You are on page 1of 43

School of Mathematics

Algebra 2

Teaching Block 2, 2023/24


Contents
1 The basics 1

2 Maps, ideals and quotient rings 8

3 Integral domains 21

4 Special rings 28

5 Gauss’ Lemma and applications 40

6 Irreducible polynomials 47

7 Fields 53

8 Finite fields 65

9 Ruler and compass constructions 73


What is algebra?
Algebra is one of the cornerstones of modern mathematics, with deep connections
to all areas of our subject. The word algebra comes from the Arabic word al-jabr,
which first appeared in the work of the highly influential Persian mathematician
and astronomer al-Khwarizmi in the 9th century (the word roughly translates as
the “reunion of broken parts”). Through Latin translations, this eventually became
algebra and its usage in English can be traced back to the 15th century.
Abstract structures arise naturally throughout mathematics, in many different
forms. For example, if X is any “mathematical object” (e.g. a set, a polygon in the
plane, a vector space, a graph, etc.) then we can often study the object by consid-
ering its symmetries, which we can think of more formally as bijective maps X ! X
preserving the core structure of X (e.g. a linear structure if X is a vector space, or the
adjacency relation if X is the set of vertices of a graph). By composing maps, we ob-
tain an operation on the set of these symmetries, which satisfies various properties
that do not depend on the choice of X.
In algebra, we are interested in studying these structures (sets equipped with
operations) in an abstract setting. The objects and operations may look rather dif-
ferent initially, but they share many common properties. We will focus on some of
the main algebraic structures, defined in terms of axioms, which arise in this way.
Here abstract algebra provides a powerful and uniform language in which to study
these structures, allowing us to develop a general theory which can then be applied
to any given example.
The abstract structures that arise in this way come in many different flavours:
groups, rings, fields, vector spaces, algebras, semigroups, monoids, magmas, etc. ...
The first three listed here are the most fundamental:

• Groups, which you met last year, have a single binary operation and examples
include Cn (cyclic group of order n) and Sn (symmetric group of degree n).

• We also have rings, which have two binary operations (denoted addition and
multiplication). Examples include Z and R[x].

• Fields are a special type of ring, where division (by nonzero elements) is also
available. Examples include Q, R and C.

In this course, we will introduce and develop the general theory of rings and
fields. Concrete examples will be used repeatedly to illustrate the main definitions,
ideas and results. An interesting application to ruler and compass constructions
will be presented in the final chapter.
1 The basics
We begin by introducing the concept of a ring. First we present the abstract defi-
nition and then we highlight some of their basic properties. Perhaps most impor-
tantly of all, we also introduce several concrete examples, which we will work with
throughout the course.

1.1 Rings

Definition 1.1. A ring is a non-empty set R equipped with two binary operations
R⇥R ! R

(a, b) 7! a + b (addition)
(a, b) 7! a ⇥ b (multiplication)

that satisfy all the following properties (for all a, b, c 2 R):

(1) Commutativity: a + b = b + a and a ⇥ b = b ⇥ a

(2) Associativity: (a + b) + c = a + (b + c) and (a ⇥ b) ⇥ c = a ⇥ (b ⇥ c)

(3) Distributivity: a ⇥ (b + c) = (a ⇥ b) + (a ⇥ c)

(4) Zero element: There exists 0 2 R such that a + 0 = a for all a 2 R

(5) Unity element: There exists 1 2 R such that a ⇥ 1 = a for all a 2 R

(6) Additive inverses: For all a 2 R, there exists a 2 R such that a + ( a) = 0.

Notation. If R is a ring and a, b 2 R, then we usually write ab for a ⇥ b.

Remark 1.2. Let (R, +, ⇥) be a ring.

(a) (R, +) is an abelian group.

(b) The elements 0 and 1 in parts (4) and (5) are unique (see Lemma 1.4).

(c) Warning: Some books will give a more general definition of a ring, where
multiplication is not necessarily commutative, and there may not be a unity
element. In this setting, a ring as defined above is typically called a commuta-
tive ring with unity.

Rings are very natural structures that arise throughout mathematics. Let’s look
at some concrete examples.

1
Example 1.3.

(1) The integers Z = {. . . , 2, 1, 0, 1, 2, . . .} is the most natural and fundamental


example of a ring, with the usual addition and multiplication operations (so 0
and 1 are the zero and unity elements, respectively).

(2) Similarly, the sets of rational, real and complex numbers, denoted Q, R and C,
are all examples of rings, with the usual operations. (Throughout the course,
‘with the usual operations’ is always implied when working with these rings.)

(3) The set


R[x] = {an xn + an 1 xn 1
+ · · · + a1 x + a0 : ai 2 R, n > 0}

of polynomials in the indeterminate x with real coefficients is a ring with re-


spect to the usual operations of addition and multiplication of polynomials.

(4) Let R be the set of all functions from R to R and define addition and multipli-
cation as follows (for all f , g 2 R, a 2 R):

( f + g)(a) = f (a) + g(a), ( f g)(a) = f (a)g(a).

Then R is a ring. Similarly, the set of continuous functions, or differentiable


functions (from R to R), are also rings.

(5) There is also the ring {0} with one element, where 0 + 0 = 0 ⇥ 0 = 0. This
trivially satisfies all the ring axioms (with 0 = 0 and 1 = 0). This is called the
zero ring.

Next we use the ring axioms to establish some basic properties.

Lemma 1.4. Let R be a ring with zero element 0 and unity element 1. Then the following
properties are satisfied:

(i) If a + b = a for some a, b 2 R, then b = 0. In particular, the zero element is unique.

(ii) If b 2 R and ab = a for all a 2 R, then b = 1. In particular, R has a unique unity


element.

(iii) If a, b, c 2 R and a + b = a + c, then b = c. Therefore, additive inverses are unique.

(iv) We have 0a = 0 for all a 2 R.

(v) We have ( 1)a = a for all a 2 R, where 1 is the additive inverse of 1.

(vi) We have 0 = 1 if and only if R = {0} is the zero ring.

2
Proof. We leave most of this as an exercise, but let’s give details for (iii). For a, b, c 2 R
we have

a+b = a+c
=) ( a) + (a + b) = ( a) + (a + c)
=) ( a + a) + b = ( a + a) + c
=) 0+b = 0+c
=) b=c

In particular, if a + b = 0 and a + c = 0 then b = c and thus a is unique.

Notation. For a, b 2 R and n 2 N = {1, 2, 3, . . .}, we will adopt the following conve-
nient notation:

• We write a b for a + ( b).

• We write an for a ⇥ · · · ⇥ a (n times); we also write a0 = 1.

• We write na for a + · · · + a (n terms), and na for ( a) + · · · + ( a), etc. (note


that na = n( a)).

Let R be a ring and fix an element a 2 R. By definition, a has an additive inverse,


but it may not have a multiplicative inverse (that is, there may not be an element
b 2 R such that ab = 1). This leads us nicely to our next definition.

Definition 1.5. Let R be a ring and let a 2 R. If there exists an element b 2 R such
that ab = 1, then a is called a unit. We write R⇥ for the set of units in R.

Exercise. As an exercise, show that if a is a unit then there is a unique element b 2 R


such that ab = 1. This element is usually denoted a 1 and we refer to it as the
multiplicative inverse of a.

Exercise. As another exercise, prove that R⇥ is an abelian group under the multi-
plication operation inherited from R. Also show that the product of two non-units
in R is a non-unit (i.e. both R⇥ and R \ R⇥ are closed under multiplication).

Example 1.6. Let’s consider the set of units for the rings in Example 1.3.

R R⇥
Z {±1}
Q Q \ {0}
R R \ {0}
R[x] {the nonzero constant polynomials}
{functions R ! R} {the nowhere zero functions R ! R}

3
Consider the latter ring. An example of a nowhere zero function f : R ! R is given
by f (a) = a2 + 1, which has multiplicative inverse g(a) = 1/(a2 + 1) (note that g is
defined on all of R precisely because f is nowhere zero). On the other hand, the
function f (a) = a is not a unit because g(a) = 1/a is not defined at a = 0.

Notice that 1 and 1 are units in every ring, while 0 is a unit if and only if
R = {0}. So for R = Z, the set of units is as small as it can possibly be, whereas rings
such as Q and R are at the other end of the scale. This leads us to the following
important family of rings.

Definition 1.7. A ring R 6= {0} is called a field if every nonzero element of R is a


unit (that is, if R⇥ = R \ {0}).

Example 1.8. We immediately observe that Q, R and C are all fields, whereas Z, R[x]
and {functions R ! R} are not.

Example 1.9. Fix a positive integer n 2 N and consider the set

R = Z/nZ := {0, 1, . . . , n 1}.

Then the operations of addition and multiplication modulo n make R into a ring:

a + b = a + b (mod n)
a ⇥ b = ab (mod n)

with zero and unity elements 0 and 1, respectively. This is called the ring of integers
modulo n. For example, if n = 5 then we have 2 + 4 = 1 and 2 ⇥ 4 = 3 (since 6 ⌘ 1
(mod 5) and 8 ⌘ 3 (mod 5)).

Exercise.

(1) Show that (Z/10Z)⇥ = {1, 3, 7, 9}.

(2) More generally, show that (Z/nZ)⇥ = {a 2 {1, . . . , n 1} : hcf(a, n) = 1}, where
‘hcf’ is ‘highest common factor’.

(3) Prove that Z/nZ is a field if and only if n is a prime number.

1.2 Subrings

We now consider some natural ways to obtain new rings from existing ones.

Definition 1.10. Let R be a ring. A subset S of R is a subring of R if 0, 1 2 S and S


itself is a ring with respect to the addition and multiplication operations on R.

4
Notice that R is always a subring of itself, whereas {0} is a subring of R if and
only if R = {0}. We say that a subring S is proper if S 6= R.

Warning. If we fix n 2 N, then Z/nZ = {0, 1, . . . , n 1} is not a subring of Z because


the ring operations on Z/nZ are not the same as the operations on Z.
Lemma 1.11 (Subring Test). A subset S ✓ R is a subring if and only if 1 2 S and S is
closed under addition, multiplication and additive inverses.

Proof. This is a straightforward exercise.


Remark. Suppose R is a field and S ✓ R is a subring. If S is a field (that is, if S⇥ =
S \ {0}), then we call S a subfield.
Example 1.12. Natural examples of subrings are given by the number systems:

Z✓Q✓R✓C

(note that Z is a subring of Q, but not a subfield since Z⇥ = {±1}). We also have the
subring R ✓ R[x] of constant polynomials.
Exercise. Show that every subring of C contains Z (in particular, Z is the only sub-
ring of Z).
Example 1.13. Consider the set of Gaussian integers, which is defined

Z[i] = {a + bi : a, b 2 Z} ✓ C
p
with i = 1 2 C. Then Z[i] is a subring of C (check this!).
Exercise. Recall that a group G of order n is cyclic if there exists an element g 2 G
such that G = {g, g2 , . . . , gn } (in which case, we say that g is a generator of G).

Show that Z[i]⇥ = {1, 1, i, i} is a cyclic group of order 4.


p
(Hint. Use the fact that |uv|2 = |u|2 |v|2 for all u, v 2 C, where |a + bi| = a2 + b2 .)
Remark 1.14. Note that Z[i] is the smallest subring of C containing i (i.e. every
subring R ✓ C containing i must also contain Z[i]). Indeed, every subring R ✓ C
contains Z, and if it also contains i then we must have Z[i] ✓ R since a subring is
closed under addition and multiplication.

The latter observation leads us naturally to the following definition. Fix a com-
plex number a and let Z[a] be the smallest subring of C containing a. Then

Z[a] = the set of complex numbers that are generated from {1, a} using +, , ⇥
= {an a n + · · · + a1 a + a0 : ai 2 Z, n > 0}
p p
For example, Z[ 2] = {a + b 2 : a, b 2 Z} and Z[a] = {a + ba + ca 2 : a, b, c 2 Z} if
p
a = e2pi/3 , noting that ( 2)2 = 2 and a 3 = 1.

5
1.3 Product rings

Definition 1.15. Let R and S be rings. The Cartesian product

R ⇥ S = {(a, b) : a 2 R, b 2 S}

can be given a ring structure by defining addition and multiplication coordinate-


wise:
(a, b) + (c, d) = (a + c, b + d), (a, b) ⇥ (c, d) = (ac, bd)
This is called the product ring (of R and S). Note that the zero element is (0, 0) and
the unity element is (1, 1).

Exercise. Check that R ⇥ S with the given operations is a ring.

Lemma 1.16. We have (R ⇥ S)⇥ = R⇥ ⇥ S⇥ .

Proof. We observe that

(a, b) 2 (R ⇥ S)⇥ () there exists (c, d) 2 R ⇥ S with (a, b)(c, d) = (1, 1)


() (ac, bd) = (1, 1)
() ac = bd = 1
() a 2 R⇥ , b 2 S⇥

and the result follows.

Note. The product ring construction extends naturally to three or more rings. In-
deed, if R1 , . . . , Rn are rings then we can define the product ring R1 ⇥ · · · ⇥ Rn in the
obvious way, with addition and multiplication defined coordinatewise.

1.4 Polynomial rings

Polynomial rings form an important family of examples, which we will work with
throughout the course. Let’s start with the formal definition.

Definition 1.17. Let R be a ring. Then

R[x] = {an xn + · · · + a1 x + a0 : ai 2 R, n > 0}

is the polynomial ring over R in the indeterminate x, with addition and multiplica-
tion defined in the usual manner, e.g.

(a2 x2 + a1 x + a0 ) + (b1 x + b0 ) = a2 x2 + (a1 + b1 )x + a0 + b0


(a2 x2 + a1 x + a0 ) ⇥ (b1 x + b0 ) = a2 b1 x3 + (a2 b0 + a1 b1 )x2 + (a1 b0 + a0 b1 )x + a0 b0

6
We say that a polynomial

f = an xn + · · · + a1 x + a0 2 R[x]

with an 6= 0 has degree n, denoted ∂ f = n, and we call an and an xn the leading


coefficient and leading term of f , respectively. In addition, f is constant if f = a0 ,
and it is monic if an = 1. Similarly, f is linear if ∂ f = 1, and quadratic if ∂ f = 2, etc.
Note that the zero and unity elements in R[x] are the constant polynomials 0 and 1,
respectively.

It is important to note that we may view R as a subring of R[x] by identifying R


with the set of constant polynomials in R[x].

Example 1.18. Let R = Z/2Z = {0, 1} be the ring of integers modulo 2. Then

R[x] = {all polynomials in x with coefficients in R}


= {0, 1, x, x + 1, x2 , x2 + 1, x2 + x, x2 + x + 1, . . .}

is an infinite ring. To illustrate the operations on R, note that

(x + 1) + (x2 + 1) = x2 + x (since 1 + 1 = 0 in R)
(x + 1)2 = x2 + 2x + 1 = x2 + 1
(x + 1)4 = (x2 + 1)2 = x4 + 2x2 + 1 = x4 + 1

The polynomial ring construction can be extended naturally to two or more in-
determinates. For example, we can consider the ring R[x, y] of polynomials in x
and y with coefficients in R. This extends in the obvious way to polynomial rings
R[x1 , . . . , xn ] with n indeterminates.

Note. R[x, y] = S[y] is the ring of polynomials in y with coefficients in S = R[x].

7
2 Maps, ideals and quotient rings

2.1 Isomorphisms

We begin by defining what it means for two rings to be ‘essentially the same’.

Definition 2.1. Let R, S be rings. A map j : R ! S is a (ring) isomorphism if

(1) j is a bijection (i.e. injective and surjective); and

(2) For all a, b 2 R, we have

j(a + b) = j(a) + j(b)


j(ab) = j(a)j(b)

If an isomorphism j : R ! S exists, then we say that R and S are isomorphic rings


and we write R ⇠
= S.

Note. The existence of an isomorphism R ! S means that the elements of R and S


can be lined up in a way that is compatible with the respective addition and multi-
plication operations on R and S. It follows that R and S have identical ring-theoretic
properties; they only differ in the choice of labels for their elements. For example,
the groups (R, +) and (S, +) are isomorphic, and so are (R⇥ , ⇥) and (S⇥ , ⇥).

Exercise. Let j : R ! S be an isomorphism.

(1) Show that j(0) = 0, j(1) = 1 and j( a) = j(a) for all a 2 R.

(2) Show that the inverse map f 1 : S ! R is an isomorphism, hence R ⇠


= S if and
only if S ⇠
= R.

Example 2.2.

(1) The identity map j : R ! R, a 7! a, is an isomorphism, so R ⇠


= R.

(2) Let S = {(a, a) : a 2 R} ✓ R ⇥ R. Then S is a subring of R ⇥ R and the map


j : R ! S, a 7! (a, a), is an isomorphism.

(3) The map j : C ! C, j(a + bi) = a bi, is an isomorphism.


p p
(4) Let R = Z[ 2] = {a + b 2 : a, b 2 Z} and define
( ! )
a 2b
S= : a, b 2 Z .
b a

8
As an exercise, show that S is a ring (with respect to the usual addition and
multiplication operations on matrices), and then prove that R and S are iso-
morphic.

(5) Consider the set R = {False,True} with operations

a + b = a XOR b, ab = a AND b,

where XOR denotes the “exclusive or” logic operation. We can compute addi-
tion and multiplication by inspecting the relevant truth tables:

XOR F T AND F T
F F T F F F
T T F T F T

Then R is a ring and it is isomorphic to Z/2Z = {0, 1}, by mapping False to 0


and True to 1.

2.2 Homomorphisms

Let’s now consider more general maps between rings.

Definition 2.3. Let R, S be rings. A map j : R ! S is a (ring) homomorphism if we


have j(1) = 1 and

j(a + b) = j(a) + j(b), j(ab) = j(a)j(b)

for all a, b 2 R.

In Lemma 2.7 below we will show that the additional properties

j(0) = 0, j(a b) = j(a) j(b)

follow from the definition of a homomorphism. But note that j(1) = 1 is part of the
definition.

Note. An isomorphism is simply a bijective homomorphism.

Example 2.4. As a simple example, notice that if S is a subring of R, then the natural
inclusion map j : S ! R (sending a 2 S to a 2 R) is a homomorphism. For instance,
we have homomorphisms Z ! Q, Q ! R and R ! C. Similarly, there is a natural
homomorphism R ! R[x], where we view R as the subring of constant polynomials
in R[x].

9
Example 2.5. Fix a positive integer n. Then the map

j : Z ! Z/nZ, a 7! a (mod n)

is a surjective homomorphism.

Example 2.6. Let R be a ring. Then the map

j : R[x] ! R
f = an xn + · · · + a0 7! f (0) = a0

is a homomorphism, and so is f 7! f (c) for any c 2 R. These maps are sometimes


called evaluation homomorphisms.

Lemma 2.7. Let R, S be rings, j : R ! S a homomorphism. Then the following properties


are satisfied:

(i) j(0) = 0.

(ii) j( a) = j(a) for all a 2 R.

(iii) The image im(j) = {j(a) : a 2 R} is a subring of S.

(iv) If y : S ! T is another homomorphism, then the composition y j : R ! T is also a


homomorphism.

Proof. We leave part (i), (ii) and (iv) as an exercise. For (iii), first note that 1 2 im(j)
since j(1) = 1. Next suppose c, d 2 im(j), say c = j(a) and d = j(b). Then we
have c + d = j(a + b), cd = j(ab) and c = j( a) by (ii), so im(j) is closed under
addition, multiplication and additive inverses. Now apply Lemma 1.11.

Definition 2.8. Let R, S be rings and j : R ! S a homomorphism. Then

ker(j) = {a 2 R : j(a) = 0}

is called the kernel of j.

Warning: Recall that the kernel of a linear map of vector spaces is a subspace, and
the kernel of a homomorphism of groups is a subgroup. However, the kernel of a
ring homomorphism is not necessarily a subring.

Example 2.9. Consider the homomorphism j : Z ! Z/2Z, a 7! a (mod 2). Then

ker(j) = {a 2 Z : a ⌘ 0 (mod 2)} = 2Z

10
is the set of even integers, which is not a subring because it does not contain 1.
Similarly, if R is a ring and we take the homomorphism

j : R[x] ! R, f = an xn + · · · + a0 7! f (0) = a0

then
ker(j) = { f 2 R[x] : f (0) = 0} = {xg : g 2 R[x]}
is the set of polynomials in R[x] with zero constant term, which is not a subring of
R[x].

Lemma 2.10. A homomorphism j : R ! S is injective if and only if ker(j) = {0}.

Proof. Suppose j is injective and a 2 ker(j). Then j(a) = 0 = j(0), so a = 0 and thus
ker(j) = {0}.
Now assume ker(j) = {0} and j(a) = j(b) for some a, b 2 R. Then

0 = j(a) + ( j(b)) = j(a) j(b) = j(a b)

and thus a b 2 ker(j). But we are assuming ker(j) = {0}, so a b = 0 and we


deduce that a = b. Therefore, j is injective.

Example 2.11. Consider R = Z/4Z = {0, 1, 2, 3} and S = Z/2Z = {0, 1}. There is a
homomorphism R ! S defined by a 7! a (mod 2) (with kernel {0, 2}). Suppose there
is a homomorphism y : S ! R. Then

2 = 1 + 1 = y(1) + y(1) = y(1 + 1) = y(0) = 0

which is a contradiction (2 and 0 are not equal in R), so there are no homomorphisms
from S to R. Similarly, if n > 2 then there is a natural homomorphism Z ! Z/nZ
taking a 7! a (mod n), but there are no homomorphisms Z/nZ ! Z.

Exercise. Prove that if R is any ring, then there exists a unique homomorphism
Z ! R. Also prove that there is a unique homomorphism Z/nZ ! Z/nZ for all n 2 N.

We can formalise the observation in the previous example in the following way.
Recall that if R is a ring and r 2 R, then we write 2r and 3r2 for the ring elements
r + r and r2 r2 r2 , respectively, where r2 = r ⇥ r.

Lemma 2.12. Let R, S be rings and let j : R ! S be a homomorphism. Suppose r 2 R


satisfies an equation of the form

an rn + · · · + a1 r + a0 = 0

with ai 2 Z. Then j(r) 2 S satisfies the same equation, that is

an j(r)n + · · · + a1 j(r) + a0 = 0.

11
Proof. Just apply j to both sides of the first equation and use the basic properties of
homomorphisms discussed above.

Example 2.13. As an application, we see that there are no homomorphisms from C


to R because i 2 C satisfies the equation x2 + 1 = 0, but there is no such element in R.
Similarly, there are no homomorphisms from R to Q (check this!).

2.3 Ideals

In order to understand the kernels of homomorphisms, we introduce the concept


of an ideal.

Definition 2.14. Let R be a ring and let I ✓ R be a non-empty subset. Then I is an


ideal of R if the following properties are satisfied:

(1) I is closed under addition (i.e. if a, b 2 I, then a + b 2 I).

(2) We have ra 2 I for all r 2 R and all a 2 I.

Some immediate observations:

• I = {0} are I = R are both ideals of R; we say that I is proper if I 6= R.

• Condition (2) implies that I contains 0 and is closed under additive inverses.
It also implies that I is closed under multiplication.

• Note that R is the only ideal containing 1, so no proper ideal is a subring (more
generally, see Lemma 2.31).

Lemma 2.15. If j : R ! S is a homomorphism, then ker(j) is an ideal of R.

Proof. Set I = ker(j) and first observe that j(0) = 0 (see Lemma 2.7(i)), so 0 2 I and
I is non-empty. If a, b 2 I and r 2 R, then

j(a + b) = j(a) + j(b) = 0 + 0 = 0


j(ra) = j(r)j(a) = j(r)0 = 0

and thus I satisfies both conditions in Definition 2.14.

Example 2.16. Although 2Z = {2a : a 2 Z} is not a subring of Z, it is an ideal. Sim-


ilarly, nZ is an ideal of Z for all n > 0. Later on (see Theorem 2.26), we will prove
that every ideal of Z is of the form nZ for some n > 0.

12
2.4 Quotient rings

Suppose we are given a ring R and an ideal I. Here we explain how to use these two
ingredients to construct a new ring, denoted R/I, which is called the quotient ring
(of R by I). In order to do this, we need to first define the elements of R/I and then
we need to equip this set with appropriate addition and multiplication operations.

Definition 2.17. Let I be an ideal of a ring R and fix an element r 2 R. Then the coset
of I containing r is the subset

r + I = {r + i : i 2 I}

of R, and we sometimes refer to r as a coset representative.

Example 2.18. Suppose R = Z and I = 4Z. Then

0 + I = {. . . , 4, 0, 4, 8, . . .} = 4 + I = 8 + I = · · ·
1 + I = {. . . , 3, 1, 5, 9, . . .} = 7 + I = 17 + I = · · ·
2 + I = {. . . , 2, 2, 6, 10, . . .} = 10 + I = 6 + I = · · ·
3 + I = {. . . , 1, 3, 7, 11, . . .} = 97 + I = 51 + I = · · ·

and we see that there are exactly four distinct cosets of I in R.

Remark 2.19. Given R and I as above, we can define a relation on R: for r, s 2 R,

r ⇠ s () r s 2 I.

It is easy to check that this is an equivalence relation on R (that is, ⇠ is reflexive,


symmetric and transitive) and we see that the equivalence class of r, namely [r] =
{s 2 R : r ⇠ s}, is precisely the coset r + I. In particular, we deduce that if r + I and
s + I are cosets, then either r + I = s + I or (r + I) \ (s + I) is the empty set (that is, the
distinct cosets form a partition of R). Moreover,

r + I = s + I () r = s + i for some i 2 I

As an important special case, r + I = 0 + I if and only if r 2 I. Also note that


r + i + I = r + I for all i 2 I.

Notation. Let R be a ring and let I be an ideal. We write R/I for the set of distinct
cosets of I in R.

We now define addition and multiplication operations on R/I to turn this set
into a ring, which we call the quotient ring of R by I.

13
Theorem 2.20. Let R be a ring and let I be an ideal. Then R/I is a ring with respect to the
operations

(a + I) + (b + I) = (a + b) + I
(a + I)(b + I) = ab + I

Here 0 + I and 1 + I are the zero and unity elements, respectively, and (a + I) = ( a) + I.

Proof. The main issue here is to show that the given operations are well-defined. To
explain what this means, notice that the operations are defined in terms of specific
coset representatives. But we know that coset representatives are not unique (in
general), so we need to show that the operations do not depend on the choices we
make.
So suppose we have c + I = a + I and d + I = b + I, so c = a + i and d = b + j for
some i, j 2 I. Then

(c + I) + (d + I) = (c + d) + I = (a + i + b + j) + I = (a + b) + (i + j) + I = (a + b) + I

since i + j 2 I. Similarly,

(c + I)(d + I) = cd + I = (a + i)(b + j) + I = ab + (bi + a j + i j) + I = ab + I

since bi + a j + i j 2 I (by definition of an ideal, this is the sum of three elements in I).
This shows that the operations are well-defined.
All of the ring properties for R/I are now easy to verify since they are inherited
from R. For example, to check distributivity let a + I, b + I, c + I 2 R/I and observe
that

(a + I)((b + I) + (c + I)) = (a + I)((b + c) + I) = a(b + c) + I


= ab + ac + I
= (ab + I) + (ac + I)
= (a + I)(b + I) + (a + I)(c + I),

where we are using the fact that a(b + c) = ab + ac by distributivity in R. We leave


the remaining properties to be checked as an exercise.

Definition 2.21. Let R be a ring with an ideal I. Then the map p : R ! R/I sending
each a 2 R to the coset a + I is called the quotient map.

Exercise. Show that the quotient map p : R ! R/I is a surjective homomorphism


with ker(p) = I. Deduce that every ideal of R arises as the kernel of a surjective
homomorphism R ! S for some ring S.

14
Recall that if j : R ! S is a homomorphism of rings, then im(j) is a subring of
S and ker(j) is an ideal of R. Therefore, we can construct R/ ker(j). The following
fundamental result reveals that the rings R/ ker(j) and im(j) are in fact isomorphic.

Theorem 2.22 (Homomorphism Theorem). Let j : R ! S be a ring homomorphism.


Then R/ ker(j) ⇠
= im(j).

Proof. Set I = ker(j) and define a map

y : R/I ! im(j)
a + I ! j(a)

We first have to check that this map is well-defined. So suppose b + I = a + I, in


which case b = a + i for some i 2 I. Then

y(b + I) = j(b) = j(a + i) = j(a) + j(i) = j(a) + 0 = j(a)

and the map is well-defined. We now check the remaining properties:

• Homomorphism: Let a + I, b + I 2 R/I. Then

y((a + I)(b + I)) = y(ab + I) = j(ab) = j(a)j(b) = y(a + I)y(b + I)

and thus y is compatible with multiplication. Similarly, it is compatible with


addition and we note that y(1 + I) = j(1) = 1.

• Surjective: Every element s 2 im(j) is of the form s = j(a) for some a 2 R, so


y(a + I) = s.

• Injective: If y(a + I) = 0, then j(a) = 0 and thus a 2 I, which implies that


a + I = 0 + I. Therefore, ker(y) is the zero ideal and y is injective by Lemma
2.10.

This completes the proof of the theorem.

Example 2.23. Let’s look at some concrete examples.

(1) Let j : R ! R be the identity homomorphism, so j(a) = a for all a 2 R. Then


ker(j) = {0} and im(j) = R, so R/{0} ⇠
= R by Theorem 2.22.

(2) Consider the map j : R ! {0}. Here ker(j) = R and im(j) = {0}, so R/R ⇠
= {0}.

(3) Take j : R[x] ! R with f 7! f (0). Then im(j) = R and

ker(j) = { f 2 R[x] : f (0) = 0} = {xg : g 2 R[x]} = xR[x]

is the set of polynomials with zero constant term. Therefore, Theorem 2.22
implies that R[x]/xR[x] ⇠
= R.

15
(4) Let n be a positive integer and consider the natural map j : Z ! Z/nZ defined
by a 7! a (mod n). Here j is surjective and ker(j) = nZ, in which case Theorem
2.22 implies that Z/nZ ⇠= Z/nZ. This explains our choice of notation for the
ring of integers modulo n, and it allows us to view Z/nZ in two different ways:

• The set {0, 1, . . . , n 1} with addition and multiplication modulo n; or


• The quotient ring {0 + nZ, 1 + nZ, . . . , n 1 + nZ} of Z by nZ, where

(a + nZ) + (b + nZ) = (a + b) + nZ, (a + nZ)(b + nZ) = ab + nZ,

recalling that c + nZ = d + nZ () c d 2 nZ () c ⌘ d (mod n).

Example 2.24. Consider the homomorphism

y : Z ! Z/2Z ⇥ Z/3Z, a 7! (a (mod 2), a (mod 3))

and note that ker(y) = 6Z (since 2|a and 3|a if and only if 6|a). In detail,

ker(y) = {a 2 Z : y(a) = (0, 0)}


= {a 2 Z : a (mod 2) = 0 and a (mod 3) = 0}
= {a 2 Z : 2|a and 3|a}
= {a 2 Z : 6|a}
= 6Z.

By Theorem 2.22, it follows that

Z/6Z ⇠
= im(y) ✓ Z/2Z ⇥ Z/3Z.

But both sets Z/6Z and Z/2Z ⇥ Z/3Z have exactly 6 elements, so y is surjective and
thus Z/6Z ⇠= Z/2Z ⇥ Z/3Z. As a consequence, for every (m, n) 2 Z/2Z ⇥ Z/3Z, there
exists some a 2 Z such that a (mod 2) = m and a (mod 3) = n.

This generalises as follows.

Theorem 2.25 (Chinese Remainder Theorem). Let a, b > 1 be coprime integers. Then

Z/abZ ⇠
= Z/aZ ⇥ Z/bZ.

Proof. We can simply repeat the argument given above.

Remark. If a and b are not coprime, then the rings Z/abZ and Z/aZ ⇥ Z/bZ are not
isomorphic. As an exercise, verify this for a = b = 2.

16
Next we use Theorem 2.22 to classify all the ideals of Z. To do this, we need to
recall two properties of Z and N from Introduction to Proofs and Group Theory:

• Division with remainder: If a, n 2 Z, then there exist integers q, r such that a =


nq + r and 0 6 r < n.

• Well-ordering principle: Every non-empty set of positive integers contains a


least element.

Theorem 2.26. Let I be an ideal of R = Z. Then I = nZ for some integer n > 0.

Proof. If I = {0} then I = 0Z, so we may assume I contains a nonzero integer. In fact,
since I is closed under additive inverses, we may assume that I contains a positive
integer. By applying the well-ordering principle, let n be the smallest positive inte-
ger in I and note that nZ ✓ I. If a 2 I then by division with remainder we may write
a = nq + r, where 0 6 r < n. Then r = a nq 2 I and thus r = 0 by the minimality of
n. Therefore, a = nq 2 nZ and thus I ✓ nZ, hence I = nZ as required.

2.5 More on ideals

Definition 2.27. Let R be a ring with ideals I and J. We define

I \ J = {a 2 R : a 2 I and a 2 J}
I + J = {i + j : i 2 I, j 2 J}
( )
`
IJ = Â ik jk : ik 2 I, jk 2 J, ` > 1
k=1

Lemma 2.28. Let I, J ✓ R be ideals. Then I \ J, I + J and IJ are all ideals of R.

Proof. The proof is left as an exercise.

Exercise. Suppose R = Z, I = mZ and J = nZ, where m, n > 1. Show that

I \ J = lcm(m, n)Z, I + J = hcf(m, n)Z, IJ = mnZ,

where ‘lcm’ means ‘lowest common multiple’. In addition, give an example to show
that if I and J are ideals of R, then I [ J need not be an ideal.

Definition 2.29. Let R be a ring and let a1 , . . . , an 2 R. Then

(a1 , . . . , an ) = Ra1 + · · · + Ran = {r1 a1 + · · · + rn an : r1 , . . . , rn 2 R}

is called the ideal of R generated by a1 , . . . , an . If I is an ideal of R of the form


I = (a1 , . . . , an ) for some ai 2 R, then we say that I is finitely generated.

17
As an exercise, check that (a1 , . . . , an ) as defined above really is an ideal of R.

Remark. Note that (a1 , . . . , an ) is the smallest ideal of R containing each ai (that is, if
I is any ideal of R containing each ai , then (a1 , . . . , an ) ✓ I).

Example 2.30.

(1) By Theorem 2.26, every ideal of Z is of the form (n) = nZ.

(2) The ideal I = { f 2 R[x, y] : f (0, 0) = 0} = (x, y) = Rx + Ry of R[x, y] is finitely


generated.

Recall that a unit is an element r 2 R such that rs = 1 for some s 2 R, and recall
that we write R⇥ for the set of units in R.

Lemma 2.31. Let I ✓ R be an ideal. Then I = R if and only if I contains a unit.

Proof. If I = R then I contains every unit in R. Conversely, suppose I contains a unit


r with rs = 1 for some s 2 R. Then by definition of an ideal, we have rs 2 I, so 1 2 I
and thus a = a1 2 I for all a 2 R. Therefore, I = R.

This leads us nicely into the following theorem. Recall that a nonzero ring is a
field if and only if every nonzero element is a unit.

Theorem 2.32. A ring R 6= {0} is a field if and only if {0} and R are the only ideals of R.

Proof. If R is a field then R⇥ = R \ {0} and so every nonzero ideal of R contains a unit
and is therefore equal to R by Lemma 2.31.
Conversely, let us assume {0} and R are the only ideals of R and fix a nonzero
element a 2 R. Set I = (a) = {ar : r 2 R}, which is a nonzero ideal of R. So by our
assumption, I = R and thus 1 2 I, hence 1 = ab for some b 2 R and therefore a 2 R⇥ .
It follows that R⇥ = R \ {0} and R is a field.

Example 2.33. We know that the rings R = Q, R and C are all fields and so their
only ideals are {0} and R. On the other hand, the following rings R all have proper
nonzero ideals (an example is given in the table) and so they are not fields:

R I
Z 2Z
R[x] xR[x]
R⇥R R ⇥ {0}
Z/4Z {0, 2}

18
Corollary 2.34. Let j : R ! S be a homomorphism, where R is a field and S 6= {0}. Then
j is injective.

Proof. Set I = ker(j). Since R is a field and I is an ideal, we have I = {0} or R by


Theorem 2.32, but I = R implies that j(1) = 0 and thus 0 = 1 in S, which is not
possible since S is nonzero. Therefore, I = {0} and j is injective by Lemma 2.10.

Example 2.35. The natural inclusion map Q ! R is a ring homomorphism. How-


ever, the previous corollary implies that there are no ring homomorphisms R ! Q
since R is uncountable but Q is countable (indeed, recall that there are no injective
maps from an uncountable set to a countable set).

Let j : R ! S be a ring homomorphism. For a subset A ✓ R we define the image


of A as follows:
j(A) = {j(a) : a 2 A} ✓ S.

Similarly, we define the pre-image (or pull-back) of a subset B ✓ S by setting

1
j (B) = {r 2 R : j(r) 2 B} ✓ R.

The latter notation should not be confused with an inverse map (indeed, j need not
be a bijection).
The following result is sometimes called the Correspondence Theorem.

Theorem 2.36. Let j : R ! S be a surjective ring homomorphism and let A and B be ideals
of R and S, respectively. Then the following hold:

(i) j(A) is an ideal of S.

(ii) j 1 (B) is an ideal of R containing ker(j).

(iii) There is a bijective correspondence

{ideals of R containing ker(j)} ! {ideals of S}


A 7! j(A)
j 1 (B) [ B

Proof. First consider (i). We have that j(0) = 0, so 0 2 j(A). Next let c, d 2 j(A) and
s 2 S, say c = j(x) and d = j(y) with x, y 2 A. In addition, since j is surjective we
have s = j(r) for some r 2 R. Then using the fact that A is an ideal of R we get

c + d = j(x) + j(y) = j(x + y) 2 j(A)


sc = j(r)j(x) = j(rx) 2 j(A)

19
and thus j(A) is an ideal of S. Part (ii) is similar and is left as an exercise. Note that
j 1 (B) contains ker(j) since 0 2 B.
Finally, let us turn to part (iii). First observe that the two maps in the statement
of part (iii) are well-defined by parts (i) and (ii). To verify that they define a bijective
correspondence, it suffices to show that

1 1
j (j(A)) = A, j(j (B)) = B

for every ideal A of R containing ker(j) and every ideal B of S. Note that the in-
clusions A ✓ j 1 (j(A)) and j(j 1 (B)) ✓ B are immediate from the definitions of
image and pre-image.
Let A be an ideal of R containing ker(j) and suppose x 2 j 1 (j(A)). Then j(x) 2
j(A), so j(x) = j(a) for some a 2 A and we have x a 2 ker(j) ✓ A. Therefore,
x 2 A and thus j 1 (j(A)) ✓ A, hence equality. A very similar argument shows that
j(j 1 (B)) = B for every ideal B of S.

Exercise. Give an example to show that the surjectivity hypothesis in Theorem 2.36
is essential for the validity of part (i). Is it also needed for part (ii)?

Example 2.37. Let’s look at a concrete example. Set R = Z, S = Z/8Z and consider
the natural map j : R ! S, a 7! a (mod 8), which is clearly surjective. Recall that
every ideal of R is of the form nZ for some n > 0 and note that mZ ✓ nZ if and only
if n divides m. We now inspect the ideals of R containing ker(j) = 8Z and the ideals
of S:

Z ! {0, 1, 2, 3, 4, 5, 6, 7}
2Z ! {0, 2, 4, 6}
4Z ! {0, 4}
8Z ! {0}

On the left we have the ideals of Z containing 8Z, and on the right we have the
complete list of ideals of Z/8Z.

Corollary 2.38. Let R be a ring with an ideal I. Then there is a bijective correspondence

{ideals of R containing I} ! {ideals of R/I}


A 7! A/I := {a + I : a 2 A}

Proof. Apply Theorem 2.36 with respect to the quotient map j : R ! R/I.

20
3 Integral domains
In this chapter (and the next) we will introduce several important families of rings,
which we will study for the remainder of the course. We begin by introducing the
family of integral domains.

3.1 Field of fractions

Definition 3.1. Let R be a ring. A nonzero element a 2 R is called a zero divisor if


ab = 0 for some nonzero b 2 R.

Definition 3.2. A ring R 6= {0} is an integral domain (ID for short) if it has no zero
divisors.

In other words, R 6= {0} is an integral domain if and only if for all a, b 2 R,

ab = 0 =) a = 0 or b = 0.

In particular, if R is an integral domain and a 6= 0, then

ab = ac =) a(b c) = 0 =) b c = 0 =) b = c.

That is, we can cancel nonzero elements.

Example 3.3. The rings Q, R, C, Z, Z[i] and Z/5Z are all integral domains, whereas
Z/4Z, Z/6Z and R ⇥ R are not. For example, in Z/6Z we have 2 ⇥ 3 = 0, and in R ⇥ R
we have (1, 0)(0, 1) = (0, 0).

Remark 3.4. Note that if j : R ! S is a ring isomorphism, then the following hold:

(a) j 1 : S ! R is also an isomorphism.

(b) j(R⇥ ) = S⇥ (since ab = 1 () j(a)j(b) = j(1) = 1).

(c) j maps zero divisors to zero divisors (indeed, ab = 0 () j(a)j(b) = 0).

So, if R ⇠
= S are isomorphic rings, then R is a field if and only if S is a field. And
similarly, R is an integral domain if and only if S is an integral domain.

Lemma 3.5. Let R be a nonzero ring.

(i) If R is a field, then R is an integral domain.

(ii) R is an integral domain if and only if for all nonzero a 2 R, the map r : R ! R, b 7! ab,
is injective.

21
(iii) If R is a finite integral domain, then R is a field.

Note that the map r in part (ii) is just a function from a set to itself (it is a ring
homomorphism if and only if a = 1).

Proof. For part (i), if ab = 0 and a 6= 0, then 0 = a 1 (ab) = b (here we are using the
fact that every nonzero element in a field has a multiplicative inverse).
Next consider part (ii). First assume R is an integral domain and a 2 R is nonzero.
Then r(b) = r(c) implies that ab = ac, so a(b c) = 0 and thus b = c since R has no
zero divisors. This shows that r is injective. On the other hand, if a 2 R is nonzero
and we assume r is injective, then ab = 0 implies that r(b) = r(0) = 0 and thus b = 0,
hence R is an integral domain.
For part (iii), fix a nonzero element a 2 R and consider the injective map r : R ! R
in part (ii). Since R is finite, this map must be surjective and thus r(b) = ab = 1 for
some b 2 R. It follows that R⇥ = R \ {0} and R is a field.

Lemma 3.6. Let R be a ring.

(i) R is an integral domain if and only if the polynomial ring R[x] is an integral domain.

(ii) Every subring of an integral domain is an integral domain.

Proof. This is a straightforward exercise.

Note that part (ii) of Lemma 3.6 implies that every subring of a field is an integral
domain (note that a subring of a field need not be a field, e.g. Z is a subring of Q).

Example 3.7. The subrings R, Q, Z and Z[i] of the field C are all integral domains.
p p p p
So is Z[ 2, 3, p], which is the smallest subring of R containing 2, 3 and p.

Lemma 3.6(ii) implies that every subring of a field is an integral domain. Con-
versely, we can show that every integral domain R is a subring of some field, and
there exists a unique smallest field with this property, which is called the field of
fractions of R. This field is defined in the following theorem, using exactly the same
idea that is used to formally construct Q from Z.

Theorem 3.8. Let R be an integral domain and set

X = {(a, b) : a, b 2 R, b 6= 0}.

Define a relation on X by setting (a, b) ⇠ (c, d) if and only if ad = bc.

a
(i) ⇠ is an equivalence relation. Write b for the equivalence class containing (a, b).

22
(ii) Let K be the set of distinct equivalence classes and define addition and multiplication
on K as follows:
a c ad + bc a c ac
+ = , = .
b d bd b d bd
0
Then K is a field with zero element 1 and unity element 11 .

Proof. For (i), first observe that ⇠ is reflexive since ab = ba, and symmetric because
ad = bc implies that cb = da. Finally, if (a, b) ⇠ (c, d) and (c, d) ⇠ (e, f ) then ad f =
bc f = bde and thus a f = be since d 6= 0 and R is an integral domain, so (a, b) ⇠ (e, f )
and we have transitivity.
Now consider (ii). First we need to show that the given operations are well-
defined. That is, if (a, b) ⇠ (a0 , b0 ) and (c, d) ⇠ (c0 , d 0 ), then

ad + bc a0 d 0 + b0 c0 ac a0 c0
= , = .
bd b0 d 0 bd b0 d 0
This is entirely straightforward. For example, we have ab0 = ba0 and cd 0 = dc0 , so

(ba0 )(dc0 ) = (bd)(a0 c0 ) =) (ab0 )(cd 0 ) = (bd)(a0 c0 ) =) (ac)(b0 d 0 ) = (bd)(a0 c0 )

ac a0 c0
and thus bd = b0 d 0 . Similarly, one can check that addition is well-defined.
We now need to verify all the ring axioms – this is also straightforward, but
rather tedious so we omit the details.
Finally, we need to check that every nonzero element in K is a unit. Again, this
a b
is straightforward because if b 2 K is nonzero then a, b 6= 0, so a 2 K and we have
ab ab 1 a
ba = ba = 1, which is the unity element in K. Therefore, b is a unit and we conclude
that K is a field.

Definition 3.9. Let R be an integral domain. The field K constructed in Theorem 3.8
is called the field of fractions of R, denoted F (R).

Lemma 3.10. Let R be an integral domain with field of fractions K = F (R). Then
na o
: a2R
1
is a subring of K isomorphic to R.

a
Proof. First observe that the map j : R ! K, a 7! 1 is a ring homomorphism with
im(j) = { a1 : a 2 R}. The latter is a subring of K by Lemma 2.7(iii) and j is injective
because
a b
= () a1 = b1 () a = b.
1 1
Therefore, Theorem 2.22 implies that R ⇠
= im(j) and the result follows.

23
Note. If R ✓ S are integral domains then F (R) ✓ F (S).

Note. Suppose R is an integral domain and R ✓ K, where K is a field. For 0 6= b 2 K,


let b 1 2 K be the multiplicative inverse of b and consider the subset

1
S = {ab : a, b 2 R, b 6= 0} ✓ K.

Then it is easy to check that S is a subring of K. In fact, we see that S is the smallest
subfield of K containing R, and the natural map
a 1
j : F (R) ! S, 7! ab
b
is a well-defined isomorphism.
So if we can embed our given integral domain R in some field K, then (up to
isomorphism) F (R) is simply the smallest subfield of K containing R.

Example 3.11. We conclude by reviewing some specific integral domains and their
field of fractions.

• We have F (Z) = Q.
a ab 1
• If R is a field, then F (R) = R since b = 1 .

• As an exercise, show that F (Z[i]) = Q[i] = {a + bi : a, b 2 Q}.


a
• Let R = Z[1/2] = 2n : a 2 Z, n > 0 be the smallest subring of Q containing
1/2. Then Z ✓ R ✓ Q and thus Q = F (Z) ✓ F (R) ✓ F (Q) = Q, so F (R) = Q.

• Suppose K is a field and R = K[x]. Then



f (x)
F (R) = K(x) = : f , g 2 K[x], g 6= 0
g(x)

is called the field of rational functions over K in the indeterminate x.

3.2 Maximal and prime ideals

Here we introduce two special families of ideals and we study their connection with
integral domains and fields.

Definition 3.12. Let R be a ring and let I be a proper ideal.

(1) I is a maximal ideal if whenever J is an ideal of R with I ✓ J ✓ R, then either


J = I or J = R.

24
(2) I is a prime ideal if whenever ab 2 I for a, b 2 R, either a 2 I or b 2 I.

Example 3.13. Let R be a nonzero ring and let I = {0} be the zero ideal.

• I is a maximal ideal () R and I are the only ideals of R () R is a field (see


Theorem 2.32)

• I is a prime ideal () R has no zero divisors () R is an integral domain.

Example 3.14. Let R = Z and recall that every ideal of Z is of the form nZ for some
n > 0 (see Theorem 2.26). As an exercise, show that

• nZ is a maximal ideal if and only if n is a prime number.

• nZ is a prime ideal if and only if n is prime or n = 0.

For example, if n = ab is composite with a, b > 1, then ab 2 nZ but a 62 nZ and b 62 nZ,


so nZ is not a prime ideal.

The following key result provides an alternative characterisation of maximal


and prime ideals in terms of quotient rings.

Theorem 3.15. Let I ✓ R be an ideal. Then

(i) I is maximal if and only if R/I is a field; and

(ii) I is prime if and only if R/I is an integral domain.

Proof. First consider (i). As noted in Corollary 2.38, there is a bijective correspon-
dence between the ideals of R/I and the ideals of R containing I. So if I is maximal
then R/I has exactly two ideals (namely, the zero ideal and R/I itself), so R/I is a
field by Theorem 2.32. And conversely, if R/I is a field then R and I are the only
ideals of R containing I, so I is maximal.
Now consider part (ii). First assume I is a prime ideal and (a + I)(b + I) = 0 + I
for some a + I, b + I 2 R/I. Then ab + I = 0 + I, so ab 2 I and thus a 2 I or b 2 I since I
is prime. Therefore, either a + I = 0 + I or b + I = 0 + I, so R/I is an integral domain.
The converse is very similar and we omit the details.

Since every field is an integral domain (see Lemma 3.5(i)), we obtain the follow-
ing corollary:

Corollary 3.16. Every maximal ideal is a prime ideal.

25
Example 3.17. Let R = Z and fix a prime number p and a composite number n.

I R/I
Z not prime, not maximal {0} not an integral domain, not a field
{0} prime, not maximal Z integral domain, not a field
pZ prime and maximal Z/pZ integral domain and field
nZ not prime, not maximal Z/nZ not an integral domain, not a field

Example 3.18. Let’s consider three examples involving polynomial rings:

(1) Let R = R[x] and consider the homomorphism j : R ! R, f 7! f (0). This is sur-
jective with kernel I = xR[x], so Theorem 2.22 implies that R/I ⇠
= R. Therefore,
R/I is a field, so I is both maximal and prime.

(2) Let R = R[x, y]. Here we have a surjective homomorphism j : R ! R, f 7! f (0, 0)


with kernel I = xR + yR, hence R/I ⇠ = R and once again I is both maximal and
prime.

(3) Let R = R[x, y] and consider the homomorphism j : R ! R[y], sending f to


f (0, y). Then j is surjective with kernel I = xR, so R/I ⇠
= R[y] is an integral do-
main (see Lemma 3.6(i)) but not a field. Therefore, I is prime but not maximal.
For example, I ( xR + yR ( R.

Theorem 3.19. Every nonzero ring R has a maximal ideal.

Proof. This proof is non-examinable. Let X be the set of proper ideals of R and note
that X is non-empty (e.g. it contains {0}). Then X is a partially ordered set (poset for
short) with respect to inclusion. That is, we have a relation A ✓ B defined on some
pairs A, B 2 X which is reflexive (since A ✓ A), antisymmetric (if A ✓ B and B ✓ A,
then A = B) and transitive (if A ✓ B and B ✓ C, then A ✓ C).
Let Y = {Ia : a 2 Z} be a totally ordered subset of X (this simply means that for all
A, B 2 Y , either A ✓ B or B ✓ A). Define
[
J= Ia
a2Z

and observe that J is a proper ideal of R containing every ideal in Y (note that 1 62 J
since each Ia is proper). This means that every totally ordered subset Y of X has a
maximal element (i.e. there is an element B 2 X such that A ✓ B for all A 2 Y ).
At this point, we can now invoke Zorn’s Lemma, which is equivalent to the Axiom
of Choice (this will be discussed in detail in the 3rd year unit Set Theory). As a con-
sequence, we deduce that X has a maximal element. In other words, R has a proper

26
ideal I such that if I ✓ J ✓ R with J 2 X (so that J 6= R), then J = I. In particular, I is
a maximal ideal of R and the proof is complete.

This has the following immediate consequence.

Corollary 3.20. Every proper ideal of a ring is contained in a maximal ideal.

Proof. Let R be a ring and let I be a proper ideal. We may as well assume I is not
maximal. We can then repeat the proof of Theorem 3.19, working with the set X
of proper ideals of R containing I. Alternatively, we can apply Theorem 3.19 to
the quotient ring R/I, noting that J/I is a maximal ideal of R/I if and only if J is a
maximal ideal of R containing I (see Remark 3.21 below).

Remark 3.21. Let R be a ring and let I be a proper ideal. By Corollary 2.38, there is
a bijective correspondence

{ideals of R containing I} ! {ideals of R/I}


A 7! A/I

This correspondence is inclusion preserving (that is, if A and B are both ideals of
R containing I with A ✓ B, then A/I ✓ B/I as ideals of R/I). Hence, the same map
induces a bijection from the set of maximal ideals of R containing I and the set of
maximal ideals of R/I. For example, if R = Z and I = 8Z, then I is contained in a
unique maximal ideal of R, namely A = 2Z, hence R/I = Z/8Z has a unique maximal
ideal, namely A/8Z = {0, 2, 4, 6} (cf. Example 2.37).

27
4 Special rings
In this chapter we introduce several special families of rings that have many prop-
erties in common with the ring of integers.

Throughout this chapter, we assume R is an integral domain.

4.1 Primes and irreducibility

Definition 4.1. If a, b 2 R we say that b divides a, written b|a, if a = bc for some c 2 R.


Equivalently, if a 2 (b), where (b) = Rb is the ideal of R generated by b.

Remark 4.2.

(1) If b 6= 0 then the element c in Definition 4.1 is unique (indeed, if a = bc = bd,


then b(c d) = 0 and thus c = d since R is an integral domain).

(2) Note that if a 2 R⇥ is a unit, then b|a if and only if b 2 R⇥ .

Definition 4.3.

(1) Elements a, b 2 R are said to be associates if a|b and b|a. This is equivalent to
the condition (a) = (b). It is also equivalent to a = bu for some unit u 2 R⇥ .

(2) An element p 2 R is irreducible if p 6= 0, p is not a unit and

p = ab =) a is a unit or b is a unit.

(3) An element p 2 R is prime if p 6= 0, p is not a unit and

p|ab =) p|a or p|b.

Equivalently, (p) is a nonzero prime ideal (ab 2 (p) ) a 2 (p) or b 2 (p)).

Suppose p 2 R is nonzero and a non-unit. If p is not irreducible, then we will


sometimes say that p is reducible.
Note. A field does not contain any irreducible or prime elements (since every
nonzero element is a unit, by definition).

Remark. As an exercise, show that if u 2 R is a unit and p 2 R is irreducible, then


the product up is also irreducible. In particular, every associate of an irreducible
element is irreducible. And similarly for prime elements.

28
Example 4.4.

(1) For R = Z we have R⇥ = {±1} and thus n and n are associates for all n. The
prime and irreducible elements in R coincide with the usual prime numbers
(and their negatives).

(2) For R = Q[x] we have R⇥ = Q \ {0} (the nonzero constant polynomials), so the
associates of f 2 R are the polynomials of the form c f with c 2 Q \ {0}.

It follows that the irreducible elements in R are the polynomials of positive


degree that cannot be expressed as the product of two other positive degree
polynomials in R.

Lemma 4.5. Every prime element in R is irreducible.

Proof. Let p 2 R be a prime element such that p = ab with a, b 2 R (note that a|p
and b|p). Then p|ab and thus p|a or p|b. So either p|a and a|p, in which case p and
a are associates and b must be a unit, or the same situation occurs with a and b
interchanged. We conclude that p is irreducible.

In general, the converse of Lemma 4.5 is false and we present a concrete example
to show this.
p
Example 4.6. Set R = {a + b 5 : a, b 2 Z} ✓ C. Clearly R contains 1 and it is easy to
see that it is closed under addition, multiplication and additive inverses. Therefore,
p
R is a subring of C and hence an integral domain (note that R = Z[ 5] in terms of
our earlier notation). We first claim that R⇥ = {±1}.
p p p
Suppose a + b 5 2 R⇥ , say (a + b 5)(c + d 5) = 1 and thus
p p
|a + b 5|2 |c + d 5|2 = (a2 + 5b2 )(c2 + 5d 2 ) = 1,
p p
where |a + b 5| = a2 + 5b2 is the usual modulus of a complex number. Since
a, b, c, d 2 Z, we deduce that a2 + 5b2 = c2 + 5d 2 = 1 is the only possibility and thus
p
a, c 2 {±1} and b = d = 0, so a + b 5 = ±1 as claimed (and clearly 1 and 1 are
indeed units in R).
Next we claim that the following elements
p p
2, 3, 1 + 5, 1 5
p p
are all irreducible. Starting with 2, let us assume 2 = (a + b 5)(c + d 5), so 4 =
2 2 2 2 2 2 2 2
p
(a + 5b )(c + 5d ) and thus a + 5b 2 {1, 2, 4}. If a + 5b = 1 then a + b 5 = ±1
p
is a unit as above. Similarly, if a2 + 5b2 = 4 then c2 + 5d 2 = 1 and c + d 5 = ±1 is a

29
unit. And since the equation a2 + 5b2 = 2 has no integer solutions, we deduce that
2 is irreducible. A very similar argument applies in the other three cases: |3|2 = 9 =
p
3 ⇥ 3, |1 ± 5|2 = 6 = 2 ⇥ 3, and neither a2 + 5b2 = 2 nor a2 + 5b2 = 3 has integer
solutions.
However, none of these elements are prime. For example, 2|6 and we have 6 =
p p
(1 + 5)(1 5), but 2 does not divide either of these two factors, so 2 is not
p
prime. And similarly, 3 is not prime. Also, 1 ± 5 divides 6 = 2 · 3, but it does not
divide 2 or 3, so it is not prime.

4.2 Unique factorisation domains

Definition 4.7. We say that an integral domain R is a unique factorisation domain


(UFD for short), if the following properties are satisfied:

(1) Every nonzero r 2 R has a factorisation of the form

r = up1 p2 · · · pn

where u 2 R⇥ is a unit, each pi 2 R is irreducible and n > 0.

(2) This factorisation is unique up to multiplication by units. That is, if r can also
be written as
r = vq1 q2 · · · qm

where v 2 R⇥ and each qi 2 R is irreducible (with m > 0), then n = m, and after
reordering we have that pi and qi are associates for i = 1, 2, . . . , n.

Example 4.8.

(1) Z is a UFD by the Fundamental Theorem of Arithmetic. For instance, we have

12 = 1 · 2 · 2 · 3 = 1 · ( 2) · ( 3) · 2 = ( 1) · 2 · 2 · ( 3)

and all of these factorisations satisfy the conditions in parts (1) and (2) of Def-
inition 4.7 (recall that n and n are associates in Z).

(2) Trivially, every field is a UFD (there are no irreducible elements in a field).

(3) The polynomial ring Z[x] is a UFD (see Theorem 5.10).

30
(4) Q[x] is also a UFD (see (5) below). Here x + 1, x 1, x2 + 1 2 Q[x] are irreducible
and we have the following factorisations of x4 1:

x4 1 = (x 1)(x + 1)(x2 + 1)
✓ ◆
1 2 1
= (2x 2)(3x + 3) x +
6 6
✓ ◆
1 2 1
= 6(x 1)(x + 1) x + .
6 6

all of which satisfy the uniqueness condition in Definition 4.7.

(5) More generally, K[x] is a UFD for any field K. In fact, if K is a field, then
K[x1 , x2 , . . . , xn ] is a UFD for all n > 1 (we will prove this in the next chapter as
a consequence of Gauss’ Lemma; see Corollary 5.11).
p
(6) The ring Z[ 5] from Example 4.6 is not a UFD. For instance, we have
p p
6 = 2 · 3 = (1 + 5) · (1 5)

which are two distinct factorisations into non-associate irreducibles (note that
the factors are non-associate since the only units in this ring are ±1).

Lemma 4.9. Let R be a UFD and let p 2 R. Then p is prime if and only if p is irreducible.

Proof. By Lemma 4.5, if p is prime, then p is irreducible. So let us assume p is


irreducible, hence p is nonzero and a non-unit. Suppose p|ab, say ab = pc. Since R
is a UFD, we can factor a, b, c into irreducibles and then compare the factorisations:

a1 a2 · · · ax b1 b2 · · · by = up c1 c2 · · · cz

where u is a unit and each ai , bi , ci is irreducible. By uniqueness of factorisation, p is


associate to one of the ai or bi , so either p|a or p|b and thus p is prime.

For the remainder of Section 4.2, let R be a UFD and consider nonzero elements
a, b 2 R. By factorising a and b, we may write

a = upa11 · · · pak k , b = vpb11 · · · pbk k ,

where u, v 2 R⇥ , ai , bi > 0 for all i (setting p0i = 1) and {p1 , . . . , pk } is a set of pair-
wise non-associate irreducible elements in R (that is, if i 6= j then pi and p j are not
associates). For example, if R = Z, a = 12 and b = 42, then

a = ( 1) · 22 · 31 · 70 , b = 1 · 21 · 31 · 71 .

31
Definition 4.10. In terms of the above notation, we define the highest common
factor of a and b to be the element

hcf(a, b) = pc11 · · · pckk 2 R,

where ci = min{ai , bi } for all i. This is uniquely defined up to multiplication by a


unit. In other words, the ideal of R generated by hcf(a, b) is uniquely defined.

In the above example with R = Z, we have c1 = c2 = 1 and c3 = 0, so

hcf( 12, 42) = 21 · 31 · 70 = 6.

And if we factorise using the irreducible element 2, rather than 2, then we get

a = ( 1) · ( 2)2 · 31 · 70 , b = ( 1) · ( 2)1 · 31 · 71 , hcf( 12, 42) = ( 2)1 · 31 · 70 = 6.

This explains why hcf(a, b) is only defined up to multiplication by a unit.


Remark 4.11.

(1) If b = 0 then we define hcf(a, b) = a for all nonzero a 2 R.

(2) If b 2 R⇥ is a unit, then hcf(a, b) = 1 for all a 2 R.

(3) Inductively, we can extend the definition of hcf to three or more elements:

hcf(a1 , a2 , . . . , an ) = hcf(hcf(a1 , a2 ), a3 , . . . , an ).

Lemma 4.12. Let a, b 2 R be nonzero and set h = hcf(a, b). Then the following hold:

(i) There exist elements c, d 2 R such that a = hc, b = hd and hcf(c, d) = 1.

(ii) If d 2 R divides a and b, then d|h.

(iii) hcf(ad, bd) = hd for all nonzero d 2 R.

Proof. As above, write a = upa11 · · · pak k and b = vpb11 · · · pbk k , so h = pc11 · · · pckk , up to mul-
tiplication by a unit, where ci = min{ai , bi }.
For (i), set c = upa11 c1
· · · pak k ck
and d = vpb11 c1
· · · pbk k ck
, so a = hc and b = hd.
Then hcf(c, d) = pe11 · · · pekk with ei = min{ai ci , bi ci } = 0 for all i, so hcf(c, d) = 1 as
required. For (ii), observe that every irreducible factor of d divides a, so we have
d = wpd11 · · · pdk k and di 6 ai for all i. Similarly, di 6 bi for all i, so di 6 ci = min{ai , bi }
and thus d divides h.
Finally, consider (iii). Write d = wpd11 · · · pdk k and note that

ad = (uw)pa11 +d1 · · · pak k +dk , bd = (vw)pb11 +d1 · · · pbk k +dk .

Then hcf(ad, bd) = pe11 · · · pekk with ei = min{ai + di , bi + di } = ci + di for all i, hence
hcf(ad, bd) = hd as required.

32
a
Remark. Let R be a UFD with field of fractions F = F (R) and let b 2 F be nonzero,
with h = hcf(a, b). Then Lemma 4.12(i) implies that a = hc and b = hd with hcf(c, d) =
a hc c a
1, so b = hd = d and thus every nonzero element in F can be written in the form b
with hcf(a, b) = 1.

4.3 Principal ideal domains

Definition 4.13. We say that an integral domain R is a principal ideal domain (PID
for short) if every ideal I of R is principal, which means that I is generated by one
element, i.e. I = (a) = Ra for some a 2 I.

In this setting, we say that a is a generator of I and we note that generators are
unique up to multiplication by units.

Example 4.14.

(1) Z is a PID by Theorem 2.26.

(2) Every field R is a PID since the only ideals are {0} = (0) and R = (1).

(3) Q[x] and Z[i] are both PIDs (this will be proved later – see Section 4.5).

(4) As an exercise, show that Z[3i] = {a + 3bi : a, b 2 Z} is not a PID.

(5) The polynomial rings Z[x] and Q[x, y] are not PIDs (see Examples 4.16 and 4.17).

The following lemma can be useful for showing that a given ring is not a PID.

Lemma 4.15. Let R be a PID and let p 2 R be nonzero and a non-unit. The following are
equivalent:

(i) p is irreducible.

(ii) (p) is a prime ideal (i.e. p is a prime element).

(iii) (p) is a maximal ideal.

Proof. We already have that (iii) =) (ii) =) (i) as these statements are true for all
integral domains (see Corollary 3.16 and Lemma 4.5). So we just need to show that
(i) =) (iii).
So let us assume p is irreducible. Note that (p) is a proper ideal since p is not a
unit. Suppose we have (p) ✓ I ✓ R for some ideal I. Since R is a PID, we have I = (a)
for some a 2 R and thus p = ab for some b 2 R. Since p is irreducible, either a is a
unit, in which case I = R, or b is a unit and thus I = (p). We conclude that (p) is a
maximal ideal.

33
Example 4.16. Let R = Z[x] and set I = (x). Consider the map j : R ! Z, f 7! f (0).
This is a surjective homomorphism with kernel I, so Theorem 2.22 implies that
R/I ⇠
= Z. Therefore, R/I is an integral domain and so I = (x) is a prime ideal (and
hence x is prime), but R/I is not a field and thus I is not a maximal ideal. Therefore,
Lemma 4.15 implies that R is not a PID.

Example 4.17. Let R = Q[x, y] and I = (x). By considering the map j : R ! Q[y],
f 7! f (0, y), we see that R/I ⇠
= Q[y] is an integral domain but not a field. So once
again, Lemma 4.15 implies that R is not a PID.

Remark 4.18. This argument applies more generally:

• If n > 2 and F is a field, then F[x1 , . . . , xn ] is not a PID.

• If n > 1 and R is an integral domain, but not a field, then R[x1 , . . . , xn ] is not a
PID.

So the only polynomial rings that are PIDs are those of the form F[x], where F is
a field (and indeed every ring of this form is a PID by Theorem 4.26 below). This
shows that PIDs are quite rare.

Lemma 4.19. Let R be a PID and let I1 ✓ I2 ✓ · · · be an increasing chain of ideals of R.


Then there exists n 2 N such that Im = In for all m > n.

Proof. Let I be the union of all the ideals in the chain. Then I is an ideal of R (check
this!), so I = (a) for some a 2 R. Therefore, a 2 In for some n and thus I = (a) ✓ In ✓ I.
We conclude that Im = In for all m > n.

Theorem 4.20. Every PID is a UFD.

Proof. Let R be a PID. We need to show that R satisfies properties (1) and (2) in
Definition 4.7. Fix a nonzero element r 2 R.

Existence. We need to show that r = up1 p2 · · · pn for some u 2 R⇥ and irreducible


elements pi 2 R. The claim is clear if r is a unit, or if r is irreducible, so assume r is
reducible.
First we claim that r = p1 r1 , where p1 is irreducible and r1 is a non-unit. Since
r is reducible, write r = a1 b1 with a1 , b1 non-units. If a1 is irreducible, then we are
done. If not, then a1 = a2 b2 and r = a2 b2 b1 , where a2 , b2 and b1 are non-units. We
can now keep repeating this process to obtain

r = a1 b1 = a2 b2 b1 = a3 b3 b2 b1 = . . .

34
where · · · |a3 |a2 |a1 |r and so we have a strictly ascending chain of ideals
(r) ( (a1 ) ( (a2 ) ( . . .
By Lemma 4.19, this process cannot continue indefinitely and thus
r = an (bn · · · b1 )
for some n, where an is irreducible and the product bn · · · b1 is a non-unit (recall that
the product of two or more non-units is a non-unit). This justifies the claim.
We now show that r is a product of irreducibles. By the previous claim, we
have r = p1 r1 with p1 irreducible and r1 62 R⇥ . If r1 is irreducible, then we are done.
Otherwise, write r1 = p2 r2 with p2 irreducible and r2 62 R⇥ (again, this decomposi-
tion exists by the previous claim). If r2 is irreducible, then r = p1 p2 r2 is a product
of irreducibles, otherwise r2 = p3 r3 with p3 irreducible and we can repeat the pro-
cess. As before, Lemma 4.19 implies that this process must eventually terminate, in
which case rn 1 = pn is irreducible for some n and we conclude that r = p1 p2 · · · pn is
a product of irreducibles.

Uniqueness. Now consider property (2) in Definition 4.7. Suppose


r = vq1 q2 · · · qm
where v is a unit and each q j irreducible. By Lemma 4.15, each pi and q j is prime.
Since p1 |vq1 q2 · · · qm , it follows that p1 |v or p1 |q j for some j, but the first possibility
can be ruled out since p1 is not a unit. So p1 |q j for some j. Since q j is irreducible, it
follows that q j = u1 p1 for some unit u1 , whence
r = up1 p2 · · · pn = vu1 p1 q1 · · · q j 1 q j+1 · · · qm

and we can cancel p1 in both expressions (recall that R is an integral domain) to


obtain
up2 · · · pn = vu1 q1 · · · q j 1 q j+1 · · · qm .

We now repeat with p2 and so on. In this way, we deduce that n = m (otherwise,
we will end up with an equation of the form a = bc, where a, b are units and c is a
non-unit, which is not possible). In addition, we can re-order q1 , . . . , qn so that pi |qi
for all i, which means that pi and qi are associates. This shows that property (2) in
Definition 4.7 is satisfied and the result follows.

For example, we know that Z is a PID by Theorem 2.26, so it is also a UFD by


Theorem 4.20 (earlier, we noted that Z is a UFD by appealing to the Fundamental
Theorem of Arithmetic).
Note that the converse of Theorem 4.20 is false. For example, Z[x] and Q[x, y] are
both UFDs, but neither ring is a PID.

35
4.4 More on polynomial rings

Let R be a ring and recall that

R[x] = {an xn + · · · + a1 x + a0 : ai 2 R, n > 0}

is the polynomial ring over R in the indeterminate x. Also recall that if

f = an xn + · · · + a1 x + a0 2 R[x]

and an 6= 0, then we call n the degree of f , denoted ∂ f . By convention, we set


∂ f = • if f is the zero polynomial.

Terminology. Suppose f 2 R[x], where R is a subring of T . Then we can view as an


element of R[x] or T [x]; in the former, we refer to ‘ f over R’, and in the latter ‘ f over
T ’. For example, x2 3 2 Z[x] is irreducible over Z, but reducible over R.

Lemma 4.21. Let R be an integral domain and let f , g 2 R[x]. Then

(i) ∂ ( f g) = ∂ f + ∂ g; and

(ii) ∂ ( f + g) 6 max{∂ f , ∂ g}, with equality if ∂ f 6= ∂ g.

Proof. Set ∂ f = n, ∂ g = m and let an xn and bm xm be the leading terms of f and g, re-
spectively. Since R is an integral domain, it follows that an bm 6= 0 and thus an bm xn+m
is the leading term of the product f g, which gives (i). Part (ii) follows immediately
from the definition of addition in R[x].

Corollary 4.22. If R is an integral domain, then R[x]⇥ = R⇥ .

Proof. Here we are viewing R as a subring of R[x], corresponding to the constant


polynomials. In particular, the inclusion R⇥ ✓ R[x]⇥ is clear. On the other hand, if
f 2 R[x]⇥ , say f g = 1 for some g 2 R[x], then Lemma 4.21(i) implies that ∂ f = ∂ g = 0,
so f , g 2 R and thus f 2 R⇥ . The result follows.

Corollary 4.23. If K is a field and f 2 K[x] has degree 1, then f is irreducible over K.

Proof. Suppose otherwise, say f = gh, where neither g nor h is a unit in K[x]. Since
K ⇥ = K \ {0}, the previous result implies that ∂ g, ∂ h > 1, so ∂ f = ∂ g + ∂ h > 2, which
is a contradiction.

Note. The condition that K is a field is needed in Corollary 4.23. For example,
f = 2x 2 2 Z[x] has degree 1, but it is not irreducible over Z since f = 2(x 1) is a

36
product of two non-units in Z[x] (note that 2 2 Z[x] is not a unit since Z[x]⇥ = Z⇥ =
{±1}).

Next we establish division with remainder in K[x] when K is a field.


Lemma 4.24. Let K be a field and let f , g 2 K[x] be nonzero. Then there exist q, r 2 K[x]
such that f = qg + r and ∂ r < ∂ g.

Proof. Let n > 0 be the degree of g. We proceed by induction on the degree of f .


To be precise, we need to show that the following statement is true for every non-
negative integer m:

P(m) : If f 2 K[x] has degree m, then there exist q, r 2 K[x] such that
f = qg + r and ∂ r < ∂ g.

Suppose m = 0. If n > 1 then we can take q = 0 and r = f . Otherwise, both f and


g are constant polynomials and we can take q = f g 1 and r = 0. This shows that
P(0) holds, which is the base case for the induction.
Now assume m > 0 and assume that P(`) holds for all ` < m. If m < n then we can
take q = 0 and r = f , so we may assume m > n. Write f = axm + · · · and g = bxn + · · · ,
where a, b 6= 0, and consider

h= f (ab 1 )xm n g 2 K[x].

Note that ∂ h 6 m. In fact, the coefficient of xm is a (ab 1 )b = 0, so ∂ h < m and by


the inductive hypothesis we have h = q1 g + r1 for some q1 , r1 2 K[x] with ∂ r1 < ∂ g.
Therefore,
f = (ab 1 )xm n g + h = (ab 1 xm n
+ q1 )g + r1
and so we can take q = ab 1 xm n+q
1 and r = r1 .
Example 4.25. Take K = Z/2Z = {0, 1} and R = K[x]. Let f = x3 + x2 + 1 and g = x2 + 1,
so m = 3, n = 2 and a = b = 1 in the notation above. Then

h= f (ab 1 )xg = x2 x + 1 = x2 + x + 1 = 1 · g + x

so q1 = 1 and r1 = x. Therefore, q = ab 1 x + 1 = x + 1 and r = x. Note that we can


also find q and r via polynomial long division:

x+1
x2 + 1 x3 + x2 + 1
x3 + x (= x(x2 + 1))
x2 + x + 1 (= (x3 + x2 + 1) (x3 + x))
x2 + 1 (= 1(x2 + 1))
x (= (x2 + x + 1) (x2 + 1))

37
which shows that x3 + x2 + 1 = (x + 1)(x2 + 1) + x.

Theorem 4.26. If K is a field, then K[x] is a PID.

Proof. Let I ✓ K[x] be an ideal. If I = {0}, then I = (0) is principal. Now assume
I 6= {0} and choose a nonzero element g 2 I of minimal degree (as in the proof of
Theorem 2.26, here we are using the well-ordering principle). We claim that I = (g).
To see this, let f 2 I and use Lemma 4.24 to write f = gq + r with ∂ r < ∂ g. Then
r= f gq 2 I, so r = 0 by the minimality of ∂ g and thus f = gq 2 (g). This justifies
the claim and we conclude that K[x] is a PID.

Since every PID is a UFD (see Theorem 4.20), we get the following corollary.

Corollary 4.27. If K is a field, then K[x] is a UFD.

4.5 Euclidean domains

In Theorem 2.26 we showed that Z is a PID. And in Theorem 4.26 we showed that
K[x] is a PID when K is a field. By inspecting the proofs, we see that they both rely
on the existence of an appropriate ‘absolute value’, or ‘degree function’ that allows
us to do division with remainder. This leads us naturally to the following definition
(here we write N0 for N [ {0}).

Definition 4.28. An integral domain R is a Euclidean domain (ED for short) if there
exists a degree function
d : R \ {0} ! N0

satisfying the following properties:

(1) For a, b 2 R, b 6= 0, there exist q, r 2 R such that a = qb + r, and either r = 0 or


d (r) < d (b).

(2) d (a) 6 d (ab) for all nonzero a, b 2 R.

Example 4.29. Some examples of degree functions are the following (so in each case,
the given ring is a Euclidean domain):

• For Z we can take d (a) = |a| (absolute value).

• For K[x] with K a field, a degree function is d ( f ) = ∂ f .

• If K is a field, we can take d (a) = 1 for all a 2 K \ {0}.

38
Note. The degree function of an ED is not necessarily unique. For example, both
d (a) = |a| and d 0 (a) = a2 are degree functions on Z.

Theorem 4.30. Every ED is a PID.

Proof. We can repeat the proof of Theorem 4.26: take an ideal I 6= {0}, choose a
nonzero element g 2 I so that d (g) is minimal and show that I = (g).

Exercise. For R = Z[i], show that d (x) = |x|2 is a degree function and thus Z[i] is an
ED. So by the previous result, Z[i] is also a PID and a UFD.

Remark 4.31. To summarise, we have

fields ✓ EDs ✓ PIDs ✓ UFDs ✓ integral domains ✓ rings

and all of these inclusions are strict. To justify the latter statement, it just remains
to show that there exists a PID that is not an ED. This is beyond the scope of this
p
course, but a concrete example is the ring Z[a], where a = (1 + 19)/2 2 C:

Oscar A. Cámpoli, A principal ideal domain that is not a Euclidean domain, Amer. Math.
Monthly, vol. 95, no. 9 (Nov. 1988), 868–871.

It may also be helpful to keep the following in mind, where R is an integral


domain and 0 6= p 2 R is a non-unit:

R is an ID: (p) maximal =) p prime =) p irreducible


3.16 4.5
R is a UFD: (p) maximal =) p prime () p irreducible
4.9
R is a PID: (p) maximal () p prime () p irreducible
4.15

39

You might also like