4.

4 Noetherian Rings
Recall that a ring A is Noetherian if it satisfies
the following three equivalent conditions:

(1) Every nonempty set of ideals of A has a

maximal element (the maximal condition);
(2) Every ascending chain of ideals is stationary
(the ascending chain condition (a.c.c.));
(3) Every ideal of A is finitely generated.

806
Later in this section we will prove
Hilberts Basis Theorem
which says that a polynomial ring in one
indeterminate over a Noetherian ring is itself
Noetherian.
In particular, by iteration, the polynomial ring
F [x1, . . . , xn] over a field F is Noetherian.

It will follow quickly that all finitely generated

rings are Noetherian.

807
But first we will prove that all proper ideals in
Noetherian rings have primary decompositions,
and simplify the First Uniqueness Theorem
concerning the uniqueness of associated prime ideals.

Call an ideal I of a ring A irreducible if, for all

ideals J , K of A ,


I = J K = I = J or I = K .

Lemma: Every ideal of a Noetherian ring is a

finite intersection of irreducible ideals.

808
Proof: Suppose the set

= { J
A | J is not a finite
intersection of irreducible ideals }

is nonempty. By the maximal condition, has a

maximal element M .
Certainly M is not irreducible, so M = J K
for some ideals J , K such that

J 6= M 6= K .
809
But M J and M K , so, by maximality of
M in ,

J 6 and K 6 .

Hence J and K are both finite intersections of

irreducible ideals,
so M = J K also is such an intersection,
contradicting that M .
Hence = , and the Lemma is proved.
810
 Lemma: Every proper irreducible ideal of a
Noetherian ring is primary.

Proof: Let A be a Noetherian ring. If I is a

proper ideal then I is irreducible iff the zero ideal
of A/I is irreducible.

It suffices then to suppose A is nonzero and {0}

is irreducible, and prove that {0} is primary.

Let x, y A such that xy = 0 .

811
 We show x = 0 or y n = 0 for some n 1 .

Consider the chain of ideals

Ann (y) Ann (y 2) . . . Ann (y n) . . .

which is stationary, since A is Noetherian. Hence

(n 1) Ann (y n) = Ann (y n+1) = . . .

812
 We show hxi hy ni = {0} .

Suppose z hxi hy ni . Then

z = vx = wy n (v, w A) ,

so

wy n+1 = (wy n)y = vxy = v0 = 0 ,

so
813
 w Ann (y n+1) = Ann (y n) ,
yielding
z = wy n = 0 .
This proves hxi hy ni = {0} . By irreducibility of
{0} , we get

hxi = {0} or hy ni = {0} ,

whence x = 0 or y n = 0 , and we are done.

814
The previous two lemmas prove:

Theorem: Every proper ideal of a Noetherian

ring has a primary decomposition.

We can refine the First Uniqueness Theorem for

primary decompositions, in this context, but first
prove:

Proposition: Every ideal of a Noetherian ring

contains a power of its radical.

815
Proof: Let I be an ideal of a Noetherian ring A ,
so
r(I) = h x1 , . . . , xk i

for some x1, . . . , xk A . Then

ni
(i = 1, . . . , k)(ni 1) xi I.

Put
k
X
m = (ni 1) + 1 .
i=1

816
Observe that
k

m j1 jk
X
r(I) = h x 1 . . . xk | ji = m i .
i=1

k
X
But if ji = m then j n for some
i=1
{1, . . . , k}) .

m
Hence
each generator of r(I) lies in I , so
m
r(I) I.
817
 Corollary: The nilradical is nilpotent in a
Noetherian ring.

Proof: If A is Noetherian, then

N = r({0})

is the nilradical of A , so, by the previous

Proposition,
N m {0} for some m 1 ,
whence equality holds, which proves N is nilpotent.
818
Exercises: (1) Find an example of a Noetherian
ring whose Jacobson radical does not equal the
nilradical.
(2) Show that if a ring satisfies the d.c.c. on
ideals then the nilradical and Jacobson radical are
equal.
(3) Find an example of an ideal I of a ring
A which does not contain a power of its radical
r(I) (so necessarily A is not Noetherian).

819
 Theorem: Let A be Noetherian, Q and M
ideals of A with M maximal. TFAE
(i) Q is M -primary;
(ii) r(Q) = M ;
(iii) Mn Q M (n > 0) .

Proof: (i) = (ii): follows by definition.

(ii) = (i): follows (regardless of whether A is

Noetherian) by an earlier result (page 684).
820
(ii) = (iii): follows from the previous Proposition.

(iii) = (ii): if M n Q M for some n > 0

then

M = r(M n) r(Q) r(M ) = M ,

so r(Q) = M .

Our final observation about primary decompositions

of ideals in Noetherian rings is a refinement of the
First Uniqueness Theorem (page 701):
821
 Theorem: Let A be a Noetherian ring and I
a proper ideal. Then the prime ideals belonging
to I are the prime ideals in

{ (I : x) | x A } .

n
\
Proof: Let I = Qi be a minimal primary
i=1
decomposition of I , and put

Pi = r(Qi) (i) .
822
Then, by the First Uniqueness Theorem,

{ P1 , . . . , Pn } = { prime ideals P |
(x A) P = r(I : x) }

If x A and (I : x) is prime, then

r(I : x) = (I : x)

is prime, so (I : x) {P1, . . . , Pn} .

Conversely, let i {1, . . . , n} . By the earlier

823
Proposition (page 815),

(m 1) Qi Pim .
\
Put R = Qj . Then
j6=i

R Pim R Pim R Qi = I .

Let m0 be the least integer such that

m0
R Pi I.
824
If m0 = 0 then

R I R,

so I = R , which contradicts minimality of the

primary decomposition of I .

Hence m0 1 , and so we may choose some

x R P m01 \ I .

In particular x R\I , so x 6 Qi . As in the

825
proof of the First Uniqueness Theorem (see pages
706-707), we have

(I : x) = (Qi : x)

and
r(I : x) = r(Qi : x) = Pi .
Certainly then

(I : x) r(I : x) Pi .
826
Also,
m01
Pi x R Pi Pi = R P m0 I ,

so Pi (I : x) , whence equality holds.

This shows

{ P1 . . . , Pn } = { prime ideals (I : x) | x A } ,

and completes the proof of the Theorem.

827
We now investigate the preservation of the
property of being Noetherian under certain natural
constructions.

We have already observed (on page 766) that

homomorphic images of Noetherian rings are

Noetherian,

and (on page 766) that

828
 a finitely generated module over a Noetherian
ring is Noetherian.

Theorem: Let A be a subring of a ring B .

Suppose that A is Noetherian and B is finitely
generated as an A-module.
Then B is a Noetherian ring.

829
Proof: By the immediately preceding observation,
B is a Neotherian A-module.
But all ideals of B are also A-submodules of B
(though not necessarily conversely).
Since A-submodules satisfy the a.c.c., so do ideals
of B , so B is a Noetherian ring.

Example: The ring Z[i] of Gaussian integers is a

finitely generated Z-module, and Z is Noetherian.
By the previous Theorem, Z[i] is a Noetherian ring.
830
 Theorem: Rings of fractions of Noetherian
rings are Noetherian.

Proof: Let A be a Noetherian ring and S a

multiplicatively closed subset.
Let J
S 1A , so
1
J = S I (I
A) .

since all ideals of S 1A are extended. But A is

Noetherian, so

I = h x1 , . . . , xn i
831
for some x1, . . . xn A , whence

J = h x1/1 , . . . , xn/1 i.

Thus all ideals of S 1A are finitely generated, which

shows S 1A is Noetherian.

Hilberts Basis Theorem: Let A be a

Noetherian ring. Then the polynomial ring
A[x] is Noetherian.

832
Proof: We prove that all ideals of A[x] are finitely
generated. Consider {0} = 6 J
A[x] , and put

I = { leading coefficients of polynomials in J } .

It is easy to show that I
A , so

I = h a1 , . . . , an i

for some a1, . . . , an A , since A is Noetherian.

833
Then, for each i = 1, . . . , n , there is a polynomial

pi(x) = aixdi + ( terms of lower degree )

in J , for some di 0 . Put

J = h p1(x) , . . . , pn(x) i

and
d = max { d1 , . . . , dn } .
834
Let
M = { q(x) A[x] | degree of q(x) d } .

Claim: J = (J M ) + J .

Clearly (J M ) + J J . Conversely let

0 6= p(x) J . Then
p(x) = axm + ( terms of lower degree )
for some m 0 .
835
 If m d then p(x) J M .

Suppose m > d . Since a I

Pn
a = i=1 ui ai (u1, . . . , un A) .

Put
Pn
q(x) = p(x) i=1 ui xmdi pi(x) .

Then q(x) is an element of J of degree < m .

836
By an inductive hypothesis,

q(x) (J M ) + J ,

so
n
X
p(x) = q(x) + ui xndi pi(x)
i=1

(J M ) + J .
837
This proves J (J M ) + J , whence equality
holds, and the Claim is proved.

Regarded as an A-module, M is finitely

generated, by 1 , x , . . . , xd .

But A is Noetherian, so M is a Noetherian

A-module (by the Theorem on page 766).

But then J M , being an A-submodule of M

838
must be finitely generated as an A-module by, say,

q1(x) , . . . , qk (x) .

The ideal of A[x] generated by

q1(x) , . . . , qk (x) , p1(x) , . . . , pn(x)

therefore contains J M and J , and is contained

in J ,
so equals (J M ) + J = J .
839
Thus J is finitely generated, completing the proof
that A[x] is Noetherian.

Corollary: If A is Noetherian then so is

A[x1, . . . , xn] for every n 1 .

Proof: immediate by iterating Hilberts Basis

Theorem.

840
 Corollary: If A is a Noetherian ring and B
is a finitely generated A-algebra, then B is a
Noetherian ring.
In particular, every finitely generated ring, and
every finitely generated algebra over a field, is
Noetherian.

Proof: If A is Noetherian and B is generated as

an A-algebra by

b1 , . . . , bn ,
841
then B is a homomorphic image of the polynomial
ring A[x1, . . . , xn] under the map

p(x1, . . . , xn) 7 p(b1, . . . , bn) ,

so must be Noetherian,

since A[x1, . . . , xn] is Noetherian (by the previous

Corollary),

and homomorphic images of Noetherian rings are

Noetherian.
842
The last statement of the Corollary follows because
every field is Noetherian,
and every ring is an algebra over its subring generated
by 1 ,
which is isomorphic to either Zn , for some n , or
Z,

both of which are Noetherian.

843
The next (tricky) result says that, under certain
conditions, an intermediate ring B , sandwiched
between well-behaved rings A and C is itself
well-behaved.
C
B

844
Theorem: Suppose A B C is a chain
of subrings and that

(i) A is Noetherian;

(ii) C is finitely generated as an A-algebra;

(iii) C is finitely generated as a B-module.

Then B is finitely generated as an A-algebra.

845
Proof: Let
x1 , . . . , xm generate C as an A-algebra,
and
y1 , . . . , yn generate C as a B-module.

Then

(i = 1, . . . , m) ( bi1 , . . . , bin B )
n
X
xi = bik yk
k=1

846
and

(i, j {1, . . . , m}) ( cij1 , . . . , cijn B )

n
X
yi yj = cijk yk .
k=1

Let B0 be the A-subalgebra of B generated by

 
bik , cijk | i, j {1, . . . , m} , k {1, . . . , n} .

By the previous Corollary, B0 is a Noetherian ring.

847
Further
A B0 B C
is a chain of subrings.

Claim: C is generated by y1 , . . . , yn
as a B0-module.

Each element c of C can be expressed as a

polynomial in x1, . . . , xm with coefficients from
A.
848
But each xi and each product yiyj is a linear
combination of y1, . . . , yn with coefficients from
B0 ,

so that, after multiplying out, we can express c as a

linear combination of y1, . . . , yn with coefficients
from B0 , which proves the Claim.

By the Claim and the observation that B0 is

Noetherian, we conclude that

C is a Noetherian B0-module.
849
But B is a B0-submodule of C , so
B is a Noetherian B0-module,
so is finitely generated.
But B0 is finitely generated as an A-algebra, so,
finally,
B is also finitely generated as an A-algebra,
and the Theorem is proved.

850