EUROPEAN JOURNAL OF PURE AND APPLIED MATHEMATICS
Vol. 7, No. 1, 2014, 109-113
ISSN 1307-5543 – www.ejpam.com

On Distinguishing Local Finite Rings from Finite Rings Only by
Counting Elements and Zero Divisors

Marcos J. González

Departamento de Matemáticas, Universidad Simón Bolívar, Caracas 1080-A, Venezuela

Abstract. The purpose of this short communication is to prove the following: Let R be a finite associa-
tive ring with unit. Then R is local if and only if |R|= pn and |Z(R)|= pm for some prime number p and
integers 1 ≤ m < n. For the commutative case, this have been recently discovered by Behboodi and
Beyranvand [1, Theorem 3]. We will also present yet another proof for the commutative case.

2010 Mathematics Subject Classifications: 16B99, 13A99, 68R10

Key Words and Phrases: Finite ring, Zero-divisor, Local rings

1. Introduction

The purpose of this paper is to prove the following:

Theorem 1. Let R be a finite associative ring with unit. Then R is local if and only if |R| = pm

and |Z(R)|= pn for some prime number p and integers 1≤ n <m.

This allows us to recognize local rings out of finite rings only by counting the number of
elements in the ring and the number of zero divisors. This result is inspired by the corre-
sponding result Behboodi and Beyranvand [1, Theorem 3, pp. 306–307] where R is assumed
to be commutative.

More precisely, they proved the following:

Theorem 2. Let R be a finite commutative ring with unit. Then R is local if and only if |R|= pm

and |Z(R)|= pn for some prime number p and integers 1≤ n <m.

We will also provide another proof for this theorem which differs both from the proof of
Theorem 1 and the original proof presented in [1].

Email address: mago@usb.ve

http://www.ejpam.com 109 c© 2014 EJPAM All rights reserved.



M. González / Eur. J. Pure Appl. Math, 7 (2014), 109-113 110

2. Preliminaries

Almost all the facts we require for the both proofs can be found in McDonald’s monograph
[4]. We use a different notation though, similar to the one used in [1].

The letter R always denotes a finite associative ring with identity. An element r ∈ R is said
to be invertible is there exists s ∈ R such that rs= sr= 1. The set of all invertible elements of
R forms a group denoted by R×. An element r ∈ R is said to be a left (right, resp.) zero divisor
if there exists s ∈ R such that sr = 0 (rs = 0, resp.) and r is said to be a two-sided zero divisor
if it is both a left and a right zero divisor. The set of all zero divisors is denoted by Z(R). For R
a finite ring with identity, any zero divisor is a two-sided zero divisor and every element that
is not a zero divisor is invertible (see Ganesan [2, Theorem 5, p.245]). Consequently,

(i) |R×|= |R|− |Z(R)|, and

(ii) If R= R1⊕ · · · ⊕Rt is a direct sum of rings, then R× = R×
1
⊕ · · · ⊕R×t .

The symbol J(R) denotes the Jacobson radical, or briefly the radical, of R, which can be equiv-
alently defined as:

(a) the intersection of all left maximal ideals of R, or

(b) the intersection of all right maximal ideals of R, or

(c) the set of all elements r ∈ R such that 1+ sr is invertible for every s ∈ R.

Consequently, 1+ J(R) is a subgroup of the group R× of invertible element in R. A finite ring
R is semi-simple if and only if J(R) = {0}. For any finite ring R, the quotient ring R/J(R) is
semi-simple. A finite ring R is local if and only if R/J(R) is a finite field, called the residue field.
Notice that a semi-simple ring is local if and only if it is a field. Finite fields are denoted by
Fq, where q= |Fq| is a power of some prime. If Fq is the residue field of some finite local ring
having pn elements, then q= pr, for some 1≤ r≤ n.

In 1969, Raghavendra [5] proved the following:

Proposition 1. Let R be a finite ring with multiplicative identity 1 6= 0 and suppose that the set
of zero divisors Z(R) forms an additive group. Then

(i) Z(R) is the Jacobson radical of R;

(ii) |R|= pnr, and |J|= p(n−1)r for some prime number p, and some positive integers n,r;

(iii) Z(R)n = {0};

(iv) the characteristic of the ring R is pk for some integer k with 1 ≤ k ≤ n, where p is the
same prime number as in (ii); and

(v) if the characteristic of the ring R is pn (i.e., the largest possible choice of pk in (iv)), then
R is commutative.

We also refer to Gilmer [3]where alternative proofs for some of these results can be found,
as well as examples showing that all possible values of n,r and k in the above proposition
already appear in the commutative case.



M. González / Eur. J. Pure Appl. Math, 7 (2014), 109-113 111

3. Proof of Theorem 1

In this section, we shall provide a proof for Theorem 1. This also provides a new proof for
Theorem 2. Our idea is to reduce the statement to the case in which R is semi-simple. It is
natural to replace R by R/J(R) and the next lemma tells us that, if we do so, the hypotheses
still hold for R/J(R).

Lemma 1. Let p be a prime number, and let R be a finite ring with identity such that |R| = pm

and |Z(R)|= pn, for some 0≤ n <m, and let S := R/J(R). Then, |S|= pm−k and |Z(S)|= pn−k,
where 0≤ k≤ n is the integer determined by |J(R)|= pk.

Let π : R → S = R/J(R) be the quotient map. Then the restriction π|R× : R× → S× is a
surjective group homomorphism. Since ker π|R× = 1+J(R), we have that R×/(1+J(R)) ∼= S×.
Consequently,

|S×|=
|R×|

|1+ J(R)|
=

|R|− |Z(R)|

|J(R)|
=

|R|

|J(R)|
−

|Z(R)|

|J(R)|
= pm−k− pn−k,

which implies that |Z(S)|= |S|− |S×|= pn−k. This concludes the proof of the lemma.
Now we can prove Theorem 1. The forward direction follows from Proposition 1, (ii). For

the converse, let R be such that |R|= pm and |Z(R)|= pn. Suppose first that J(R) = {0}. In this
case, R is semi-simple and the Wedderburn-Artin theorem (cf. [4, Theorem VIII.4, pp.128–
130]) asserts that R is a direct sum of full matrix rings over fields. More precisely, there exist
finite fields Fq1 , . . . ,Fqt having qs = p

ds elements, for every s= 1, . . . , t and positive integers
n1, . . . ,nt, such that

R=Mn1

�

Fq1
�

⊕ · · · ⊕Mnt

�

Fqt
�

= Rn1,q1
⊕ · · · ⊕Rnt,qs

, (1)

where Rn,q :=Mn(Fq). Now clearly |Mn(Fq)| = qn
2

and the group of invertible elements in
Mn(Fq) is the general linear group GLn(Fq) which has

(qn− 1)(qn−q) . . . (qn−qn−1) = qn(n−1)/2
n∏
k=1

(qk− 1)

elements (cf. [4, Theorem VIII.19, p.156]). Since |R| = pm, we have that qs = p
ds , for every

s= 1, . . . , t (that is, all the qs’s are powers of the same prime) and the numbers ds satisfy the
equation m=

∑t
s=1n

2
sds. Since R× = R×n1,q1

⊕ · · · ⊕R×nt,qt
, we have that

|Z(R)|=|R|− |R×|= |R|− |R×n1,q1
|× · · · × |R×nt,qt

|

=pm−

t∏
s=1

 

q
ns(ns−1)/2
s

ns∏
k=1

(qks − 1)

!

=

t∏
s=1

q
n2
s
s −

 

t∏
s=1

q
ns(ns−1)/2
s

! 

t∏
s=1

ns∏
k=1

(qks − 1)

!



M. González / Eur. J. Pure Appl. Math, 7 (2014), 109-113 112

=

t∏
s=1

q
ns(ns−1)/2
s

 

t∏
s=1

q
ns(ns+1)/2
s −

t∏
s=1

ns∏
k=1

(qks − 1)

!

,

and hence |Z(R)| is a power of p if and only if ∆ :=
∏t
s=1q

ns(ns+1)/2
s −

∏t
s=1

∏ns

k=1(q
k
s − 1)

is a power of p. Since ∆ ≡ ±1 mod p, and ∆ 6= 1 unless n1 = · · · = nt = 1, and t = 1, we
conclude that |Z(R)| is a power of p if and only if R = Fq1 . For the general case, we have
that S = R/J(R) is semi-simple and, according to Lemma 1, both |S| and Z(S) are powers of
p. Consequently, by applying the first part of the proof, S is a field and hence R is a local
ring.

4. New Proof of Theorem 2

In this section, we shall provide a new proof for Theorem 2 based on the fact that, for
the case in which R is a finite commutative ring, the Jacobson radical J(R) coincides with the
ideal of all the nilpotent elements of the ring R.

Let R be finite commutative ring with identity of order |R|= pm and having
|Z(R)| = pn zero divisors. We want to prove that R is local, for which it is enough to prove
that Z(R) = J(R). Moreover, since J(R) ⊂ Z(R), it is enough to prove that |J(R)| = |Z(R)|.
Since |R×| = |R|− |Z(R)| = (pm−n − 1)pn, and the integers pm−n − 1 and pn are relatively
prime, there exists a Sylow p-subgroup of |R×|, say G, having order pn (cf. Wielandt [6]). We
claim that G ⊂ 1+ J(R). Indeed, the characteristic of R is a power of p, and in particular p
is nilpotent, so p ∈ J(R). We have that (a+ b)p = ap + bp + c, where c is nilpotent, whence
c ∈ J(R). Let g ∈ G be an arbitrary element. Then gp

n

= 1 (because pn is the order of G).
Hence we can conclude that (g− 1)p

n

= gp
n

− 1+ d = d ∈ J(R) (even for the case p = 2

because 1 = −1+ p) and so g− 1 ∈ J(R) (because, if some power of a ∈ R belongs to J(R),
then a ∈ J(R)). We have proved that g ∈ 1+ J(R), for every g ∈ G, that is, G ⊂ 1+ J(R), as
claimed. Finally, pn = |G| ≤ |1+ J(R)| = |J(R)| ≤ |Z(R)| = pn and hence |J(R)| = pn = |Z(R)|,
whence J(R) = Z(R).

5. Final Remarks

Concerning the proof of Theorem 1, the present version of Wedderburn-Artin Theorem
given by [4, Theorem VIII.4, pp.128–130], asserts that there exist finite fields Fq1 , . . . ,Fqt
and some positive integers n1, . . . ,nt such that

R=Mn1

�

Fq1
�

⊕ · · · ⊕Mnt

�

Fqt
�

. (2)

The fact that Fqs has qs = p
ds elements, for every s= 1, . . . , t, is a consequence of two facts:

(i)
�

�

�Mns

�

Fqs
�

�

�

� divides |R| for every s= 1, . . . , t, and

(ii) the assumption that |R|= pm.



REFERENCES 113

Finally, we have that

pm = |R|=
�

�

�Mn1

�

Fq1
�

�

�

�× · · · ×
�

�

�Mnt

�

Fqt
�

�

�

�=

t∏
s=1

pn
2
sds

and hence we obtain the condition m=
∑t
s=1n

2
sds.

ACKNOWLEDGEMENTS The author is indebted to Professor Marek Wójtowicz for making
many important observations regarding the composition of this paper.

References

[1] M. Behboodi and R. Beyranvand. On the Structure of Commutative Rings with pk1
1
· · ·pknn

(1 ≤ k ≤ 7) zero divisors. European Journal of Pure and Applied Mathematics, 3(2):303–
316, 2010.

[2] N. Ganesan. Properties of rings with a finite number of zero-divisors II. Mathematische
Annale, 161:241–246, 1965.

[3] R. Gilmer. Zero-divisors in commutative rings. The American Mathematical Monthly,
93(5):382–387, 1986.

[4] B. R. McDonald. Finite Rings with Identity. Marcel Dekker, New York, 1974.

[5] R. Raghavendran. Finite associative rings. Compositio Mathematica, 21:195–229, 1969.

[6] H. Wielandt. Ein Beweis für die Existenz der Sylowgruppen. Archiv der Mathematik,
10(1):401–402, 1959.