B2.

2: Commutative Algebra
Nikolay Nikolov

Hilary term, 2018

All rings in this course will be assumed commutative and containing an

identity element. For a ring R we denote by R[t1 , . . . , tk ] the polynomial ring
in indeterminates ti with coefficients in R. A subset S of R is said to be
multiplicatively closed if 1 ∈ S and whenever x, y ∈ S then xy ∈ S.

Zorn’s Lemma
A partial order ≤ on a set X is a reflexive transitive relation such that a ≤ b
and b ≤ a implies a = b.
A chain C in a partially ordered set X is a subset C ⊆ X which is totally
ordered, i.e. for any x, y ∈ C we have a ≤ b or b ≤ a. The following result is
known as Zorn’s Lemma. It is equivalent to the Axiom of choice and also to
the Well-ordering principle.

Lemma 1 (Zorn’s Lemma) Let (X, ≤) be a partially ordered set such that
every chain of elements of X has an upper bound in X. Then X has a
maximal element.

A typical application of Zorn’s lemma is the existence of maximal ideals

in any unital ring R: Let X be the set of all ideals of R different from R
ordered by inclusion. Note that X is not empty since {0} ∈ X. If C is a
chain in X we easily check that ∪C ∈ X and so the condition of the lemma
is satisfied. Therefore X has maximal elements, i.e. maximal ideals.

1
1 Introduction
Commutative algebra has developed under the unfluence of two major sub-
jects: Algebraic Number Theory and Algebraic Geometry.
Recall that an ideal I of a ring R is prime if R/I is a domain, or equiva-
lently whenever the complement R\I is multiplicatively closed.
The main object of study in Algebraic Number theory is the ring of
integers O of a finite extension field K of Q. The ring O is an example of
a Dedekind domain: all nonzero prime ideals are maximal (in fact of finite
index in O), and moreover every ideal of O has a unique factorization into a
product of prime ideals.
The main object of study of (Affine) Algebraic geometry are the affine
algebraic varieties (which we will call algebraic sets in this course).
Let F be a field, k ∈ N and let R := k[t1 , . . . , tk ] be the polynomial ring
in k variables ti and let F k , denote the k-dimensional vector space of row
vectors.
Let Y ⊆ R be a collection of polynomials from R and define

V(S) := {x = (xi ) ∈ F k | f (x) = 0 ∀f ∈ S}

This is just the subset in F k of common zeroes for all polynomials in S (it
may happen of course that this is the empty set).
It is easy to see that V(S) = V(I) where I = hSi is the ideal generated
by S in R.

Definition 2 A set U ⊆ F k is an algebraic set if U = V(S) for some S ⊆ R

(equivalently U = V(I) for some ideal I of R).

We may consider an opposite operation associating an ideal to each subset

of F k .

Definition 3 Let Z ⊆ F k be any subset. Define

I(Z) := {f (t1 , . . . , tk ) ∈ R | f (x) = 0 ∀x ∈ Z}.

Thus I(Z) is the set of polynomials which vanish on all of Z. It is clear

that I(Z) is an ideal of R.

2
Proposition 4 For ideals I ⊆ I ′ ⊆ R and subsets Z ⊆ Z ′ ⊆ F k we have
(1) V(I(Z)) ⊇ Z, moreover there is equality if Z is an algebraic set.
(2) I(V(I)) ⊇ I,
(3) V(I) ⊇ V(I ′ ),
(4) I(Z) ⊇ I(Z ′ ).

Proof. Exercise.
The above proposition shows that I and V are order reversion maps
between the set of ideals of R and the algebraic subsets of F k . Moreover (1)
shows that V is surjective while I is injective. Understanding the relationship
between an algebraic set Z and the ideal I(Z) is the beginning of algebraic
geometry which we will address in Section 4.

2 Noetherian rings and modules

Let R be a ring and let M be an R-module. Recall that M is said toPbe finitely
generated if there exist elements m1 , . . . , mk ∈ M such that M = ki=1 Rmi .

Lemma 5 The following three conditions on M are equivalent.

(a) Any submodule of M is finitely generated.
(b) Any nonempty set of submodules of M has a maximal element under
inclusion.
(c) Any ascending chain of submodules N1 ≤ N2 ≤ N3 ≤ · · · eventually
becomes stationary.

Proof. (c) implies (b) is easy.

(b) implies (a): Let N be a submodule of M and let X be the collection
of finitely generated submodules of N . X contains {0} and so by (b) there
is a maximal element N0 ∈ X. We claim that N0 = N . Otheriwise there is
some x ∈ N \N0 and then N0 + Rx is a finitely generated submodule of N
which is larger than N , contradiction. So N0 = N is finitely generated.
(a) implies (c): Let N1 ≤ N2 ≤ · · · be an ascending chain of submodules
and let N := ∪∞ i=1 Ni . Then N is a submodule of M which is finitely generated
by (a). Suppose N is generated by elements x1 , . . . , xn . For each xi there is
some Nki such that xi ∈ Nki . Take k = maxi {ki }. We see that all xi ∈ Nk
and so N = Nk . Therefore the chain becomes stationary at Nk . 

3
Definition 6 An R-module M is said to be Noetherian if it satisfies any of
the three equivalent conditions of Lemma 5.

Proposition 7 Let N ≤ M be two R-modules. Then M is Noetherian if

and only if both N and M/N are Noetherian.

Proof. Problem sheet 1, Q4. 

As a consequence we see that M n := M ⊕ M ⊕ · · · ⊕ M is Noetherian for
any Noetherian module M .

Definition 8 A ring R is Noetherian if R is a Noetherian R-module.

Examples of Noetherian rings are fields, Z, PIDs and (as we shall see
momentarily) polynomial rings over fields. An example of a ring which
is not Noetherian is the polynomial ring of infinitely many indeterminates
Z[t1 , t2 , . . .].

Proposition 9 A homomorphic image of a Noetherian ring is Noetherian.

Proof. Let f : A → B be a surjective ring homomorphism with A Noethe-

rian. Then B ≃ A/ ker f and the ideals of B are in 1 − 1 correspondence
with the ideals of A containing ker f . Now A satisfies the ascending chain
condition on its ideals and therefore so does A/ ker f ≃ B.

Proposition 10 Let R be a Noetherian ring. Then an R-module M is

Noetherian if and only if M is finitely generated as an R-module.

Proof. If M is Noetherian then clearly M is finitely generated as a

module. Conversely, soppose that M = ki=1 Rmi for some mi ∈ M . Then
P
M is a homomorphic image of the free R-module Rk withP basis: Define the
k
module homomorphism f : R → M by f (r1 , . . . , rk ) := i ri mi . Since R
and Rk are Noetherian modules so is M ≃ Rk / ker f .
The main result of this section is

Theorem 11 (Hilbert’s Basis Theorem) Let R be a Noetherian ring. Then

the polynomial ring R[t] is Noetherian.

4
 Let A ≤ B be two rings. We say that B is finitely generated as A-algebra
(or that B is finitely generated as a ring over A) if there exists elements
b1 , . . . , bk ∈ B such that B = A[b1 , . . . , bk ] meaning that B is the smallest
ring containing A and all bi . This is equivalent to the existence of a surjective
ring homomorphism f : A[t1 , . . . , tk ] → B which is the identity on A and
f (ti ) = bi for each i.

Corollary 12 Let R be a Noetherian ring and suppose S ≥ R is a ring which

is finitely generated as R-algebra. Then S is a Noetherian ring.

Proof. The above discussion shows that S is a homomorphic image of the

polynomial ring R[t1 , . . . , tk ] and with Theorem 11 and induction on k we
deduce that R[t1 , . . . , tk ] is a Noetherian ring. Therefore S is a Noetherian
ring.
In particular this implies that the polynomial ring F [t1 , . . . , tk ] is a Noethe-
rian ring for any field F . This has the following central application to alge-
braic geometry.

Corollary 13 Let X ⊆ F [t1 , . . . , tk ] be any subset. Then there is a finite

subset Y ⊆ X such that V(X) = V(Y ).

Proof. Let I = hXi be the ideal generated by X in R = F [t1 , . . . , tk ]. Since

R is a Noetherian ring the ideal I is finitely generated and hence I = hY i
for some finite subset Y of X. Then V(X) = V(Y ). 

Proof of Theorem 11.

It is enough to show that any ideal I of R[t] is finitely generated. If I = {0}

this is clear. Suppose I is not zero. Let M be the ideal of R generated by
all leading coefficients of all non-zero polynomials in I. Then M is finitely
generated ideal and hence there are some P polynomials p1 , . . . , pk ∈ I such that
pi has leading coefficient ci and M = i Rc
N

i . Let N = max{deg pi | 1 ≤ i ≤
k} and let K = I ∩ R ⊕ Rt ⊕ · · · ⊕ Rt . Note that K is an R-submodule
of the Noetherian R-module RN and hence K is finitely generated as an
R-module, say by elements a1 , . . . , as ∈ K ⊂ I. Let J be the ideal of R[t]
generated by a1 , . . . , as , p1 , . . . , pk . We claim that J = I. Clearly J ≤ I and
it remains to prove the converse. Let f ∈PI and argue by induction on deg f
that f ∈ J. If deg f ≤ N then f ∈ K = i Rai and so f ∈ J. Suppose that

5
 P
deg f > N . Let a ∈ M be the leading coefficient of f . We have a = j rj cj
for some rj ∈ R. Consider the polynomial g := f − j rj tdeg f −deg pj pj and
P
note that deg g < deg f . Since g ∈ I we can assume from the induction
hypothesis that g ∈ J. Therefore f ∈ J. Hence I = J is finitely generated
ideal of R[t]. Therefore R[t] is a Noetherian ring. 

3 The Nilradical
A prime ideal P of a ring is said to be minimal if P does not contain another
prime ideal Q ⊂ P .

Theorem 14 Let R be a Noetherian ring. Then R has finitely many minimal

prime ideals and every prime ideal contains a minimal prime ideal.

Proof. Let’s say that an ideal I of R is good if I ⊇ P1 · · · Pk for some

prime ideals Pi , not necessarily distinct. We claim that all ideals of R are
good. Otherwise let X be the set of bad ideals and since R is Noetherian
there is a maximal element of X, call it J. Clearly J is not prime. So there
exist elements x, y outside J such that xy ∈ J. Let S = J + Rx, T = J + Ry,
we have ST ⊆ J and both S and T are strictly larger than J and hence
must be good ideals. Therefore P1 · · · Pk ⊆ S, P1′ · · · Pl′ ⊆ T for some prime
ideals Pi , Pi′ of R. But then P1 · · · Pk P1′ · · · Pl′ ⊆ T S ⊆ J and so J is good,
contradiction. So all ideals of R are good an in particular {0} is good and
so P1 · · · Pk = 0 for some prime ideals Pi . Let Y be the set of minimal ideals
from the set {P1 , . . . , Pk }. We claim that Y is the set of all minimal prime
ideals of R. Indeed if I is any prime ideal, then P1 · · · Pk ⊆ I and so Pi ⊆ I
for some i, justifying our claim. This also proves the second statement of the
theorem. 
Let I be any ideal of a Noetherian ring R. By appying the above the-
orem to the quotient ring R/I we deduce that there is a finite collection
{P1 , . . . , Pn } of prime ideals Pi of R which are minimal subject to I ⊆ Pi .
We will refer to {P1 , . . . , Pn } as the minimal primes of the ideal I.
An element x ∈ R is nilpotent if xn = 0 for some n. An ideal I is said to
be nilpotent if I n = 0 for some n ∈ N.
The set {x ∈ R| x nilpotent} of all nilpotent elements of R is an ideal of
R (exercise).

6
Definition 15 The nilradical of a ring R denoted by nilrad(R) is the set of
all nilpotent elements of R.

The nilradical may L∞not be nilpotent: consider the ideal generated by

k
t1 , t2 , . . . in the ring k=1 R[tk ]/(tk ) . However

Proposition 16 Let I be an ideal of a ring R consisting of nilpotent ele-

ments (such ideal is called a nil ideal). Suppose that I is finitely generated
as an ideal. Then I is nilpotent.

Proof. Let xi ∈ I such that I = hx1 , . . . , xk i = Rx1 + Rx2 + · · · Rxk .

Let xni i = 0 for some integers ni ∈ N and take m = n1 + · · · nk . Now
X
I n = (Rx1 + Rx2 + · · · Rxk )n ⊆ Rxs11 · · · xskk
s1 +···+sk =n

where the sum is over all tuples si subject to ki=1 si = n. We must have
P
s
at least one j such that sj ≥ nj and then xj j = 0. Therefore the right hand
side above is the zero ideal and so I n = 0.

Corollary 17 The nilradical of a Noetherian ring is nilpotent.

There is another very useful characterization of the nilradical.

Theorem 18 (Krull’s theorem) For any ring R, nilrad(R) is the inter-

section of all prime ideals of R.

Proof. If x is nilpotent and P is a prime ideal then xn = 0 ∈ P for some

n and so x ∈ P . So nilrad(R) ⊆ X := ∩{P | P prime ideal of R}. For the
converse suppose that x is not nilpotent. Let S = {xn |n ≥ 0}, then S is a
multiplicatively closed subset of R avoiding 0. By problem sheet 1 Q1 there
is a prime ideal P such that P ∩ S = ∅. So x 6∈ X. Thus X ⊆ nilrad(R) and
so nilrad(R) = X.

Definition 19 Let I be an ideal of R. The radical of I is defined to be

rad(I) := {x ∈ R | xn ∈ I, for some n ∈ N}.

7
 So by definition rad(I)/I = nilrad(R/I) from where we see by Theorem
18 the first part of the following.

Corollary 20 Let I be an ideal of a ring R. Then

(1) rad(I) = ∩{P | P prime ideal of R with I ⊆ P }
(2) If R is Noetherian then rad(I) = P1 ∩ · · · ∩ Pk for some prime ideals
Pi of R. There exists some n ∈ N such that rad(I)n ⊆ I.

Proof. It remains to prove (2). By considering R/I and applying The-

orem 14 we deduce that there are finitely many prime ideals, say P1 , . . . , Pk
minimal subject to I ⊆ Pi and every prime ideal Q above I contains some
Pi . It is now clear that r(I) = P1 ∩ · · · ∩ Pk . The last part follows from
Corollary 17 applied to the nil ideal r(I)/I of the Noetherian ring R/I.

3.1 Connection with algebraic sets

Recall the definitions of the maps V and I from the Introduction. The
following Proposition is an easy exercise.

Proposition 21 Let Ij , j = 1, 2, . . . be ideals of the polynomial ring R =

F [t1 , . . . , tk ]. Then
P
(1) V( j Ij ) = ∩j V(Ij ).
(2) V(I1 ∩ I2 ) = V(I1 I2 ) = V(I1 ) ∪ V(I2 ).
(3) rad I(Z) = I(Z) for any subset Z ⊆ F k .

When studying algebraic sets it is natural first to express them as union

of ’simpler’ algebraic sets. For example the algebraic set W = V(t1 t2 ) can be
written as W = L1 ∪ L2 , a union of the two lines Li = V(ti ), i = 1, 2. This
leads us to consider algebraic sets which cannot be decomposed further and
we make the following definition.

Definition 22 A non-empty algebraic set W is said to be irreducible if

whenever W = W1 ∪ W2 for some algebraic sets W1 , W2 then W1 = W
or W2 = W .

Proposition 23 An algebraic set W is irreducible if and only if I(W ) is a

prime ideal.

8
Proof. Suppose I(W ) is a prime ideal and W = W1 ∪W2 with each Wi 6= W .
Then I(Wi ) is strictly larger than I(W ) and we can take fi ∈ I(Wi )\I(W ).
Then the polynomial f1 f2 vanishes on both W1 and W2 hence it vanishes
on W and so f1 f2 ∈ I(W ). Thus I(W ) is not a prime ideal, contradiction.
Therefore W must be irreducible.
We leave the converse as an exercise in Problem sheet 2.

Theorem 24 Every algebraic set is a union of finitely many irreducible al-

gebric sets.

Proof. Problem sheet 2.

Suppose W is an algebraic set and W = V1 ∪ · · · ∪ Vn where Vi are
irreducible algebraic sets and n is minimal possible. Then Vi 6⊆ Vj for any
i, j otherwise we may omit Vi from the union. Now I(W ) = ∩ni=1 I(Vi ). If
P is a prime ideal containing I(W ) then P must contain at least one of the
ideals Pj := I(Vi ). It follows that P1 , . . . , Pn are precisely the minimal primes
of the ideal I(W ). Since Vi = V(Pi ) it follows that the irreducible sets Vi in
the minimal decomposition W = V1 ∪ · · · ∪ Vn are determined uniquely by
W and we refer to them as the irreducible components of W .
It remains to determine the relationship between the algebraic set W =
V(I) and the ideal I(W ). This is the topic of the next section.

4 The Nullstellensatz
We start with a technical result.

Proposition 25 Let A ⊆ B ⊆ C be three rings with A Noetherian. Sup-

pose that C is finitely generated as an A-algebra and also that C is finitely
generated as a B-module. Then B is finitely generated as A-algebra.
Pn
Proof. Suppose that C = i=1 Byi for some yi ∈ C. Let x1 , . . . xm generate
C as A-algebra. We have
n
X
xi = bij yj (1 ≤ i ≤ m)
j=1

9
 n
X
yj yk = bjkl yl (1 ≤ j, k ≤ n)
l=1

for some bij , bjkl ∈ B. Let B0 be the subring of B generated by A and all
the elements bij , bjkl . Then B0 is finitely generated as A-algebra and hence
by Theorem 11 P B0 is a Noetherian ring. We have A ⊆ B0 ⊆ B ⊆ C.
Let M = B0 + ni=1 B0 yi . By the definition of B0 it follows that A ⊆ M
and xi M ⊆ M for all i = 1, . . . , m. Therefore C = M . So C is finitely
generated as B0 -module and in particular C is a Noetherian B0 -module. Its
submodule B is therefore also a Noetherian B0 -module and hence it is finitely
as a B0 -module. In particular there are elements ls ∈ B such
generated P
that C = rs=1 B0 li . Then the set of all bij , bjkl , ls for all possible i, j, k, l, s
generates B as an A-algebra. 

4.1 Field extensions

Let F ⊆ E be two fields. By [E : F ] we denote dimF E, the dimension of E
as a vector space over F and we say that that the extension E/F is finite if
[E : F ] is finite. The following is mostly part A material.

Proposition 26 Let E/F be a field extension such that E = F (x) for some
element x ∈ E (meaning that E is the smallest field containing F and x).
The following are equivalent
(1) E/F is a finite extension.
(2) x is algebraic over F .
(3) E is generated by x as an F -algebra.
(4) E is finitely generated as an F -algebra.

Proof. The equivalence of (1),(2) and (3) is part A material. Clearly(3)

implies (4). It remains to prove that (4) implies (2).
Suppose that x is not algebraic but transcendental over F . Then E =
F (x) is the field of rational functions in the variable x. Suppose E is gen-
erated as F -algebra by the elements gi = pi (x)/qQ i (x), i = 1, . . . , k where
pi , qi ∈ F [x] are polynomials in x. Let r(x) = ki=1 qi and consider the
element a = 1/(xr(x) + 1) ∈ E. Then

a = f (g1 , . . . , gk )

10
for some polynomial f ∈ F [t1 , . . . , tk ]. By muliplying with appropriate power
of r to clear the denominators on RHS we reach the equation a = s(x)/r(x)n
for some n ∈ N and polynomial s(x) ∈ F [x]. Thus r(x)n = s(x)(xr(x) + 1)
which is impossible since xr(x) + 1 is coprime to r(x)n .

Theorem 27 (weak Nullstellensatz) Let F ⊆ E be two fields such that

E is finitely generated as an algebra over F . Then E/F is a finite extension.

Proof. Suppose E = F [x1 , . . . , xk ] and argue by induction on k. The case

k = 1 is the above Proposition 26. Assuming the result is true for k − 1
consider the sequence of fields F ⊆ F ′ ⊆ E where F ′ = F (x1 ). We have
that E is finitely generated as F ′ -algebra by k − 1 elements and hence by the
induction hypothesis E/F ′ is finite. So E is finitely generated as F ′ -module
and by Propostion 25 F ′ is finitely generated as F -algebra. Now Proposition
26 gives that F ′ /F is finite and therefore [E : F ] = [E : F ′ ][F ′ : F ] is finite.


Corollary 28 Let F be a field and let R be a finitely generated F -algebra.

Let M be a maximal ideal of R. Then dimF R/M is finite.

Proof. R/M is a field which is finitely generated as F -algebra.

The next corollary describes the maximal ideals of polynomial rings over
algebraically closed fields. First we need some notation.
Let F be a field and let R = F [t1 , . . . , tk ] be a polynomial ring. Let M
denote the set of maximal ideals of R and define a map µ : F k → M by
k
X
µ(a1 , . . . , ak ) := R(ti − ai ) = ht1 − a1 , . . . , tk − ak i
i=1

It is easy to check that µ(a1 , . . . , ak ) ∈ M and that the map µ is injective.

Corollary 29 Assume that the field F is algebraically closed. Then µ is

bijective.

Proof. It remains to show that µ is surjective. Let M be a maximal ideal

of R. By Corollary 28 R/M is a finite field extension of F , and since F is
algebraically closed, it follows that R/M ≃ F and so dimF R/M = 1. This
implies M + F = R. In particular for each ti there exists ai ∈ F such that
ti − ai ∈ M . Then µ(a1 , . . . , ak ) ⊆ M and hence M = µ(a1 , . . . , ak ).

11
Corollary 30 Let R be a polynomial ring over algebraically closed field F .
Let I be an ideal of R. Then V(I) = ∅ if and only if I = R. Moreover a ∈ F k
belongs to V(I) if and only if I ⊆ µ(a).

Proof. If R = I then 1 ∈ I and so V(I) = ∅. Conversely if I 6= R there

is a maximal ideal M ∈ M such that I ⊆ M . By Corollary 29 M = µ(a)
for some a ∈ F k . Notice that I{a} = µ(a). So if f ∈ I then f ∈ µ(a) and
hence f (a) = 0. Thus a ∈ V(I) and so V(I) 6= ∅. The second part follows
by the same argument. 
So the points of the algebraic set V(I) correspond to the maximal ideals
of R which contain I.
It remains to identify I(V(I)).

Theorem 31 (The Nullstellensatz) Let F be an algebraically closed field

and let R = F [t1 , . . . , tk ]. Let I be an ideal of R. Then
I(V(I)) = rad(I).

Proof. Let W = V(I). Let f ∈ rad(I) then f n ∈ I for some n ∈ N and

so f n is zero on W . Hence f vanishes on W and so f ∈ I(V(I). Conversely
suppose f ∈ I(V(I). We want to prove that f ∈ rad(I). If f = 0 this is clear,
so assume f 6= 0. Consider the polynomial ring S := R[z] = F [t1 , . . . , tk , z]
where we have added an extra indeterminate variable z. Let J be the ideal
of S generated by I together with the polynomial zf − 1. Observe that
V(J) = ∅: if the tuple (a, y) ∈ F k+1 (with a ∈ F k ) belongs to V(J) then
a ∈ W but then f (a) = 0 so zf − 1 = −1 is not zero. Hence by Corollary 30
we must have J = S. Therefore there are polynomials g, g1 , . . . , gm ∈ S and
f1 , . . . fm ∈ I such that
g(zf − 1) + g1 f1 + · · · + gm fm = 1
This is an identity of polynomials in variables t1 , . . . , tk , z. In particular it
ramains true when we substitute z = 1/f . Then gi become polynomials in
t1 , . . . , tk and 1/f . Bringing everything under a common denominator f n we
reach
g1′ f1 + · · · gm
′
fm
=1
fn
for some gi′ ∈ R. This implies f n = m ′
P
i=1 gi fi ∈ I since all fi ∈ I. Thus
f ∈ rad(I) and the Theorem is proved. 

12
Corollary 32 Let F and R be as in Theorem 31 and let I be an ideal of R.
Then rad(I) is an intersection of maximal ideals of R.

Proof. Let U be the intersection of all maximal ideals of R which contain

I. Clearly rad(I) ⊆ U (since rad(I) is the intersection of all prime ideals of
R which contain I).
Suppose now f 6∈ rad(I). By Theorem 31 we have f 6∈ I(V(I)) and so
there is some a ∈ V(I) such that f (a) 6= 0 and in particular f 6∈ µ(a). On
the other hand I ⊆ µ(a) and so µ(a) is a maximal ideal of R which contains
I. So f 6∈ U . Thus U ⊆ rad(I) and so we have equality U = rad(I). 
This leads us to the following definition.

Definition 33 The Jacobson radical J(R) of a ring R is defined to be the

intersection of all maximal ideals of R.

Clearly nilrad(R) ⊆ J(R).

Definition 34 A ring R is said to be a Jacobson ring if J(R/I) = rad(I)/I =

nilrad(R/I) for each ideal I of R. Equivalently R is a Jacobson ring if each
prime ideal of R is an intersection of maximal ideals.

So in Corollary 32 we have proved that F [t1 , . . . , tk ] is a Jacobson ring

whenever F is an algebraically closed field. In fact more is true: any finitely
generated algebra over a field is a Jacobson ring. We will prove this later
once we have developed a new tool: the notion of integral ring extensions.

5 The Cayley-Hamilton Theorem, Nakayama’s

lemma
Theorem 35 Let R be a ring and let M be a finitely generated R-module.
Let I be an ideal of R and φ : M → M be an endomorphism of M such that
φ(M ) ⊆ IM . There exist a1 , . . . , an ∈ I such that the module homomorphism

φn + a1 φn−1 + · · · + an = 0

as a map on M .

13
 of M , so M = ni=1 Rxi . There
P
Proof. Let x1 , . . . , xn ∈ M be generators
Pn
exist ci,j ∈ I such that φ(xi ) = j=1 cji xj . Let C = (ci,j ) and consider C
as a matrix in Mn (R[t]). Let p = p(t) = det(t · In − C) be the characteristic
polynomial of C and note that p(t) = tn + a1 tn−1 + · · · + an where ai ∈ I since
ai is a polynomial in the coefficients ci,j of C. From the Cayley-Hamilton
theorem in part A we have p(C) = 0 and hence p(φ) is the zero map on M
because φ acts as C on M .

Corollary 36 (Nakayama’s Lemma) Let M be a finitely generated R-

module and let I be an ideal of R such that M = IM . Then there exists
x ∈ I such that (1 + x)M = 0.

Proof. Take φ = IdM in Theorem 35. Then P there exist ai ∈ I such that
(1 + a1 + · · · + an )M = 0 and we can take x = ni=1 ai .
The above corollary has an important special case (which is sometimes
also stated as Nakayama’s lemma).

Corollary 37 Let R be a ring and M be a finitely generated R-module such

that M = JM , where J = J(R) is the Jacobson radical of R. Then M = {0}.

Proof. Problem sheet 3.

Corollary 38 Let M be a finitely generated R-module and let J = J(R).

Let N be a submodule of M such that M = N + JM . Then M = N .

Proof. Apply Corollary 37 to the module M/N . 

These results is particularly useful for local rings.

Definition 39 A ring R is a local ring if R has a unique maximal ideal.

It is clear that if R is a local ring with maximal ideal I then I = J(R) is

the Jacobson radical of R. We have that the elements of R\I are the units
of R. The last corollary then implies that in order to generate a Noetherian
module M over a local ring R is is sufficient to generate the quotient M/IM .
In turn M/IM is a vector space over the field R/I and the problem of
generating M reduces to linear algebra in M/IM .

14
6 Localization
Now we describe a technique which often helps to simplify arguments and
reduce them to the case of local rings. Let R be a domain, that is a ring
without zero divisors. Let Y be a multiplicativley closed subset of R which
contains 1 and such that 0 6∈ Y . Let E be the field of fractions of R.

Definition 40 We define

S := Y −1 R := {ry −1 | r ∈ R, y ∈ Y } ⊆ E.

For an ideal I of R we define e(I) := SI = {xy −1 | x ∈ I, y ∈ Y }.

It is easy to check that S = Y −1 R is a ring and that e(I) is an ideal of S.

For example when R = Z and Y = {2k | k = 0, 1, 2 . . .} then Y −1 R is the
ring of rational numbers with denominators which are a power of 2. Now if
I = 3Z then e(I) = 3S = { 3n 2k
| n ∈ Z, k = 0, 1, 2, . . .}.
For an ideal J of S we define c(J) := R ∩ J, this is an ideal of R, the
contraction of the ideal J.
Let R and S denote the set of ideals of R and S respectively. We can
regard e : R → S and c : S → R as maps between R and S. Let Rc denote
the set {J ∩ R |J ∈ S}, the image of the contraction map c.

Proposition 41
(1) The maps c and e are mutually inverse bijections between S and
Rc . Both c and e respect inclusion and intersection of ideals. In addition e
respects sums of ideals.
(2) The prime ideals in Rc are precisely the prime ideals P of R such that
P ∩ Y = ∅.
(3) e maps prime ideals from Rc to prime ideals of S, c maps prime ideals
of S to prime ideals of R.

Proof. Part (1) is an easy exercise. For part (2), suppose P = c(J) is a
contracted prime ideal of R. If y ∈ P ∩ Y then y ∈ J but y −1 ∈ S and so
1 ∈ J, giving J = S and P = R ∩ S = R contradiction. So P ∩ Y = ∅.
Conversely if P is a prime ideal of R such that P ∩ Y = ∅ then let J = e(P )
and consider c(J) = P ∩ J. Clearly P ⊆ c(J). Suppose x ∈ c(J), thus x ∈ R

15
and x = py −1 for some p ∈ P and y ∈ Y . Hence p = xy with y 6∈ P , hence
x ∈ P because P is prime. Therefore P = c(J) = ce(P ) proving (2).
For part (3): If J is a prime ideal of S then c(J) = J ∩ R is a prime ideal
of R.
Now suppose P is a prime ideal of R with P ∩ Y = ∅. We want to
show that e(P ) = SP = Y −1 P is a prime ideal of S. Suppose r1 , r2 ∈ R,
y1 , y2 ∈ Y with (r1 y1−1 )(r2 y2−1 ) ∈ e(P ). Hence r1 r2 (y1 y2 )−1 = py −1 for some
p ∈ P, y ∈ Y . This gives y1 y2 p = yr1 r2 ∈ P and then either r1 ∈ P or r2 ∈ P
since P is prime and y 6∈ P . Hence either r1 /y1 ∈ e(P ) or r2 /e2 ∈ e(P ).
Therefore e(P ) is a prime ideal. 

Corollary 42 Suppose Y = R\P for some prime ideal P of R. Let S :=

Y −1 R. Then S has precisely one maximal ideal, namely e(P ) = SP . The
prime ideals of S correspond bijectively via c to the prime ideals of R con-
tained in P .

Proof. Let M be a maximal ideal of S. Now M = ec(M ) and c(M ) = R∩M

is a prime ideal of R disjoint from Y , hence c(M ) ⊆ P . Thus M = ec(M ) ⊆
e(P ) and by maximality M = e(P ). So e(P ) is the unique maximal ideal of
S. The rest of the claims follow from Proposition 41 (2) and (3).

Corollary 43 If R is Noetherian then S = Y −1 R is also Noetherian.

Proof. A strictly ascending chain of ideals in S contracts to a strictly

ascending chain of ideals in Rc . 

Definition 44 When P is a prime ideal of R and Y = R\P we write RP

for Y −1 R and call this the localization of R at P . By Corollary 42 RP is a
local ring whose prime ideals correspond bijectively to the prime ideals of R
contained in P .

For example when R = Z and P = 2Z then Z2Z is the ring of rational

numbers with odd denominators which has a unique maximal ideal 2Z2Z .

Proposition 45 Let I and J be ideals in a domain R. Suppose that IRM ⊆

JRM for each maximal ideal M of M . Then I ⊆ J.

16
 Proof. Suppose for the sake of contradiction that there is some a ∈ I\J
and let L := {x ∈ R | xa ⊆ J}. Then L is a proper ideal of R since 1 6∈ L and
so there is some maximal ideal M of R with L ⊆ M . Now a ∈ IRM ⊆ JRM
and so a = xy −1 with x ∈ J and y 6∈ M . But then ay = x ∈ J and so
y ∈ L ⊆ M , contradiction. Hence I ⊆ J. 
The above proposition is useful when we want to prove equality of two
ideals I and J of a ring R: it is sufficient to show IRM = JRM for each
maximal ideal M and the problem reduces to working in the local ring RM
which is usually much easier to understand.

7 Integrality
Let R ⊆ S be two rings.
Definition 46 An element x ∈ S is said to be integral over R if x is the
root of a monic polynomial with coefficients in R, that is
xn + a1 xn−1 + · · · an−1 x + an = 0 (1)
for some ai ∈ R.
The ring S is said to be integral over R, if every element of S is integral
over R. We also say that R ⊆ S is an integral extension.

Proposition 47 Let R ⊆ S be an integral extension and suppose that S is

a domain. Let I be a non-zero ideal of S. Then I ∩ R 6= {0}.

Proof. Let x ∈ I\{0} and let x satisfy (1) with n minimal possible. We
can write this as xh(x) = −an where h(x) = xn−1 + · · · + an−1 . Then an 6= 0
because S is a domain and both x and h(x) are not zero. Since x ∈ I we
have an ∈ I ∩ R. 

Proposition 48 Let x ∈ S. Then x is integral over R if and only if there

is a finitely generated R-module M ⊆ S such that 1 ∈ M and xM ⊆ M .

Proof. Suppose x is integral over R and satisfies (1). We can take M =

P n−1 j
j=0 x R.
Conversely if M is a finitely generated module with xM ⊆ M by Theorem
35 there is a monic polynomial f (t) ∈ R[t] such that f (x)M = {0}. Since
1 ∈ M we see that f (x) = 0 and x is integral over R. .

17
Definition 49 The integral closure of R in S is the set of all elements of S
which are integral over R.

Corollary 50 Let C be the integral closure of R in S. Then C is a subring

of S.

Proof. Let x, y ∈ C and let n and m be the degreesPof thePmonic polynomials

n−1 m−1 i j
with roots x and y respectively. We set M := i=0 j=0 x y R. Then
1 ∈ M , xM ⊆ M , yM ⊆ M and so (x + y)M ⊆ M and xyM ⊆ M .
Proposition 48 now gives that x + y and xy ∈ C. 

Proposition 51 Let R ⊆ S ⊆ T be three rings such that S is integral over

R and T is integral over S. Then T is integral over R.

Proof. Let x ∈ T and let ai ∈ S such that xn + a1 xn−1 + · · · + an = 0.

Let S ′ := R[a1 , . . . , an ] ⊆ S. Since each ai is integral over R the argument
of Proposition 48 gives that S ′ is a finitely
P generated R-module. Let B be a
′ ′
finite set of generators of S , so S = b∈B Ra.
Now consider
n−1
X n−1 X
X
M := S ′ [x] = S ′ xi = Rbxi .
i=0 i=0 b∈B

n−1 i
We have 1 ∈ M , xM ⊆ M and M is generated by the finite set ∪i=0 xB
as an R-module. So by Proposition 48 x is integral over R. Therefore T is
integral over R. 
When R ⊆ S is an integral extension there is a close relationship between
the prime ideals of S and the prime ideals of R.

Proposition 52 Let R ⊆ S be an integral extension.

(a) If S is a field then R is a field.
(b) If R is a field and S is a domain then S is a field.
(c) Let P be a prime ideal of S and let Q := R ∩ P . Then P is a maximal
ideal of S if and only if Q is a maximal ideal of R.

Proof. (a) Let x ∈ R\{0} and let x−1 ∈ S satisfy the equation

x−n + a1 x−n+1 + · · · + an = 0

18
with ai ∈ R. This gives x−1 = −(a1 + a2 x + · · · an xn−1 ) and so x−1 ∈ R.
(b) Let x ∈ S\{0} and let
xn + a1 xn−1 + · · · + an = 0
with ai ∈ R and n minimal possible. Then an 6= 0 and we can rewrite the
above equation as xy = −an where y = xn−1 + a1 xn−2 + · · · + an−1 ∈ S. Since
R is a field the element an is invertible in R and thus −ya−1
n is an inverse for
x in S. So S is a field.
(c) We have R/Q = R/(P ∩ R) ≃ (R + P )/P ⊆ S/P . Since S is integral
over R by reducing the equation (1) modulo P we deduce that S/P is integral
extension of R/Q. Note that S/P is a domain since P is a prime ideal of S.
Now by parts (a) and (b) S/P is a field if and only if R/Q is a field. 

Proposition 53 Let R ⊆ S be an integral extension. Let Q be a prime ideal

of R.
(a) There exists a prime ideal P of S such that P ∩ R = Q.
(b) Suppose P1 ⊆ P2 are two prime ideals of S such that P1 ∩ R = P2 ∩ R.
Then P1 = P2 .

Proof. (a) Let Y = R\Q and note that Y is multiplicatively closed subset of
R. Choose an ideal P of S maximal subject to the condition P ∩ Y = ∅, such
an ideal P exists by Zorn’s Lemma. Then P is a prime ideal of S by Problem
sheet 1. From the choice of P we have R∩P ⊆ Q. Suppose there exists x ∈ Q
with x 6∈ P . Then P + Sx is an ideal strictly bigger than P and therefore
there exists z ∈ (P + Sx) ∩ Y . We can write z = p + sx where p ∈ P, s ∈ S.
The element s is integral over R and therefore sn + a1 sn−1 + · · · + an = 0 for
some ai ∈ R. This gives
(xs)n + a1 x(xs)n−1 + · · · + an xn = 0
We have xs ≡ z mod P and and therefore
z n + a1 xz n−1 + · · · + an xn ∈ P ∩ R ⊆ Q.
Since x ∈ Q this implies z n ∈ Q but z 6∈ Q and Q is a prime ideal of R,
contradiction. Therefore P ∩ R = Q.
(b) Let Q := P1 ∩ R = P2 ∩ R and consider the integral extension R/Q ⊆
S/P1 . The ring S/P1 is a domain with ideal P2 /P1 such that (P2 /P1 ) ∩
(R/Q) = Q/Q = {0}R/Q . By Proposition 47 we must have that P2 /P1 is the
zero ideal, hence P1 = P2 . 

19
Theorem 54 Let R ⊆ S be an integral extension and let Q1 < Q2 < · · · <
Qk be a chain of prime ideals of R. There exists a chain P1 < P2 < · · · < Pk
of prime ideals of S such that Pi ∩ R = Qi for i = 1, . . . , k.

Proof. We use induction on k, the case of k = 1 being Proposition 53. For

the inductive step it is sufficient to prove the following:
Given prime ideals Q1 ⊆ Q2 of R and a prime ideal P1 of S with P1 ∩ R =
Q1 then there exists a prime ideal P2 ⊇ P1 such that P2 ∩ R = Q2 .
Let R̄ = R/Q1 , S̄ = S/P1 . Now Q̄2 := Q2 /Q1 is a prime ideal of R̄ and
S̄ is integral over R̄. By Proposition 53 there is a prime ideal P̄2 of S̄ such
that P̄2 ∩ R̄ = Q̄2 .
There is a prime ideal P2 of S with P2 ⊇ P1 such that P̄2 = P2 /P1 and
we claim that P2 ∩ R = Q2 . From the choice of P̄2 we have (P2 ∩ R) + P1 =
P2 ∩ (R + P1 ) = Q2 + P1 . Taking intersection with R we obtain

P2 ∩ R = ((P2 ∩ R) + P1 ) ∩ R = (Q2 + P1 ) ∩ R = Q2 .

This completes the induction step. 

Theorem 54 and Proposition 53 (b) together give the following.

Corollary 55 Let R ⊆ S be an integral extension. A strictly increasing

chain of prime ideals of S intersects R in a strictly increasing chain of prime
ideals of R. Conversely any strictly increasing chain of prime ideals of R is
the intersection of R with some strictly increasing chain of prime ideals of
S.

8 Krull dimension
Let F be an algebraically closed field. We want to define a notion of di-
mension to every algebraic set, which generalizes the dimension of the vector
space F k .

Definition 56 Let V ⊆ F k be an irreducible algebraic set. The dimension

dim V of V is the largest integer n such that there is a strictly increasing
chain
∅ 6= Vn ⊂ Vn−1 ⊂ · · · ⊂ V0 = V (2)

20
of irreducible algebraic sets Vi .
More generally when V is reducible we set dim V to be the largest dimen-
sion of an irreducible component of V .

For example if V = {a} is a single point in F k then dim V = 0. We

will prove later that that dim V is always finite and in fact dim V ≤ k with
equality if and only if V = F k .
Let Pi = I(Vi ) where Vi are the irreducible sets of (2). Then P0 ⊂ P1 ⊂
· · · ⊂ Pn is a strictly increasing chain of prime ideals of the polynomial ring
R = F [t1 , . . . , tk ]. This leads to the following definition.

Definition 57 Let R be a ring. The Krull dimension of R denoted by dim R

is the largest n such that there is a chain
P0 ⊂ P1 ⊂ · · · ⊂ Pn (3)
of prime ideals Pi of R. We set dim R = ∞ if there is no such integer n.

So we see that for an algebraic set V ⊆ F k we have dim V = dim R/I(V )

where R = F [t1 , . . . , tk ].
Corollary 55 now implies the following.

Proposition 58 Let R ⊆ S be an integral extension. Then dim R = dim S.

A word of warning: The dimension of a Noetherian ring does not have to

be finite (an example is sketched in the 2015 Exam paper C2.3, Q3).

Definition 59 Let P be a prime ideal of a ring R. The height, ht(P ) of P

is defined to be the largest integer n such that there is chain
P0 ⊂ · · · ⊂ Pn = P
of prime ideals Pi terminating at P .

So dim R is the maximum of the heights of its prime ideals. It turns out
that ht(P ) < ∞ for every prime ideal P of a Noetherian ring R but we won’t
prove this here.
Our next aim is to prove that dim F [t1 , . . . tk ] = k. We will prove a
more general result about the dimension of F - algebras. First we need more
definitions.

21
Definition 60 Let F ⊆ E be a field extension. Elements x1 , . . . xk ∈ E are
said to be algebraically dependent over F if there is a non-zero polynomial
f ∈ F [t1 , . . . , tk ] such that f (x1 , . . . , xk ) = 0.
We say that x1 , . . . , xk are algebraically independent (also said to be tran-
scendental) over F if they are not algebraically dependent.

Definition 61 With F ⊆ E as above the set X := {x1 , . . . , xn } is a tran-

scendence basis for E over F if X is a maximal algebraically independent
subset of E.

The notion of transcendence basis is defined even for infinite sets but we
won’t need this here.
It is clear that if E = F (c1 , . . . , cm ) is finitely generated as a field over F
then there is a finite subset X ⊆ {c1 , . . . , cm } which is a transcendence basis
for E/F . What needs proving is the analogue of fundamental property of
bases of a vector space:

Proposition 62 Any two transcendence bases for E over F have the same
size.

Proof. Let X = {x1 , . . . , xn } and Y = {y1 , . . . , ym } be two transcen-

dence bases for E over F . Suppose that m > n.
By the maximality of X we have that E is algebraic over L := F (x1 , . . . , xn )
and therefore there is a non-zero polynomial f ∈ F [t1 , . . . , tn+1 ], such that
f (x1 , . . . , xn , y1 ) = 0. We can further assume that f has as small degree as
possible. Now f 6∈ F [tn+1 ], because y1 is transcendental over F . Hence there
is some tj with j ≤ n, say t1 , which appears in a nonzero monomial of f .
Rewriting f as a polynomial in t1 with coefficients in F [t2 , . . . , tn+1 ] gives
that x1 is algebraic over the subfield L1 := F (y1 , x2 , . . . , xn ). Hence L is al-
gebraic over L1 and E is algebraic over L, therefore E is algebraic extension
of L1 .
Now y2 is algebraic over L1 and in the same way we deduce that there
is some xj , j > 1, for example x2 such that x2 is algebraic over L2 :=
F (y1 , y2 , x3 , . . . , xn ) and as a consequence E is algebraic over L2 . We can
continue in the same way replacing each successive xi with yi until we reach
the situation where E is algebraic over the subfield Ln := F (y1 , . . . , yn ). But

22
then yn+1 is algebraic over Ln which is contradiction to Y being transcen-
dental over F . So n ≥ m and by exchanging the roles of X and Y we get
m ≥ n and therefore m = n. 

Definition 63 Let F ≤ E be a field extension. The transcendence degree

tr.degF E of E over F is the cardinality of a transcendence basis for E over
F.
More generally for a domain R which is a finitely generated algebra over
a field F we set tr.degF R = tr.degF E, where E is the field of fractions of R.

The following result, known as Noether’s Normalization Lemma is very

useful in simplifying many proofs by reducing them to polynomial rings.

Lemma 64 (Noether’s Normailization Lemma) Let R = F [y1 , . . . , yn ]

be a finitely generated as an algebra over a subfield F . Assume that R is a
domain. There exists an algebraically independent set {x1 , . . . , xk } ⊂ R over
F such that R is integral over F [x1 , . . . , xk ].

Proof. We will prove the lemma in the case when the field F is infinite.
We argue by induction on n, the case n = 0 being trivially true. Suppose
the lemma is true for R[y1 , . . . , yn−1 ] and we can find x1 , . . . , xs , (s ≤ n − 1),
which are algebraically independent over F and such that F [y1 , . . . , yn−1 ] is
integral over F [x1 , . . . , xs ].
Suppose first that x1 , . . . , xs , yn are still algebraically independent. Take
k = s + 1, xs+1 = yn . Now y1 , . . . , yn are integral over F [x1 , . . . , xk ] and we
are done.
So we may assume that f (x1 , . . . , xs , yn ) = 0 for some nonzero polynomial
f ∈ F [t1 , . . . , ts+1 ] in s + 1. Let g be the sum of all monomials of highest de-
gree m in f . Since F is infinite there exist ci ∈ F \{0} with g(c1 , . . . , cs+1 ) 6=
0. Since g is homogeneous we can consider g(c1 , . . . cs+1 )/cm s+1 and by replac-
ing each ci by ci /cs+1 we may assume cs+1 = 1. Let b := g(c1 , . . . , cs , 1) ∈
F \{0}.
Let zi := xi − ci yn . We have

0 = f (x1 , . . . , xs , yn ) = f (z1 +c1 yn , . . . , zs +cs yn , yn ) = bynm +h(z1 , . . . , zs , yn ),

where h is a polynomial whose degree in yn is at most m−1. Dividing by b 6= 0

we conclude that yn is integral over R′ := F [z1 , . . . , zs ]. Since xi ∈ R′ [yn ] it

23
follows that F [x1 , . . . , xs ] is integral over R′ and hence R is integral over R′ .
The ring R′ is generated as an F -algebra by s < n elements and so by the
induction hypothesis there exist elements x′1 , . . . x′k which are algebraically
independent set over F and R′ is integral over F [x′1 , . . . , x′k ]. In turn R is
integral over R′ and by Proposition 51 R is integral over F [x′1 , . . . , x′k ]. 

Proposition 65 Let R be a domain which is finitely generated as an algebra

over a field F . Let P 6= {0} be a prime ideal of R. Then tr.degF R >
tr.degF R/P .

Proof. Let k = tr.degF R̄ and let {x̄1 , . . . , x̄k } be a transcendence basis of

R̄ over F . Choose elements xi ∈ R such that x̄i = xi + P and note that
{x1 , . . . , xk } are algebraically independent over F . Hence tr.degF R ≥ k.
Suppose for the sake of contradiction that tr.degF R = k. This implies that
R is algebraic over its subring L := F [x1 , . . . , xk ]. Suppose R = L[y1 , . . . , yn ]
for some elements yi ∈ R. Note that since x̄1 , . . . , x̄k are algebraically in-
dependent we have L ∩ P = {0}. Let Y = L\{0} this is a multiplicatively
closed subset of R such that Y ∩ P = ∅. Consider the localization S := Y −1 R
and let T := e(P ) = SP . By Proposition 41 we have that T 6= {0} is a prime
ideal of S. Let E := Y −1 L = F (x1 , . . . , xk ) be the field of fractions of L.
Then S = Y −1 R = E[y1 , . . . , yn ]. Each of the elements yi is algebraic over
E and so S is a finite field extension of E. This contradicts the fact that
T = e(P ) is a nonzero prime ideal of S.
Therefore tr.degF R̄ < tr.degF R as claimed. 

Theorem 66 Let R be a domain which is finitely generated as an algebra

over its subfield F . Then dim R = tr.degF R.

Proof. By Theorem 64 there exists an algebraically independent set

{x1 , . . . xk } ⊂ R such that R is integral over its subring K := R[x1 , . . . , xk ].
We have dim R = dim K and since the field of fractions of R is algebraic over
R(x1 , . . . , xk ) we have k = tr.degF R. Now consider the chain of ideals of K

{0} = P0 ⊂ P1 ⊂ · · · ⊂ Pk ,

where Pi = hx1 , . . . , xi i. Since K is a polynomial ring over xi , each Pi is a

prime ideal of K and so dim R = dim K ≥ k.

24
 Suppose for the sake of contradiction that dim R > k and let {0} = P0 ⊂
P1 ⊂ · · · Pk+1 be a chain of k + 2 non-zero ideals of R. Let Ri := R/Pi , this
is a domain which is a finitely generated algebra over F and by Proposition
65 we have tr.degF R > tr.degF R1 > · · · > tr.degRk+1 ≥ 0. So tr.degR > k,
contradiction. Hence dim R = k = tr.degF R. 

Corollary 67 Let F be a field, and R = F [t1 , . . . , tk ] be a polynomial ring.

Then dim R = k.

Corollary 68 Let F be an algebraically closed field and let V ⊆ F k be an

algebraic set. Then dim V ≤ k with equality if and only if V = F k .

Proof. We have I(F k ) = {0} and so dim F k = dim F [t1 , . . . tk ] = k.

Now suppose V ⊂ F k is a proper algebraic set of dimension l. We may
replace V with an irreducible component and so without loss of generality
may assume that V is irreducible. Let P = I(V ). Since V 6= F k the prime
ideal P is not zero. But then

l = dim V = dim F [t1 , . . . , tk ]/P < k

by Proposition 65. 

9 Noetherian rings of small dimension. Dedekind

domains
We can apply the theory developed so far to study the Noetherian rings of
dimension 0 and 1.

Theorem 69 Let R be a Noetherian ring of dimension 0. Then R/nilrad(R)

is isomorphic to a finite direct product of fields.

Proof. By Proposition 14 R has finitely many minimal prime ideals,

say P1 , . . . , Pn and nilradR = ∩ni=1 Pi is nilpotent. Let Qj := ∩i6=i Pi . Since
dim R = 0 each Pj is a maximal ideal of R and so Q Pi 6⊆ Pi for any i 6= j.
Given j for each i 6= j choose ai ∈ Pi \Pj and then i6=j ai belongs to Qj but

25
not to Pj . So Qj 6⊆ Pj and hence Qj + Pj = R. This holds for any j and by
the Chinese remainder theorem
n
R R YR
= ≃ .
nilradR ∩i Pi P
i=1 i

Now each R/Pi is a field by the maximality of Pi . 

Conversely, a ring R such that nilradR is a nilpotent finitely generated
ideal and R/nilradR is a direct product of fields, is a Noetherian ring of
dimension 0. We leave the proof as an exercise.
We now move to Noetherian rings of dimension 1.

Definition 70 Let R be a domain. We say that R is integrally closed if R

is integerally closed in its field of fractions E, that is any x ∈ E which is
integral over R must satisfy x ∈ R.

For example Z and more generally any PID is an integrally closed domain,
see the proof of Proposition 72 below.

Definition 71 A Noetherian domain R is said to be a Dedekind domain if

dim R = 1 and R is integrally closed.

Examples of Dedekind domains are all principal ideal domains.

Proposition 72 Let R be a PID which is not a field. Then R is a Dedekind

domain.

Proof. Any nonzero prime ideal P of R is maximal, thus a maximal chain

of prime ideals is provided by {0} ⊂ P . Hence dim R = 1. It remains to
show that R is integrally closed. Let K be the field of fractions of R and
let x = yz −1 ∈ K be integral over R where y, z ∈ R and z 6= 0. We will
prove that x ∈ R. We may assume that y and z are coprime elements of R.
Suppose x satisfies the equaition (1) with ai ∈ R. Multiply by z n to clear
denominators and reach y n + a1 y n−1 z + · · · + an z n = 0. This gives that z
divides y n and since y and z are assumed coprime it follows that z is a unit
of R. Thus x = yz −1 ∈ R and R is integrally closed. 

26
 A rich source of Dedekind domains is provided by Algebraic Number
Theory.
Let E/Q be a finite field extension of Q and let R be the integral closure of
Z in E. Then R is a domain and since R is integral over Z we have dim R =
dim Z = 1. Moreover it can be proved that (R, +) is a finitely generated
abelian group, thus R is a Noetherian Z-module, hence a Noetherian R-
module and hence R is a Noetherian ring. An important characterisation of
Dedekind domains is that their ideals have unique factorization property.

Theorem 73 Let R be a Dedekind domain. Then any nonzero ideal I is a

product of prime ideals. This factorization is unique up to reordering of the
prime ideals.

Proof. Let I 6= {0} be an ideal of R. Let P1 , . . . , Pn be the minimal

primes of I. Choose some Pi and consider the localization RPi . Then RPi
is a local Noetherian domain of dimension 1. Since R is integrally closed so
is RPi by Problem sheet 4. Now we can apply the last problem in Sheet 4
which gives that RPi is a PID in which every nonzero ideal is a power of its
maximal ideal e(Pi ) = Pi RP1 . Hence there is an integer ni ∈ N such that
e(I) = IRPi = e(Pi )ni .
Let J = P1n1 · · · Pknk . Now observe that for j 6= i we have Pj RPi = RPi and
so JRPi = (Pi )ni RPi = IRPi . On the other hand if Q is a non-zero prime ideal
different from any of the Pi then I 6⊆ Q and so IRQ = RQ = JRQ . Therefore
IRM = JRM for every maximal prime ideal M of R. By Proposition 45 we
have I = J is a product of prime ideals. The same arguments shows that the
integers ni and the prime ideals Pi are uniquely determined by I. 
There is a converse to Theorem 73: A domain all of whose ideals are
product of prime ideals is necessarily a Dedekind domain. We won’t prove
this here, instead we shall prove some other results.
Let I and J be two ideals of R. We say that I divides J if J = IT for
some ideal T of R.

Proposition 74 Let R be a Dedekind domain and I and J two ideals of R.

Then I divides J if and only if J ⊆ I.

Proof. If I Q QmJ ⊆siI. Conversely suppose J ⊆ I. We

divides J then clearly
m ni
can write J = i=1 Pi and I = i=1 Pi for some integers ni , si ≥ 0 and
prime ideals Pi .

27
 Then JRPi = Pini RPi ⊆QPisi RPi = IRPi . Therefore ni ≥ si for each i. Let
ui = ni − si and put U := m ui
i=1 Pi . We have U I = J and so I divides J. 

Proposition 75 Let R be a Dedekind domain. Then every ideal of R can

be generated by at most 2 elements.

Proof. Let a ∈ I\{0} and let J = Ra. We can factorize J = Q m si

Q
i=1 Pi
m ni
for some prime ideals Pi and si ∈ N. Since J ⊆ I we must have I = i=1 Pi
for some integers 0 ≤ ni ≤ si .
We have I/J ≃ m ni si
Q
i=1 Pi /Pi by the Chinese Remainder theorem. Let
us choose bi ∈ Pi \Pi2 . This gives Rbni i + Pisi = Pini and so each Pini /Pisi is
a principal ideal in R/Pisi . Hence I/J is a principal ideal generated say by
b + J in the ring R/J. Then I = Rb + J = ha, bi. 

9.1 Fractional ideals and the ideal class group

We know that a PID is a Dedekind domain, but not every Dedekind domain
R is a PID. How can we measure the failure of R to be a PID?

Definition 76 Let R be a Dedekind domain with field of fractions K. A

fractional ideal of K is a subset of the form αI where α ∈ K\{0} and I is a
nonzero ideal of R. Denote by F the set of all fractional ideals of K.

It is clear that if αI and βJ are fractional ideals of K then so is their

product αβIJ. The fractional ideal R plays the role of identity since αI ·R =
αI for each αI ∈ F. We now show that every fractional ideal has an inverse,
thus making F into abelian group.

Theorem 77 Let L ∈ F be a fractional ideal of K. Then L has an inverse

L−1 , namely a fractional ideal L−1 ∈ F such that LL−1 = R.

Proof. Suppose L = αI for an ideal I of R and nonzero α ∈ K.

Choose any nonzero element x ∈ I. Since Rx ⊆ I by Proposition 74 we
must have Rx = IJ for some ideal J of R. Define L−1 := α−1 x−1 J. Then
L−1 L = α−1 αx−1 IJ = x−1 IJ = x−1 xR = R. 
This shows that F is an abelian group under multiplication. We have the
subgroup of principal ideals P := {αR| α ∈ K\{0}} and so we can define

28
Definition 78 The ideal class group of a Dedekind domain is the quotient
C := F/P of fractional ideals modulo principal ideals.

Thus R is a PID if and only if C = {0}. One of the major results in Algebraic
number theory is that |C| is finite when R is a ring of integers. The proof
relies on geometric arguments specific to rings of integers, in particular their
realization as a lattice in Euclidean space and lies outside the scope of this
course.

29