Communications in Algebra

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/lagb20

A new class of irreducible polynomials

Jitender Singh & Sanjeev Kumar

To cite this article: Jitender Singh & Sanjeev Kumar (2021) A new class of irreducible polynomials,
Communications in Algebra, 49:6, 2722-2727, DOI: 10.1080/00927872.2021.1881789

To link to this article: https://doi.org/10.1080/00927872.2021.1881789

Published online: 17 Feb 2021.

Submit your article to this journal

Article views: 74

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=lagb20
COMMUNICATIONS IN ALGEBRAV
R

2021, VOL. 49, NO. 6, 2722–2727

https://doi.org/10.1080/00927872.2021.1881789

A new class of irreducible polynomials

Jitender Singha and Sanjeev Kumarb
a
Department of Mathematics, Guru Nanak Dev University, Amritsar, India; bDepartment of Mathematics, SGGS
College, Chandigarh, India

ABSTRACT ARTICLE HISTORY

In this article, we propose a few sufficient conditions on polynomials hav- Received 12 September 2019
ing integer coefficients all of whose zeros lie outside a closed disk cen- Communicated by Scott
tered at the origin in the complex plane and deduce the irreducibility over Chapman
the ring of integers.
KEYWORDS
Integer coefficients;
irreducibility criterion;
irreducibility of polynomials

2010 MATHEMATICS
SUBJECT
CLASSIFICATION
Primary 12E05; 11C08

1. Introduction
Testing polynomials for irreducibility over a given domain is an arduous task. Of particular inter-
est are the polynomials having integer coefficients for which some well–known classical irreduci-
bility criteria due to Sch€
onemann, Eisenstein, and Dumas exist (see [2, 3, 7] and for an insightful
historical account of Sch€ onemann and Eisenstein criteria, see [1]). Recently, the elegant criteria
established in [4, 6] turn out to be extremely significant keeping in view their intimate connec-
tion with prime numbers. Moreover, the notion of locating the zeros of the given polynomial
being tested for irreducibility is quite informative (see [5]). In this regard, one can infer that if
for each zero f of g 2 Z½x, jfj  r holds for some r > 0, then each zero h of f ¼ gðx  cÞ is given
by h ¼ f þ c which on applying the triangle inequality yields jhj > 1 for any integer c whose
absolute value exceeds r þ 2. Also, the translational invariance of irreducibility of polynomials in
the ring Z½x ensures the irreducibility of g vis–a–vis from that of f. Proceeding in this manner,
one can frame the following irreducibility criterion from that of the one given in [4, Theorem 1].
Theorem A. Let f 2 Z½x be such that f ð0Þ ¼ 6pd for some positive integer d and prime p - d. If
each zero h of f satisfies jhj>d, then f is irreducible in Z½x:
Proof of Theorem A. If possible, let f ðxÞ ¼ f1 ðxÞf2 ðxÞ, where f1 and f2 are non–constant polyno-
mials in Z½x: By hypothesis on f, f ð0Þ ¼ f1 ð0Þf2 ð0Þ ¼ 6pd which shows that p divides exactly
one of the factors f1 ð0Þ or f2 ð0Þ: Assume without loss of generality that pjf2 ð0Þ: Then jf1 ð0Þj  d:
On the other hand if c 6¼ 0 is the leading coefficient of f1, then we may write

CONTACT Jitender Singh [email protected] Department of Mathematics, Guru Nanak Dev University, Amritsar-
143005, India.

We dedicate this article to our families for their love, gregarious, moral, and unconditional support in every sphere of
our life.
ß 2021 Taylor & Francis Group, LLC
 COMMUNICATIONS IN ALGEBRAV
R
2723

Y
f1 ðxÞ ¼ c ðx  hÞ, (1)
h

where the product runs Q over all zeros of f1. By the hypothesis on zeros of f we must have from
(1) that jf1 ð0Þj ¼ jcj h jhj > jcjddegf1  d, a contradiction. w

In Theorem A, the primality of jf ð0Þj=d is necessary to deduce the irreducibility. In an attempt

to weaken the hypothesis, we confront the following natural question: Given jhj > d for each zero
h of f, is it still possible to recover the irreducibility of f if instead jf ð0Þj=d is a prime power?
Nevertheless, under certain mild conditions on the coefficients of f, we show that the answer to
the above question is in the affirmative.
Recall that a polynomial f having integer coefficients is primitive if the greatest common div-
isor of all its coefficients is 1. Our main results are the following:
Theorem 1. Let f ¼ a0 þ a1 x þ    þ an xn 2 Z½x be a primitive polynomial such that each zero h
of f satisfies jhj > d, where a0 ¼ 6pk d for some positive integers k and d, and a prime p - d. If j 2
f1, :::, ng is such that gcdðk, jÞ ¼ 1, pk jgcdða0 , a1 , :::, aj1 Þ and for k > 1, p - aj , then f is irreducible
in Z½x:
Theorem 2. Let f ¼ a0 þ a1 x þ    þ an xn 2 Z½x be a primitive polynomial such that each zero h
of f satisfies jhj > d, where an ¼ 6pk d for some positive integers k and d, and a prime p - d. Let j 2
f1, :::, ng be such that gcdðk, jÞ ¼ 1, pk jgcdðanjþ1 , anjþ2 , :::, an Þ and for k > 1, p - anj . If ja0 =qj 
jan j where q is the smallest prime divisor of a0, then f is irreducible in Z½x:
To prove Theorems 1 and 2, elementary divisibility theory for integers is devised. The cogent
techniques involved in the proofs are of independent interest as well. Further, the notations speci-
fied below are imperative and shall be used in the sequel.
Notations. For f ðxÞ ¼ f1 ðxÞf2 ðxÞ, unless otherwise specified, we write f ¼ a0 þ a1 x þ    þ
an xn 2 Z½x; f1 ¼ b0 þ b1 x þ    þ bm xm and f2 ¼ c0 þ c1 x þ    þ cnm xnm are non–constant
polynomials in Z½x: Define further that
bmþ1 ¼ bmþ2 ¼    ¼ bn ¼ 0; cnmþ1 ¼ cnmþ2 ¼    ¼ cn ¼ 0,
so that we may write
at ¼ b0 ct þ b1 ct1 þ    þ bt c0 , for each t ¼ 0, 1, :::, n: (2)

2. Proofs of Theorems 1 and 2

To prove Theorems 1 and 2, we first prove the following crucial result.
Lemma 3. Let f ¼ a0 þ a1 x þ    þ an xn , f1 ¼ b0 þ b1 x þ    þ bm xm , and f2 ¼ c0 þ c1 x þ    þ
cnm xnm be non–constant polynomials in Z½x such that f ðxÞ ¼ f1 ðxÞf2 ðxÞ. Suppose there exists a
prime number p and positive integers k  2 and j  n such that pk jgcdða0 , a1 , :::, aj1 Þ, pkþ1 - a0 ,
and gcdðk, jÞ ¼ 1. If pjb0 and pjc0 , then pjaj :
Proof of Lemma 3. In view of the hypothesis that pjb0 and pjc0 , there exists a positive integer ‘ 
k  ‘ such that p‘ jb0 and pk‘ jc0 , where ‘ and k  ‘ are highest powers of p appearing in b0 and
c0 respectively. To proceed, we define the nonnegative integer j such that j ¼ ðj  2Þ=2 if j is
even and j ¼ ðj  1Þ=2 if j is odd. We now arrive at the following cases:
Case I: ‘ < k  ‘: In this case, we have the following subcases:
Subcase I: pjbi for all i ¼ 0, :::, j: Using the expressions for ai and a2i successively for each i ¼
0, :::, j, we find that p divides c0, c1, :::, cj : If ai and bi are the highest powers of p appearing in
bi and ci respectively, then a0 ¼ ‘ and b0 ¼ k  ‘: We claim that ai  ‘ and bi  k  ‘ for all i 
j: For proof, we consider a1 ¼ b0 c1 þ b1 c0 which tells us that
2724 J. SINGH AND S. KUMAR

‘ þ b1  k, b1  k  2‘ þ a1 ,
which further give a1  ‘ and b1  k  ‘ with a1 < b1 since ‘ < k  ‘: Then pk jða2  b1 c1 Þ ¼
b0 c2 þ b2 c0 which for the similar reasons shows that a2  ‘ and b2  k  ‘ with a2 < b2 :
Continuing in this manner, suppose for some positive integer i
< j that the following have been
proved successively
ai  ‘, bi  k  ‘, ai < bi , for each i ¼ 0, 1, :::, i
: (3)
Then consider ai
þ1 ¼ b0 ci
þ1 þ ðb1 ci
þ    þ bi
c1 Þ þ bi
þ1 c0 , where from (3) we get p‘ jbi and
pk‘ jci
þ1i for each i ¼ 1, :::, i
so that pk jbi ci
þ1i : Consequently, pk jðb1 ci
þ    þ bi
c1 Þ: Also, by
the hypothesis, pk jai
þ1 : So we get pk jðai
þ1  b1 ci
     bi
c1 Þ ¼ b0 ci
þ1 þ bi
þ1 c0 : This proves
that ai
þ1  ‘ and bi
þ1  k  ‘ with ai
þ1 < bi
þ1 since ‘ < k  ‘: With this, we conclude that
ai  ‘, bi  k  ‘, ai < bi for all i ¼ 0, :::, j: (4)
To proceed further, we first assume that j ¼ ðj  2Þ=2: Using (4) in the expression for aj1 in
(2), we have
pk jðaj1  b0 cj1      bðj4Þ=2 cðjþ2Þ=2  bðjþ2Þ=2 cðj4Þ=2      bj1 c0 Þ
¼ bðj2Þ=2 cj=2 þ bj=2 cðj2Þ=2 ,
which shows that pk2‘ jcj=2 : Consequently
pjfb0 cj þ    þ bðj2Þ=2 cðjþ2Þ=2 þ bj=2 cj=2 þ bðjþ2Þ=2 cðj2Þ=2 þ    þ bj1 c0 g ¼ aj ,
where the equality follows from (2).
For j ¼ ðj  1Þ=2 we have from (4) and (2) that
p‘ jðb0 cj þ b1 cj1 þ    þ bðj1Þ=2 cðjþ1Þ=2 þ bðjþ1Þ=2 cðj1Þ=2 þ    þ bj c0 Þ ¼ aj :
Subcase II: There exists a smallest positive integer i  j for which p - bi : From the Subcase I, p‘
divides each of b0, :::, bi1 and pk‘ divides each of c0, :::, ci1 : Let qj be the positive integer, such
that iqj  j  1 < ð1 þ qj Þi: Let bs denote the highest power of p appearing in cs for i  s  j  1:
We will show that btiþr ¼ k  ðt þ 1Þ‘, for each t ¼ 1, :::, qj and r ¼ 0, :::, i  1:
To proceed, we first observe from (2) that
b0 ct ¼ at  Cðc0 , c1 , :::, ct1 Þ, (5)
where Cðc0 , :::, ct1 Þ is the integer combination of c0 , :::, ct1 which we define as follows:
Cðc0 Þ ¼ 0; Cðc0 , c1 , :::, ct1 Þ ¼ bt c0 þ bt1 c1 þ    þ b1 ct1 for t > 1: (6)
Since p jct for each t ¼ 0, :::, i  1, it follows from (6) that p jCðc0 , :::, ci1 Þ, which in view
k‘ k‘

of (5) and the fact that pk jai gives bi ¼ k  2‘ since p - bi : Suppose we have proved successively
that biþr ¼ k  2‘ for 0  r < i  1: Then pk‘ jðbiþr c0 þ    þ bi cr Þ and pk‘ jðbiþ1 cr þ    þ
b1 ciþr Þ so that from (6), we get pk‘ jCðc0 , :::, ciþr Þ, which in view of (5) gives pk2‘ jciþrþ1 or
biþrþ1  k  2‘: Since p - bi , we must also have biþrþ1  k  2‘: So, biþrþ1 ¼ k  2‘: This proves
the claim for t ¼ 1 and all r ¼ 0, :::, i  1:
Now suppose that btiþr ¼ k  ðt þ 1Þ‘ for each t ¼ 0, :::, t
and r ¼ 0, :::, i  1 for some posi-
tive integer t
 qj : Then we have
as ¼ a0 ; btiþs ¼ k  ðt þ 1Þ‘ for s ¼ 0, :::, i  1; t ¼ 0, :::, t
: (7)
For convenience, we define
hðsÞ ¼ bs cið1þt
Þþrs , s ¼ 0, :::, ið1 þ t
Þ þ r: (8)
From (7) and (8), we have for r ¼ 0 and each s ¼ 0, :::, i  1




p‘þkð1þt Þ‘ jhðsÞ; pkð1þt Þ‘ jhði þ sÞ; pkt ‘ jhð2i þ sÞ; :::; pk‘ jhðið1 þ t
Þ þ sÞ, (9)
 COMMUNICATIONS IN ALGEBRAV
R
2725

Also, from (6) and (8), we have

X
i1 X
2i1 XÞ1
ið1þt XÞþr
ið1þt
Cðc0 , :::, cið1þt
Þþr1 Þ ¼ hðsÞ þ hðsÞ þ    þ hðsÞ þ hðsÞ
s¼1 s¼i s¼it
s¼ið1þt
Þ

X
i1 X
i1 X
r
¼ hðsÞ þ fhði þ sÞ þ    þ hðit
þ sÞg þ hðið1 þ t
Þ þ sÞ (10)
s¼1 s¼0 s¼0
X
i1 X
t
X
i1 X
r
¼ hðsÞ þ hðis0 þ sÞ þ hðið1 þ t
Þ þ sÞ:
s¼1 s0 ¼1 s¼0 s¼0
kð1þt
Þ‘
Using (9) in (10) for r ¼ 0, we get p jCðc0 , :::, cið1þt
Þ1 Þ: Consequently, from (5), we have



pkð1þt Þ‘ jðaið1þt
Þ  Cðc0 , :::, cið1þt
Þ1 ÞÞ ¼ b0 cið1þt
Þ : This further gives pkð2þt Þ‘ jcið1þt
Þ : Thus,
bið1þt
Þþr ¼ k  ð2 þ t
Þ‘ > 0 (11)
holds for r ¼ 0. In view of (11), the assertion in (9) holds for r ¼ 1, using which further in (10)
proves (11) for r ¼ 1. Suppose then that (11) holds for each r ¼ 0, :::, r
for some positive integer
r
< i  1: Then in view of (11), we have that (9) holds for r ¼ r
: Using this further in (10)
proves that (11) holds for r ¼ r
þ 1: This proves the claim. So, pkð1þqj Þ‘ jcs , where k > ð1 þ qj Þ‘
for all s ¼ 0, :::, j  1 which in view of (2) proves
pkð1þqj Þ‘ jðb0 cj þ b1 cj1 þ ::: þ bi cji þ    þ bj c0 Þ ¼ aj :
Case II: ‘ ¼ k  ‘: Here k is even. Then j is odd since gcdðk, jÞ ¼ 1: In this case, we use the fact
that for any two integers a and b, and prime p, if pjða þ bÞ and pjab, then pja and pjb:
In view of the above fact, we have from the expressions for a1 and a2 in (2) that pjb1 and pjc1 :
Similarly from the expressions for a2 and a4 in (2), we get pjb2 and pjc2 : Continuing this way,
having proved that p divides each of the integers b0, c0, b1, c1, :::, bðj3Þ=2 , cðj3Þ=2 , it follows from
the expressions for aðj1Þ=2 and aj1 in (2) that pjbðj1Þ=2 and pjcðj1Þ=2 : So in view of (2), we get
the following:
pjðb0 cj þ    þ bðj1Þ=2 cðjþ1Þ=2 þ bðjþ1Þ=2 cðj1Þ=2 þ    þ bj c0 Þ ¼ aj :
This completes the proof of Lemma 3. w

Remark. Proof of Lemma 3 becomes considerably short if one assumes gcdðk, j!Þ ¼ 1: In that
case, the condition gcdðk, j!Þ ¼ 1 implies k > j and k  t‘ > 0 for each t ¼ 1, :::, j: Consequently
in view of (2), one immediately finds recursively that
pkðt1Þ‘ jðat1  b1 ct2  b2 ct3      bt1 c0 Þ ¼ b0 ct1 , t ¼ 1, :::, j: (12)
So from (12) it follows that pjct for each t ¼ 0, :::, j  1 which in view of (2) and the fact that
pjb0 yields the desired conclusion pjðb0 cj þ b1 cj1 þ    þ bj c0 Þ ¼ aj :

Proof of Theorem 1. If possible, assume that f ðxÞ ¼ f1 ðxÞf2 ðxÞ where f1 and f2 are as in the nota-
tions. Then in view of (4), we have
a0 ¼ b0 c0 ¼ 6pk d; am ¼ bm cnm : (13)
Since each zero h of f satisfies jhj > d, we must have jb0 =bm j > d and jc0 =cnm j > d which fur-
ther give jb0 j > d and jc0 j > d:
If p - c0 , then pk jb0 and consequently the second equality in (13) yields jc0 j  d, a contradic-
tion. On the other hand, if pjb0 and pjc0 , then k > 1 which in view of Lemma 3 gives the desired
contradiction pjaj : w
2726 J. SINGH AND S. KUMAR

Proof of Theorem 2. Suppose the contrary that f ðxÞ ¼ f1 ðxÞf2 ðxÞ where f1 and f2 are as in the
notations. Then b0 c0 ¼ a0 and bm cnm ¼ an ¼ 6pk d: Since each zero h of f satisfies jhj > d, we
must have jb0 =bm j > d and jc0 =cnm j > d: If p - bm then pk jcnm so that jbm j  d, and we have
ja0 =an j ¼ jb0 =bm j  jc0 =cnm j > jb0 =djd ¼ jb0 j  q,
which contradicts the hypothesis.
On the other hand, if pjbm and pjcnm , then k  2 which on using Lemma 3 yield the desired
contradiction, that is, pjanj : w

Remark. In view of Theorems 1 and 2, the hypothesis on zeros of f is not required in the case
when j ¼ n, wherein the hypothesis on a0 is also not required in Theorem 2 and we then have:

Theorem B. Let f ¼ a0 þ a1 x þ    þ an xn 2 Z½x be a primitive polynomial. For a prime p and

positive integers k and n, if gcdðk, nÞ ¼ 1, pk jgcdða0 , a1 , :::, an1 Þ, p - an , and pkþ1 - a0 , then f is irre-
ducible in Z½x:
Theorem B is well known and is generally proved using Newton polygons (see [2]). However
here, we provide an alternative proof based on Lemma 3.
Proof of Theorem B. Assume that f ðxÞ ¼ f1 ðxÞf2 ðxÞ where f1 and f2 are as in the notation. In view
of Lemma 3, it suffices to show that pjb0 and pjc0 in order to get the desired contradiction. Since
pja0 ¼ b0 c0 , we may assume without loss of generality that pjb0 : Since p - an ¼ bm cnm , we have
p - bm and p - cnm : So, there exists a least positive integer t  m such that p - bt : This in view of
(2) yields the following:
pjðat  b0 ct  b1 ct1      bt1 c1 Þ ¼ bt c0 ,
so that pjbt c0 , which further gives pjc0 : w

3. Examples
1. For a prime p, positive integers n and k with gcdðk, jÞ ¼ 1, consider the polynomial
Xj, k ¼ pkþ1 ð1 þ x þ x2 þ    þ xj1 Þ þ ðpk  1Þxj þ pk1 xjþ1 ð1 þ x þ    þ xnj1 Þ: (14)
We will show that each zero f of Xj, k satisfies jfj > 1: Observe that
ðx  1ÞXj, k ¼ pkþ1 þ ðpkþ1  pk þ 1Þxj þ ðpk  pk1  1Þxjþ1 þ pk1 xnþ1 : (15)
so that the coefficients of x , xjþ1 , and xnþ1 in ðx  1ÞXj, k are all positive. If jfj < 1, then from
j

(15), we have
pkþ1 ¼ ðpkþ1  pk þ 1Þfj þ ðpk  pk1  1Þfjþ1 þ pk1 fnþ1
 ðpkþ1  pk þ 1Þjfjj þ ðpk  pk1  1Þjfjjþ1 þ pk1 jfjnþ1 (16)
< ðpkþ1  pk þ 1Þ þ ðpk  pk1  1Þ þ pk1 ¼ pkþ1 ,
which is absurd. So we must have jfj  1:
If jfj ¼ 1 for some zero f of Xj, k , then f ¼ eit for some real number t. Now from (16),
ðp  pk þ 1Þð1  ejit Þ þ ðpk  pk1  1Þð1  eðjþ1Þit Þ þ pk1 ð1  eðnþ1Þit Þ ¼ 0, which on com-
kþ1

paring real parts gives

ðpkþ1  pk þ 1Þ sin 2 fðjt=2Þg þ ðpk  pk1  1Þ sin 2 fðj þ 1Þt=2g þ pk1 sin 2 fðn þ 1Þt=2g ¼ 0
which is possible only if jt, ðj þ 1Þt, ðn þ 1Þt 2 2pZ: Thus we have fj ¼ fjþ1 ¼ fnþ1 ¼ 1, which
 COMMUNICATIONS IN ALGEBRAV
R
2727

give f ¼ 1. But from (14), Xj, k ð1Þ > 0 which again leads to a contradiction. We conclude that
each zero f of Xj, k satisfies jfj > 1:
Clearly Xj, k satisfies rest of the hypotheses of Theorem 1. So Xj, k is irreducible in Z½x:
2. For a prime p, positive integers k, n, m < p, and j  n with gcdðk, jÞ ¼ 1, the polynomial
Yj, k, m ¼ pk ðn þ x þ x2 þ    þ xnj1 Þ þ mxnj þ pk xnjþ1 ð1 þ    þ xj1 Þ
satisfies the hypotheses of Theorem 2. So Yj, k, m is irreducible in Z½x:
3. Let d be a positive integer and f ¼ a0 þ a1 x þ    þ an xn 2 Z½x such that
ja0 j > ja1 jd þ ja2 jd2 þ    þ jan jdn :
Then for jxj  d, we have
jf ðxÞj  ja0 j  ja1 jjxj      jan jjxjn > ja0 j  ja1 jd  ja2 jd2      jan jdn > 0,
which shows that each zero h of f satisfies jhj > d: Now imposing the conditions of Theorem 1
or Theorem 2 on f, the irreducibility of f in Z½x is immediate.

ORCID
Jitender Singh http://orcid.org/0000-0003-3706-8239

References
[1] Cox, D. A. (2011). Why Eisenstein proved the Eisenstein criterion and why Sch€ onemann discovered it first.
Amer. Math. Monthly 118(1):3–21.
[2] Dumas, G. (1906). Sur quelques cas d’irreductibilite des polynomes a coefficients rationnels. J. Math. Pure
Appl. 2:191–258.
[3] €
Eisenstein, G. (1850). Uber die irreduzibilitat und einige andere eigenschaften der gleichungen. J. Reine
Angew. Math. 39:160–179.
[4] Girstmair, K. (2005). On an irreducibility criterion of M. Ram Murty. Amer. Math. Monthly 112(3):
269–270. DOI: 10.2307/30037445.
[5] Perron, O. (1907). Neue kriterien f€
ur die irreduzibilit€at algebraischer gleichungen. J. Reine Angew. Math.
132:288–307.
[6] Ram Murty, M. (2002). Prime numbers and irreducible polynomials. Amer. Math. Monthly 109(5):452–458.
DOI: 10.2307/2695645.
[7] Sch€onemann, T. (1846). Von denjenigen Moduln, welche Potenzen von Primzahlen sind. J. Reine Angew.
Math. 32:93–105.