefore p contains a nonzero polynomial, and even an irreducible
polynomial , because it is prime. According to (4.1), the ideal ( ) is prime. If ( ) =p,
then
tr deg
k
,p
16.3
> tr deg
k
,( )
16.2
= tr deg
k
1.
which contradicts the hypothesis.
2
THEOREM 16.5. Let be neither zero nor a unit, and let p be a prime ideal that is
minimal among those containing ( ); then
tr deg
k
,p =tr deg
k
1.
16 DIMENSION THEORY FOR FINITELY GENERATED k-ALGEBRAS 75
We rst need a lemma.
LEMMA 16.6. Let be an integrally closed integral domain, and let 1 be a nite extension
of the eld of fractions 1 of . If 1 is integral over , then Nm
1{1
, and
divides Nm
1{1
in the ring |.
PROOF. Let X
i
a
i-1
X
i-1
a
0
be the minimum polynomial of over 1. Then
r divides the degree n of 1,1, and Nm
1{1
() =a
n
r
0
(see 5.40 of my notes Fields and
Galois Theory). Moreover, a
0
lies in by (6.10). From the equation
0 =(
i-1
a
i-1

i-2
a
1
) a
0
we see that divides a
0
in |, and therefore it also divides Nm
1{1
.
2
PROOF (OF THEOREM 16.5). Write rad( ) as an irredundant intersection of prime ideals
rad( ) =p
1
. . . p
i
(see 13.9). Then V(a) =V(p
1
) L LV(p
i
) is the decomposition
of V(a) into its irreducible components. There exists an m
0
V(p
1
)
_
i_2
V(p
i
) and an
open neighbourhood D(h) of m
0
disjoint from
_
i_2
V(p
i
). The ring
h
(resp.
h
,S
-1
p)
is an integral domain with the same transcendance degree as (resp. ,p) in fact, with
the same eld of fractions. In
h
, rad(
(
1
) =rad( )
e
=p
e
1
. Therefore, after replacing
with
h
, we may suppose that rad( ) is prime, say, equal to p.
According to the Noether normalization theorem (6.26), there exist algebraically inde-
pendent elements .
1
. . . . . .
d
in such that is a nite k.
1
. . . . . .
d
|-algebra. Note that
J = tr deg
k
. According to the lemma,
0
def
= Nm( ) lies in k.
1
. . . . . .
d
|, and we shall
show that pk.
1
. . . . . .
d
| =rad(
0
). Therefore, the homomorphism
k.
1
. . . . . .
d
|,rad(
0
) ,p
is injective. As it is also nite, this implies that
tr deg
k
,p =tr deg
k
k.
1
. . . . . .
d
|,rad(
0
)
16.2
= J 1.
as required.
By assumption is nite (hence integral) over its subring k.
1
. . . . . .
d
|. The lemma
shows that divides
0
in , and so
0
( ) p. Hence (
0
) pk.
1
. . . . . .
d
|, which
implies
rad(
0
) pk.
1
. . . . . .
d
|
because p is radical. For the reverse inclusion, let g pk.
1
. . . . . .
d
|. Then g rad( ),
and so g
n
= h for some h , m N. Taking norms, we nd that
g
ne
=Nm( h) =
0
Nm(h) (
0
).
where e is the degree of the extension of the elds of fractions, which proves the claim.
2
COROLLARY 16.7. Let p be a minimal nonzero prime ideal in ; then tr deg
k
(,p) =
tr deg
k
() 1.
PROOF. Let be a nonzero element of p. Then is not a unit, and p is minimal among the
prime ideals containing .
2
17 PRIMARY DECOMPOSITIONS 76
THEOREM 16.8. The length J of any maximal (i.e., nonrenable) chain of distinct prime
ideals
p
d
p
d-1
p
0
(44)
in is tr deg
k
(). In particular, every maximal ideal of has height tr deg
k
(), and so the
Krull dimension of is equal to tr deg
k
().
PROOF. From (16.7), we nd that
tr deg
k
() =tr deg
k
(,p
1
) 1 = =tr deg
k
(,p
d
) J.
But p
d
is maximal, and so ,p
d
is a nite eld extension of k. In particular, tr deg
k
(,p
d
) =
0.
2
EXAMPLE 16.9. Let (X. Y ) and g(X. Y ) be nonconstant polynomials with no common
factor. Then kX. Y |,( ) has Krull dimension 1, and so kX. Y |,(. g) has dimension zero.
EXAMPLE 16.10. We classify the prime ideals p in =kX. Y |. If ,p has dimension
2, then p =(0). If ,p has dimension 1, then p = ( ) for some irreducible polynomial
of (by 16.4). Finally, if ,p has dimension zero, then p is maximal. Thus, when k
is algebraically closed, the prime ideals in kX. Y | are exactly the ideals (0), ( ) (with
irreducible), and (X a. Y b) (with a. b k).
REMARK 16.11. Let be a nitely generated k-algebra (not necessarily an integral do-
main). Every maximal chain of prime ideals in ending in xed prime ideal p has length
tr deg
k
(,p), and so the Krull dimension of is max(tr deg
k
(,p)) where p runs over the
minimal prime ideals of . In the next section, we show that a noetherian ring has only
nitely many minimal prime ideals, and so the Krull dimension of is nite.
If .
1
. . . . . .
n
is an algebraically independent set of elements of such that is a nite
k.
1
. . . . . .
n
|-algebra, then dim =m.
REMARK 16.12. Let be a discrete valuation ring with maximal ideal (). Then X|
is a noetherian integral domain of Krull dimension 2, and (X 1) is a maximal ideal in
X| of height 1 (cf. 14.10).
17 Primary decompositions
In this section, is an arbitrary commutative ring.
DEFINITION 17.1. An ideal q in is primary if it is proper and
ab q, b q == a
n
q for some n _1.
Thus, a proper ideal q in is primary if and only if all zero-divisors in ,q are nilpotent. A
radical ideal is primary if and only if it is prime. An ideal (m) in Z is primary if and only if
m is a power of a prime.
PROPOSITION 17.2. The radical of a primary ideal q is a prime ideal containing q, and
it is contained in every other prime ideal containing q (i.e., it is the smallest prime ideal
containing p).
17 PRIMARY DECOMPOSITIONS 77
PROOF. Suppose that ab rad(q) but b rad(q). Then some power, say a
n
b
n
, of ab lies
in q, but b
n
q, and so a rad(q). This shows that rad(q) is primary, and hence prime
(because it is radical).
Let p be a second prime ideal containing q, and let a rad(q). For some n, a
n
q p,
which implies that a p.
2
When q is a primary ideal and p is its radical, we say that q is p-primary.
PROPOSITION 17.3. Every ideal q whose radical is a maximal ideal m is primary (in fact,
m-primary); in particular, every power of a maximal ideal m is m-primary.
PROOF. Every prime ideal containing q contains its radical m, and therefore equals m. This
shows that ,a is local with maximal ideal m,a. Therefore, every element of ,a is either
a unit, and hence is not a zero-divisor, or it lies in m,a, and hence is nilpotent.
2
PROPOSITION 17.4. Let : T be a homomorphism of rings. If q is a p-primary ideal
in T, then q
c
def
=
-1
(q) is a p
c
-primary ideal in .
PROOF. The map ,q
c
T,q is injective, and so every zero-divisor in ,q
c
is nilpotent.
This shows that q
c
is primary, and therefore rad(q
c
)-primary. But (see 2.10), rad(q
c
) =
rad(q)
c
=p
c
, as claimed.
2
LEMMA 17.5. Let q and p be a pair of ideals in such that q p rad(q) and
ab q == a p or b q. (45)
Then p is a prime ideal and q is p-primary.
PROOF. Clearly q is primary, hence rad(q)-primary, and rad(q) is prime. By assumption
p rad(q), and it remains to show that they are equal. Let a rad(q), and let n be the
smallest positive integer such that a
n
q. If n =1, then a q p; on the other hand, if
n >1, then a
n
=aa
n-1
q and a
n-1
q, and so a p by (45).
2
PROPOSITION 17.6. A nite intersection of p-primary ideals is p-primary.
PROOF. Let q
1
. . . . . q
n
be p-primary, and let q = q
1
. . . q
n
. We show that the pair of
ideals q p satises the conditions of (17.5).
Let a p; since some power of a belongs to each q
i
, a sufciently high power of it will
belong to all of them, and so p rad(q).
Let ab q but a p. Then ab q
i
but a p, and so b q
i
. Since this is true for all i ,
we have that b q.
2
The minimal prime ideals of an ideal a are the minimal elements of the set of prime
ideals containing a.
DEFINITION 17.7. A primary decomposition of an ideal a is a nite set of primary ideals
whose intersection is a. A primary decomposition S of a is minimal if
(a) the prime ideals rad(q), q S, are distinct, and
(b) no element of S can be omitted, i.e., for no q
0
S is q
0

_
{q [ q S, q =q
0
].
17 PRIMARY DECOMPOSITIONS 78
If a admits a primary decomposition, then it admits a minimal primary decomposition,
because Proposition 17.6 can be used to combine primary ideals with the same radical, and
any q
i
that fails (b) can simply be omitted. The prime ideals occurring as the radical of an
ideal in a minimal primary decomposition of a are said to belong to a.
PROPOSITION 17.8. Suppose that a =q
1
q
n
where q
i
is p
i
-primary for i =1. . . . . n.
Then the minimal prime ideals of a are the minimal elements of the set {p
1
. . . . . p
n
].
PROOF. Let p be a prime ideal containing a, and let q
t
i
be the image of q
i
in the integral
domain ,p. Then p contains q
1
q
n
, and so q
t
1
q
t
n
=0. This implies that, for some i ,
q
t
i
=0, and so p contains q
i
. Now (17.2) shows that p contains p
i
.
2
In particular, if a admits a primary decomposition, then it has only nitely many minimal
prime ideals, and so its radical is a nite intersection of prime ideals.
For an ideal a in and an element . , we let
(a: .) ={a [ a. a].
It is again an ideal in , which equals if . a.
LEMMA 17.9. Let q be a p-primary ideal and let . q. Then (q: .) is p-primary (and
hence rad(q: .) =p).
PROOF. For any a (q: .), we know that a. q and . q, and so a p. Hence (q: .) p.
On taking radicals, we nd that rad(q: .) =p. Let ab (q: .). Then .ab q, and so either
a p or .b q (because q is p-primary); in the second case, b (q: .) as required.
2
THEOREM 17.10. Let a =q
1
. . . q
n
be a minimal primary decomposition of a, and let
p
i
=rad(q
i
). Then
{p
1
. . . . . p
n
] ={rad(a: .) [ . . rad(a: .) prime].
In particular, the set {p
1
. . . . . p
n
] is independent of the choice of the minimal primary
decomposition.
PROOF. For any a ,
(a: a) =(
_
q
i
: a) =
_
(q
i
: a).
and so
rad(a: a) =rad
_
(q
i
: a)
(17.9)
=
_
oq
i
p
i
. (46)
If rad(a: a) is prime, then it equals one of the p
i
(otherwise, for each i there exists an
a
i
p
i
p, and a
1
a
n

_
oq
i
p
i
but not p, which is a contradiction). Hence RHSLHS.
For each i , there exists an a
_
}yi
q
}
q
i
because the decomposition is minimal, and (46)
shows that rad(a: a) =p
i
.
2
THEOREM 17.11. In a noetherian ring, every ideal admits a primary decomposition.
The theorem is a consequence of the following more precise statement, but rst we need
a denition: an ideal a is said to be irreducible if
a =bc (b, c ideals) == a =b or a =c.
17 PRIMARY DECOMPOSITIONS 79
PROPOSITION 17.12. Let be a noetherian ring.
(a) Every ideal in can be expressed as a nite intersection of irreducible ideals.
(b) Every irreducible ideal in is primary.
PROOF. (a) Suppose that (a) fails, and let a be maximal among the ideals for which it
fails. Then, in particular, a itself is not irreducible, and so a =bc with b and c properly
containing a. Because a is maximal, both b and c can be expressed as nite intersections of
irreducible ideals, but then so can a.
(b) Let a be irreducible in , and consider the quotient ring
t
def
= ,a. Let a be a
zero-divisor in
t
, say ab =0 with b =0. We have to show that a is nilpotent. As
t
is
noetherian, the chain of ideals
((0): a) ((0): a
2
)
becomes constant, say, ((0): a
n
) =((0): a
n1
)) = . Let c (a
n
) (b). Then c (b)
implies ca =0, and c (a
n
) implies that c =Ja
n
for some J . Now
(Ja
n
)a =0 =J (0: a
n1
) =(0: a
n
) =c =0.
Hence (a
n
) (b) = (0). Because a is irreducible, so also is the zero ideal in
t
, and it
follows that a
n
=0.
2
A p-primary ideal a in a noetherian ring contains a power of p by Proposition 3.16. The
next result proves a converse when p is maximal.
PROPOSITION 17.13. Let m be a maximal ideal of a noetherian ring. Any proper ideal a of
that contains a power of a maximal ideal m is m-primary.
PROOF. Suppose that m
i
a, and let p be a prime ideal belonging to a. Then m
i
a p,
so that m p, which implies that m =p. Thus m is the only prime ideal belonging to a,
which means that a is m-primary.
2
EXAMPLE 17.14. We give an example of a power of a prime ideal p that is not p-primary.
Let
=kX. Y. 7|,(Y
2
X7) =k.. ,. :|.
The ideal (X. Y ) in kX. Y. 7| is prime and contains (Y
2
X7), and so the ideal p =(.. ,)
in is prime. Now .: =,
2
p
2
, but one checks easily that . p
2
and : p, and so p
2
is
not p-primary.
REMARK 17.15. Let a be an ideal in a noetherian ring, and let b =
_
n_1
a
n
. We give
another proof that ab =b (see p. 13). Let
ab =q
1
. . . q
x
. rad(q
i
) =p
i
.
be a minimal primary decomposition of ab. We shall show that b ab by showing that
b q
i
for each i .
If there exists a b bq
i
, then
ab ab q
i
,
from which it follows that a p
i
. We know that p
i
i
q
i
for some r (see 3.16), and so
b =
_
a
n
a
i
p
i
i
q
i
,
which is a contradiction. This completes the proof.
18 DEDEKIND DOMAINS 80
NOTES. In a future version, Ill explain the geometric signicance of these statements. Also, Ill
include the statements for modules, and explain how to deduce them from the statements for rings by
a trick of Nagata. (When M is an -module, M can be made into a ring by setting the product
of any two elements of M equal to zero.). See also mo3910.
18 Dedekind domains
Discrete valuation rings
It follows from the elementary theory of principal ideal domains that the following conditions
on a principal ideal domain are equivalent:
(a) has exactly one nonzero prime ideal;
(b) up to associates, has exactly one prime element;
(c) is local and is not a eld.
A ring satisfying these conditions is called a discrete valuation ring.
EXAMPLE 18.1. The ring Z
(])
def
={
n
n
Q[ n not divisible by ] is a discrete valuation ring
with () as its unique nonzero prime ideal. The units in Z
(])
are the nonzero elements m,n
with neither m nor n divisible by , and the prime elements are those of the form unit.
In a discrete valuation ring with prime element , nonzero elements of can be
expressed uniquely as u
n
with u a unit and m_0 (and m>0 unless the element is a unit).
Every nonzero ideal in is of the form (
n
) for a unique m N. Thus, if a is an ideal in
and p denotes the (unique) maximal ideal of , then a =p
n
for a well-dened integer
m_0.
Recall that, for an -module M and an m M, the annihilator of m
ann(m) ={a [ am=0].
It is an ideal in , which is proper if m =0. Suppose that is a discrete valuation ring,
and let c be a nonzero element of . Let M =,(c). What is the annihilator of a nonzero
element b (c) of M? Fix a prime element of , and let c =u
n
, b =
n
with u and
units. Then n <m (else b (c) =0 in M), and
ann(b (c)) =(
n-n
).
Thus, a b for which ann(b (c)) is maximal, is of the form
n-1
, and for this choice
ann(b (c)) is a prime ideal generated by
c
b
. We shall exploit these observations in the
proof of the next proposition, which gives a criterion for a ring to be a discrete valuation
ring.
PROPOSITION 18.2. An integral domain is a discrete valuation ring if and only if
(a) is Noetherian,
(b) is integrally closed, and
(c) has exactly one nonzero prime ideal.
PROOF. The necessity of the three conditions is obvious, and so let be an integral domain
satisfying (a), (b), and (c). We have to show that every ideal in is principal. As a rst step,
we prove that the nonzero prime ideal is principal. Note that (c) implies that is a local ring.
18 DEDEKIND DOMAINS 81
Choose an element c , c =0, c =unit, and consider the -module M
def
=,(c). For
each nonzero element m of M,
ann(m) ={a [ am=0]
is a proper ideal in . Because is Noetherian, we can choose an m so that ann(m) is
maximal among these ideals. Write m=b (c) and p =ann(b (c)). Note that c p, and
so p =0, and that
p ={a [ c[ab].
I claim that p is prime. If not there exist elements ., , such that ., p but neither .
nor , p. Then ,b (c) is a nonzero element of M because , p. Consider ann(,b (c)).
Obviously it contains p and it contains ., but this contradicts the maximality of p among
ideals of the form ann(m). Hence p is prime.
I claim that
b
c
. Otherwise b =c
b
c
(c), and m=0 (in M).
I claim that
c
b
, and p =(
c
b
). By denition, pb (c), and so p
b
c
, and it is an
ideal in . If p
b
c
p, then
b
c
is integral over (by 6.1, since p is nitely generated), and
so
b
c
(because of condition (b)), but we know
b
c
. Thus p
b
c
= (by (c)), and this
implies that p =(
c
b
).
Let =
c
b
, so that p =(). Let a be a proper ideal of , and consider the sequence
a a
-1
a
-2
.
If a
-i
=a
-i-1
for some r, then
-1
(a
-i
) =a
-i
, and
-1
is integral over (by
6.1), and so lies in this is impossible ( is not a unit in ). Therefore the sequence
is strictly increasing, and (again because is Noetherian) it cant be contained in . Let
m be the smallest integer such that a
-n
but a
-n-1
. Then a
-n
p, and so
a
-n
=. Hence a =(
n
).
2
Dedekind domains
DEFINITION 18.3. A Dedekind domain is an integral domain , not equal to a eld, such
that
(a) is Noetherian,
(b) is integrally closed, and
(c) every nonzero prime ideal is maximal (i.e., has Krull dimension 1).
Thus Proposition 18.2 says that a local integral domain is a Dedekind domain if and only
if it is a discrete valuation ring.
PROPOSITION 18.4. Let be a Dedekind domain, and let S be a multiplicative subset of
. Then S
-1
is either a Dedekind domain or a eld.
PROOF. Condition (c) says that there is no containment relation between nonzero prime
ideals of . If this condition holds for , then (5.4) shows that it holds for S
-1
. Conditions
(a) and (b) follow from the next lemma.
2
PROPOSITION 18.5. Let be an integral domain, and let S be a multiplicative subset of .
(a) If is Noetherian, then so also is S
-1
.
(b) If is integrally closed, then so also is S
-1
.
18 DEDEKIND DOMAINS 82
PROOF. (a) Let a be an ideal in S
-1
. Then a =S
-1
(a) (see 5.4), and so a is generated
by any (nite) set of generators for a.
(b) Let be an element of the eld of fractions of (=eld of fractions of S
-1
) that
is integral over S
-1
. Then

n
a
1

n-1
a
n
=0, some a
i
S
-1
.
For each i , there exists an s
i
S such that s
i
a
i
. Set s =s
1
s
n
S, and multiply
through the equation by s
n
:
(s)
n
sa
1
(s)
n-1
s
n
a
n
=0.
This equation shows that s is integral over , and so lies in . Hence =(s),s S
-1
.
(See also 6.13.)
2
COROLLARY 18.6. For any nonzero prime ideal p in a Dedekind domain , the localization

p
is a discrete valuation ring.
PROOF. We saw in (5.7) that
p
is local, and the proposition implies that it is Dedekind.
2
The main result concerning Dedekind domains is the following.
THEOREM 18.7. Every proper nonzero ideal a in a Dedekind domain can be written in the
form
a =p
i
1
1
p
i
s
x
with the p
i
distinct prime ideals and the r
i
> 0; the ideals p
i
are exactly the prime ideals
containing a, and the exponents r
i
are uniquely determined.
PROOF. The primary ideals in a Dedekind domain are exactly the powers of prime ideals,
and so this follows from the preceding section. (For an elementary proof, see my notes on
algebraic number theory.)
2
REMARK 18.8. Note that
r
i
>0 a
p
i
=
p
i
a p
i
.
COROLLARY 18.9. Let a and b be ideals in ; then
a b a
p
b
p
for all nonzero prime ideals p of . In particular, a =b if and only if a
p
=b
p
for all p.
PROOF. The necessity is obvious. For the sufciency, factor a and b
a =p
i
1
1
p
i
m
n
. b =p
x
1
1
p
x
m
n
. r
i
. s
i
_0.
Then
a
p
i
b
p
i
r
i
_s
i
.
(recall that
p
i
is a discrete valuation ring) and r
i
_s
i
all i implies a b.
2
18 DEDEKIND DOMAINS 83
COROLLARY 18.10. Let be an integral domain with only nitely many prime ideals;
then is a Dedekind domain if and only if it is a principal ideal domain.
PROOF. Assume is a Dedekind domain. After (18.7), to show that is principal, it
sufces to show that the prime ideals are principal. Let p
1
. . . . . p
n
be these ideals. Choose
an element .
1
p
1
p
2
1
. According to the Chinese Remainder Theorem (2.12), there is an
element . such that
. .
1
mod p
2
1
. . 1 mod p
i
. i =1.
Now the ideals p
1
and (.) generate the same ideals in
p
i
for all i , and so they are equal in
(by 18.9).
2
COROLLARY 18.11. Let a b =0 be two ideals in a Dedekind domain; then a =b(a)
for some a .
PROOF. Let b =p
i
1
1
p
i
m
n
and a =p
x
1
1
p
x
m
n
with r
i
. s
}
_ 0. Because b a, s
i
_ r
i
for
all i . For 1 _ i _ m, choose an .
i
such that .
i
p
x
i
i
, .
i
p
x
i
1
i
. By the Chinese
Remainder Theorem, there is an a such that
a .
i
mod p
i
i
i
, for all i.
Now one sees that b(a) =a by looking at the ideals they generate in
p
for all p.
2
COROLLARY 18.12. Let a be an ideal in a Dedekind domain, and let a be any nonzero
element of a; then there exists a b a such that a =(a. b).
PROOF. Apply (18.11) to a (a).
2
COROLLARY 18.13. Let a be a nonzero ideal in a Dedekind domain; then there exists a
nonzero ideal a
+
in such that aa
+
is principal. Moreover, a
+
can be chosen to be relatively
prime to any particular ideal c, and it can be chosen so that aa
+
=(a) with a any particular
element of a (but not both).
PROOF. Let a a, a =0; then a (a), and so we have
(a) =p
i
1
1
p
i
m
n
and a =p
x
1
1
p
x
m
n
. s
i
_r
i
.
If a
+
=p
i
1
-x
1
1
p
i
m
-x
m
n
, then aa
+
=(a).
We now show that a
+
can be chosen to be prime to c. We have a ac, and so (by 18.11)
there exists an a a such that a =ac (a). As a (a), we have (a) =a a
+
for some ideal
a
+
(by the above argument); now, ac aa
+
=a, and so c a
+
=. (Otherwise c a
+
p
some prime ideal, and ac aa
+
=a(c a
+
) ap =a.)
2
In basic graduate algebra courses, it is shown that
a principal ideal domain = is a unique factorization domain.
The converse is false because, for example, kX. Y | is a unique factorization domain in
which the ideal (X. Y ) is not principal, but it is true for Dedekind domains.
18 DEDEKIND DOMAINS 84
PROPOSITION 18.14. A Dedekind domain that is a unique factorization domain is a princi-
pal ideal domain.
PROOF. In a unique factorization domain, an irreducible element can divide a product
bc only if divides b or c (write bc =q and express each of b, c, and q as a product of
irreducible elements). This means that () is a prime ideal.
Now let be a Dedekind domain with unique factorization. It sufces to show that each
nonzero prime ideal p of is principal. Let a be a nonzero element of p. Then a factors into
a product of irreducible elements (see 4.3) and, because p is prime, it will contain one of
these irreducible factors . Now p () (0), and, because () is a nonzero prime ideal,
it is maximal, and so equals p.
2
Modules over Dedekind domains (sketch).
The structure theorem for nitely generated modules over principal ideal domains has an
interesting extension to modules over Dedekind domains. Throughout this subsection, is a
Dedekind domain.
First, note that a nitely generated torsion-free -module M need not be free. For
example, every fractional ideal is nitely generated and torsion-free but it is free if and only
if it is principal. Thus the best we can hope for is the following.
THEOREM 18.15. Let be a Dedekind domain.
(a) Every nitely generated torsion-free -module M is isomorphic to a direct sum of
fractional ideals,
M ~a
1
a
n
.
(b) Two nitely generated torsion-free -modules M ~a
1
a
n
and N ~b
1

b
n
are isomorphic if and only if m=n and

a
i

b
i
modulo principal ideals.
Hence,
M ~a
1
a
n
~ a
1
a
n
.
Moreover, two fractional ideals a and b of are isomorphic as -modules if and only if they
dene the same element of the class group of .
The rank of a module M over an integral domain 1 is the dimension of 1
T
M as a
1-vector space, where 1 is the eld of fractions of 1. Clearly the rank of M ~a
1
a
n
is m.
These remarks show that the set of isomorphism classes of nitely generated torsion-free
-modules of rank 1 can be identied with the class group of . Multiplication of elements
in Cl() corresponds to the formation of tensor product of modules. The Grothendieck
group of the category of nitely generated -modules is Cl() Z.
THEOREM 18.16 (INVARIANT FACTOR THEOREM). Let M N be nitely generated torsion-
free -modules of the same rank m. Then there exist elements e
1
. .... e
n
of M, fractional
ideals a
1
. .... a
n
, and integral ideals b
1
b
2
... b
n
such that
M =a
1
e
1
a
n
e
n
. N =a
1
b
1
e
1
a
n
b
n
e
n
.
The ideals b
1
, b
2
, ..., b
n
are uniquely determined by the pair M N, and are called the
invariant factors of N in M.
The last theorem also yields a description of nitely generated torsion -modules.
19 DIMENSION THEORY FOR NOETHERIAN RINGS 85
NOTES. We sketch a proof of (18.15a). Let be a Dedekind domain, and let M be nitely generated
torsion-free -module. Then
p
M is free, hence projective, for every nonzero prime ideal p in
(because
p
is principal ideal domain), and this implies that M is projective. Therefore there is a
nonzero homomorphism M , whose image is an ideal a in . As a is projective, there exists
a section to the map M a, and so M ~ a M
1
for some submodule M
1
of M. Now M
1
is
projective because it is a direct summand of a projective module, and so we can repeat the argument
with M
1
. This process ends because M is noetherian.
NOTES. The Jordan-H older and Krull-Schmidt theorems fail for nitely generated projective mod-
ules over non-principal Dedekind domains. For example, suppose that has a nonprincipal ideal a
of order 2 in the class group. Then aa ~, contradicting both theorems.
19 Dimension theory for noetherian rings
Let be a noetherian ring and let p be a prime ideal in . Let
p
=S
-1
where S =p.
We begin by studying extension and contraction of ideals with respect to the homomorphism

p
(cf. 2.9). Recall (5.7) that
p
is a local ring with maximal ideal p
e
def
=p
p
. The
ideal
_
p
n
_
ec
={a [ sa p
n
for some s S]
is called the nth symbolic power of p, and is denoted p
(n)
. If m is maximal, then m
(n)
=m
n
(see 5.8).
LEMMA 19.1. The ideal p
(n)
is p-primary.
PROOF. According to Proposition 17.3, the ideal (p
e
)
n
is p
e
-primary. Hence (see 17.4),
((p
e
)
n
)
c
is (p
e
)
c
-primary. But p
ec
=p (see 5.4), and
(((p
e
)
n
)
c
2.10
= ((p
n
)
e
)
c
def
=p
(n)
. (47)
LEMMA 19.2. Consider ideals a p
t
p with p
t
prime. If p
t
is a minimal prime ideal of a,
then p
te
is a minimal prime ideal of a
e
(extension relative to
p
).
PROOF. If not, there exists a prime ideal p
tt
=p
te
such that p
te
p
tt
a
e
. Now, by (5.4),
p
t
=p
tec
and p
ttc
=p
tec
, and so
p
t
=p
tec
p
ttc
a
ec
a
contradicts the minimality of p
t
.
2
THEOREM 19.3 (KRULLS PRINCIPAL IDEAL THEOREM). Let be a noetherian ring. For
any nonunit b , the height of a minimal prime ideal p of (b) is at most one.
PROOF. Consider
p
. According to Lemma 19.2, p
e
is a minimal prime ideal of
(b)
e
=(
b
1
), and (5.4) shows that the theorem for
p
p
e
(
b
1
) implies it for p (b).
Therefore, we may replace with
p
, and so assume that is a noetherian local ring with
maximal ideal p.
Suppose that p properly contains a prime ideal p
1
: we have to show that p
1
p
2
==
p
1
=p
2
.
19 DIMENSION THEORY FOR NOETHERIAN RINGS 86
Let p
(i)
1
be the rth symbolic power of p
1
. The only prime ideal of the ring ,(b) is
p,(b), and so ,(b) is artinian (apply 7.6). Therefore the descending chain of ideals
_
p
(1)
1
(b)
_
,(b)
_
p
(2)
1
(b)
_
,(b)
_
p
(3)
1
(b)
_
,(b)
eventually becomes constant: there exists an s such that
p
(x)
1
(b) =p
(x1)
1
(b) =p
(x2)
1
(b) = . (48)
We claim that, for any m_s,
p
(n)
1
(b)p
(n)
1
p
(n1)
1
. (49)
Let . p
(n)
1
. Then
. (b) p
(n)
1
(48)
= (b) p
(n1)
1
.
and so . = ab .
t
with a and .
t
p
(n1)
1
. As p
(n)
1
is p
1
-primary (see 19.1) and
ab =..
t
p
(n)
1
but b p
1
, we have that a p
(n)
1
. Now . =ab .
t
(b)p
(n)
1
p
(n1)
1
as claimed.
We next show that, for any m_s,
p
(n)
1
=p
(n1)
1
.
As b p, (49) shows that p
(n)
1
,p
(n1)
1
=p
_
p
(n)
1
,p
(n1)
1
_
, and so p
(n)
1
,p
(n1)
1
=0 by
Nakayamas lemma (3.9).
Now
p
x
1
p
(x)
1
=p
(x1)
1
=p
(x2)
1
=
and so p
x
1

_
n_x
p
(n)
1
. Note that
_
n_x
p
(n)
1
(47)
=
_
n_x
((p
e
1
)
n
)
c
=(
_
n_x
(p
e
1
)
n
)
c
3.15
= (0)
c
.
and so for any . p
x
1
, there exists an a p
1
such that a. =0. Let . p
1
; then a.
x
=0
for some a p
1
p
2
, and so . p
2
(because p
2
is prime). We have shown that
p
1
=p
2
, as required.
2
In order to extend Theorem 19.6 to non principal ideals, we shall need a lemma.
LEMMA 19.4. Let p be a prime ideal in a noetherian ring , and let S be a nite set of
prime ideals in , none of which contains p. If there exists a chain of distinct prime ideals
p p
d-1
p
0
.
then there exists such a chain with p
1
not contained in any ideal in S.
PROOF. We rst prove this in the special case that the chain has length 2. Suppose that
p p
1
p
0
are distinct prime ideals and that p is not contained in any prime ideal in S.
According to Proposition 2.7, there exists an element
a p(p
0
L
_
{p
t
S]).
19 DIMENSION THEORY FOR NOETHERIAN RINGS 87
As p contains (a) p
0
, it also contains a minimal prime ideal p
t
1
of (a) p
0
. Now p
t
1
,p
0
is a minimal prime ideal of the principal ideal ((a) p
0
),p
0
in ,p
0
, and so has height 1,
whereas the chain p,p
0
p
1
,p
0
p
0
,p
0
shows that p,p
0
has height at least 2. Therefore
p p
t
1
p
0
are distinct primes, and p
t
1
S because it contains a. This completes the proof
of the special case.
Now consider the general case. On applying the special case to p p
d-1
p
d-2
, we
see that there exists a chain of distinct prime ideals p p
t
d-1
p
d-2
such that p
t
d-1
is not
contained in any ideal in S. Then on applying the special case to p
t
d-1
p
d-2
p
d-1
, we
we see that there exists a chain of distinct prime ideals p p
t
d-1
p
t
d-2
p
d-2
such that
p
t
d-2
is not contained in any ideal in S. Repeat the argument until the proof is complete.
2
THEOREM 19.5. Let be a noetherian ring. For any proper ideal a = (a
1
. . . . . a
n
), the
height of a minimal prime ideal of a is at most m.
PROOF. For m=1, this was just proved. Thus, we may suppose that m_ 2 and that the
theorem has been proved for ideals generated by m1 elements. Let p be a minimal prime
ideal of a, and let p
t
1
. . . . . p
t
t
be the minimal prime ideals of (a
2
. . . . . a
n
). Each p
t
i
has height
at most m1. If p is contained in one of the p
t
i
, it will have height _m1, and so we may
suppose that it isnt.
Let p have height J. We have to show that J _m. According to the lemma, there exists
a chain of distinct prime ideals
p =p
d
p
d-1
p
0
. J _1.
with p
1
not contained in any p
t
i
, and so Proposition 2.7 shows that there exists a
b p
1

_
i
i=1
p
t
i
.
We next show that p is a minimal prime ideal of (b. a
2
. . . . . a
n
). Certainly p contains a
minimal prime ideal p
t
of this ideal. As p
t
(a
2
. . . . . a
n
), p contains one of the p
t
i
s, but, by
construction, it cannot equal it. If p =p
t
, then
p p
t
p
i
are distinct ideals, which shows that p
def
= p,(a
2
. . . . . a
n
) has height at least 2 in

def
=
,(a
2
. . . . . a
n
). But p is a minimal ideal in

of the principal ideal (a
1
. . . . . a
n
),(a
2
. . . . . a
n
),
which contradicts Theorem 19.3. Hence p is minimal, as claimed.
But now p,(b) is a minimal prime ideal of (b. a
2
. . . . . a
n
) in 1,(b), and so the height
of p,(b) is at most m1 (by induction). The prime ideals
p,(b) =p
d
,(b) p
d-1
,(b) p
1
,(b)
are distinct, and so J 1 _m1. This completes the proof that J =m.
2
The height of an ideal a in a noetherian ring is the minimum height of a prime ideal
containing it,
ht(a) = min
pa, p prime
ht(p).
The theorem shows that ht(a) is nite.
The following provides a (strong) converse to Theorem 19.5.
19 DIMENSION THEORY FOR NOETHERIAN RINGS 88
THEOREM 19.6. Let be a noetherian ring, and let a be a proper ideal of of height r.
Then there exist r elements a
1
. . . . . a
i
of a such that, for each i _r, (a
1
. . . . . a
i
) has height
i .
PROOF. If r =0, then we take the empty set of a
i
s. Thus, suppose that r _ 1. There are
only nitely many prime ideals of height 0, because such an ideal is a minimal prime ideal
of (0), and none of these ideals can contain a because it has height _ 1. Proposition 2.7
shows that there exists an
a
1
a
_
{prime ideals of height 0].
By construction, (a
1
) has height at least 1, and so Theorem 19.3 shows it has height exactly
1.
This completes the proof when r =1, and so suppose that r _2. There are only nitely
many prime ideals of height 1 containing (a
1
) because such an ideal is a minimal prime
ideal of (a
1
), and none of these ideals can contain a because it has height _2. Choose
a
2
a
_
{prime ideals of height 1 containing (a
1
)].
By construction, (a
1
. a
2
) has height at least 2, and so Theorem 19.5 shows that it has height
exactly 2.
This completes the proof when r =2, and when r > 2 we can continue in this fashion
until it is complete.
COROLLARY 19.7. Every prime ideal of height r in a noetherian ring arises as a minimal
prime ideal for an ideal generated by r elements.
PROOF. According to the theorem, an ideal a of height r contains an ideal (a
1
. . . . . a
i
) of
height r. If a is prime, then it is a minimal ideal of (a
1
. . . . . a
i
).
2
COROLLARY 19.8. Let be a commutative noetherian ring, and let a be an ideal in that
can be generated by n elements. For any prime ideal p in containing a,
ht(p,a) _ht(p) _ht(p,a) n.
PROOF. The rst inequality follows immediately from the correspondence between ideals
in and in ,a.
Denote the quotient map
t
def
= ,a by a a
t
. Let ht(p,a) = J. Then there
exist elements a
1
. . . . . a
d
in such that p,a is a minimal prime ideal of (a
t
1
. . . . . a
t
d
). Let
b
1
. . . . . b
n
generate a. Then p is a minimal prime ideal of (a
1
. . . . . a
d
. b
1
. . . . . b
n
), and hence
has height _J n.
2
We now use dimension theory to prove a stronger version of generic atness (10.13).
THEOREM 19.9 (GENERIC FREENESS). Let be a noetherian integral domain, and let T
be a nitely generated -algebra. For any nitely generated T-module M, there exists a
nonzero element a of such that M
o
is a free
o
-module.
20 REGULAR LOCAL RINGS 89
PROOF. Let J be the eld of fractions of . We prove the theorem by induction on the
Krull dimension of J

T, starting with the case of Krull dimension 1. Recall that this

means that J

T =0, and so a1
B
=0 for some nonzero a . Then M
o
=0, and so
the theorem is trivially true (M
o
is the free
o
-module generated by the empty set).
In the general case, an argument as in (10.14) shows that, after replacing , T, and M
with
o
, T
o
, and M
o
for a suitable a , we may suppose that the map T J

T
is injective we identify T with its image. The Noether normalization theorem (6.26)
shows that there exist algebraically independent elements .
1
. . . . . .
n
of J

T such that
J

T is a nite J.
1
. . . . . .
n
|-algebra. As in the proof of (10.13), there exists a nonzero
a such that T
o
is a nite
o
.
1
. . . . . .
n
|-algebra. Hence M
o
is a nitely generated

o
.
1
. . . . . .
n
|-module.
As any extension of free modules is free
24
, Proposition 3.5 shows that it sufces to
prove the theorem for M
o
=
o
.
1
. . . . . .
n
|,p for some prime ideal p in
o
.
1
. . . . . .
n
|. If
p = 0, then M
o
is free over
o
(with basis the monomials in the .
i
). Otherwise, J

(
o
.
1
. . . . . .
n
|,p) has Krull dimension less than that of J

T, and so we can apply the

induction hypothesis.
2
20 Regular local rings
Throughout this section, is a noetherian local ring with maximal ideal m and residue eld
k. The Krull dimension J of is equal to the height of m, and
ht(m)
(19.5)
_ minimum number of generators of m
(3.11)
= dim
k
(m,m
2
).
When equality holds, the ring is said to be regular. In other words, dim
k
(m,m
2
) _ J,
and equality holds exactly when the ring is regular.
For example, when has dimension zero, it is regular if and only if its maximal ideal
can be generated by the empty set, and so is zero. This means that is a eld; in particular,
it is an integral domain. The main result of this section is that all regular rings are integral
domains.
LEMMA 20.1. Let be a noetherian local ring with maximal ideal m, and let c mm
2
.
Denote the quotient map
t
def
=,(c) by a a
t
. Then
dim
k
m,m
2
=dim
k
m
t
,m
t2
1
where m
t
def
=m,(c) is the maximal ideal of
t
.
PROOF. Let e
1
. . . . . e
n
be elements of m such that {e
t
1
. . . . . e
t
n
] is a k-linear basis for m
t
,m
t2
.
We shall show that {e
1
. . . . . e
n
. c] is a basis for m,m
2
.
As e
t
1
. . . . . e
t
n
span m
t
,m
t2
, they generate the ideal m
t
(see 3.11), and so m=(e
1
. . . . . e
n
)
(c), which implies that {e
1
. . . . . e
n
. c] spans m,m
2
.
Suppose that a
1
. . . . . a
n1
are elements of such that
a
1
e
1
a
n
e
n
a
n1
c 0 mod m
2
. (50)
24
If M
t
is a submodule of M such that M
tt
def
=M,M
t
is free, then M ~M
t
M
tt
.
20 REGULAR LOCAL RINGS 90
Then
a
t
1
e
t
1
a
t
n
e
t
n
0 mod m
t2
,
and so a
t
1
. . . . . a
t
n
m
t
. It follows that a
1
. . . . . a
n
m. Now (50) shows that a
n1
c m
2
.
If a
n1
m, then it is a unit in , and c m
2
, which contradicts its denition. Therefore,
a
n1
m, and the relation (50) is the trivial one.
2
PROPOSITION 20.2. If is regular, then so also is ,(a) for any a mm
2
; moreover,
dim =dim,(a) 1.
PROOF. With the usual notations, (19.8) shows that
ht(m
t
) _ht(m) _ht(m
t
) 1.
Therefore
dim
k
(m
t
,m
t2
) _ht(m
t
) _ht(m) 1 =dim
k
(m,m
2
) 1 =dim
k
(m
t
,m
t2
).
Equalities must hold throughout, which proves that
t
is regular with dimension dim1.
2
THEOREM 20.3. Every regular noetherian local ring is an integral domain.
PROOF. Let be a regular local ring of dimension J. We have already noted that the
statement is true when J =0.
We next prove that is an integral domain if it contains distinct ideals a p with a =(a)
principal and p prime. Let b p, and suppose that b a
n
=(a
n
) for some n _ 1. Then
b =a
n
c for some c . As a is not in the prime ideal p, we must have that c p a, and
so b a
n1
. Continuing in this fashion, we see that b
_
n
a
n
3.15
= {0]. Therefore p ={0],
and so is an integral domain.
We now assume J _ 1, and proceed by induction on J. Let a mm
2
. As ,(a) is
regular of dimension J 1, it is an integral domain, and so (a) is a prime ideal. If it has
height 1, then the last paragraph shows that is an integral domain. Thus, we may suppose
that, for all a mm
2
, the prime ideal (a) has height 0, and so is a minimal prime ideal
of . Let S be the set of all minimal prime ideals of recall (17) that S is nite. We
have shown that mm
2

_
{p [ p S], and so m m
2
L
_
{p [ p S]. It follows from
Proposition 2.7 that either mm
2
(and hence m=0) or m is a minimal prime ideal of ,
but both of these statements contradict the assumption that J _1.
2
COROLLARY 20.4. A regular noetherian local ring of dimension 1 is a principal ideal
domain (with a single nonzero prime ideal).
PROOF. Let be a regular local ring of dimension 1 with maximal ideal m, and let a be
a nonzero proper ideal in . The conditions imply that m is principal, say m =(t ). The
radical of a is m because m is the only prime ideal containing a, and so a m
i
for some
r (by 3.16). The ring ,m
i
is local and artinian, and so a =(t
x
) m
i
for some s _1 (by
7.8). This implies that a =(t
x
) by Nakayamas lemma (3.9).
2
THEOREM 20.5. Let be a regular noetherian local ring.
(a) For any prime ideal p in , the ring
p
is regular.
(b) The ring is a unique factorization domain (hence is integrally closed).
PROOF. Omitted for the moment.
2
The best proof uses homological methods. See May, RegularLocal.pdf or Matsumura
1986 19.3, 20.3.
21 COMPLETIONS 91
21 Completions
Let be a ring and a an ideal in . For any -module, we get an inverse system of quotient
maps
M,aM M,a
2
M M,a
n
M
whose limit we dene to be the a-adic completion

M of M:

M
def
=lim

M,a
n
M.
For example, the a-adic completion of is

def
=lim

n
,a
n
.
We now explain why this is called the completion. Let M be an -module. A ltration
on M is a sequence of submodules
M =M
0
M
n
.
LEMMA 21.1. Let (M
n
)
nN
be a ltration on an -module M. There is a unique topology
on M such that, for each . M, the set {. M
n
[ n N] is a fundamental system of
neighbourhoods for .. The completion

M of M relative to this topology is canonically
isomorphic to lim

M,M
n
.
PROOF. The rst statement is obvious. For the second, recall that

M consists of the
equivalence classes of Cauchy sequences in M. Let (m
n
)
nN
be a Cauchy sequence. For
each n, the image of m
i
in M,M
n
becomes constant for large i let m
n
denote the constant
value. The family ( m
n
)
nN
depends only on the equivalence class of the Cauchy sequence
(m
n
)
nN
, and
(m
n
)| ( m
n
):

M lim

M,M
n
is an isomorphism.
2
Let be a ring and let a be an ideal in . A ltration (M
n
)
nN
on an -module M is
an a-ltration if aM
n
M
n1
for all n. An a-ltration is stable if aM
n
=M
n1
for all
sufciently large n.
LEMMA 21.2. Any two stable a-ltrations on an -module M dene the same topology on
M.
PROOF. It sufces to show that a stable a-ltration (M
n
)
nN
denes the a-adic topology
on M. As aM
n
M
n1
for all n, we have that a
n
M M
n
for all n. For some n
0
,
aM
n
=M
n1
for all n _n
0
, and so M
nn
0
=a
n
M
n
0
a
n
M.
2
LEMMA 21.3 (ARTIN-REES). If is noetherian and M is nitely generated, then, for any
-submodule M
t
of M, the ltration (M
t
a
n
M)
nN
on M
t
is a stable a-ltration.
PROOF. Omitted for the moment.
2
PROPOSITION 21.4. For every noetherian ring and ideal a, the functor M

M is exact
on nitely generated -modules.
21 COMPLETIONS 92
PROOF. Let
0 M
t
M M
tt
0
be an exact sequence of -modules. For each n, the sequence
0 M
t
a
n
M a
n
M a
n
M
tt
0
is exact, and so
0 M
t
,(M
t
a
n
M) M,a
n
M M
tt
,a
n
M
tt
0
is exact. On passing to the inverse limit, we obtain an exact sequence
0 lim

n
M
t
,(M
t
a
n
M)

M

M
tt
0,
but the last three lemmas show that lim
n
M
t
,(M
t
a
n
M) is the a-adic completion of M
t
.
2
PROPOSITION 21.5. For every ideal a in a noetherian ring and nitely generated -
module M, the homomorphism
amam:

M

M
is an isomorphism.
PROOF. In other words, when is noetherian, the functors M

M and M

M
agree on nitely generated -modules M. This is obvious for M =, and it follows for
nitely generated free -module because both functors take nite direct sums to direct sums.
Choose a surjective homomorphism
n
M, and let N be its kernel. The exact sequence
0 N
n
M 0
gives rise to a exact commutative diagram

M 0
0

N

n

M 0
o :
b
Because the middle vertical arrow is an isomorphism, the arrow b is surjective. But M
is arbitrary, and so the arrow a is also surjective, which implies that the arrow b is an
isomorphism.
2
PROPOSITION 21.6. For every noetherian ring and ideal a, the a-adic completion

of
is a at -algebra.
PROOF. It follows from (21.4) and (21.5) that

is exact on nitely generated -

modules, but this implies that it is exact on all -modules.
2
REFERENCES 93
References
BOURBAKI, N. AC. Alg` ebre Commutative.

El ements de math ematique. Hermann; Masson, Paris.
Chap. IIV Masson 1985; Chap. VVII Hermann 1975; Chap. VIII-IX Masson 1983; Chap. X
Masson 1998.
CARTIER, P. 2007. A primer of Hopf algebras, pp. 537615. In Frontiers in number theory, physics,
and geometry. II. Springer, Berlin. Preprint available at IHES.
KRULL, W. 1938. Dimensionstheorie in stellenringen. J. Reine Angew. Math. 179:204226.
MATSUMURA, H. 1986. Commutative ring theory, volume 8 of Cambridge Studies in Advanced
Mathematics. Cambridge University Press, Cambridge.
NAGATA, M. 1962. Local rings. Interscience Tracts in Pure and Applied Mathematics, No. 13.
Interscience Publishers, New York-London.
NORTHCOTT, D. G. 1953. Ideal theory. Cambridge Tracts in Mathematics and Mathematical Physics,
No. 42. Cambridge, at the University Press.
RAYNAUD, M. 1970. Anneaux locaux hens eliens. Lecture Notes in Mathematics, Vol. 169. Springer-
Verlag, Berlin.
Index

, 9
algebra, 3
nite, 3
nitely generated, 3
nitely presented, 3
symmetric, 40
tensor, 40
algebraically independent, 31
annihilator, 10
axiom of dependent choice, 9
belong to, 78
coefcient
leading, 11
components
irreducible, 61
conductor, 71
content of a polynomial, 16
contraction
of an ideal, 7
Cramers rule, 23
decomposition
minimal primary, 77
primary, 77
Dedekind domain, 81
degree
of a polynomial, 17
total, 17
dimension
Krull, 12
directed, 34
discrete valuation ring, 80
domain
unique factorization, 14
element
integral over a ring, 23
irreducible, 14
prime, 14
extension
of an ideal, 7
faithfully at, 41
at, 41
generate
an algebra, 3
height, 87
of a prime ideal, 12
homomorphism
nite, 3
nite type, 3
of algebras, 3
ideal, 4
generated by a subset, 4
irreducible, 78
maximal, 5
minimal prime, 77
primary, 76
prime, 4
principal, 4
radical, 6
idempotent, 3
trivial, 3
identity element, 2
integral closure, 25
integral domain, 3
integrally closed, 26
normal, 26
k(p), 4
lemma
Gausss, 16
Nakayamas, 11
Zariskis, 54
limit
direct, 34
map
bilinear, 37
module
artinian, 32
faithful, 24
nitely presented, 47
noetherian, 9
monomial, 17
multiplicative subset, 5
saturated, 23
nilpotent, 6
nilradical, 6
orthogonal idempotents, 3
complete set of, 3
polynomial
primitive, 16
primary, 77
radical
94
INDEX 95
Jacobson, 6
of an ideal, 5
relations
between generators, 47
relatively prime, 8
ring
artinian, 32
Jacobson, 62
local, 6
noetherian, 9
reduced, 6
regular local, 89
set
directed, 34
spectrum, 62
subring, 3
symbolic power, 85
system
direct, 34
tensor product
of algebras, 38
of modules, 37
theorem
Chinese remainder, 8
generic atness, 45
generic freeness, 88
Hilbert basis, 11
invariant factor, 84
Krull intersection, 13
Krulls principal ideal, 85
modules over Dedekind domain, 84
Noether normalization, 31
Nullstellensatz, 55
strong Nullstellensatz, 55
unique factorization of ideals, 82
topological space
irreducible , 60
noetherian, 59
quasi-compact, 59
topology
Zariski, 57
unit, 2