Communications in Algebra

ISSN: 0092-7872 (Print) 1532-4125 (Online) Journal homepage: https://www.tandfonline.com/loi/lagb20

On clean and regular elements of

noncommutative ring extensions

Ebrahim Hashemi, Maryam Hamidizadeh & Abdollah Alhevaz

To cite this article: Ebrahim Hashemi, Maryam Hamidizadeh & Abdollah Alhevaz (2019): On
clean and regular elements of noncommutative ring extensions, Communications in Algebra, DOI:
10.1080/00927872.2018.1513013

To link to this article: https://doi.org/10.1080/00927872.2018.1513013

Published online: 20 Feb 2019.

Submit your article to this journal

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

https://doi.org/10.1080/00927872.2018.1513013

On clean and regular elements of noncommutative

ring extensions
Ebrahim Hashemi, Maryam Hamidizadeh, and Abdollah Alhevaz
Faculty of Mathematical Sciences, Shahrood University of Technology, Shahrood, Iran

ABSTRACT ARTICLE HISTORY

Let R be an associative ring with identity and a be an endomorphism of R. Received 10 February 2018
In this article, we are interested to study the some of relations between a Revised 23 July 2018
ring R and that of D. A. Jordan’s construction of the ring A(R,a) as well as Communicated by
the skew Laurent polynomial ring R½x; x1 ; a. The main propose of this art- V. A. Artamonov
icle is to characterize the unit elements, the idempotent elements, von KEYWORDS
Neumann regular elements, p-regular elements, von Neumann local ele- Idempotent element;
ments and also the clean elements of the skew Laurent polynomial ring as regular element; unit
well as Jordan’s construction of the ring A(R,a). Applying these character- element; skew Laurent
izations, one can easily get some nice radical-theoretic properties of the polynomial rings; Jordan
mentioned classes of rings. construction

2010 MATHEMATICS
SUBJECT
CLASSIFICATION
Primary: 16S36; Secondary:
16U60; 16U99

1. Introduction and definitions

Throughout this article, all rings are associative with identity and a is a monomorphism of R
such that að1Þ ¼ 1. We adopt the notations U(R), Nil(R), Nil ðRÞ; Nil ðRÞ, L-rad(R), and J(R) to
represent the set of all unit elements, the set of all nilpotent elements, the lower nilradical (i.e.,
the prime radical), the upper nilradical (i.e., the sum of all nil ideals), the Levitzki radical (i.e.,
the sum of all locally nilpotent ideals), and the Jacobson radical of a ring R, respectively. For
some undefined notations in radical theory, we refer the reader to the book [7]. For a ring R, the
polynomial ring with an indeterminate x over R is denoted by R½x. Given an endomorphism a
on R, we let R½x; a denote the skew polynomial ring, consisting the polynomials in x with coeffi-
cients in R written on the left, subject to the skewing condition xr ¼ aðrÞx, for all r 2 R.
Similarly, for an automorphism a of a ring R, the skew Laurent polynomial ring R½x; x1 ; a,
which contains R½x; a as a subring, arises as the localization of R½x; a with respect to the re set
fxj gj0 , and it consists of elements of the form xj rxi , for some r 2 R and nonnegative integers i
and j, where the addition is defined as usual and the multiplication is defined by the rule, xr ¼
aðrÞx and rx1 ¼ x1 aðrÞ for all r 2 R.
Recall that an element a 2 R is called von Neumann regular, if there exists b 2 R such that
a ¼ aba. If every element of R is von Neumann regular, then R is said to be
von Neumann regular ring. Also an element a 2 R is said to be p-regular if there exists a positive
integer n and an element b 2 R, such that an ban ¼ an . If every element of R is p-regular, then R

CONTACT Ebrahim Hashemi [email protected] Faculty of Mathematical Sciences, Shahrood University of

Technology, Shahrood, P.O. Box 316-3619995161, Iran.
ß 2019 Taylor & Francis Group, LLC
2 E. HASHEMI ET AL.

is said to be p-regular. Let IðRÞ ¼ fa 2 Rja2 ¼ ag and vnrðRÞ ¼ fa 2 Rja is von Neumann regu-
lar} and also prðRÞ ¼ fa 2 Rja is p-regular}. Thus IðRÞ  vnrðRÞ  prðRÞ. Also a ring R is called
Boolean, whenever I(R) ¼ R. Thus a Boolean ring is von Neumann regular and a von Neumann
regular ring is p-regular.
Following [6], an element a 2 R is called to be a von Neumann local element if either a 2
vnrðRÞ or 1a 2 vnrðRÞ, and a ring R is von Neumann local if every element of R is von
Neumann local. Following Nicholson [18], an element a of R is called clean if it can be expressed
as the sum of an idempotent and a unit in R. A ring R is called clean if every element of R is
clean. Let vnlðRÞ ¼ fa 2 Rja is von Neumann localg and clnðRÞ ¼ fa 2 Rja is cleang. Clearly
units and idempotents are clean elements, and a is clean if and only if 1a is clean. It follows
that every local ring is clean. More generally, Camillo and Yu [4, Theorem 9] showed that a ring
is semiperfect if and only if it is clean and contains no infinite family of orthogonal idempotents.
Moreover, they showed that every unit-regular ring is clean [4, Theorem 5].
Let X be a subset of R. The left ideal generated by X, the right ideal generated by X and the
ideal generated by X is denoted by hXi‘ ; hXir and hXi, respectively.
Recall that a ring R is reduced if it does not contain nonzero nilpotent elements. According to
Krempa [15], an endomorphism a of a ring R is called rigid if aaðaÞ ¼ 0 implies a ¼ 0 for a 2 R.
R is called an a-rigid ring [12] if there exists a rigid endomorphism a of R. Hirano [11] used the
term a-reduced for a reduced ring R with a rigid automorphism a of R. Clearly, the endomor-
phism a of an a-rigid ring is a monomorphism. Hong et al. [12] proved that any rigid endomor-
phism of a ring is a monomorphism and a-rigid rings are reduced.
As observed by Hong et al. [13], for an endomorphism a of a ring R, an a-ideal I is called to
be an a-rigid ideal if aaðaÞ 2 I implies a 2 I for a 2 R. A remarkable connections between the
a-rigid ideals of R and the related ideals of some ring extensions is shown by Hong et al. [13].
Following [1, 10], a ring R is called a-compatible if for each a; b 2 R, ab ¼ 0 if and only if
aaðbÞ ¼ 0. Consider that if R is a-compatible, then a is injective. Note that a-compatible rings
are a generalization of a-rigid rings. The authors in [10, Lemma 2.2] proved that R is a-rigid if
and only if R is a-compatible and reduced. According to [8], an ideal I of R is called an a-com-
patible ideal of R if for each a; b 2 R; ab 2 I if and only if aaðbÞ 2 I.
According to Cohn [5], a ring R is called reversible, if ab ¼ 0 implies ba ¼ 0, for a; b 2 R. A
ring R is semicommutative if for all, a; b 2 R, ab ¼ 0 implies aRb ¼ 0. In a commutative ring, the
set of nilpotent elements coincides with the intersection of all prime ideals. This property is also
possessed by certain noncommutative rings, which are known as 2-primal rings. A ring R is said
2-primal if Nil ðRÞ ¼ NilðRÞ (see [3]). Research on 2-primal rings was initiated by G. Shin [19].
Shin proved in [19, Proposition 1.11] that a ring R is 2-primal if and only if every minimal prime
ideal P of R is completely prime (i.e., R/P is a domain). A ring R is weakly 2-primal if
Nil(R) ¼ L-rad(R) and also a ring R is called NI if NilðRÞ ¼ Nil ðRÞ. Generally, we have reduced
) reversible ) semicommutative ) 2-primal ) weakly 2-primal ) NI. In general, these impli-
cations are irreversible. It is defined a ring R to be abelian if each idempotent element of R is
central. Note that semicommutative rings are abelian.
In [9], the authors showed some results about the constant products of elements in Ore exten-
sion rings, when the base ring is right duo. They also determined all the unit elements, the idem-
potent elements, the von Neumann regular elements, the p-regular elements, the von Neumann
local elements and the clean elements of the full Ore extension ring R½x; a; d, when the coeffi-
cient ring R is a right duo ring which is ða; dÞ-compatible.
In this article, we continue the ongoing effort to investigate some element-wise properties of
the noncommutative ring extensions of an associative ring R. We consider two important classes
of rings, the Jordan’s construction of the ring A(R,a) and the skew Laurent polynomial ring
R½x; x1 ; a. Our main aim is to determine the unit elements, the idempotent elements, the von
 COMMUNICATIONS IN ALGEBRAV
R
3

Neumann regular elements, the p-regular elements, the von Neumann local elements and also the
clean elements of the ring A(R,a) and R½x; x1 ; a, when the base ring R is a-rigid.

2. Some properties of the Jordan’s construction of the ring AðR; aÞ

In this section, we consider some relations between a ring R and that of D. A. Jordan’s construc-
tion of the ring A(R,a) (see [14], for more details), when a is a monomorphism of R. Let A(R,a)
or A be the subset fxi rxi j r 2 R; i  0g of the skew Laurent polynomial ring R½x; x1 ; a. For
each j  0, we have xi rxi ¼ xðiþjÞ aj ðrÞxðiþjÞ . It then follows that the set of all such elements
forms a subring of R½x; x1 ; a with
xi rxi þ xj rxj ¼ xðiþjÞ ðaj ðrÞ þ ai ðsÞÞxðiþjÞ
and
ðxi rxi Þðxj sxj Þ ¼ xðiþjÞ aj ðrÞai ðsÞxðiþjÞ ;

for r; s 2 R and i; j  0. Note that a is actually an automorphism of AðR; aÞ. We have

R½x; x1 ; a ffi A½x; x1 ; a, by way of an isomorphism which maps xi rxj to ai ðrÞxji .
We start with the following simple lemma which will be useful in the sequel.
Lemma 2.1. Let a be a monomorphism of a ring R. Then

(1) IðAÞ ¼ fxi rxi jr 2 IðRÞ and i  0g.

(2) UðAÞ ¼ fxi rxi jr 2 R; i  0 and an ðrÞ 2 UðRÞ for some n  0g.
(3) NilðAÞ ¼ fxi axi 2 Aja 2 NilðRÞ and i  0g.
Proof. (1) and (3) are clear and (2) is proved in [14, Proposition 3.1(i)]. w

Remark 2.2. If a is a monomorphism of R and aðeÞ ¼ e for each idempotent element e 2 R, then
IðAÞ ¼ IðRÞ, by Lemma 2.1(1).
Now by Lemma 2.1, we can find some of the relation between properties of R and A.
Lemma 2.3. Let a be a monomorphism of a ring R. Then we have the following assertions:

(1) A is abelian if and only if R is abelian.

(2) A is semicommutative if and only if R is semicommutative.
(3) A is reversible if and only if R is reversible.
(4) A is reduced if and only if R is reduced.

Proof
(1) ( Suppose that R is an abelian ring and xi rxi 2 IðAÞ, where r 2 R and i is a nonnegative integer.
Then r 2 IðRÞ, by Lemma 2.1. For each xj sxj 2 A, we have ðxi rxi Þðxj sxj Þ ¼
xðiþjÞ aj ðrÞai ðsÞxiþj . Since aðrÞ is an idempotent element of R and R is an abelian ring, it follows
that ðxi rxi Þðxj sxj Þ ¼ xðjþiÞ ai ðsÞaj ðrÞxjþi ¼ ðxj sxj Þðxi rxi Þ. Therefore, A is abelian.
) It is clear.
(2) ( Let R be semicommutative and ðxi rxi Þðxj sxj Þ ¼ 0, for some r; s 2 R and nonnegative
integer i, j. Then xðiþjÞ aj ðrÞai ðsÞxiþj ¼ 0 and so aj ðrÞai ðsÞ ¼ 0. Hence, ajþt ðrÞaiþt ðsÞ ¼ 0,
for each t  0. Since R is semicommutative, we have ajþk ðrÞaiþj ðbÞaiþk ðsÞ ¼ 0, for each k 
0 and b 2 R. Hence
xðiþkþjÞ ajþk ðrÞajþi ðbÞaiþk ðsÞxiþkþj ¼ 0, which implies that
ðxi rxi Þðxk bxk Þðxj sxj Þ ¼ 0, for each b 2 R and k  0. Therefore A is a semicommutative ring.
) It is clear.
4 E. HASHEMI ET AL.

(3) It can be proved by the same method as the proof of (2) with the definition of revers-
ible rings.
(4) It follows from Lemma 2.1. w
In the following, we determine the von Neumann regular elements, the p-regular elements, the
von Neumann local elements and the clean elements of A.

Remark 2.4. It is clear that if a is a monomorphism of R and t 2 UðRÞ, then aðtÞ 2 UðRÞ.

Theorem 2.5. Let R be an abelian ring and a a monomorphism of R. Then

(1) vnrðAÞ ¼ fxi ðerÞxi 2 Aje 2 IðRÞ; r 2 R; i  0 and an ðrÞ 2 UðRÞ for some n  0g.
(2) vnlðAÞ ¼ fxi axi 2 Aji  0 and a ¼ er or a ¼ 1er; where e 2 IðRÞ; r 2 R and an ðrÞ 2
UðRÞ for some n  0g.

Proof. (1) First assume that xi axi 2 A such that a ¼ er for some e 2 IðRÞ; r 2 R and an ðrÞ 2
UðRÞ for some nonnegative integer n. Then xi exi 2 IðRÞ and xi rxi 2 UðAÞ, by Lemma 2.1. By
Lemma 2.3, A is abelian, so xi axi 2 vnrðAÞ, by [9, Proposition 4.2], as desired.
Now let xi axi 2 vnrðAÞ. By Lemma 2.3, A is abelian. Then xi axi ¼ ðxk exk Þðxj rxj Þ, where
x rx 2 UðAÞ and xk exk 2 IðAÞ, by [9, Proposition 4.2], so e 2 IðRÞ and an ðrÞ 2 UðRÞ for some
j j

nonnegative integer n, by Lemma 2.1. We have xi axi ¼ ðxk exk Þðxj rxj Þ ¼ xðkþjÞ aj ðeÞak ðrÞxkþj
(). There are the following three cases:

Case (1): If k > j, then by Eq. (), xi axi ¼ xðkþjÞ aj ðeÞaj ðakj ðrÞÞxkþj . Consider akj ðrÞ ¼ s, so
xi axi ¼ xðkþjÞ aj ðesÞxkþj ¼ xk esxk ðÞ, by [14, p. 436]. We have the following two subcases:
Subcase (1): If n  kj, then an ðrÞ ¼ anðkjÞ ðakj ðrÞÞ ¼ anðkjÞ ðsÞ, so am ðsÞ 2 UðRÞ, where m ¼
nðkjÞ is a nonnegative integer. Thus by Eq. ðÞ; xi axi ¼ xk esxk , for some e 2 IðRÞ; s 2 R
and am ðsÞ 2 UðRÞ for some nonnegative integer m, as wanted.
Subcase (2): If n<kj, then by Remark 2.4, s ¼ akj ðrÞ 2 UðRÞ and so am ðsÞ 2 UðRÞ for each
m  1. Hence by Eq. ðÞ; xi axi ¼ xk esxk , where e 2 IðRÞ; s 2 R and am ðsÞ 2 UðRÞ for each
m  1, as desired.
Case (2): If k < j, then by Eq. ðÞ; xi axi ¼ xðkþjÞ ak ðajk ðeÞÞak ðrÞxkþj ¼ xðkþjÞ ak ðajk ðeÞrÞxkþj .
Put ajk ðeÞ ¼ e0 . Then xi axi ¼ xðkþjÞ ak ðe0 rÞxkþj ¼ xj e0 rxj , by [14, p. 436], where e0 2 IðRÞ; r 2
R and an ðrÞ 2 UðRÞ for some nonnegative integer n, as desired.
Case (3): If k ¼ j, then by replacing j instead of k in Eq. (), we have
xi axi ¼ xðjþjÞ aj ðerÞxjþj ¼ xj ðerÞxj , where e 2 IðRÞ; r 2 R and an ðrÞ 2 UðRÞ for some n  0.
Therefore the result follows.
(2) It follows from (1). w

Theorem 2.6. Let R be a semicommutative ring and a a monomorphism of R. Then prðAÞ ¼

fxi ðer þ wÞ xi 2 Aj e 2 IðRÞ; i  0; w 2 NilðRÞ; r 2 R and an ðrÞ 2 UðRÞ for some n  0g.

Proof. First let xi axi 2 A, where a ¼ er þ w; e 2 IðRÞ, w 2 NilðRÞ; r 2 R and an ðrÞ 2 UðRÞ for
some n  0. Then xi axi ¼ xi ðer þ wÞxi ¼ xi erxi þ xi wxi , where xi erxi 2 vnrðAÞ and
xi wxi 2 NilðAÞ, by Lemma 2.1 and Theorem 2.5. By Lemma 2.3, A is semicommutative, hence
xi axi 2 prðAÞ, by [9, Theorem 5.5], as wanted.
Now assume that xi axi 2 prðAÞ. Since by Lemma 2.3, A is semicommutative, then
x ax ¼ xj bxj þ xk cxk ¼ xðjþkÞ ðak ðerÞ þ aj ðcÞÞxjþk ,
i i
for some xj bxj 2 vnrðAÞ and
k k
x cx 2 NilðAÞ, by [9, Theorem 5.5]. By Lemma 2.1 and Theorem 2.5, c 2 NilðRÞ; e 2 IðRÞ; r 2 R
and an ðrÞ 2 UðRÞ for some nonnegative integer n. There are the following two cases:
 COMMUNICATIONS IN ALGEBRAV
R
5

Case (1): If j  k, then xi axi ¼ xðjþkÞ ðak ðerÞ þ ak ðajk ðcÞÞÞxjþk ¼ xj ðer þ ajk ðcÞÞxj , where
ajk ðcÞ 2 NilðRÞ, e 2 IðRÞ; r 2 R and an ðrÞ 2 UðRÞ for some nonnegative integer n, as wanted.
Case (2): If j < k, then xi axi ¼ xðjþkÞ ðaj ðakj ðerÞÞ þ aj ðcÞÞxjþk ¼ xk ðakj ðeÞakj ðrÞ þ cÞxk . Put
akj ðrÞ ¼ s and akj ðeÞ ¼ e0 . Hence
xi axi ¼ xðjþkÞ aj ðe0 s þ cÞxjþk ¼ xk ðe0 s þ cÞxk (**)
and e0 2 IðRÞ. If n  kj, then anðkjÞ ðsÞ ¼ anðkjÞ ðakj ðrÞÞ ¼ an ðrÞ, so am ðsÞ 2 UðRÞ, where
m ¼ nðkjÞ  0. If kj>n, then s ¼ akj ðrÞ ¼ aðkjÞn ðan ðrÞÞ 2 UðRÞ, since an ðrÞ 2 UðRÞ.
Therefore xi axi ¼ xk ðe0 s þ cÞxk , where e0 2 IðRÞ; c 2 NilðRÞ; s 2 R and there exists a nonnega-
tive integer m such that am ðsÞ 2 UðRÞ, as wanted. w

Theorem 2.7. Let a be a monomorphism of a ring R. Then

clnðAÞ ¼ fxi ðe þ rÞxi 2 Aj e 2 IðRÞ; i  0; r 2 R and an ðrÞ 2 UðRÞ for some n  0g.

Proof. Suppose that xi axi 2 A, where a ¼ e þ r; e 2 IðRÞ, r 2 R and an ðrÞ 2 UðRÞ for some
n  0. Then xi axi ¼ xi ðe þ rÞxi ¼ xi exi þ xi rxi such that xi exi 2 IðRÞ and xi rxi 2 UðRÞ, by
Lemma 2.1, which implies that xi axi 2 clnðAÞ.
Now let xi axi 2 clnðAÞ. Then xi axi ¼ xj exj þ xk rxk , where xj exj 2 IðAÞ and
k k
x rx 2 UðAÞ, so e 2 IðRÞ; r 2 R and an ðrÞ 2 UðRÞ for some n  0, by Lemma 2.1. We have
xi axi ¼ xðjþkÞ ðak ðeÞ þ aj ðrÞÞxjþk . Now by a similar way as used in the proof of Theorem 2.6,
one can complete the proof. w

3. Idempotent elements and unit elements of skew Laurent polynomial

ring R½x; x 21 ; a
In this section, we determine the relation between the idempotent elements and unit elements of
R and that of skew Laurent polynomial ring R½x; x1 ; a, when a is a rigid automorphism of R.
NotePthat if a is an automorphism of R, then every element f 2 R½x; x1 ; a has the form
f ¼ m i¼0 ai x
iþn
, where n is an integer, m  0 and ai 2 R.
Lemma 3.1. Let a be an automorphism of R. Then R is a-compatible if and only if R
is a1 -compatible.
Proof. First suppose that R is a-compatible and ab ¼ 0 for some a; b 2 R. Then
 
ab ¼ 0 () aa a1 ðbÞ ¼ 0 () aa1 ðbÞ ¼ 0:
Thus R is a1 -compatible.
Now let R be a1 -compatible and ab ¼ 0 for some a; b 2 R. Then
ab ¼ 0 () aa1 ðaðbÞÞ ¼ 0 () aaðbÞ ¼ 0::
Therefore R is a-compatible. w

The following lemma will be beneficial for our main results.

Lemma 3.2. Let a be a compatible automorphism of R and a; b 2 R. Then we have the follow-
ing assertions:

(1) If ab ¼0, then aan ðbÞ ¼ 0 ¼ an ðaÞb, for each integer n.

(2) If ak ðaÞb ¼ 0 for some integer k, then ab ¼0.
(3) If ab ¼0, then an ðaÞam ðbÞ ¼ 0, for each m; n 2 Z.
(4) If ab ¼0, then axm b ¼ 0 in R½x; x1 ; a, for each integer m.
6 E. HASHEMI ET AL.

(5) If axm b ¼ 0 in R½x; x1 ; a, for some integer m, then ab ¼0.

Proof. It follows from Lemma 3.1 and [10, Lemma 2.1]. w

Let R be a ring with an endomorphism a. According to [17, Definition 2.1], R is called a skew-
Armendariz ring, if for polynomials f ¼ a0 þ a1 x þ    þ an xn and g ¼ b0 þ b1 x þ    þ bm xm in
R½x; a, fg ¼ 0 if and only if ai bj ¼ 0 for each 0 i n; 0 j m. If a is an endomorphism of R,
then R is a-rigid if and only if R is reduced and skew-Armendariz, by [17, Proposition 2.3].

Proposition 3.3. Let R be a semicommutative ring and a a compatible automorphism of R.

Then NilðR½x; x1 ; aÞ ¼ NilðRÞ½x; x1 ; a.

Proof. Since R is a-compatible, Nil(R) is an a-compatible ideal of R. Hence R  ¼ R=NilðRÞ is

P 1
 -rigid, by [10, Lemma 2.2]. Suppose that f ¼ m
a a
i¼0 i x iþn
2 NilðR½x; x P ; aÞ. Then f k ¼ 0, for
some positive integer k. Thus ðf Þ ¼ 0,
k
where f ¼ m
i¼0 ai x
iþn
2 R=NilðRÞ
½x; x1 ;   x1 ; a . Thus, by [17, Theorem 5.1], R
a  ¼ R½x;  is skew-Armendariz and so ðai Þk ¼ 0 for
each 0 i m, by [17, Propositions 3.1 and 3.3], which implies that ai 2 NilðRÞ for each
0 i m. Therefore f 2 NilðRÞ½x; x1 ;P a and the desired inclusion is now clear.
For the reverse inclusion, let f ¼ m i¼0 ai x
iþn
2 NilðRÞ½x; x1 ; a. Then ai 2 NilðRÞ for each
0 i m. Since R is semicommutative, hence R is weakly 2-primal. It means that
NilðRÞ ¼ LradðRÞ. Hence there exists a positive integer k such that fa0 ; a1 ; :::; am gk ¼ 0. By
using Lemma 3.2, we have f k ¼ 0 which implies that f 2 NilðR½x; x1 ; aÞ and the result fol-
lows. w
Pm
Note that if f ¼ i¼0 ai xiþn is an idempotent element of R½x; x1 ; a, then n 0.
Proposition 3.4. Let R be a semicommutative ring and a a compatible automorphism of R. Let f ¼
P m
i¼0 ai x
iþn
be an idempotent element of R½x; x1 ; a. Then ai 2 NilðRÞ, for each 0 i m; i 6¼ n
and there exists an idempotent element e 2 R such that an ¼ e in R=NilðRÞ.

Proof. Since R is semicommutative and a-compatible, Nil(R) is an a-compatible ideal of R. Hence

 ¼ R=NilðRÞ is an a -rigid ring, by [10, Lemma 2.2], so R
R  is skew-Armendariz, by [17, Theorem
5.1]. Since f ¼ f 2 R=NilðRÞ½x; x1 ; a , by using [17, Proposition 3.2] one can easily show that
2

f ¼ an 2 R=NilðRÞ and ai ¼ 0 for each i 6¼ n. Then an ¼ e in R=NilðRÞ, for some idempotent
e 2 R, by [16, Theorem 21.28]. w

In the following, we specify the relation between idempotent elements of R and R½x; x1 ; a,
when R is semicommutative and a is a compatible automorphism of R.
Theorem 3.5. Let R be a semicommutative ring and a a compatible automorphism of R.
Then IðR½x; x1 ; aÞ ¼ IðRÞ.
P
Proof. Assume that f ¼ m i¼0 ai x
iþn
is an idempotent element of R½x; x1 ; a. By Proposition 3.4,
ai 2 NilðRÞ, for each 0 i m; i 6¼ n, and there exist an P idempotent element e 2 R and a nil-
potent element w 2 R such that an ¼ e þ w. Let f 0 ¼ w þ m i¼0;i6¼n ai x
iþn
. Then f ¼ f 0 þ e. By
Proposition 3.3 we have NilðR½x; x ; aÞ ¼ NilðRÞ½x; x ; a. Then f 2 NilðR½x; x1 ; aÞ. By a simi-
1 1 0

lar way as used in the proof of [9, Theorem 3.3], one can show that f 0 ¼ 0. Therefore f ¼ an 2
IðRÞ and the proof is complete. w

Corollary 3.6. Let R be a semicommutative ring and a a compatible automorphism of R. Then

R½x; x1 ; a is an abelian ring.
 COMMUNICATIONS IN ALGEBRAV
R
7

Proof. Since R is a-compatible, we have aðeÞ ¼ e for each idempotent e 2 R. Now the result fol-
lows from Theorem 3.5. w

Now we intend to prove a significant lemma, which can be useful for finding the unit elements
of R½x; x1 ; a.
Lemma 3.7. Let R be an a-rigid ring. Let f ¼ a0 þ a1 x þ    þ an xn and g ¼ b0 þ b1 x þ    þ
m x be nonzero elements of R½x; a with fg ¼ cx for some nonnegative integer k and c 2 R. Then
m k
bP
iþj¼k ai a ðbj Þ ¼ c and ai bj ¼ 0, for each i þ j 6¼ k.
i

Proof. We proceed by induction on k. If k ¼ 0, then the result follows from [9, Lemma 2.3]. Now,
let k ¼ 1. Then a0 b0 ¼ 0 ¼ b0 a0 , since R is reduced by [10, Lemma 2.2]. Thus
 
b0 cx ¼ b0 fg ¼ b0 a1 þ b0 a2 x þ    þ b0 an xn1 xðb0 þ b1 x þ    þ bm xm Þ
   (3.1)
¼ b0 a1 þ b0 a2 x þ    þ b0 an xn1 aðb0 Þ þ aðb1 Þx þ    þ aðbm Þxm x:
Let f 0 ¼ b0 a1 þ b0 a2 x þ    þ b0 an xn1 and g 0 ¼ aðb0 Þ þ aðb1 Þx þ    þ aðbm Þxm . Then by Eq.
(3.1), f 0 g 0 ¼ b0 c, which implies that b0 a1 aðb0 Þ ¼ b0 c and b0 ai aðb0 Þ ¼ 0 for i 6¼ 1, by [9, Lemma
2.3]. Since R is a-compatible and reduced, we have ai b0 ¼ 0 ¼ ai aðb0 Þ, for each i 6¼ 1. Thus
cx ¼ fg ¼ a1 aðb0 Þx þ f ðb1 x þ    þ bm xm Þ, so f ðb1 x þ    þ bm xm Þ ¼ ðca1 aðb0 ÞÞx, which implies
that f ðb1 þ    þ bm xm1 Þ ¼ ca1 aðb0 Þ. Therefore a0 b1 þ a1 aðb0 Þ ¼ c and ai bj ¼ 0 for each
i þ j 6¼ 1, by [9, Lemma 2.3].
Now, let k  2 and suppose that the result is true for all the smaller values than k. Also let
fg ¼ cxk . Then a0 b0 ¼ 0 ¼ b0 a0 , this means that
 
b0 cxk ¼ b0 fg ¼ b0 a1 þ b0 a2 x þ    þ b0 an xn1 xðb0 þ b1 x þ    þ bm xm Þ
   (3.2)
¼ b0 a1 þ b0 a2 x þ    þ b0 an xn1 aðb0 Þ þ aðb1 Þx þ    þ aðbm Þxm x:
Pn1 0 i
Let f 0 ¼ i¼0 ai x , where a0i ¼ b0 aiþ1 and g 0 ¼ aðb0 Þ þ aðb1 Þx þ    þ aðbm Þxm .
Then by Eq. (3.2), f 0 g 0 ¼ b0 cxk1 . By induction hypothesis we get a0i aðb0 Þ ¼ 0, for each
i 6¼ k1, which implies that b0 ai aðb0 Þ ¼ 0 for each i 6¼ k. Since R is a-compatible and reduced,
ai b0 ¼ 0 ¼ ai aðb0 Þ for each i 6¼ k. Thus cxk ¼ fg ¼ ak ak ðb0 Þxk þ f ðb1 x þ    þ bm xm Þ, so
f ðb1 x þ    þ bm xm Þ ¼ ðca ak ðb0 ÞÞxk , it means that f ðb1 þ b2 x þ    þ bm xm1 Þ ¼ ðcak ak
Pk m1
ðb0 ÞÞx . ConsiderP
P
k1
g ¼ j¼0 b0j xj , where b0j ¼ bjþ1 . By using induction hypothesis, we have
00
i 0 0
iþj¼k1 ai a ðbj Þ ¼ iþj¼k ai a ðbj Þ ¼ cak a ðb0 Þ and ai bj ¼ 0 for each i þ j 6¼ k1, so ai bj ¼ 0
i k

for each i þ j 6¼ k. Hence a0 bk þ a1 aðbk1 Þ þ a2 a ðbk2 Þ þ    þ ak1 ak1 ðb1 Þ þ ak ak ðb0 Þ ¼ c, and
2

the proof is complete. w

Now we determine the relation between unit elements of R and that of R½x; x1 ; a, when R is
reversible and a is a compatible automorphism of R.
Proposition
Pm 3.8. Let R be a reversible ring and a a compatible
Ps automorphism of R. Then f ¼
1
i¼0 a i x iþn
2 UðR½x; x ; aÞ if and only
P if there is g ¼ P
j¼0 b j x jþr
2 R½x; x1 ; a such that ai bj 2
NilðRÞ for each i þ j 6¼ ðr þ nÞ and iþj¼ðrþnÞ ai a ðbj Þ; iþj¼ðrþnÞ bj ajþr ðai Þ 2 UðRÞ.
iþn

P P
Proof. First let f ¼ m i¼0 ai x
iþn
2 UðR½x; x1 ; aÞ. Then there is g ¼ sj¼0 bj xjþr 2 R½x; x1 ; a such
that fg ¼ gf ¼ 1. Hence ða0 þ a1 x þ    þ am xm Þxn ðb0 þ b1 x þ    þ bs xs Þxr ¼ f1 xn g1 xr ¼ 1, which
implies that f1 ðan ðb0 Þ þ an ðb1 Þx þ    þ an ðbs Þxs Þ ¼ f1 g2 ¼ xðrþnÞ . Since f1 ; g2 2 R½x; a, we have
n þ r 0. Put  ¼ R=NilðRÞ. Then f1 g2 ¼ xp . By [10, Lemma 2.2], R  is
P ðr þ nÞ ¼ p and R

a -rigid, so a aiþn ðb Þ ¼ 
1 and a an ðb Þ ¼ 
0 for each i þ j ¼
6 p, by Lemma 3.7. Thus
P iþj¼p i j i j P
iþj¼p ai a
iþn
ðbj Þ1 2 NilðRÞ  JðRÞ, which implies that iþj¼p ai aiþn ðbj Þ 2 UðRÞ. Since Nil(R) is
an a-compatible ideal of R and ai an ðbj Þ 2 NilðRÞ for each i þ j 6¼ p, hence ai bj 2 NilðRÞ for each
i þ j 6¼ p, as desired.
8 E. HASHEMI ET AL.

P gf ¼ 1, by a similar way as used in the previous paragraph, one can show

Since
that iþj¼p bj ajþr ðai Þ 2 UðRÞ. P
P Conversely, iþn assume P that there exists g ¼ sj¼0 bj xjþr 2 R½x; x1 ; a such that
iþj¼ðrþnÞ ai a ðbj Þ; iþj¼ðrþnÞ bj a ðai Þ 2 UðRÞ and ai bj 2 NilðRÞ for each i þ j 6¼ ðr þ nÞ.
jþr

Put ðr þ nÞ ¼ p. Then we have P fg ¼ ða0 þ a1 x P

þ  þ
Pam x Þðanþi
m n
ðb0 Þ þ an ðb1 Þx þ   
þa ðbs Þx Þx . It means that fg ¼ ð iþj¼p ai a ðbj ÞÞ þ ð k6¼p ð iþj¼k ai a ðbj ÞÞxk Þxp . Since
n s p nþi

Nil(R) is an a-compatible ideal of R and ai bj 2 NilðRÞ for each i þ j 6¼ p;

P
iþj6¼p ai a
nþi
ðbj Þ 2 NilðRÞ, which implies that
X X     
ai anþi bj xk xp 2 NilðRÞ x; x1 ; a :
k6¼p iþj¼k

Hence, by Proposition 3.3 we have

X X      
ai anþi bj xk xp 2 Nil R x; x1 ; a :
k6¼p iþj¼k
P
Now, since ð iþj¼p ai anþi ðbj ÞÞ 2 UðRÞ, we get fg 2 UðR½x; x1 ; aÞ, as desired.
By a similar argument as used in the previous paragraph, one can also show that
gf 2 UðR½x; x1 ; aÞ. Therefore f 2 UðR½x; x1 ; aÞ and the result follows. w

Corollary
P 3.9. Let R be a reversible ring and a a compatible automorphism of R. If
f ¼ m i¼0 i x
a iþn
2 UðR½x; x1 ; aÞ, then ai aj 2 NilðRÞ for each i 6¼ j.
P P
Proof. Since f ¼ m
P i¼0 ai x
iþn
2 UðR½x; x1 ; aÞ, there is g ¼ sj¼0 bj xjþr 2 R½x; x1 ; a such that
iþj¼ðrþnÞ ai a
iþn
ðbj Þ 2 UðRÞ and Pai bj 2 NilðRÞ for each i þ j 6¼ ðr þ nÞ, by Proposition 3.8. Put
ðr
P þ nÞ ¼ p. Then we have a
iþj¼p i aiþn
ðb j ÞaðcÞ ¼ 1, for some c 2 R, which implies that
a aiþn
Piþj¼p i iþn j0 ðb a1ðiþnÞ
ðcÞÞ ¼ 1. Consider b j a1ðiþnÞ
ðcÞ ¼ b0j , for each i þ j ¼ p. Then
iþj¼p ai a ðbj Þ ¼ 1. Since Nil(R) is an a-compatible ideal of R and aP i bj 2 NilðRÞ, for each
i þ j 6¼ p, it follows that ai b0j 2 NilðRÞ for each i þ j 6¼ p. Then from a0 ð iþj¼p ai aiþn ðb0j Þ1Þ 2
NilðRÞ we have a20 an ðb0p Þa0 2 NilðRÞ. Similarly, we have a21 anþ1 ðb0p1 Þa1 2 NilðRÞ, so
a0 a1 2 NilðRÞ. By a similar argument as used in the above, one can show that ai aj 2 NilðRÞ, for
each i 6¼ j, as desired. w

Pm
Corollary 3.10. Let a be a rigid automorphism of R. If f ¼ i¼0 ai xiþn 2 UðR½x; x1 ; aÞ, then
there exist cj 2 R such that ai cj 2 IðRÞ, for each i þ j ¼ m.
P P
Proof.
P Since f ¼ m i¼0 ai x
iþn
2 UðR½x; x1 ; aÞ, there is g ¼ sj¼0 bj xjþr 2 R½x; x1 ; a, such that
Piþj¼ðrþnÞiþnai a ðbj Þ 2 UðRÞ, by Proposition 3.8. Put ðr þ nÞ ¼ p. Then we have
iþn

Piþj¼p ai aiþn ðbj ÞaðzÞ ¼ 1, for some z 2 R, since a is automorphism. Hence

a
iþj¼p i a ðb j a1ðiþnÞ
ðzÞÞ ¼ 1. Let a iþn
ðb j a 1ðiþnÞ
ðzÞÞ ¼ cj , for each i þ j ¼ p. Hence
a0 cp þ a1 cp1 þ    þ ap c0 ¼ 1: (3.3)
By Corollary 3.9, ai aj ¼ 0 for each i 6¼ j, since R is reduced. Then for each i  p þ 1 by multi-
plying ai from left-hand side to Eq. (3.3) we conclude that ai ¼ 0. Hence p ¼ m. Multiplying Eq.
(3.3) by ai cj from left-hand side yields that ai cj 2 IðRÞ, for each i þ j ¼ m. Therefore the result
follows. w

By using Corollaries 3.9 and 3.10, we can simplify Proposition 3.8, where a is a rigid automor-
phism of R.
 COMMUNICATIONS IN ALGEBRAV
R
9

Theorem
P 3.11. Let a be a rigid automorphism of R. Let
1 1
f ¼ m a
i¼0 i x iþn
2 R½x; x ; a
P with a 0 ¼
6 0 ¼
6
P ma . Then f 2 UðR½x; x ; aÞ if and only if there is
dj 2 R for 0 j m with iþj¼m ai dj ¼ iþj¼m dj ai ¼ 1 and ai dj ¼ 0 for each i þ j 6¼ m.
P
Proof.
Ps First assume that f ¼ m i¼0 ai x
iþn
is a unit element of R½x; x1 ; a. Then there is g ¼
1
j¼0 bj x
jþr
2 R½x; x ; a such that
X X
ai aiþn ðbj Þ; bj ajþr ðai Þ 2 UðRÞ
iþj¼ðrþnÞ iþj¼ðrþnÞ

and ai bj ¼ 0 for each i þ j 6¼ ðr þ nÞ, by Proposition 3.8. Put ðr þ nÞ ¼ p. Since a is an auto-
morphism of R, there are b0j ; b00j 2 R for 0 j p such that
   
a0 an b0p þ a1 anþ1 b0p1 þ    þ ap anþp b00 ¼ 1 (3.4)

and
b000 ar ðap Þ þ b001 arþ1 ðap1 Þ þ    þ b00p arþp ða0 Þ ¼ 1: (3.5)
By Corollary 3.9, ai aj ¼ 0 for each i 6¼ j, since R is reduced. Then for each i  p þ 1 by multiply-
ing ai from left-hand side to Eq. (3.4) we conclude that ai ¼ 0. Hence p ¼ m. By a similar argu-
ment one can show that p ¼ s.
Claim 1. For each i þ j ¼ m; anþi ðb0j Þai 2 IðRÞ.
Proof of Claim 1. By a similar way as used in the proof of Corollary 3.10 one can prove it.
Claim 2. For each i þ j ¼ m; anþi ðb0j Þai ¼ ai anþi ðb0j Þ.
Proof of Claim 2. Since R is abelian and anþi ðb0j Þai ; ai anþi ðb0j Þ 2 IðRÞ, for each i þ j ¼ m, hence
ða ðb0j Þai ai anþi ðb0j ÞÞ2 ¼ 0. Thus ai anþi ðb0j Þ ¼ anþi ðb0j Þai , for each i þ j ¼ m, since R is reduced.
nþi

Claim 3. For each 0 i m; haðai Þi ¼ hai i‘ ; haðai Þi ¼ hai ir and haðai Þi ¼ hai i.
Proof of Claim 3. First we show that haðai Þi‘ ¼ hai i‘ . Since R is semicommutative, multiplying
Eq. (3.4) by ai on the right-hand side yields ai anþi ðb0j Þai ¼ ai , for each i þ j ¼ m, by Corollary
3.9. Thus ai ½anþi ðb0j Þai 1 ¼ 0 and so ai ½anþiþ1 ðb0j Þaðai Þ1 ¼ 0, since R is a-compatible and
að1Þ ¼ 1. This means that

ai anþiþ1 b0j aðai Þ ¼ ai (3.6)

for each i þ j ¼ m, so ai 2 haðai Þi‘ . Hence hai i‘  haðai Þi‘ .

By a similar discussion as above, one can show that aðai Þ½anþi ðb0j Þai 1 ¼ 0 for each i þ j ¼ m.
Therefore

aðai Þanþi b0j ai ¼ aðai Þ (3.7)

for each i þ j ¼ m, which implies that aðai Þ 2 hai i‘ and so haðai Þi‘  hai i‘ .
Hence haðai Þi‘ ¼ hai i‘ .
By a same way as used above, one can show that hai ir ¼ haðai Þir . Therefore hai i ¼ haðai Þi.
Claim 4. For each 0 i m; hai i ¼ ha1 ðai Þi.
Proff of Claim 4. First we prove that ha1 ðai Þi‘ ¼ hai i‘ . By Eq. (3.6), we have
a ðai Þanþi ðb0j Þai ¼ a1 ðai Þ, which implies that a1 ðai Þ 2 hai i‘ , so ha1 ðai Þi‘  hai i‘ . Also by Eq.
1

(3.7), it follows that ai anþi1 ðb0j Þa1 ðai Þ ¼ ai , so ai 2 ha1 ðai Þi‘ . It means that hai i‘  ha1 ðai Þi‘ .
Hence hai i‘ ¼ ha1 ðai Þi‘ .
By a similar discussion as used in the previous paragraph, we get hai ir ¼ ha1 ðai Þir .
Therefore hai i ¼ ha1 ðai Þi.
Claim 5. For each k; t 2 Z and 0 i m, hak ðai Þi ¼ hat ðai Þi.
Proof of Claim 5. It follows from Claim 3 and Claim 4.
10 E. HASHEMI ET AL.

P P
Claim 6. For each 0 j m there are dj ; d0j 2 R such that iþj¼m ai dj ¼ P 0
iþj¼m dj ai ¼ 1.
nþi 0
Proof of Claim 6. In Eq. (3.4), put dj ¼ a ðbj Þ, for each 0 j m. Then iþj¼m ai dj ¼ 1, by
Eq. (3.4).
By Claim 5, we have han ðai Þi‘ ¼ ham ðai Þi‘ ¼ hai i‘ for each m; n 2 Z. Thus there exists d00j 2
00
R forP0 j m such that P a 00ðai00Þ ¼ d j ai for each i þ j ¼ m, so 00by00 Eq. 0 (3.5), we get
rþj
00 rþj
1 ¼ iþj¼m Pb j a ðai Þ ¼ iþj¼m b j d j ai . Consider b j d j ¼ dj for 0 j m.
Therefore iþj¼m d0j ai ¼ 1.
Claim 7. For each i þ j 6¼ m; ai dj ¼ 0 ¼ ai d0j .
Proof of Claim 7. Since R is a-rigid and ai bj ¼ 0 for each i þ j 6¼ m, it follows that ai dj ¼ 0 ¼
ai d0j for each i þ j 6¼ m, by the definition of dj and d0j in Claim 6, as wanted.
Claim 8. For each 0 j m; dj ¼ d0j .
Proof of Claim 8. Since R is semicommutative, by multiplying d00 from right-hand side to
a0 dm þ a1 dm1 þ    þ am d0 ¼ 1 and using Claim 7 we conclude that ap d0 d00 ¼ d00 . Also by multi-
plying d0 from right-hand side to d00 am þ d01 am1 þ    þ d0m a0 ¼ 1 we have d00 am d0 ¼ d0 . By
Corollary 3.10, ap d0 is an idempotent of R. Hence d00 ap d0 ¼ ap d0 d00 , since R is abelian.
Therefore d0 ¼ d00 .
0 0 0 0
By repeating this P method for dP 1 ; d2 ; :::; dm and d1 ; d2 ; :::; dm , we obtain dj ¼ dj for each
0 j m. Therefore iþj¼m ai dj ¼ iþj¼m dj ai ¼ 1 and ai dj ¼ 0, for each i þ j 6¼ m.
Conversely, since a is an automorphism of R, for each i þ j ¼ m there exists bj 2 R such that
anþi ðbj Þ ¼ dj . Then ai bj ¼ 0 for each P i þ j 6¼ m, since R is a-rigid and ai dj ¼ 0, for each
i þ j 6¼ m. Put g ¼ g1 xðnþmÞ , where g1 ¼ m j
j¼0 bj x . Hence
X    X  X   
fg ¼ fg1 xðnþmÞ ¼ ai anþi bj þ ai anþi bj xk xm
iþj¼m k6¼m iþj¼k
X  X  X  
nþi 
¼ a i dj þ ai a bj xk xm ¼ 1;
iþj¼m k6¼m iþj¼k

since ai bj ¼ 0 for each i þ j 6¼ m and R is a-rigid. Thus f is a left inverse of g. Now we find a
right inverse
P for g.
Since iþj¼m ai dj ¼ 1 and ai bj ¼ 0, for each i þ j 6¼ m, hence by a similar argument as used
in the proof of Claims 3, 4, and 5, one can show that hat ðdj Þir ¼ hak ðdj Þir , for each
k; t 2 Z and 0 j m. Then for each P i þ j ¼ m there existsPrj 2 R such that dj ¼ bj rj , since
hbj ir ¼ haðnþiÞ ðdj Þir ¼ hdj ir . Then iþj¼m bj ðrj ai Þ ¼ 1, since iþj¼m dj ai ¼ 1 and dj ¼ bj rj . For
each i þ j ¼ m, let rj ai ¼ a0i . Then a0i bj ¼ 0, for each i þ j 6¼ m. For P each i þ j ¼ m there exists
b0i 2 R such that ajðnþmÞ ðb0i Þ ¼!a0i , since a is automorphism.
! Put f 0
¼ m 0 iþn
i¼0 bi x . Then
X m Xm
gf 0 ¼ bj xj xðnþmÞ b0i xi xn
j¼0 i¼0
X  0   X  X   
jðmþnÞ
¼ bj a bi þ bj ajðmþnÞ b0i xk xm
iþj¼m k6¼m iþj¼k
X  X  X  
¼ bj a0i þ bj a0i xk xm ¼ 1;
iþj¼m k6¼m iþj¼k
since bj a0i ¼ 0, for each i þ j 6¼ m. Thus f 0 is a right inverse for g. Therefore f 2 UðR½x; x1 ; aÞ,
as wanted. w

4 Clean and regular elements of skew Laurent polynomial rings

In this section, we investigate the von Neumann regular elements, the p-regular elements, the
von Neumann local elements and the clean elements in skew Laurent polynomial ring
R½x; x1 ; a, by studying these elements in A½x; x1 ; a, when the base ring R is a-rigid.
 COMMUNICATIONS IN ALGEBRAV
R
11

1
Pm 1
Theorem 4.1. Let R be an a-rigid P x ; aÞ ¼ f i¼0 eai x 2 A½x; x ; aje 2
P ring. Then vnrðR½x;
iþn

IðRÞ and thereexist dj 2 A with iþj¼m ai dj ¼ 1 ¼ iþj¼m dj ai and ai dj ¼ 0 foreach i þ j 6¼ mg:

Proof. Since R is a reduced ring which is a-compatible, A is also reduced and a-compatible, by
Lemma 2.3 and [10, Lemma 3.11]. Then A½x; x1 ; a is abelian, by Corollary 3.6. Hence
vnrðA½x; x1 ; aÞ ¼ UðA½x; x1 ; aÞIðA½x; x1 ; aÞ, by [9, Proposition 4.2]. By Theorem 3.5 and
Remark 2.2, we have IðA½x; x1 ; aÞ ¼ IðAÞ Pm ¼ IðRÞ. Hence, by Theorem 3.11 the von Neumann
elements of A½x; x1 ; a are the
P form of P i
i¼0 ea x iþn
2 A½x; x1 ; a such that e 2 IðRÞ and there is
dj 2 A for 0 j m with iþj¼m ai dj ¼ iþj¼m dj ai ¼ 1 and ai dj ¼ 0 for each i þ j 6¼ m. Now
the result follows from the fact that R½x; x1 ; a ffi A½x; x1 ; a. w

P
Theorem 4.2. be an a-rigid ring. Then f 2 vnlðR½x; x1 ; aÞ if and only if f ¼P m
PmLet R iþn i¼0 eai x
iþn

or f ¼ 1 þ i¼0 eai x , where e 2 IðRÞ; ai 2 A and there exist dj 2 A such that

P iþj¼m ai dj ¼
d a
iþj¼m j i ¼ 1 and a d
i j ¼ 0 for each i þ j ¼
6 m:

Proof. Since R½x; x1 ; a ffi A½x; x1 ; a is an abelian ring, hence by [9, Theorem 6.1(2)], f 2
vnlðA½x; x1 ; aÞ if and only if f ¼ eu or f ¼ 1 þ eu, for some
u 2 UðA½x; x1 ; aÞ and e 2 IðA½x; x1 ; aÞ. Now the result follows from Theorems 3.5 and 3.11. w

Lemma 4.3. Let R be an abelian ring and Nil(R) be a two-sided ideal of R. Then for each a 2 R,
the following statements are equivalent:
(1) a 2 prðRÞ.
(2) a ¼ b þ w, for some b 2 vnrðRÞ and w 2 NilðRÞ.
(3) a ¼ ue þ w, for some u 2 UðRÞ; e 2 IðRÞ and w 2 NilðRÞ.
(4) a þ NilðRÞ 2 vnrðR=NilðRÞÞ.

Proof. ð2Þ () ð3Þ It follows from [9, Theorem 4.2], since R is abelian.
ð1Þ ) ð3Þ Since R is abelian and Nil(R) is a two-sided ideal of R, then the result follows from
[2, Corollary 2].
ð2Þ ) ð4Þ It follows from [9, Theorem 4.1(8)], since R is an abelian ring.
ð4Þ ) ð1Þ Since R is abelian and Nil(R) is a two-sided ideal of R, hence the result follows
from [2, Theorem 3]. w

Theorem 4.4. Let R be a semicommutative

P ring and a a compatible automorphism. Then f 2
prðR½x; x1 ; aÞ if and onlyPif f ¼ m ðea
i¼0 P i þ wi Þxiþn , where e 2 IðRÞ, ai 2 A; wi 2 NilðRÞ and
there exist dj 2 A such that iþj¼m ai dj ¼ iþj¼m dj ai ¼ 1 and ai dj ¼ 0 for each i þ j 6¼ m:

Proof. By Corollary 3.6, A½x; x1 ; a is abelian. Also by Proposition 3.3,

NilðAÞ½x; x1 ; a ¼ NilðA½x; x1 ; aÞ, so NilðA½x; x1 ; aÞ is a two-sided ideal of A½x; x1 ; a. This
means that
prðA½x; x1 ; aÞ ¼ vnrðA½x; x1 ; aÞ þ NilðA½x; x1 ; aÞ;

by Lemma 4.3. Now the result follows from Theorem 4.1 and the fact that
R½x; x1 ; a ffi A½x; x1 ; a w

P
Theorem 4.5. Let R be an a-rigid ring. Then P clnðR½x; x1 ; aÞ
P ¼ fe þ mi¼0 ai x
iþn
2A
1
½x; x ; aj e 2 IðRÞ; ai 2 A; and there exist dj 2 A with iþj¼m ai dj ¼ iþj¼m dj ai ¼ 1 and ai dj
¼ 0 for each i þ j 6¼ mg:
12 E. HASHEMI ET AL.

Proof. Since IðA½x; x1 ; aÞ ¼ IðAÞ ¼ IðRÞ, hence the result follows from Theorem 3.11 and the
fact that R½x; x1 ; a ffi A½x; x1 ; a. w

Acknowledgements
The authors would like to thank the referee for a careful reading of the paper and for all of the constructive com-
ments, which have greatly improved the presentation of the article.

References
[1] Annin, S. (2002). Associated primes over skew polynomial rings. Comm. Algebra. 30(5):2511–2528.
[2] Badawi, A. (1997). On abelian p-regular rings. Comm. Algebra 25(4):1009–1021.
[3] Birkenmeier, H. E., Heatherly, G. F., Lee, E. K. (1993). Completely prime ideals and associated radicals. In:
S. K. Jain and S. T. Rizvi, eds. Proceeding of the Biennial Ohio State-Denison Conference (Granville, USA,
1992). Hackensack, NJ: World Scientific, pp. 102–129.
[4] Camillo, V. P., Yu, H.-P. (1994). Exchange rings, units and idempotents. Comm. Algebra. 22:4737–4749.
[5] Cohn, P. M. (1999). Reversible rings. Bull. London Math. Soc. 31(6):641–648.
[6] Contessa, M. (1984). On certain classes of MP rings. Comm. Algebra. 12:1447–1469.
[7] Gardner, B. J., Wiegandt, R. (2004). Radical Theory of Rings. New York: Marcel Dekker.
[8] Hashemi, E. (2006). Compatible ideals and radicals of Ore extensions. New York J. Math. 12:349–356.
[9] Hashemi, E., Hamidizadeh, M., Alhevaz, A. (2017). Some types of ring elements in Ore extensions over
noncommutative rings. J. Algebra Appl. 16(8):1750201 (17 pp.).
[10] Hashemi, E., Moussavi, A. (2005). Polynomial extensions of quasi-Baer rings. Acta Math. Hungar.
107(3):207–224.
[11] Hirano, Y. (1999). On the uniqueness of rings of coefficients in skew polynomial rings. Publ. Math.
Debrecen. 54(3–4) 489–495.
[12] Hong, C. Y., Kim, N. K., Kwak, T. K. (2000). Ore extensions of Baer and p.p.-rings. J. Pure Appl. Algebra
151(3):215–226.
[13] Hong, C. Y., Kwak, T. K., Rizvi, S. T. (2005). Rigid ideals and radicals of Ore extensions. Algebra Colloq.
12(3):399–412.
[14] Jordan, D. A. (1982). Bijective extensions of injective ring endomorphisms. J. London Math. Soc.
2(3):435–448.
[15] Krempa, J. (1996). Some examples of reduced rings. Algebra Colloq. 3(4):289–300.
[16] Lam, T. Y. (2001). A First Course in Noncommutative Rings. 2nd ed. (Graduate Texts in Mathematics, 131).
New York: Springer-Verlag.
[17] Nasr-Isfahani, A. R. (2014). Radicals of skew polynomial and skew Laurent polynomial rings over skew
Armendariz rings. Comm. Algebra 42(3):1337–1349.
[18] Nicholson, W. K. (1977). Lifting idempotents and exchange rings. Trans. Amer. Math. Soc. 229:269–278.
[19] Shin, G. Y. (1973). Prime ideals and sheaf representation of a pseudo symmetric ring. Trans. Amer. Math.
Soc. 184:43–60.