10 Irreducible Polynomials Copyright © 2023, Gregory G.

Smith
Last Updated: 31 March 2023

Although the irreducibility depends on the coefficients, irre‑

ducible polynomials are much like positive prime integers. In
some ways, they are even simpler.

10.0 Factoring polynomials

When are polynomial rings unique factorization domains? To
answer this question, we need an auxiliary invariant.

Definition 10.0.0. Let 𝑅 be a unique factorization domain and

consider a polynomial 𝑓 ∶= 𝑎𝑚 𝑥𝑚 + 𝑎𝑚−1 𝑥𝑚−1 + ⋯ + 𝑎1 𝑥 + 𝑎0 in
To define the content, we need to

𝑅[𝑥]. The content of the polynomial 𝑓 is defined to be

know that greatest common divisors
exist. The greatest common divisor, if

cont(𝑓) ∶= gcd(𝑎𝑚 , 𝑎𝑚−1 , … , 𝑎0 ) .

it exists, is unique only up to multipli‑
cation by a unit. Hence, the content of

The polynomial 𝑓 is primitive if cont(𝑓) = 1.

a polynomial is an equivalence class.

Lemma 10.0.1 (Gauss). Let 𝑅 be a unique factorization domain. For

any two polynomials 𝑓 and 𝑔 in 𝑅[𝑥], we have
cont(𝑓 𝑔) = cont(𝑓) cont(𝑔) .
In particular, when 𝑓 and 𝑔 are primitive, the product 𝑓 𝑔 also is.

Proof. We write 𝑓 = cont(𝑓) 𝑓ˆ and 𝑔 = cont(𝑔) 𝑔 ˆ where 𝑓ˆ and 𝑔 ˆ

are primitive polynomials in 𝑅[𝑥]. As 𝑓 𝑔 = cont(𝑓) cont(𝑔) 𝑓 𝑔 ˆ ˆ,
ˆ ˆ
it suffices to verify that the product 𝑓 𝑔 is a primitive polynomial.
Let 𝑓ˆ = 𝑎0 + 𝑎1 𝑥 + ⋯ + 𝑎𝑚 𝑥𝑚 and 𝑔 ˆ = 𝑏0 + 𝑏1 𝑥 + ⋯ + 𝑏𝑛 𝑥𝑛 for
some 𝑎0 , 𝑎1 , … , 𝑎𝑚 , 𝑏0 , 𝑏1 , … , 𝑏𝑛 in 𝑅. Suppose that the coefficients
of 𝑓ˆ 𝑔
ˆ have a common divisor 𝑑 which is not a unit. If the element
𝑝 in 𝑅 were an irreducible divisor of 𝑑, then 𝑝 must divide all the
coefficients of 𝑓ˆ 𝑔ˆ. Since 𝑓ˆ and 𝑔 ˆ are primitive, 𝑝 does not divide
all the coefficients of 𝑓ˆ or 𝑔 ˆ. Let 𝑎𝑗 be the first coefficient of 𝑓ˆ not
divisible by 𝑝 and let 𝑏𝑘 be the first coefficient of 𝑔 ˆ not divisible by
𝑝. Consider the coefficient of 𝑥𝑗+𝑘 in 𝑓ˆ 𝑔 ˆ; it has the form
𝑎𝑗 𝑏𝑘 + (𝑎𝑗+1 𝑏𝑘−1 + 𝑎𝑗+2 𝑏𝑘−2 + ⋯ ) + (𝑎𝑗−1 𝑏𝑘+1 + 𝑎𝑗−2 𝑏𝑘+2 + ⋯ ) .
By hypothesis, 𝑝 divides this sum. Moreover, all the terms in the
first parenthesis are divisible by 𝑝 (because 𝑝 divides 𝑏𝑖 for all
𝑖 < 𝑗) and all terms in the second parenthesis are divisible by
𝑝 (because 𝑝 divides 𝑎𝑖 for all 𝑖 < 𝑘). It follows that 𝑝 divides
𝑎𝑗 𝑏𝑘 . Since ⟨𝑝⟩ is prime ideal, the element 𝑝 divides either 𝑎𝑗 or
𝑏𝑘 contrary to our choice of 𝑎𝑗 and 𝑏𝑘 . This contradiction shows
that no irreducible element divides all the coefficients of 𝑓ˆ 𝑔ˆ and,
therefore, the product 𝑓 𝑔ˆ ˆ is primitive.

Replacing the coefficient domain by its fields of fraction does

not alter irreduciblity.
94 Rings First Copyright © 2023 by Gregory G. Smith

Lemma 10.0.2. Let 𝑅 be a unique factorization domain and let 𝐾 be its

⦁ For any nonzero polynomial 𝑓 in the ring 𝐾[𝑥], we have 𝑓 = 𝑐 𝑓ˆ

field of fractions.

where 𝑐 ∈ 𝐾 and 𝑓ˆ is a primitive polynomial in 𝑅[𝑥]. Moreover, this

factorization is unique up to multiplication by in unit of 𝑅.
⦁ Let 𝑓 be a polynomial in 𝑅[𝑥] having positive degree. When 𝑓 is
irreducible in 𝑅[𝑥], the polynomial 𝑓 is also irreducible in 𝐾[𝑥].

Proof. Finding a common denominator 𝑑 for the coefficients of

the polynomial 𝑓, we obtain 𝑓 = ( 𝑑1 ) 𝑓˜ where 𝑓˜ is a polynomial in
˜ it follows that 𝑓 = 𝑐 𝑓ˆ where 𝑓ˆ is a
𝑅[𝑥]. Setting 𝑐 ∶= 𝑑1 cont(𝑓),
primitive polynomial in 𝑅[𝑥]. Suppose that 𝑓 = ( 𝑎𝑏 ) 𝑔 for some the
fraction 𝑎𝑏 in 𝐾 and some primitive polynomial 𝑔 in 𝑅[𝑥]. It follows
that 𝑎 𝑑 𝑔 = 𝑏 cont(𝑓) 𝑓.ˆ Taking the content of both sides yields
𝑢 𝑎 𝑑 = 𝑏 cont(𝑓) for some unit 𝑢 in 𝑅. We deduce that 𝑢 𝑔 = 𝑓. ˆ
Since cont(𝑓) divides 𝑓, the polynomial 𝑓 is primitive in 𝑅[𝑥].
Suppose that 𝑓 is reducible in 𝐾[𝑥]. It follows that 𝑓 = 𝑔1 𝑔2
for some polynomials 𝑔1 and 𝑔2 in 𝐾[𝑥] having positive degree.
The first part implies that, for any index 𝑗, we have 𝑔𝑗 = 𝑐𝑗 ℎ𝑗 for
some 𝑐𝑗 ∈ 𝐾 and some primitive polynomial ℎ𝑗 in 𝑅[𝑥]. Hence,
𝑓 = 𝑐1 𝑐2 ℎ1 ℎ2 and the product ℎ1 ℎ2 is primitive by Lemma 10.0.1.
The first part implies 𝑓 and ℎ1 ℎ2 differ up to multiplication by a
unit of 𝑅, which contradicts the irreducibility of 𝑓 in 𝑅[𝑥].

Theorem 10.0.3. For any unique factorization domain 𝑅, the

polynomial ring 𝑅[𝑥] is also a unique factorization domain.

Proof. Let 𝐾 be the field of fractions for the domain 𝑅. Consider

a nonzero polynomial 𝑓 in the ring 𝑅[𝑥]. As 𝐾[𝑥] is a principal

domain. Hence, we can write 𝑓 = 𝑝1 𝑝2 ⋯ 𝑝𝑟 where each 𝑝𝑗 is an

ideal domain, Corollary 9.1.5 shows that it a unique factorization

irreducible polynomial in 𝐾[𝑥]. Lemma 10.0.2 implies that, for all

1 ⩽ 𝑖 ⩽ 𝑟, we have 𝑝𝑗 = 𝑐𝑗 𝑞𝑗 for some 𝑐𝑗 ∈ 𝐾 and some primitive
polynomial 𝑞𝑗 in 𝑅[𝑥]. Thus, we deduce that 𝑓 = 𝑐 𝑞1 𝑞2 ⋯ 𝑞𝑟
where 𝑐 = ∏𝑗 𝑐𝑗 ∈ 𝐾. Write 𝑐 = 𝑎𝑏 for some elements 𝑎 and 𝑏 in
𝑅. Taking contents, we obtain cont(𝑏 𝑓) = cont(𝑎 𝑞1 𝑞2 ⋯ 𝑞𝑟 ) = 𝑎
by Lemma 10.0.1. We deduce that 𝑏 cont(𝑓) = 𝑎, so 𝑏 divides 𝑎
and cont(𝑓) = 𝑐 lies in 𝑅. Since each 𝑞𝑗 is irreducible in 𝐾[𝑥], it
is irreducible in 𝑅[𝑥]. The ring 𝑅 is a unique factorization domain,
so we have 𝑐 = 𝑢 𝑑1 𝑑2 ⋯ 𝑑𝑠 where each 𝑑𝑖 is irreducible in 𝑅 and
𝑢 in 𝑅 is a unit. It follows that 𝑓 = 𝑢 𝑑1 𝑑2 ⋯ 𝑑𝑠 𝑞1 𝑞2 ⋯ 𝑞𝑟 is a
factorization of 𝑓 into a product of irreducible elements in 𝑅[𝑥].

factorization: 𝑓 = 𝑢′ 𝑑1′ 𝑑2′ ⋯ 𝑑𝑡′ 𝑞′1 𝑞′2 ⋯ 𝑞′𝑘 where each 𝑞𝑗′ is prim‑
It remains to check uniqueness. Suppose that we have a second

itive polynomial in 𝑅[𝑥] and 𝑑𝑗′ is irreducible element in 𝑅. Since

this is also a factorization in 𝐾[𝑥], it is unique, so 𝑟 = 𝑘 and 𝑞𝑗′ = 𝑞𝑗

unit in 𝐾[𝑥], then they also differ by a unit in 𝑅[𝑥]. Furthermore,

(up to units and reordering). If primitive polynomials differ by a
Copyright © 2023 by Gregory G. Smith Irreducible Polynomials 95

we have cont(𝑓) = 𝑢′ 𝑑1′ 𝑑2′ ⋯ 𝑑𝑡′ = 𝑢 𝑑1 𝑑2 ⋯ 𝑑𝑠 so 𝑠 = 𝑡 and

𝑑𝑗′ = 𝑑𝑗 (up to units and reordering).

Example 10.0.4. The ring ℤ[𝑥] is a unique factorization domain,

but not a principal ideal domain.

Corollary 10.0.5. For any nonnegative integer 𝑛 and any unique factor‑
ization domain 𝑅, the polynomial ring 𝑅[𝑥1 , 𝑥2 , … , 𝑥𝑛 ] is also a unique
factorization domain.

Proof. We proceed by induction on 𝑛. When 𝑛 = 0, the assertion

is trivial. Since Theorem 10.0.3 establishes the induction step, the
claim follows.

Exercises

prime integers in the following way: if 𝑝1 , 𝑝2 , … , 𝑝𝑘 are positive

Problem 10.0.6. Euclid proves that there are infinitely many

prime integers, then any prime factor of 1 + 𝑝1 𝑝2 ⋯ 𝑝𝑘 must be

different from 𝑝𝑗 for any 1 ⩽ 𝑗 ⩽ 𝑘.

the form 4 𝑛 − 1 is infinite.

(i) Adapt this argument to show that the set of prime integers of

(ii) Adapt this argument to show that, for any field 𝕂, there are
infinitely many monic irreducible polynomials in 𝕂[𝑥].

Problem 10.0.7. Let 𝑅 be a principal ideal domain and let 𝐾 be its

(i) Suppose 𝑅 = ℤ. Write 𝑟 = 24 ∈ ℚ in the form 𝑟 = 𝑏3 + 𝑎8 for

field of fractions.
7

some integers 𝑎 and 𝑏.

(ii) Let 𝑔 ∶= 𝑝 𝑞 ∈ 𝑅 where 𝑝 and 𝑞 are coprime. Prove that every
fraction 𝑓/𝑔 ∈ 𝐾 can written in the form
𝑓 𝑢 𝑣
= +
𝑔 𝑞 𝑝
for some elements 𝑢 and 𝑣 in 𝑅.
(iii) Let 𝑔 ∶= 𝑝11 𝑝22 ⋯ 𝑝𝑚𝑚 ∈ 𝑅 be the factorization of 𝑔 into
𝑒 𝑒 𝑒

irreducible elements 𝑝𝑗 , for all 1 ⩽ 𝑗 ⩽ 𝑚, such that the

relation 𝑝𝑗 = 𝑢 𝑝𝑘 for some unit 𝑢 ∈ 𝑅 implies that 𝑗 = 𝑘.
Prove that every fraction 𝑓/𝑔 ∈ 𝐾 can be written in the form

𝑓
𝑘
ℎ𝑗
= ∑ 𝑒𝑗
𝑔 𝑗=1 𝑝𝑗

for some elements ℎ1 , ℎ2 , … , ℎ𝑚 in 𝑅.

10.1 Irreducibility Criteria

Can we identify irreducible polynomials? In some situations, this
can be relatively easy.

Problem 10.1.0. Is 𝑓(𝑥) = 𝑥3 + 6𝑥2 + 7 in ℤ[𝑥] irreducible?

96 Rings First Copyright © 2023 by Gregory G. Smith

Solution. Yes. Otherwise 𝑓 would have linear factor and its root
would divide 7. However, we have 𝑓(1) = 14, 𝑓(−1) = 12, 𝑓(7) > 0,
and 𝑓(−7) = (−1)(49) + 7 < 0.

Proposition 10.1.1. Let 𝑓 = 𝑎𝑚 𝑥𝑚 + ⋯ + 𝑎1 𝑥 + 𝑎0 be a polynomial

in the ring 𝑅[𝑥] and let ⟨𝑝⟩ be a prime ideal in 𝑅 that does not contain
𝑎𝑚 . When the image of 𝑓 in (𝑅/⟨𝑝⟩)[𝑥] is irreducible, the polynomial 𝑓
is irreducible in 𝑅[𝑥].

Proof. The canonical surjection 𝜋∶ 𝑅 → 𝑅 / ⟨𝑝⟩ induces a ring

homomorphism 𝜑∶ 𝑅[𝑥] → 𝑅/ ⟨𝑝⟩ [𝑥](𝑅 / ⟨𝑝⟩)[𝑥]. When 𝑓 = 𝑔 ℎ in
the ring 𝑅, we obtain 𝜑(𝑓) = 𝜑(𝑔) 𝜑(ℎ). The assumption that the
element 𝑝 does not divide 𝑎𝑚 implies that deg(𝜑(𝑔)) = deg(𝑔) and
deg(𝜑(ℎ)) = deg(ℎ). Therefore, reducibility of the polynomial 𝑓 in
𝑅[𝑥] implies the reducibility of the image 𝜑(𝑓) in (𝑅/⟨𝑝⟩)[𝑥].

Problem 10.1.2. Is 𝑥4 + 15𝑥3 + 7 in ℚ[𝑥] irreducible?

Solution. The image of this polynomial in 𝔽5 [𝑥] is 𝑥4 + 2. Since

𝑥4 ≡ 0, 1 (mod 5), we see that 𝑥4 + 2 has no root in 𝔽5 . Suppose that
𝑥4 + 2 = (𝑥2 + 𝑎 𝑥 + 𝑏)(𝑥2 + 𝑐 𝑥 + 𝑑). It follows that 𝑎 + 𝑐 = 0,
𝑎𝑐 + 𝑏 + 𝑑 = 0, 𝑎𝑑 + 𝑏𝑐 = 0, and 𝑏𝑑 = 2. Since 𝑐 = −𝑎, we have
0 = 𝑎𝑑 + 𝑏𝑐 = 𝑎(𝑑 − 𝑏), so 𝑎 = 0 or 𝑑 = 𝑏.
⦁ Suppose that 𝑎 = 0. We have 𝑐 = 0. The equations 𝑏 + 𝑑 = 0
and 𝑏𝑑 = 2 imply that 𝑑 = −𝑏, −𝑏2 = 2, and 𝑏2 = 3. However,
02 = 0, 12 = 1, 22 = 4, 32 = 4, and 42 = 1. Hence, there is no
element 𝑏 ∈ 𝔽5 such that 𝑏2 = 3.
⦁ Suppose that 𝑏 = 𝑑. We have 𝑏2 = 2. This is again impossible
because the only perfect squares in 𝔽5 are 0, 1, and 4.
We see that the polynomial 𝑥4 + 2 is irreducible in 𝔽5 [𝑥]. Thus,
Proposition 10.1.1 shows that 𝑥4 + 15 𝑥3 + 7 is irreducible in ℤ[𝑥]
and Lemma 10.0.1 shows that it is irreducible in ℚ[𝑥].

Theorem 10.1.3 (Eisenstein Criterion). Let 𝑅 a commutative domain

and let 𝑓 ∶= 𝑎0 + 𝑎1 𝑥 + ⋯ + 𝑎𝑚 𝑥𝑚 be a primitive polynomial in 𝑅[𝑥]
of positive degree 𝑛. When there exists a prime ideal 𝑃 in 𝑅 such that
Theodor Schönemann first published
a version of this criterion in 1846.

⦁ 𝑎𝑚 ∉ 𝑃,
Gotthold Eisenstein published a

⦁ 𝑎0 , 𝑎1 , … , 𝑎𝑚−1 ∈ 𝑃, and
somewhat different version in the

⦁ 𝑎0 ∉ 𝑃 2 ,
same journal in 1850.

the polynomial 𝑓 is irreducible in 𝑅[𝑥].

Proof. Suppose that 𝑓 = 𝑔 ℎ for some polynomials 𝑔 and ℎ in

𝑅[𝑥] having positive degree. Set 𝑔 ∶= 𝑏0 + 𝑏1 𝑥 + ⋯ + 𝑏𝑗 𝑥𝑗 and
ℎ ∶= 𝑐0 + 𝑐1 𝑥 + ⋯ + 𝑐𝑘 𝑥𝑘 where deg(𝑔) = 𝑗 and deg(ℎ) = 𝑘. It
follows that 𝑎0 = 𝑏0 𝑐0 belongs to the ideal 𝑃. Since 𝑃 is a prime
ideal, we have 𝑏0 ∈ 𝑃 or 𝑐0 ∈ 𝑃. Having both 𝑏0 and 𝑐0 belong to 𝑃
would imply that 𝑎0 ∈ 𝑃 2 contradicting our hypotheses. Without
loss of generality, we may assume that 𝑏0 ∈ 𝑃 and 𝑐0 ∉ 𝑃. If
every coefficient of 𝑔 were in 𝑃, then every coefficient of 𝑓 would
Copyright © 2023 by Gregory G. Smith Irreducible Polynomials 97

also be in 𝑃 again contradicting our hypothesis. Let 𝑏𝑖 be the first

coefficient of 𝑔 such that 𝑏𝑖 ∉ 𝑃. Since
𝑎𝑖 = 𝑏𝑖 𝑐0 + 𝑏𝑖−1 𝑐1 + ⋯ + 𝑏0 𝑐𝑖 ,
we obtain the equation 𝑏𝑖 𝑐0 = 𝑎𝑖 − 𝑏𝑖−1 𝑐1 − ⋯ − 𝑏0 𝑐𝑖 . Every
element on the right side of this equation lies in 𝑃. However, this
implies that 𝑏𝑖 𝑐0 ∈ 𝑃. Because 𝑃 is a prime ideal, we deduce that
either 𝑏𝑖 ∈ 𝑃 or 𝑐0 ∈ 𝑃 which is a contradiction.

We record the following special case.

Corollary 10.1.4. Let 𝑅 be a unique factorization domain with fraction
field 𝐾 and consider 𝑓 ∶= 𝑎0 + 𝑎1 𝑥 + ⋯ + 𝑎𝑚 𝑥𝑚 in the ring 𝑅[𝑥]. When
there exists an irreducible element 𝑝 ∈ 𝑅 such that
⦁ 𝑝 does not divide 𝑎𝑚 ,
⦁ 𝑝 divides 𝑎𝑖 for all 0 ⩽ 𝑖 ⩽ 𝑚 − 1, and
⦁ 𝑝2 does not divide 𝑎0 ,
the polynomial 𝑓 is irreducible in 𝐾[𝑥].
Proof. Theorem 10.1.3 shows that the polynomial 𝑓 is irreducible
in 𝑅[𝑥] and Lemma 10.0.1 shows that 𝑓 is irreducible in 𝐾[𝑥].

Problem 10.1.5. Is 𝑥5 − 6 𝑥4 + 3 ∈ ℚ[𝑥] irreducible?

Solution. Yes, apply Corollary 10.1.4 with 𝑝 = 3.

Corollary 10.1.6. For any positive prime integer 𝑝, the polynomial

𝑓 ∶= 𝑥𝑝−1 + 𝑥𝑝−2 + ⋯ + 𝑥 + 1
is irreducible in ℚ[𝑥].
Proof. Since (𝑥 − 1) 𝑓(𝑥) = 𝑥𝑝 − 1, the ring isomorphism given by
𝑥 ↦ 𝑦 + 1 yields
𝑦𝑓(𝑦 + 1) = (𝑦 + 1)𝑝 − 1 = 𝑦𝑝 + ( 𝑝1 ) 𝑦𝑝−1 + ( 𝑝2 ) 𝑦𝑝−2 + ⋯ + ( 𝑝−1
𝑝
)𝑦.
We have ( 𝑝𝑖 ) = . When 𝑖 < 𝑝, the prime integer 𝑝 is
𝑝(𝑝−1)⋯(𝑝−𝑖+1)
𝑖!
not a factor of 𝑖!, so 𝑖! divides the product (𝑝 − 1)(𝑝 − 2) ⋯ (𝑝 − 𝑖 + 1)
which implies that ( 𝑝𝑖 ) is divisible by 𝑝. Dividing the expansion of
𝑦 𝑓(𝑦 + 1) by 𝑦 shows that 𝑓(𝑦 + 1) satisfies the hypothesis of

𝑦𝑝−1 + ( 𝑝1 ) 𝑦𝑝−2 + ( 𝑝2 ) 𝑦𝑝−3 + ⋯ + ( 𝑝−1 )

Corollary 10.1.4. Therefore, the polynomial
𝑝

is irreducible. We conclude that 𝑓 is irreducible.

Exercises
Problem 10.1.7. Let 𝑓 ∶= 𝑎3 𝑥3 + 𝑎2 𝑥2 + 𝑎1 𝑥 + 𝑎0 be a polynomial
in ℤ[𝑥] having degree 3. Assume that 𝑎0 , 𝑎1 + 𝑎2 , and 𝑎3 are all odd.
Prove that 𝑓 is irreducible in ℚ[𝑥].

𝑔 ∶= 𝑥5 + 6 𝑥4 − 12 𝑥3 + 9 𝑥2 − 3 𝑥 + 𝑘
Problem 10.1.8. Prove that the polynomial

in ℚ[𝑥] is irreducible for infinitely many integers 𝑘.

Problem 10.1.9. Prove that ℎ ∶= 𝑥5 + 𝑥4 + 𝑥 − 1 is irreducible in
ℚ[𝑥] using the Eisenstein criterion.
98 Rings First Copyright © 2023 by Gregory G. Smith

10.2 Counting Irreducibles

a method of determining the primes less than a given number 𝑛.

How do we count irreducible elements? The sieve of Eratosthenes is
The Greek polymath, Eratosthenes of

List the integers from 2 to 𝑛. The smallest entry 2 is prime. Cross

Cyrene (276BCE–194BCE), is famous

out the multiplies of 2 from our list. The smallest remaining entry
for his work on prime numbers and

3 is prime because it is not divisible by any smaller prime. Cross

for measuring the diameter of the
earth.

out the multiplies of 3. Repeat. Using this method, Table 10.1 list
the positive prime integers less than 100.
4 6 8 9 10
Table 10.1: The 25 positive prime
12 14 15 16 18 20
2 3 5 7
integers less than 100
21 22 24 25 26 27 28 30
11 13 17 19

32 33 34 35 36 38 39 40
23 29

42 44 45 46 48 49 50
31 37

52 54 55 56 57 58 60
41 43 47

62 63 64 65 66 68 69 70
51 53 59

72 74 75 76 77 78 80
61 67

81 82 84 85 86 87 88 90
71 73 79

91 92 93 94 95 96 98 99
83 89
97
The asymptotic distribution of the primes among the positive
integers has a famous description.

Definition 10.2.0. The prime‑counting function π∶ ℝ → ℕ counts the

number; π(𝑥) ∶= |{𝑝 ∈ ℕ | 𝑝 is a positive prime integer and 𝑝 ⩽ 𝑥}|.

number of positive prime integers less than or equal to some real

The logarithmic integral function li∶ (1, ∞) → ℝ is defined by

𝑑𝑦
𝑥
li(𝑥) ∶= ∫ .
0 ln(𝑦)
π(𝑥)
Assuming the Riemann hypothesis,

= 1. ◼ √𝑥 ln(𝑥)
one has
|π(𝑥) − li(𝑥)| < .
𝑥→∞ 8𝜋
Prime Number Theorem 10.2.1. We have lim
li(𝑥)
For any positive prime integer 𝑝, sieve methods also allows one
to identify the irreducible polynomials in 𝔽𝑝 [𝑥]. List all polyno‑

irreducible polynomials of degree at most 4 in 𝔽2 [𝑥].

mials by degree and then cross out products. Table 10.2 lists the

0 1 𝑥 𝑥+1
𝑥2 𝑥 +1
2
𝑥 +𝑥 2
𝑥 +𝑥+1
2
polynomials in 𝔽2 [𝑥] having small
Table 10.2: Irreducible

𝑥3 𝑥3 +1 𝑥3 +𝑥 𝑥3 +𝑥+1
𝑥3 +𝑥2 𝑥3 +𝑥2 +1 𝑥3 +𝑥2 +𝑥 𝑥3 +𝑥2 +𝑥+1
degree

𝑥4 𝑥4 +1 𝑥4 +𝑥 𝑥4 +𝑥+1
𝑥4 +𝑥2 𝑥4 +𝑥2 +1 𝑥4 +𝑥2 +𝑥 𝑥4 +𝑥2 +𝑥+1
𝑥4 +𝑥3 𝑥4 +𝑥3 +1 𝑥4 +𝑥3 +𝑥 𝑥4 +𝑥3 +𝑥+1
𝑥4 +𝑥3 +𝑥2 𝑥4 +𝑥3 +𝑥2 +1 𝑥4 +𝑥3 +𝑥2 +𝑥 𝑥4 +𝑥3 +𝑥2 +𝑥+1

Problem 10.2.2. Is 𝑥4 − 6𝑥3 + 12𝑥2 − 3𝑥 + 9 in ℤ[𝑥] irreducible?

Solution. This polynomial is irreducible because its image in 𝔽2 [𝑥]

is the irreducible polynomial 𝑥4 + 𝑥 + 1.
Copyright © 2023 by Gregory G. Smith Irreducible Polynomials 99

Remark 10.2.3. Since 𝑥2 + 𝑥 + 1 is irreducible in 𝔽2 [𝑥], Proposi‑

tion 9.0.11 implies that the quotient ring 𝐾 ∶= 𝔽2 [𝑥]/⟨𝑥2 + 𝑥 + 1⟩ is
a field. When 𝛼 denotes the image of 𝑥 in 𝐾, the set {1, 𝛼} forms a
basis of 𝐾 over 𝔽2 . The field 𝐾 has four elements: {0, 1, 𝛼, 1 + 𝛼}.

An analogue of the prime number theorem counts irreducible

polynomials over a finite field.

Theorem 10.2.4. Let 𝑝 be a positive prime integer. For some positive

integer 𝑒, set 𝑞 ∶= 𝑝𝑒 . Setting 𝑁𝑑 to be the number of monic irreducible
polynomials in 𝔽𝑞 [𝑥] of degree 𝑑, we have
One can even prove an analogue of
the Riemann hypothesis, namely that

𝑞𝑑 𝑞𝑑/2
∑ 𝑑 𝑁𝑑 = 𝑞𝑛 .
𝑁𝑑 = +𝑂( ).
𝑑 𝑑
𝑑|𝑛

Sketch of Proof. Consider the formal power series ∑𝑔 𝑡deg(𝑔) having

mials 𝑔 in the ring 𝔽𝑞 [𝑥]. The total number of monic polynomials 𝑔

integer coefficients where the summation is over all monic polyno‑

in 𝔽𝑞 [𝑥] of degree 𝑛 is 𝑞𝑛 , so we have

1
∞
∑ 𝑡deg(𝑓) = ∑ 𝑞𝑛 𝑡𝑛 = .
𝑔 𝑛=0
1 − 𝑞𝑡
The polynomial ring 𝔽𝑞 [𝑥] is a unique factorization domain. As a
consequence, we obtain
∞
∑ 𝑡deg(𝑔) = ∏(1 − 𝑡deg(𝑓) )−1 = ∏(1 − 𝑡𝑑 )−𝑁𝑑
𝑔 𝑓 𝑑=1

mials in 𝑓 in 𝔽𝑞 [𝑥]. It follows that

where the middle product runs over the monic irreducible polyno‑

1
∞
= ∏(1 − 𝑡𝑑 )−𝑁𝑑 ,
1 − 𝑞𝑡 𝑑=1
Taking logarithms gives
𝑞𝑛 𝑡 𝑛
∞ ∞
∑ = − log(1 − 𝑞 𝑡) = − ∑ 𝑁𝑑 log(1 − 𝑡𝑑 )
𝑛=1
𝑛 𝑑=1
𝑡𝑑𝑐 𝑡𝑛
∞ ∞ ∞
= ∑ ∑ 𝑑 𝑁𝑑 = ∑ ( ∑ 𝑑 𝑁𝑑 ) .
𝑑=1 𝑐=1
𝑑𝑐 𝑛=1
𝑛 𝑑𝑐=𝑛

Theorem 10.2.5. Let 𝑝 be a positive prime integer. For some positive

integer 𝑒, set 𝑞 ∶= 𝑝𝑒 . The irreducible factors of 𝑥𝑞 − 𝑥 are precisely the
monic irreducible polynomials in 𝔽𝑝 [𝑥] whose degree divides 𝑑. ◼

Example 10.2.6. In 𝔽2 [𝑥], we have

𝑥8 − 𝑥 = 𝑥(𝑥 + 1)(𝑥3 + 𝑥 + 1)(𝑥3 + 𝑥2 + 1)
𝑥16 − 𝑥 = 𝑥(𝑥 + 1)(𝑥2 + 𝑥 + 1)(𝑥4 + 𝑥 + 1)(𝑥4 + 𝑥3 + 1)(𝑥4 + 𝑥3 + 𝑥2 + 𝑥 + 1) .
Similarly, in 𝔽3 [𝑥], we have
𝑥9 − 𝑥 = 𝑥(𝑥 + 1)(𝑥 − 1)(𝑥2 + 1)(𝑥2 + 𝑥 − 1)(𝑥2 − 𝑥 − 1) .