Divisibility in Integral Domains

Subject: ALGEBRA-III
Semester-IV

Lesson Name: Divisibility in Integral Domains

Lesson Developer: Sapna Malhotra & Divya Bhambri

College/Department: Department of Mathematics,

Gargi College,
University of Delhi

&

Department of Mathematics,
St. Stephens College,
University of Delhi

1|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Table of Contents

1. Introduction ........................................................................................................................ 3
2. Irreducible and Prime elements.............................................................................. 3
3. Unique Factorization Domains ............................................................................... 12
Application of Ascending chain condition .......................................................................... 14
4. Euclidean Domains ........................................................................................................ 19
5. Exercises ............................................................................................................................. 24
6. References ............................................................................................................................ 25
7. Suggested Readings .......................................................................................................... 25

2|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

1. Introduction

In previous two chapters, we studied “how to factorize the polynomial over the ring
of integers or over a field. Several important results such as Division algorithm, Factor
Theorem, Remainder Theorem, Unique factorization in ℤ[𝑥], Gauss Lemma, several
irreducibility tests including the Eisenstein’s criterion and many more were studied. In
this Chapter, we look at factorising in a more theoretical way.

2. Irreducible and Prime elements

Definition 2.1: Let 𝑅 be a commutative ring with unity and 𝑎, 𝑏 ∈ 𝑅. Then we say that 𝑎
and 𝑏 are associates or 𝒂 is an associate of 𝒃 denoted by 𝑎~𝑏 if 𝑎 = 𝑢𝑏, for some unit
𝑢 ∈ 𝑅.

Example 2.2: In ℤ, 5 and −5 are associates because −1 ∈ ℤ which is a unit

and 5 = (−1) −5 . Thus, 5 ~ (−5).

In fact, for any 𝑥 ∈ ℤ, the pair 𝑥 and −𝑥 are associates.

Value Addition
The relation ‘~’ (associate) is an equivalence relation on 𝑅, a commutative ring with
unity.
(i) Reflexive: 𝑎 ~ 𝑎 ∀ 𝑎 ∈ 𝑅, because 𝑎 = 1. 𝑎, where 1 is the unity of 𝑅.
(ii) Symmetric: Let 𝑎 ~ 𝑏. Then, 𝑎 = 𝑢𝑏, for some unit 𝑢 ∈ 𝑅.
⇒ 𝑢−1 𝑎 = 𝑏, where 𝑢−1 ∈ 𝑅 is unit as 𝑢 is a unit.
⇒ 𝑏 = 𝑢−1 𝑎.
⇒ 𝑏 ~ 𝑎.
(iii) Transitive: Let 𝑎 ~ 𝑏 and 𝑏 ~ 𝑐.
Then 𝑎 = 𝑢𝑏 and 𝑏 = 𝑣𝑐, for some unit elements 𝑢, 𝑣 ∈ 𝑅.
As 𝑎 = 𝑢𝑏.
= 𝑢(𝑣𝑐).
= 𝑢𝑣 𝑐 , where 𝑢𝑣 ∈ 𝑅 is unit
(because 𝑢−1 , 𝑣 −1 ∈ 𝑅 and 𝑢𝑣 𝑣 −1 𝑢−1 = 1).
⇒ 𝑎 ~ 𝑐.

Definition 2.3: Let 𝑅 be a commutative ring with unity. Then 0 ≠ 𝑎 ∈ 𝑅 is said to be

irreducible if
(i) 𝑎 is non-unit.
(ii) whenever 𝑎 = 𝑏. 𝑐, then either 𝑎 is a unit or 𝑏 is a unit.

3|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Value Addition
Let 𝑅 be a commutative ring with unity. If 𝑎 is an irreducible element in 𝑅, then 𝑎 is an
irreducible polynomial in 𝑅[𝑥].

Example 2.4: In the ring ℤ, 2 is an irreducible element because 2 is non-zero and non-
unit and 2 = 1.2.

Whereas 6 is not an irreducible element in ℤ, though it is non-zero and non-unit but

6 = 2.3 where neither 2 nor 3 is unit in ℤ.

Definition 2.5: Let 𝑅 be a commutative ring with unity. A non-zero, non-unit element 𝑎
of 𝑅 is said to be prime if 𝑎 ∣ 𝑏𝑐 then either 𝑎 ∣ 𝑏 or 𝑎 ∣ 𝑐.

Example 2.6: 2 is a prime element in the ring ℤ of integers,

Whereas 6 is not a prime element in ℤ as 6 | 4.3, but 6 ∤ 4 and 6 ∤ 3.

Value Addition
 0 element of the ring is always reducible.
 From the definition of irreducible element, it is clear that in order to show that
𝑎 ∈ 𝑅 is not irreducible it is enough to find a pair of non-unit elements 𝑏, 𝑐 ∈
𝑅 such that 𝑎 = 𝑏𝑐.
 It may be observed that 𝑎 ∈ 𝑅 is not prime if there exist a pair of elements 𝑏, 𝑐 ∈ 𝑅
such that 𝑎 ∣ 𝑏𝑐 but 𝑎 ∤ 𝑏 and 𝑎 ∤ 𝑐.
 One can easily verify that every prime number in ℤ is both prime element as well
as irreducible element.
 An element 𝑎 ∈ 𝑅 is prime if and only if < 𝑎 > is a prime ideal of 𝑅.

I.Q.1 Prove that if 𝑎 and 𝑏 are associates in a commutative ring 𝑅 with unity such that 𝑎
is irreducible then 𝑏 is also irreducible.

I.Q.2 Prove that if 𝑎 and 𝑏 are associates in a commutative ring 𝑅 with unity such that 𝑎
is prime element then 𝑏 is also a prime element.

Note: In case of integers, the concept of irreducible and prime element are equivalent
but in general it is not so. We will see in Example 2.7, that in the ring ℤ −3 , 1 + −3 is
an irreducible element but not a prime element.

We define integral domains of the form

ℤ 𝑑 = {𝑎 + 𝑏 𝑑 ∶ 𝑎, 𝑏 ∈ ℤ},

4|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

where 𝑑 ≠ 1 and is not divisible by the square of a prime.

Define a function 𝑁 ∶ ℤ 𝑑 → ℤ+ ∪ 0 as

𝑁 𝑥 = 𝑁 𝑎 + 𝑏 𝑑 = 𝑎2 − 𝑑𝑏 2 , for every 𝑥 = 𝑎 + 𝑏 𝑑 ∈ ℤ 𝑑 .
It may be observed that

(i) 𝑁 𝑥 = 0 if and only if 𝑥 = 0.

Let 𝑁 𝑥 = 0, where 𝑥 = 𝑎 + 𝑏 𝑑.
⇔ 𝑁 𝑎+𝑏 𝑑 =0
⇔ 𝑎2 − 𝑑𝑏 2 = 0
⇔ 𝑎2 − 𝑑𝑏 2 = 0
⇔ 𝑎2 = 𝑑𝑏 2
⇔ 𝑎 = − 𝑑 𝑏 or = 𝑑 𝑏 .
⇔ 𝑎 + 𝑏 𝑑 = 0 or 𝑎 − 𝑏 𝑑 = 0.
If 𝑑 = 0, then we obtain 𝑎 = 0 and therefore 𝑥 = 0.
If 𝑑 ≠ 0. As 𝑎, 𝑏 ∈ ℤ , and 𝑎 + 𝑏 𝑑 = 0 or 𝑎 − 𝑏 𝑑 = 0.
This is possible only if 𝑎 = 0 and 𝑏 = 0, so 𝑥=0
Thus 𝑁 𝑥 = 0 ⇔ 𝑥 = 0.

(ii) 𝑁 𝑥𝑦 = 𝑁 𝑥 𝑁 𝑦 ∀ 𝑥, 𝑦 ∈ ℤ 𝑑 .

Let 𝑥 = 𝑎 + 𝑏 𝑑 and 𝑦 = 𝑎′ + 𝑏′ 𝑑, where 𝑎, 𝑎′ , 𝑏, 𝑏′ ∈ ℤ.

Then 𝑥𝑦 = (𝑎 + 𝑏 𝑑)(𝑎′ + 𝑏′ 𝑑)
= 𝑎𝑎′ + 𝑏𝑏 ′ 𝑑 + 𝑎𝑏 ′ + 𝑎′ 𝑏 𝑑.
Consider 𝑁 𝑥𝑦 = 𝑁( 𝑎𝑎′ + 𝑏𝑏 ′ + 𝑎𝑏 ′ + 𝑎′ 𝑏 𝑑).
= 𝑎𝑎′ + 𝑏𝑏 ′ 𝑑 2
− 𝑑 𝑎𝑏 ′ + 𝑎′ 𝑏 2
.
= |𝑎 𝑎 + 𝑑 𝑏 𝑏 + 2𝑎𝑎 𝑑𝑏𝑏 − 𝑑𝑎2 𝑏 ′ 2 − 2𝑎𝑎′ 𝑏𝑏 ′ 𝑑|.
2 ′2 2 2 ′2 ′ ′

= |𝑎2 𝑎′ 2 − 𝑎2 𝑏 ′ 2 𝑑 − 𝑑𝑏 2 𝑎′ 2 + 𝑑 2 𝑏 2 𝑏 ′ 2 |.
= |𝑎2 𝑎′ 2 − 𝑑𝑏 ′ 2 − 𝑏 2 𝑑 𝑎′ 2 − 𝑑𝑏 ′ 2 |.
= (𝑎2 − 𝑑𝑏 2 )(𝑎′ 2 − 𝑑𝑏 ′ 2 ) .
= 𝑎2 − 𝑑𝑏 2 𝑎′ 2 − 𝑑𝑏 ′ 2 .
= 𝑁 𝑥 𝑁(𝑦).

(iii) 𝑁 𝑥 = 1 if and only if 𝑥 is unit.

Let 𝑥 = 𝑎 + 𝑏 𝑑 where 𝑎, 𝑏 ∈ ℤ.
𝑁 𝑥 = 1.
iff 𝑎2 − 𝑑𝑏 2 = 1.
iff 𝑎 + 𝑏 𝑑 𝑎 − 𝑏 𝑑 = 1.
iff 𝑥 is unit.
The function 𝑵 is called the norm.

5|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

(iv) If 𝑁(𝑥) is prime, then 𝑥 is irreducible in ℤ[ 𝑑].

Let 𝑁 𝑥 = 𝑝, where 𝑝 ∈ ℤ is prime.

As 𝑁(𝑥) is prime , therefore 𝑁(𝑥) ≠ 0.

⇒ 𝑥 ≠ 0.

Also, since 𝑁(𝑥) is prime ⇒ 𝑁 𝑥 ≠1 ⇒ 𝑥 is non-unit.

Let 𝑥 = 𝑦𝑧 where 𝑦, 𝑧 ∈ ℤ[ 𝑑].

We show that either 𝑦 is unit or 𝑧 is unit.

𝑁 𝑥 = 𝑁(𝑦𝑧).

⇒ 𝑁 𝑥 = 𝑁 𝑦 𝑁(𝑧).

⇒ 𝑝 = 𝑁 𝑦 𝑁(𝑧).

As 𝑝 is prime, therefore we have

either 𝑁 𝑦 = 𝑝, 𝑁 𝑧 = 1 or 𝑁 𝑦 = 1, 𝑁 𝑧 = 𝑝.

Thus we have either 𝑁 𝑦 = 1 or 𝑁 𝑧 = 1.

⇒ either 𝑦 is unit or 𝑧 is unit.

Example 2.7 Prove that 1 + −3 is an irreducible element in ℤ −3 but not a prime

element.
Solution: We use the fact that 𝑁 𝑎 + 𝑏 −3 = 𝑎2 + 3𝑏 2 is a norm.

(i) Clearly, 1 + −3 ≠ 0.

(ii) 1 + −3 is non-unit.
If 1 + −3 is unit, then
𝑁 1 + −3 = 1.
⇒ 4 = 1, which is not possible.
Hence, 1 + −3 is non-unit.

(iii) To show that: 1 + −3 is irreducible.

Let if possible 1 + −3 is not irreducible. Then,
1 + −3 = 𝑥𝑦, where neither 𝑥 nor 𝑦 is a unit in ℤ[ −3].
Note that 𝑁 1 + −3 = 𝑁(𝑥𝑦).
⇒ 4 = 𝑁 𝑥 𝑁(𝑦).

6|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

⇒ 𝑁 𝑥 = 2 and 𝑁 𝑦 = 2.
But there are no integers 𝑎 and 𝑏 satisfying 𝑎2 + 3𝑏 2 = 2.
Hence, our assumption is wrong and therefore 1 + −3 must be irreducible.
Now, we will verify that 1 + −3 is not a prime.
As 4 = 1 + −3 . 1 − −3 .

It follows that 1 + −3 | 4 = 2.2 , but (1 + −3) ∤ 2

For if (1 + −3) | 2, then

2 = (1 + −3)(𝑎 + 𝑏 −3) for some 𝑎 , 𝑏 ∈ ℤ.
⇒ 𝑁(2) = 𝑁 1 + −3 𝑁(𝑎 + 𝑏 −3).
⇒ 4 = 4(𝑎2 + 3𝑏 2 ).
⇒ 1 = 𝑎2 + 3𝑏 2 .
⇒ 𝑎 = ±1 and 𝑏 = 0.
⇒ 2 = (1 + −3) which is not possible.
Hence, 1 + −3 is not a prime.

Next we give one more example of an irreducible element in an integral domain

that is not a prime element.

Example 2.8: Consider the integral domain ℤ −5 = {𝑎 + 𝑏 −5 ∶ 𝑎, 𝑏 ∈ ℤ}. We show that

3 is an irreducible element but not a prime element in ℤ −5 .

Solution: (i) Clearly, 3 ≠ 0.

(ii) 3 is a non-unit element in ℤ −5 .

For if 3 is unit in ℤ −5 , then

𝑁 3 =1 ⇒ 9 = 1, which is not possible.
(iii) Claim 1: 3 is an irreducible element.
Let 3 = (𝑎 + 𝑏 −5)(𝑐 + 𝑑 −5) where 𝑎, 𝑏, 𝑐, 𝑑 ∈ ℤ.

𝑁 3 =𝑁 𝑎 + 𝑏 −5 𝑐 + 𝑑 −5 .

⇒ 9 = 𝑁(𝑎 + 𝑏 −5)𝑁(𝑐 + 𝑑 −5) .

⇒ 9 = (𝑎2 + 5𝑏 2 )(𝑐 2 + 5𝑑 2 ).
⇒ 𝑎2 + 5𝑏 2 = 1, 𝑐 2 + 5𝑑 2 = 9 or 𝑎2 + 5𝑏 2 = 3, 𝑐 2 + 5𝑑 2 = 3 or 𝑎2 + 5𝑏 2 = 9, 𝑐 2 + 5𝑑 2 = 1.

If 𝑎2 + 5𝑏 2 = 1 then 𝑎 = ±1, 𝑏 = 0.
If 𝑎2 + 5𝑏 2 = 3 then 𝑐 2 + 5𝑑 2 = 3 is not possible as 𝑎, 𝑏, 𝑐, 𝑑 ∈ ℤ.
If 𝑐 2 + 5𝑑 2 = 1 then 𝑐 = ±1, 𝑑 = 0.

Therefore, if 𝑎2 + 5𝑏 2 = 1 then 𝑎 + 𝑏 −5 = ±1 which is a unit and if 𝑐 2 + 5𝑑 2 = 1, then

𝑐 + 𝑑 −5 = ±1 which is again a unit.

7|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Hence the Claim.

(iv) Claim 2: 3 is not a prime element.

We see that 3 ∣ 9 where 9 = (2 + −5)(2 − −5) ∈ ℤ −5 .

But 3 ∤ (2 + −5) and 3 ∤ (2 − −5) because if 3 ∣ (2 + −5) in ℤ −5 , then

2 + −5 = 3(𝑎 + 𝑏 −5) for some 𝑎, 𝑏 ∈ ℤ.

⇒ 2 = 3𝑎
⇒ 𝑎 = 2/3 which is not possible.
Therefore, 3 ∤ (2 + −5). Similarly we can see that 3 ∤ (2 − −5).
Hence, the Claim.

In Examples 2.7 and 2.8, we have seen that in an integral domain every
irreducible element need not be prime. The next question that arises in our mind is that
if an integral domain contains a prime element, then will it be irreducible. The answer to
this question lies in the theorem stated below.

Theorem 2.9: Let 𝑅 be an integral domain with unity. If 𝑎 ∈ 𝑅 is prime, then it is

irreducible.
Proof: Let 𝑎 ∈ 𝑅 be a prime element.

To show: 𝑎 is an irreducible element in 𝑅.

Let 𝑎 = 𝑏. 𝑐 where 𝑏, 𝑐 ∈ 𝑅.
⇒ 𝑎 ∣ 𝑏𝑐.
As 𝑎 is prime, therefore 𝑎 ∣ 𝑏 or 𝑎 ∣ 𝑐.
Let us assume that 𝑎 ∣ 𝑏.
⇒ 𝑏 = 𝑎𝑡 for some 𝑡 ∈ 𝑅.
⇒ 𝑏 = 𝑏𝑐 𝑡 = 𝑏(𝑐𝑡).
⇒ 𝑏 1 − 𝑐𝑡 = 0.
As 𝑅 is an integral domain ⇒ 𝑏 = 0 or 1 − 𝑐𝑡 = 0.
If 𝑏 = 0, then 𝑎 = 𝑏. 𝑐 = 0 which is not possible as 𝑎 being a prime element is non-zero.
Therefore, 1 − 𝑐𝑡 = 0.
⇒ 𝑐𝑡 = 1.
⇒ 𝑐 is unit.
Therefore, 𝑎 is an irreducible element. ■

Remark 2.10: The converse of the Theorem 2.9 is not true. As seen in Example 2.8, in
the integral domain ℤ −5 , 3 is an irreducible element but not a prime element.

8|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Theorem 2.11: Let 𝑅 be a principal ideal domain. Then 𝑎 ∈ 𝑅 is prime if and only if it is
irreducible.
Proof: Let 𝑎 ∈ 𝑅 be a prime element.
Since 𝑅 is a principal ideal domain therefore it is an integral domain with unity and from
Theorem 2.9, it follows that 𝑎 is an irreducible element.
Conversely,
Let 𝑎 ∈ 𝑅 be an irreducible element.
Let 𝑎 ∣ 𝑏𝑐 where 𝑏, 𝑐 ∈ 𝑅.
To show that: 𝑎 ∣ 𝑏 or 𝑎 ∣ 𝑐.

Consider 𝐼 = {𝑎𝑥 + 𝑏𝑦 ∶ 𝑥, 𝑦 ∈ 𝑅}.

Then ∅ ≠ 𝐼 ⊆ 𝑅 because 0 = 𝑎. 0 + 𝑏. 0 ∈ 𝐼, where 0 ∈ 𝑅.
We will verify that 𝐼 is an ideal of 𝑅.

Let 𝑢, 𝑣 ∈ 𝑅 be arbitrary.
Then 𝑢 = 𝑎𝑥 + 𝑏𝑦 and 𝑣 = 𝑎𝑥 ′ + 𝑏𝑦′ for some 𝑥, 𝑥 ′ , 𝑦, 𝑦′ ∈ 𝑅.
Therefore 𝑢 − 𝑣 = 𝑎 𝑥 − 𝑥 ′ + 𝑏(𝑦 − 𝑦 ′ ) ∈ 𝐼 because 𝑥 − 𝑥 ′ , 𝑦 − 𝑦 ′ ∈ 𝑅.
For 𝑟 ∈ 𝑅, 𝑢𝑟 = 𝑎 𝑥𝑟 + 𝑏(𝑦𝑟) ∈ 𝐼 as 𝑥𝑟, 𝑦𝑟 ∈ 𝑅.
Since 𝑅 is commutative ⇒ 𝑟𝑢 ∈ 𝐼.
Hence, 𝐼 is an ideal of 𝑅 which is a principal ideal domain.

Therefore, 𝐼 = < 𝑑 > for some 𝑑 ∈ 𝐼 ⊂ 𝑅.

As 𝑎 = 𝑎. 1 + 𝑏. 0 ∈ 𝐼 =< 𝑑 >.
⇒ 𝑎 = 𝑑𝑡 for some 𝑡 ∈ 𝑅.
Since 𝑎 is an irreducible element.
⇒ either 𝑑 is unit or 𝑡 is unit.

If 𝒅 is unit, then 𝑑 −1 ∈ 𝑅 ⇒ 𝑑𝑑 −1 ∈ 𝐼 ⇒ 1 ∈ 𝐼 ⇒ 1 = 𝑎𝑥 + 𝑏𝑦 for some 𝑥, 𝑦 ∈ 𝑅.

⇒ 𝑐 = 𝑎𝑐𝑥 + 𝑏𝑐𝑦.
Since 𝑎 ∣ 𝑎 ⇒ 𝑎 ∣ 𝑎𝑐𝑥
and as 𝑎 ∣ 𝑏𝑐 ⇒ 𝑎 ∣ 𝑏𝑐𝑦.
⇒ 𝑎 ∣ 𝑎𝑐𝑥 + 𝑏𝑐𝑦 = 𝑐 𝑎𝑥 + 𝑏𝑦 .
⇒ 𝑎 ∣ 𝑐.

If 𝒕 is unit, then 𝑡 −1 ∈ 𝑅 ⇒ 𝑑 = 𝑎𝑡 −1 .
As 𝑏 = 0. 𝑎 + 𝑏. 1 ∈ 𝐼 = < 𝑑 >
⇒ 𝑏 = 𝑑𝑘 for some 𝑘 ∈ 𝑅.
⇒ 𝑏 = 𝑎(𝑡 −1 𝑘).
⇒ 𝑎 ∣ 𝑏.

9|Page
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Hence, 𝑎 is a prime element of 𝑅. ■

I.Q. 3 Is ℤ −5 a principal ideal domain?

The answer to this I.Q. is No. (In view of Theorem 2.11 and Example 2.8).

Theorem 2.12: Let 𝑅 be a principal ideal domain which is not a field and 𝑝 ∈ 𝑅. Then
< 𝑝 > is a maximal ideal of 𝑅 if and only if 𝑝 is irreducible.
Proof: Let 𝐼 = < 𝑝 > be a maximal ideal of 𝑅.

To show: 𝑝 is an irreducible element of 𝑅.

(i) 𝑝 ≠ 0 .
Assume that 𝑝 = 0.
Since 𝑅 is not a field therefore we can find 0 ≠ 𝑏 ∈ 𝑅 such that 𝑏 −1 does not exists.
Let 𝐵 = < 𝑏 > be an ideal of 𝑅.
Also 𝑏 ∉ 𝐼 as 𝑏 ≠ 0.
We note that 1 ∉ 𝐵 because if 1 ∈ 𝐵 =< 𝑏 > then 1 = 𝑏𝑥 for some 𝑥 ∈ 𝑅
⇒ 𝑏 is unit which is not true.
Hence, we have 𝐼 ⊊ 𝐵 ⊊ 𝑅 contradicting the fact that 𝐼 is a maximal ideal of 𝑅.
Therefore 𝑝 must be non-zero.

(ii) 𝑝 is non-unit.
Let us assume that 𝑝 is unit.
Then 𝑝−1 ∈ 𝑅 ⇒ 𝑝𝑝−1 ∈ 𝐼 ⇒ 1 ∈ 𝐼 ⇒ 𝐼 = 𝑅,

which is not true as 𝐼 is a maximal ideal of 𝑅.

Therefore 𝑝 is not a unit.

(iii) Let 𝑝 = 𝑐𝑑, where 𝑐, 𝑑 ∈ 𝑅.

Claim: either 𝑐 is a unit or 𝑑 is a unit.

As 𝑝 = 𝑐𝑑 ⇒ 𝑝 ∈ < 𝑑 >.
Thus 𝐼 ⊆ < 𝑑 > ⊆ 𝑅.
Since 𝐼 is maximal ideal of 𝑅 ⇒ 𝐼 =< 𝑑 > or < 𝑑 > = 𝑅.

If 𝐼 = < 𝑑 > .
Then as 𝑑 ∈ < 𝑑 > = 𝐼 = < 𝑝 > ⇒ 𝑑 = 𝑝𝑦 for some 𝑦 ∈ 𝑅.
As 𝑝 = 𝑐𝑑
⇒ 𝑝 = 𝑐𝑝 𝑦 = 𝑝(𝑐𝑦) (as 𝑅 is commutative).
⇒ 𝑝 1 − 𝑐𝑦 = 0 .
⇒ 𝑝 = 0 or 1 − 𝑐𝑦 = 0 (as 𝑅 is an integral domain).

10 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

⇒ 1 − 𝑐𝑦 = 0 (as 𝑝 ≠ 0).
⇒ 𝑐𝑦 = 1.
⇒ 𝑐 is unit.

If < 𝑑 > = 𝑅.
As 1 ∈ 𝑅 ⇒ 1∈< 𝑑 > ⇒ 1 = 𝑑𝑡, 𝑡 ∈ 𝑅.
⇒ 𝑑 is unit.

Therefore, 𝑝 is an irreducible element of 𝑅.

Conversely,

Let 𝑝 be an irreducible element of 𝑅.

To show: 𝐼 = < 𝑝 > is a maximal ideal of 𝑅.

Let 𝐴 be any ideal of 𝑅 such that 𝐼 ⊆ 𝐴 ⊆ 𝑅.

If 𝐼 = 𝐴, then we are done.

So, let us assume that 𝐼 ≠ 𝐴.

As 𝑅 is a principal ideal domain, therefore there exist some 𝑎 ∈ 𝑅 such that 𝐴 = < 𝑎 >.
Also 𝑎 ∉ 𝐼, because if 𝑎 ∈ 𝐼 then < 𝑎 > ⊆ 𝐼 which in turn implies that 𝐴 = 𝐼, which is not
true.
As 𝑝 ∈ 𝐼 ⊊ 𝐴 = < 𝑎 >.
⇒ 𝑝 = 𝑎𝑏 for some 𝑏 ∈ 𝑅.
Since 𝑝 is an irreducible element of 𝑅, therefore either 𝑎 is a unit or 𝑏 is a unit.
If 𝑏 is unit, then 𝑏 −1 exists in 𝑅 and as 𝑝 = 𝑎𝑏 ⇒ 𝑝𝑏 −1 = 𝑎.
As 𝑝 ∈ 𝐼, 𝑏 −1 ∈ 𝑅 and 𝐼 is an ideal of 𝑅 ⇒ 𝑝𝑏 −1 ∈ 𝐼 ⇒ 𝑎 ∈ 𝐼, which is not true.
Therefore, 𝑏 cannot be a unit and hence 𝑎 must be a unit.
⇒ 𝑎−1 ∈ 𝑅.
As 𝑎 ∈ 𝐴, 𝑎−1 ∈ 𝑅 and 𝐴 is an ideal of 𝑅.
It follows that 𝑎𝑎−1 ∈ 𝐴 ⇒ 1 ∈ 𝐴.
⇒ 𝐴 = 𝑅.
Hence, 𝐼 = < 𝑝 > is a maximal ideal of 𝑅. ■

Corollary 2.13: Let 𝑝 ∈ 𝑅 be an irreducible element, where 𝑅 is a principal ideal domain

which is not a field, then 𝑅 < 𝑝 > is a field.
Proof: As 𝑝 is an irreducible element of 𝑅.
From the Theorem 2.12, < 𝑝 > is a maximal ideal of 𝑅.
⇒ 𝑅 < 𝑝 > is a field. ■

11 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

3. Unique Factorization Domains

We know that every integer 𝑛 > 1 can be uniquely expressed as product of primes
and in a principal ideal domain an element is prime if and only if it is irreducible. Thus,
we can say that integer 𝑛 > 1, can be uniquely expressed as product of irreducibles. The
next question that arises is whether every integral domain has this property or not?

Definition 3.1: An integral domain 𝑅 is said to be a unique factorization domain

(U.F.D) if
(i) every non-zero, non-unit element of 𝑅 can be expressed as a product of irreducible
elements in 𝑅.
(ii) the factorization into irreducibles is unique up to associates and the order in which
the factors appear.
𝑛 𝑛 𝑛 𝑚 𝑚 𝑚𝑠
In other words, if 𝑎 = 𝑝1 1 𝑝2 2 … 𝑝𝑟 𝑟 = 𝑞1 1 𝑞2 2 … 𝑝𝑠 be two factorizations of an element 𝑎 ∈ 𝑅
as a product of irreducible, where no two 𝑝𝑖 ′𝑠 are associates and no two 𝑞𝑗 ′𝑠 are
associates, then 𝑟 = 𝑠 and each 𝑝𝑖 is an associate of one and only one 𝑞𝑗 .

Example 3.2: Every field 𝐹 is a unique factorization domain as it has no non-zero, non-
unit element.

Example 3.3: The ring ℤ of integers is a U.F.D.

Note that the only units in ℤ are ±1.

If 𝑛 ∈ ℤ is non-zero and non-unit, then either 𝑛 > 0 or 𝑛 < 0.

If 𝑛 > 0.
Then by fundamental theorem of arithmetic, 𝑛 can be uniquely expressed as
𝑘 𝑘 𝑘 𝑘
𝑛 = 𝑝1 1 𝑝2 2 𝑝3 3 … 𝑝𝑚𝑚 where 𝑝𝑖 ′𝑠 are prime.

⇒ 𝑛 = 𝑝1 𝑝1 … 𝑝1 … (𝑝𝑚 𝑝𝑚 … 𝑝𝑚 ).

Since ℤ is a P.I.D and we know that in a P.I.D, an element is prime iff it is irreducible.
Hence, 𝑛 is uniquely expressed as product of irreducible elements.

If 𝑛 < 0, then 𝑛 = −𝑚, for some 𝑚 > 0.

As in case of 𝑛 > 0, 𝑚 can be expressed uniquely as product of irreducible elements.

Let 𝑚 = 𝑞1 𝑞2 … 𝑞𝑘 .

Then, 𝑛 = −𝑚 = −𝑞1 𝑞2 … . . 𝑞𝑘 .

12 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

As 𝑞1 = (−1)(−𝑞1 )), where −1 is unit.

Therefore −𝑞1 and 𝑞1 are associates and we know that associate of an irreducible is
irreducible, so 𝑞1 is irreducible.

Example 3.4: From Unique Factorization in ℤ 𝑥 , Theorem 4.36 (previous chapter), we

can conclude that the ring of polynomials ℤ[𝑥] over integers is a unique factorization
domain.

Recall from previous chapter

Theorem: Let 0 ≠ 𝑓 𝑥 ∈ ℤ[𝑥]. Then 𝑓 𝑥 = 𝑑. 𝑓1 (𝑥), where 𝑑 is the content of 𝑓(𝑥) and 𝑓1 (𝑥)
is a primitive polynomial.

Theorem: Every primitive polynomial 𝑓(𝑥) ∈ ℤ[𝑥], can be written as product of irreducible
polynomials of positive degree.

I.Q. 4: If an integral domain 𝑅 is a unique factorization domain, then what can you say
about 𝑅[𝑥]?
The answer to this I.Q. is that 𝑅[𝑥] will also be a unique factorization domain.

Let 𝑓(𝑥) be a non-zero non-unit element of 𝑅 𝑥 .

Then 𝑓(𝑥) can be written as

𝑓 𝑥 = 𝑐(𝑓)𝑓1 (𝑥)
𝑓 𝑥 = 𝑎. 𝑓1 (𝑥), where 𝑎 = 𝑐(𝑓).

Also, as every primitive polynomial can be written uniquely as product of irreducibles.

Therefore, 𝑓1 𝑥 = 𝑞1 𝑥 𝑞2 𝑥 … 𝑞𝑛 (𝑥) where each 𝑞𝑖 (𝑥) is irreducible in 𝑅[𝑥].

Further, this representation is unique.

Also, 𝑎 ∈ 𝑅 and 𝑅 is a UFD, therefore 𝑎 = 𝑑1 𝑑2 … 𝑑𝑚 where each 𝑑𝑖 is irreducible and this

representation is unique.

We know that every irreducible element of 𝑅 is also an irreducible element of 𝑅[𝑥].

So, each 𝑑𝑖 is irreducible in 𝑅[𝑥].

Thus, 𝑓 𝑥 = 𝑑1 𝑑2 … 𝑑𝑚 𝑞1 𝑥 𝑞2 𝑥 … 𝑞𝑛 (𝑥), which is a unique representation of 𝑓(𝑥) as

product of irreducible elements in 𝑅[𝑥]. Hence, the result follows.

The next Theorem states that any ascending chain of ideals in a P.I.D. must
terminate. Some authors also refer this result as Ascending Chain Condition.

Theorem 3.5: Let 𝑅 be a principal ideal domain and 𝐴1 ⊂ 𝐴2 ⊂ ⋯ be any strictly

increasing chain of ideals in 𝑅. Then this chain of ideals must be of finite length.

13 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Proof: Let 𝐴1 ⊂ 𝐴2 ⊂ ⋯ be a strictly increasing chain of ideals in the principal ideal

domain 𝑅.
Let 𝐴 = 𝑛≥1 𝐴𝑛 union of all ideals in the chain.
Claim: 𝐴 is an ideal of 𝑅.

Since each 𝐴𝑖 is an ideal of 𝑅. Therefore 0 ∈ 𝐴𝑖 for each 𝑖.

⇒ 0 ∈ 𝐴 and hence 𝐴 ≠ ∅.
Let 𝑥, 𝑦 ∈ 𝐴 be arbitrary.
Then 𝑥 ∈ 𝐴𝑖 and 𝑦 ∈ 𝐴𝑗 for some 𝑖, 𝑗.
If 𝑖 < 𝑗, then 𝐴𝑖 ⊂ 𝐴𝑗 ⇒ 𝑥, 𝑦 ∈ 𝐴𝑗 which is an ideal of 𝑅.
⇒ 𝑥 − 𝑦 ∈ 𝐴𝑗 ⊂ 𝐴.
⇒ 𝑥 − 𝑦 ∈ 𝐴.
Similarly, if 𝑗 < 𝑖, then 𝐴𝑗 ⊂ 𝐴𝑖 ⇒ 𝑥 − 𝑦 ∈ 𝐴𝑖 ⊂ 𝐴.

Let 𝑟 ∈ 𝑅 be arbitrary.
Then, as 𝐴𝑖 is an ideal of 𝑅. ⇒ 𝑎𝑟, 𝑟𝑎 ∈ 𝐴𝑖 ⊂ 𝐴.
Thus 𝐴 is an ideal of 𝑅.
Since 𝑅 is a principal ideal domain and 𝐴 is an ideal of 𝑅.
Therefore 𝐴 = < 𝑎 > for some 𝑎 ∈ 𝑅.
As 𝑎 ∈ 𝐴 = 𝑛≥1 𝐴𝑛 therefore 𝑎 ∈ 𝐴𝑘 , for some 𝑘.
⇒ < 𝑎 > ⊆ 𝐴𝑘 ⇒ 𝐴 ⊆ 𝐴𝑘 ⊆ 𝐴𝑘+1 ⊆ ⋯.
⇒ 𝐴 ⊆ 𝐴𝑘 .
Also, 𝐴𝑘 ⊆ 𝐴.
Thus, it follows that 𝐴 = 𝐴𝑘 .
Therefore 𝐴𝑘 must be the last member of the chain.

Hence, the result. ■

Application of Ascending chain condition

Theorem 3.6: Let 𝑅 be a principal ideal domain. If 𝑎 ∈ 𝑅 is a non-zero, non-unit element
then there exists an irreducible element 𝑝 such that 𝑝 ∣ 𝑎.
Proof: If 𝑅 is a field, then the result is vacuously true, as every non-zero element of a
field is a unit.

So, let us assume that 𝑅 is a P.I.D which is not a field.

Let 0 ≠ 𝑎 ∈ 𝑅 be non-unit.

Consider the ideal 𝐼1 = < 𝑎 >, the ideal generated by 𝑎.

If 𝐼1 is a maximal ideal, then 𝑎 is an irreducible element and 𝑎 ∣ 𝑎 and in this case we are
done.
Suppose 𝐼1 is not a maximal ideal, then there exists an ideal 𝐼2 of 𝑅 such that

14 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

𝐼1 ⊊ 𝐼2 ⊊ 𝑅.

Let 𝐼2 = < 𝑝2 >, where 𝑝2 ∈ 𝑅.

If 𝐼2 is a maximal ideal, then 𝑝2 will be an irreducible element and as 𝑎 ∈ 𝐼1 ⊊ 𝐼2 = < 𝑝2 >.
⇒ 𝑎 = 𝑝2 𝑥 for some 𝑥 ∈ 𝑅 ⇒ 𝑝2 ∣ 𝑎.
Therefore, in this case, the result holds.

If 𝐼2 is not a maximal ideal, then there exist an ideal 𝐼3 of 𝑅 such that

𝐼1 ⊊ 𝐼2 ⊊ 𝐼3 ⊊ 𝑅.

Proceeding like this, we get an ascending chain of ideals in 𝑅 which must terminate after
a finite number of steps say at 𝐼𝑛 = < 𝑝𝑛 > (by Theorem 3.5).

Then, 𝐼𝑛 will be a maximal ideal of 𝑅 and 𝑝𝑛 an irreducible element such that 𝑝𝑛 ∣ 𝑎.

Theorem 3.7: Every principal ideal domain is a unique factorization domain.

Proof: Let 𝑅 be a principal ideal domain and 𝑎0 ∈ 𝑅 be a non-zero, non-unit element of 𝑅.

Claim 1: 𝑎0 can be written as a product of irreducible elements in 𝑅.

If 𝑎0 is irreducible, then 𝑎0 = 𝑎0 and we are done.

Suppose that 𝑎0 is not irreducible.
Then as 𝑎0 is non-zero, non-unit element, therefore by Theorem 3.6, there exists an
irreducible element 𝑝1 such that 𝑝1 ∣ 𝑎0 .
⇒ 𝑎0 = 𝑝1 𝑎1 , for some 𝑎1 ∈ 𝑅.
⇒ < 𝑎0 > ⊆ < 𝑎1 >.

If 𝑎1 is irreducible, then we have expressed 𝑎0 as a product of irreducible elements in 𝑅.

If 𝑎1 is not irreducible, then 𝑎1 ≠ 0 because if 𝑎1 = 0 then 𝑎0 = 𝑝1 𝑎1 = 𝑝1 . 0 = 0 which is not
true. Also 𝑎1 is non-unit, because if 𝑎1 is unit then 𝑎1 and 𝑝1 are associates and as 𝑝1 is
irreducible therefore 𝑎1 is irreducible which is not true.

Hence 𝑎1 is a non-zero, non-unit element of 𝑅 which is a principal ideal domain.

By Theorem 3.6 there exists an irreducible element 𝑝2 such that 𝑝2 ∣ 𝑎1 .
⇒ 𝑎1 = 𝑝2 𝑎2 , for some 𝑎2 ∈ 𝑅.
⇒ < 𝑎1 > ⊆ < 𝑎2 >.
If 𝑎2 is irreducible, then 𝑎0 = 𝑎1 𝑝1 = 𝑝2 𝑎2 𝑝1 which is a finite product of irreducible elements
and we are done.
If 𝑎2 is not irreducible. Then proceeding as above, it can be seen that 𝑎2 is non-zero and
non-unit in 𝑅.
Continuing like this we obtain an ascending chain of ideals
< 𝑎0 > ⊆ < 𝑎1 > ⊆ < 𝑎2 > ⊆ ⋯ .

15 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Then by Theorem 3.5 this chain must be of finite length.

Therefore there exists an irreducible element 𝑎𝑛 ∈ 𝑅 such that
< 𝑎0 > ⊆ < 𝑎1 > ⊆ < 𝑎2 > ⊆ ⋯ ⊆ < 𝑎𝑛 >
and 𝑎0 = 𝑝1 𝑝2 … 𝑝𝑛 𝑎𝑛 where 𝑝1 , 𝑝2 , … , 𝑝𝑛 , 𝑎𝑛 are irreducible elements in 𝑅.
Thus, Claim 1 holds.

Claim 2: The factorization is unique up to associated and the order in which they
appear.
In order to prove this, we have to show that if 𝑎 ∈ 𝑅 is non-zero, non-unit element and
𝑎 = 𝑝1 𝑝2 … 𝑝𝑚 = 𝑞1 𝑞2 … 𝑞𝑛
are two representations of 𝑎 as product of irreducibles then 𝑚 = 𝑛 and each 𝑝𝑖 is an
associate of some 𝑞𝑗 .
We prove this by using induction on 𝑛.

For 𝑛 = 1, 𝑎 = 𝑝1 𝑝2 … 𝑝𝑚 = 𝑞1 …(1)

As 𝑞1 is irreducible, some 𝑝𝑖 must be unit. But each 𝑝𝑖 being irreducible cannot be unit.
Therefore, (1) holds only if 𝑚 = 1.
Thus, 𝑎 = 𝑝1 = 𝑞1 and 𝑝1 = 1. 𝑞1 .
This implies that 𝑝1 and 𝑞1 are associates.
Hence the result holds for 𝑛 = 1.

Let the result be true for 𝑛 − 1.

As 𝑎 = 𝑝1 𝑝2 … 𝑝𝑚 = 𝑞1 𝑞2 … 𝑞𝑛 .
⇒ 𝑝1 𝑝2 … 𝑝𝑚 = 𝑞1 (𝑞2 … 𝑞𝑛 ).
⇒ 𝑞1 ∣ 𝑝1 𝑝2 … 𝑝𝑚 …(2)

As 𝑞1 is irreducible in 𝑅 which is a principal ideal domain and we know that in a principal

ideal domain an element is prime if and only if it is irreducible.
Therefore 𝑞1 is prime.
(2) ⇒ 𝑞1 ∣ 𝑝𝑖 for some 𝑖.
Since 𝑅 is a commutative ring, we can assume without loss of generality that 𝑖 = 1.
⇒ 𝑞1 ∣ 𝑝1 .
⇒ 𝑝1 = 𝑞1 𝑢1 for some 𝑢1 ∈ 𝑅.
As 𝑝1 is irreducible, therefore either 𝑞1 is unit or 𝑢1 is unit.
But 𝑞1 being irreducible is non-unit therefore 𝑢1 must be unit.
⇒ 𝑝1 and 𝑞1 are associates.

As 𝑝1 𝑝2 … 𝑝𝑚 = 𝑞1 (𝑞2 … 𝑞𝑛 ) and 𝑝1 = 𝑞1 𝑢1 .
⇒ (𝑞1 𝑢1 )𝑝2 … 𝑝𝑚 = 𝑞1 𝑞2 … 𝑞𝑛 .
⇒ 𝑢1 𝑝2 … 𝑝𝑚 = 𝑞2 𝑞3 … 𝑞𝑛

16 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

(as 𝑞1 ≠ 0 and cancellation law holds in an integral domain for non-zero elements).
⇒ 𝑝2 ′𝑝3 … 𝑝𝑚 = 𝑞2 𝑞3 … 𝑞𝑛 , where 𝑝2′ = 𝑢1 𝑝2 is irreducible. …(3)
Because 𝑝2 is irreducible, 𝑢1 is unit which implies that 𝑝2′ and 𝑝2 are associates and
associate of an irreducible element is irreducible, therefore 𝑝2′ is irreducible.
In (3), we have two equal representations as a product of irreducible and one of the
representation on R.H.S. contains (𝑛 − 1) elements.
Therefore, by induction hypothesis, on L.H.S. also we should have (𝑚 − 1) elements.
Therefore, 𝑛−1 = 𝑚−1 ⇒ 𝑚 = 𝑛.
Also we have seen above that 𝑝1 and 𝑞1 are associates.
Similarly, we can show that 𝑝2 and 𝑞2 are associates by considering
𝑝1 𝑝2 … 𝑝𝑚 = 𝑞2 (𝑞1 𝑞3 … 𝑞𝑛 ),
and proceeding as above.
In this manner, we can see that each 𝑝𝑖 is an associate of some 𝑞𝑗 .
Therefore, by Principle of mathematical induction result holds for every 𝑛.
Hence, Claim 2 holds.

Thus, from Claim 1 and 2, it follows that 𝑅 is a unique factorization domain. ■

Theorem 3.8: Let 𝐹 be a field, then 𝐹[𝑥] is a unique factorization domain.

Proof: Since every field is a P.I.D, so 𝐹 is a P.I.D. and hence 𝐹[𝑥] is a principal ideal
domain.

So, by Theorem 3.7 we obtain that 𝐹[𝑥] is a unique factorization domain. ■

Example 3.9: Prove that ℤ[ −10] is not a unique factorization domain.

Solution: Recall that 𝑁(𝑎 + −10𝑏) = 𝑎2 + 10𝑏 2 .
As 10 = −10 . −10 = 2.5 …(1)
We will show −10, 2, 5 are irreducible in ℤ[ −10].
(i) −10 is irreducible
Let, if possible, −10 be reducible.
Then there exists 𝑥 = 𝑎 + −10𝑏, 𝑦 = 𝑐 + −10𝑑 ∈ ℤ[ −10] , 𝑎, 𝑏, 𝑐, 𝑑 ∈ ℤ ,
such that −10 = 𝑥. 𝑦 .
⇒ 𝑁 −10 = 𝑁(𝑥. 𝑦).
⇒ 10 = 𝑁 𝑥 𝑁(𝑦).
⇒ 𝑁 𝑥 = 2, 𝑁 𝑦 = 5 or 𝑁 𝑥 = 5, 𝑁 𝑦 = 2 or 𝑁 𝑥 = 1, 𝑁 𝑦 = 10 or 𝑁 𝑥 = 10, 𝑁 𝑦 = 1.

If 𝑁 𝑥 = 2, 𝑁 𝑦 = 5.
⇒ 𝑎2 + 10𝑏 2 = 2 and 𝑐 2 + 10𝑑 2 = 5 which is not possible as 𝑎, 𝑏, 𝑐, 𝑑 ∈ ℤ.

Similarly, we can see that 𝑁 𝑥 = 5, 𝑁 𝑦 = 2 is not possible.

17 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

If 𝑁 𝑥 = 1, 𝑁 𝑦 = 10.
⇒ 𝑥 is unit, which is not true.

If 𝑁 𝑥 = 10, 𝑁 𝑦 = 1.
⇒ 𝑦 is unit, which is not true.

Hence, our assumption is wrong. Therefore, 10 is irreducible.

(ii) 2 is irreducible.
Let if possible 2 be reducible.
Then there exists 𝑥 = 𝑎 + −10𝑏, 𝑦 = 𝑐 + −10𝑑 ∈ ℤ[ −10] , 𝑎, 𝑏, 𝑐, 𝑑 ∈ ℤ ,
such that 2 = 𝑥. 𝑦 .
⇒ 𝑁 2 = 𝑁(𝑥. 𝑦).
⇒ 4 = 𝑁 𝑥 𝑁(𝑦).
⇒ 𝑁 𝑥 = 2, 𝑁 𝑦 = 2 or 𝑁 𝑥 = 1, 𝑁 𝑦 = 4 or 𝑁 𝑥 = 4, 𝑁 𝑦 = 1.

If 𝑁 𝑥 = 2, 𝑁 𝑦 = 2.
⇒ 𝑎2 + 10𝑏 2 = 2 and 𝑐 2 + 10𝑑 2 = 2 which is not possible as 𝑎, 𝑏, 𝑐, 𝑑 ∈ ℤ.

If 𝑁 𝑥 = 1, 𝑁 𝑦 = 4.
⇒ 𝑥 is unit, which is not true.

If 𝑁 𝑥 = 4, 𝑁 𝑦 = 1.
⇒ 𝑦 is unit, which is not true.

Hence, our assumption is wrong. Therefore, 2 is irreducible.

(ii) 5 is irreducible.
Let, if possible, 5 be reducible.
Then there exists 𝑥 = 𝑎 + −10𝑏, 𝑦 = 𝑐 + −10𝑑 ∈ ℤ[ −10] , 𝑎, 𝑏, 𝑐, 𝑑 ∈ ℤ ,
such that 5 = 𝑥. 𝑦 .
⇒ 𝑁 5 = 𝑁(𝑥. 𝑦).
⇒ 25 = 𝑁 𝑥 𝑁(𝑦).
⇒ 𝑁 𝑥 = 5, 𝑁 𝑦 = 5 or 𝑁 𝑥 = 1, 𝑁 𝑦 = 25 or 𝑁 𝑥 = 25, 𝑁 𝑦 = 1.

If 𝑁 𝑥 = 5, 𝑁 𝑦 = 5.
⇒ 𝑎2 + 10𝑏 2 = 5 and 𝑐 2 + 10𝑑 2 = 5 which is not possible as 𝑎, 𝑏, 𝑐, 𝑑 ∈ ℤ.

If 𝑁 𝑥 = 1, 𝑁 𝑦 = 25.
⇒ 𝑥 is unit, which is not true.

If 𝑁 𝑥 = 25, 𝑁 𝑦 = 1.
⇒ 𝑦 is unit, which is not true.

18 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Hence, our assumption is wrong. Therefore, 5 is irreducible.

Also, we note that −10 and 2 are not associates because if they were then
2 = −10. 𝑢 for some unit 𝑢 in ℤ[ −10].

⇒ 𝑁 2 = 𝑁( −10. 𝑢) .
⇒ 4=𝑁 −10 𝑁(𝑢).
⇒ 4 = 10.1 .
⇒ 4 = 10 , which is not possible.
Similarly, we can see that −10 and 5 are not associates.
Thus, by (1), we have two different representations of 10 as product of irreducibles and
therefore ℤ[ −10] is not a unique factorization domain.

4. Euclidean Domains

Definition 4.1: An integral domain 𝑅 is said to be a Euclidean domain if there is a

function 𝑑 from the non-zero elements of 𝑅 to the non-negative integers such that
1. 𝑑 𝑎 ≤ 𝑑 𝑎𝑏 for every non-zero 𝑎, 𝑏 ∈ 𝑅.
2. if 𝑎, 𝑏 ∈ 𝑅, 𝑏 ≠ 0, then there exist 𝑞, 𝑟 ∈ 𝑅 such that 𝑎 = 𝑏𝑞 + 𝑟, where either 𝑟 = 0 or
𝑑 𝑟 < 𝑑(𝑏).

Example 4.2: We see that < 𝑍 , + , . >, the ring of integers is a Euclidean domain.
For any 0 ≠ 𝑎 ∈ ℤ , we define 𝑑 𝑎 = 𝑎 ( the absolute value of 𝑎).
Then 𝑑(𝑎) > 0.

Let 𝑎, 𝑏 ∈ ℤ such that 𝑎 ≠ 0, 𝑏 ≠ 0.

Then 𝑑 𝑎 = 𝑎 , 𝑑 𝑎𝑏 = 𝑎𝑏 = 𝑎 𝑏 .
As 𝑎 ≤ 𝑎 𝑏 = 𝑎𝑏 .
⇒ 𝑑 𝑎 ≤ 𝑑 𝑎𝑏 , for every 0 ≠ 𝑎, 0 ≠ 𝑏 ∈ ℤ.

Again, let 𝑎, 𝑏 ∈ ℤ such that 𝑏 ≠ 0.

Case1: If 𝑏 > 0.
By division algorithm in ℤ, there exists 𝑡, 𝑟 ∈ ℤ such that
𝑎 = 𝑏𝑡 + 𝑟 where 0 ≤ 𝑟 < 𝑏.
If 𝑟 = 0, then we are done.
If 𝑟 ≠ 0 , then as 0 < 𝑟 < 𝑏.
⇒ 𝑟 < 𝑏 .
⇒ 𝑑(𝑟) < 𝑑(𝑏).

Case2: If 𝑏 < 0, then −𝑏 > 0.

Applying division algorithm to 𝑎 and −𝑏, we obtain

19 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

𝑎 = −𝑏 𝑡 + 𝑟, where 0 ≤ 𝑟 < (−𝑏) and 𝑟, 𝑡 ∈ ℤ.

⇒ 𝑎 = −𝑡 𝑏 + 𝑟, −𝑡, 𝑟 ∈ ℤ.

If 𝑟 = 0, then we are done.

If 𝑟 ≠ 0, as 0 < 𝑟 < (−𝑏).
⇒ 𝑟 < −𝑏 = 𝑏 .
⇒ 𝑑(𝑟) < 𝑑(𝑏).

Hence, < 𝑍 , + , . > is a Euclidean domain.

Example 4.3: The ring of Gaussian integers given by

ℤ 𝑖 = 𝑎 + 𝑖𝑏 ∶ 𝑎, 𝑏 ∈ ℤ .
is an Euclidean domain.
Solution: We know that ℤ 𝑖 is an integral domain.
For 0 ≠ 𝑎 + 𝑖𝑏 ∈ ℤ 𝑖 , define 𝑑 𝑎 + 𝑖𝑏 = 𝑎2 + 𝑏 2 .

Then 𝑑(𝑎 + 𝑖𝑏) is a non-negative integer because if 0 ≠ 𝑎 + 𝑖𝑏 ∈ ℤ 𝑖 then either 0 ≠ 𝑎 or

0 ≠ 𝑏. In any case 𝑎2 + 𝑏 2 > 0.

(i) Let 0 ≠ 𝑥 = 𝑎 + 𝑖𝑏 , 0 ≠ 𝑦 = 𝑐 + 𝑖𝑑 ∈ ℤ 𝑖 be arbitrary.

Then 𝑑 𝑥𝑦 = 𝑑( 𝑎 + 𝑖𝑏 𝑐 + 𝑖𝑑 ).
= 𝑑 𝑎𝑐 − 𝑏𝑑 + 𝑖 𝑎𝑑 + 𝑏𝑐 .
2
= 𝑎𝑐 − 𝑏𝑑 + 𝑎𝑑 + 𝑏𝑐 2 .
= (𝑎2 + 𝑏 2 )(𝑐 2 + 𝑑 2 ).
= 𝑑 𝑥 𝑑(𝑦) …(1)

As 𝑦 ≠ 0 ⇒ 𝑑(𝑦) > 0 ⇒ 𝑑(𝑦) ≥ 1 (as 𝑑(𝑦) is an integer).

⇒ 𝑑 𝑥 𝑑(𝑦) ≥ 𝑑(𝑥).
Therefore, 𝑑 𝑥 ≤ 𝑑 𝑥 𝑑 𝑦 = 𝑑(𝑥𝑦).

(ii) Step1: Let 𝑥 ∈ ℤ 𝑖 be a positive integer.

Then 𝑥 = 𝑛 + 𝑖0, where 𝑛 ∈ ℤ+ .
Let 0 ≠ 𝑦 = 𝑎 + 𝑖𝑏 ∈ ℤ 𝑖 be arbitrary.
We show that there exists 𝑡, 𝑟 ∈ ℤ 𝑖 such that 𝑦 = 𝑛𝑡 + 𝑟 where 𝑟 = 0 or 𝑑(𝑟) < 𝑑(𝑥).
Applying Division algorithm to integers 𝑎, 𝑛 and 𝑏, 𝑛.
We obtain,
𝑎 = 𝑢𝑛 + 𝑟1 , 0 ≤ 𝑟1 < 𝑛 where 𝑟1 ∈ ℤ and
𝑏 = 𝑣𝑛 + 𝑟2 , 0 ≤ 𝑟2 < 𝑛 where 𝑟2 ∈ ℤ .
As 𝑟1 and 𝑛/2 are two real numbers, therefore by law of trichotomy either
𝑟1 ≤ 𝑛/2 or 𝑟1 > 𝑛/2 .

20 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

If 𝑟1 > 𝑛/2.
Then −𝑟1 < −𝑛/2.
⇒ 𝑛 − 𝑟1 < 𝑛 − 𝑛/2 .
⇒ 𝑛 − 𝑟1 < 𝑛/2.

As 𝑎 = 𝑢𝑛 + 𝑟1 .
⇒ 𝑎 = 𝑢𝑛 + 𝑛 − 𝑛 + 𝑎 + 𝑟1 .
= 𝑛 𝑢 + 1 − (𝑛 − 𝑟1 ).
= 𝑛𝑞 + 𝑘1 , where 𝑞 = 𝑛(𝑢 + 1) and 𝑘1 = −(𝑛 − 𝑟1 ).

We note that 𝑟1 < 𝑛 which implies that 𝑛 − 𝑟1 > 0.

⇒ −(𝑛 − 𝑟1 ) = 𝑛 − 𝑟1 < 𝑛/2.

If 𝑟1 ≤ 𝑛/2 .
Then 𝑎 = 𝑢𝑛 + 𝑟1 where 𝑟1 ≤ 𝑛/2.

Thus in either case, we have 𝑎 = 𝑢𝑛 + 𝑘1 , where 𝑘1 < 𝑛/2.

Similarly, for 𝑏 = 𝑣𝑛 + 𝑟2 we can find an integer 𝑘2 such that 𝑏 = 𝑣𝑛 + 𝑘2 , where 𝑘2 < 𝑛/2.

Therefore, 𝑦 = 𝑎 + 𝑖𝑏
= 𝑢𝑛 + 𝑘1 + 𝑖(𝑣𝑛 + 𝑘2 ).
= 𝑡𝑛 + 𝑟, where 𝑡 = 𝑢 + 𝑖𝑣 and 𝑟 = 𝑘1 + 𝑖𝑘2 ∈ ℤ[𝑖].

As 𝑘1 and 𝑘2 can be zero also. It follows that either 𝑟 = 0 or

𝑑 𝑟 = 𝑑 𝑘1 + 𝑖𝑘2 .
= 𝑘12 + 𝑘22 .
≤ 𝑛2 /4 + 𝑛2 /4.
= 𝑛2 /2.
< 𝑛 2 + 02 .
= 𝑑(𝑛 + 𝑖0) .
= 𝑑(𝑥).

Thus, we obtain that 𝑦 = 𝑛𝑡 + 𝑟, where either 𝑟 = 0 or 𝑑(𝑟) < 𝑑(𝑥).

Step2: Let 0 ≠ 𝑥, 𝑦 ∈ ℤ[𝑖].

Let 𝑥 denotes the conjugate of 𝑥.

Then 𝑥𝑥 is a positive integer.
Let 𝑥 = 𝑛, 𝑛 ∈ ℤ+ .

As 𝑦𝑥 ∈ ℤ[𝑖] and 𝑛 is a positive integer. Applying step1 to 𝑦𝑥 and 𝑛, there exists 𝑡, 𝑟 ∈ ℤ[𝑖]
such that

21 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

𝑦𝑥 = 𝑡𝑛 + 𝑟 where either 𝑟 = 0 or 𝑑(𝑟) < 𝑑(𝑛) …(2)

If 𝑟 = 0, then 𝑦𝑥 = 𝑡𝑛 = 𝑡𝑥𝑥.
⇒ 𝑦 = 𝑡𝑥 + 0 (as 𝑥 ≠ 0 ⇒ 𝑥 ≠ 0 and cancellation law holds in an integral domain).

If 𝑑(𝑟) < 𝑑(𝑛).

From (2), we have 𝑟 = 𝑦𝑥 − 𝑡𝑛.
⇒ 𝑑 𝑦𝑥 − 𝑡𝑛 < 𝑑 𝑥𝑥 .
⇒ 𝑑 𝑦𝑥 − 𝑡𝑥𝑥 < 𝑑 𝑥 𝑑(𝑥 ) (using (1)).
⇒ 𝑑(𝑥 (𝑦 − 𝑡𝑥)) < 𝑑 𝑥 𝑑(𝑥 ) .
⇒ 𝑑 𝑥 𝑑(𝑦 − 𝑡𝑥) < 𝑑 𝑥 𝑑(𝑥 ) .
⇒ 𝑑(𝑦 − 𝑡𝑥) < 𝑑 𝑥 .

Let −𝑡𝑥 = 𝑟0 , then 𝑑(𝑟0 ) < 𝑑(𝑥).

Thus, 𝑦 = 𝑡𝑥 + 𝑟0 , where either 𝑟0 = 0 or 𝑑(𝑟0 ) < 𝑑(𝑥).

Hence, ℤ[𝑖] is a Euclidean Domain. ■

Theorem 4.4: If 𝐹 is a field, then 𝐹[𝑥] is a Euclidean domain.

Proof: Let 𝐹 be a field.
For 0 ≠ 𝑓(𝑥) ∈ 𝐹[𝑥], define 𝑑 𝑓 𝑥 = deg 𝑓(𝑥)
Then 𝑑(𝑓 𝑥 ) is a non-negative integer.

As 𝐹 is a field, therefore 𝐹[𝑥] is an integral domain with unity.

(i) For 0 ≠ 𝑓(𝑥), 0 ≠ 𝑔(𝑥) ∈ 𝐹[𝑥] be arbitrary.

deg 𝑓 𝑥 𝑔 𝑥 = deg 𝑓 𝑥 + deg 𝑔(𝑥)

⇒ deg 𝑓(𝑥) < deg 𝑓 𝑥 𝑔 𝑥 − deg 𝑔(𝑥)
≤ deg 𝑓 𝑥 𝑔 𝑥 .

⇒ 𝑑(𝑓 𝑥 ) ≤ 𝑑(𝑓 𝑥 𝑔 𝑥 ).

(ii) Let 𝑓(𝑥), 𝑔(𝑥) ∈ 𝐹[𝑥] and 𝑔(𝑥) ≠ 0.

Applying Division algorithm to 𝑓 𝑥 , 𝑔(𝑥) in 𝐹[𝑥], there exists unique polynomials 𝑞(𝑥),
𝑟(𝑥) ∈ 𝐹[𝑥] such that
𝑓 𝑥 = 𝑔 𝑥 𝑞 𝑥 + 𝑟(𝑥), where either 𝑟 𝑥 = 0 or deg 𝑟 𝑥 < deg 𝑔(𝑥).
⇒ 𝑓 𝑥 = 𝑔 𝑥 𝑞 𝑥 + 𝑟(𝑥), where either 𝑟 𝑥 = 0 or 𝑑 𝑓 𝑥 < 𝑑(𝑔 𝑥 ).
Hence, 𝐹[𝑥] is a Euclidean domain. ■

Theorem 4.5: Every Euclidean domain is a principal ideal domain.

Proof: Let 𝑅 be an Euclidean domain.

22 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Then, 𝑅 is an integral domain.

Let I be any ideal of 𝑅.
Claim: 𝐼 is generated by an element of 𝑅.

If 𝐼 = {0}, then we are done.

Assume that 𝐼 ≠ {0}, then there exists at least one 0 ≠ 𝑦 ∈ 𝐼.

Let 0 ≠ 𝑎 ∈ 𝐼 be such that 𝑑 𝑎 < 𝑑 𝑥 , for every 0 ≠ 𝑥 ∈ 𝐼.

Claim: 𝐼 = < 𝑎 >.

Let 0 ≠ 𝑏 ∈ 𝐼 be arbitrary.
Then, by definition of Euclidean domain there exists 𝑡, 𝑟 ∈ 𝑅 such that
𝑏 = 𝑎𝑡 + 𝑟, where either 𝑟 = 0 or 𝑑(𝑟) < 𝑑(𝑎).

If 𝑟 ≠ 0, then as ∈ 𝐼 , 𝑡 ∈ 𝑅 and 𝐼 is an ideal of 𝑅 ⇒ 𝑎𝑡 ∈ 𝐼.

As 𝑏∈𝐼 ⇒ 𝑏 − 𝑎𝑡 ∈ 𝐼
⇒ 𝑟 ∈ 𝐼 where 𝑑(𝑟) < 𝑑(𝑎) which contradicts the choice of 𝑎.
Therefore, 𝑟 must be zero i.e. 𝑟 = 0.

Hence, 𝑏 = 𝑎𝑡 ∈ 𝐼 for every 0 ≠ 𝑏 ∈ 𝐼 ⇒ 𝐼 ⊆ < 𝑎 >.

If 𝑥 ∈ < 𝑎 >, then 𝑥 = 𝑎𝑘 for some 𝑘 ∈ 𝑅.

As 𝑎 ∈ 𝐼 and 𝑘 ∈ 𝑅 ⇒ 𝑎𝑘 ∈ 𝐼 ⇒ 𝑥 ∈ 𝐼.
∴ < 𝑎 > ⊆ 𝐼.
⇒ 𝐼 = < 𝑎 >.

Since 𝐼 was any ideal of 𝑅, it follows that every ideal of 𝑅 is generated by some element
of 𝑅.

Also, note that 𝑅 is an ideal of itself.

Therefore, it follows that 𝑅 = < 𝑟′ >, for some 𝑟′ ∈ 𝑅.
As 𝑟 ′ ∈ 𝑅 = < 𝑟 ′ >, it follows that 𝑟 ′ = 𝑟′𝑘 where 𝑘 ∈ 𝑅.
Let 𝑐 ∈ 𝑅 (any).
Then 𝑐 = 𝑟′𝑥 for some 𝑥 ∈ 𝑅.
⇒ 𝑐𝑘 = (𝑟 ′ 𝑥)𝑘 = 𝑥𝑟 ′ 𝑘 = 𝑥 𝑟 ′ 𝑘 = 𝑥𝑟 ′ = 𝑟 ′ 𝑥 = 𝑐, for every 𝑐 ∈ 𝑅.
⇒ 𝑘 is the unity of 𝑅.

Thus, 𝑅 is a principal ideal domain. ■

Theorem 4.6: Every Euclidean domain is a unique factorization domain.

Proof: Let 𝑅 be a Euclidean domain.
Then by Theorem 4.5, 𝑅 is a principal ideal domain.

23 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

Since every Principal ideal domain is a unique factorization domain, it follows that 𝑅 is a
unique factorization domain. ■

Remark 4.7: From the above theorems, we can summarize that

E.D ⇒ P.I.D ⇒ U.F.D.
U.F.D ⇏ P.I.D ⇏ E.D.

Example 4.8: Give an example of a U.F.D which is not a P.I.D.

Solution: We have seen, in I.Q. 9 of Chapter entitled Polynomial Rings, that ℤ[𝑥] is not a
principal ideal domain.

But ℤ[𝑥] is a UFD as observed in Example 3.4.

Example 4.9: Give an example of a P.I.D which is not an E.D.

1+ −19
Solution: ℤ is a PID which is not an ED. (Proof is out of scope of this course).
2

5. Exercises
Q 1. Show that −5 is a prime element in ℤ[ −5].

Q 2. Prove that 3 + 2𝑖 and 3𝑖 − 2 are associates in ℤ[𝑖].

Q 3. Find a prime element in ℤ6 the ring of integers modulo 6 which is not irreducible?

Q 4. Prove that if 𝑝 and 𝑞 are prime elements in an integral domain 𝑅 with unity such
that 𝑞/𝑝 then 𝑝 and 𝑞 are associates.

Q 5. Prove that 7 is an irreducible element in ℤ[ 5].

Q 6. Show that ℤ[ 3] is an Euclidean domain.

Q 7. Define unique factorization domain. Prove that ℤ[ 2] is a U.F.D.

Q 8. Prove that ℤ[ −6] is not a U.F.D.

Q 9. Give an example of a polynomial which is primitive as well as irreducible.

Q 10. Prove that ℤ5 [𝑥] is a U.F.D.

Q 11. Prove that every P.I.D is a U.F.D. but converse is not true.

Q 12. Determine whether the following elements in ℤ[𝑖] are irreducible or not

24 | P a g e
Institute of Lifelong Learning, University of Delhi
 Divisibility in Integral Domains

(i) 3 + 2𝑖
(ii) 2 + 𝑖

Q 13. Determine the units in ℤ[𝑖].

Q 14. Prove that if 𝑎 and 𝑏 are associates in ℤ[ 𝑑], then 𝑁 𝑎 = 𝑁(𝑏).

Q 15. Prove that in ℤ[ 10], 2 is not a prime element.

Q 16. Prove that ℤ[ −5] is neither a unique factorization domain nor a Euclidean
domain.

Q 17. Show that ℤ 3 = {𝑎 + 𝑏 3 ∶ 𝑎 , 𝑏 ∈ ℤ } is a Euclidean domain.

6. References
1. Joseph A. Gallian, Contemporary Abstract Algebra (4th Edition), Narosa
Publishing House.
2. David S. Dummit and Richard M. Foote, Abstract Algebra (3rd Edition),
John Wiley & Sons, Inc.

7. Suggested Readings
1. John B. Fraleigh, A First Course in Abstract Algebra (7th Edition), Pearson.
2. Joseph A. Gallian, Contemporary Abstract Algebra (7th Edition), Narosa
Publishing House.
3. Serge Lang, Algebra, (3rd Edition), Graduate Text in Mathematics,
Springer.

25 | P a g e
Institute of Lifelong Learning, University of Delhi