Open Mathematics 2020; 18: 333–352

Research Article

Junying Guo and Xiaojiang Guo*

Self-injectivity of semigroup algebras

https://doi.org/10.1515/math-2020-0023
received June 12, 2019; accepted January 18, 2020

Abstract: It is proved that for an IC abundant semigroup (a primitive abundant semigroup; a primitively
semisimple semigroup) S and a field K, if K0[S] is right (left) self-injective, then S is a finite regular
semigroup. This extends and enriches the related results of Okniński on self-injective algebras of regular
semigroups, and affirmatively answers Okniński’s problem: does that a semigroup algebra K[S] is a right
(respectively, left) self-injective imply that S is finite? (Semigroup Algebras, Marcel Dekker, 1990), for IC
abundant semigroups (primitively semisimple semigroups; primitive abundant semigroups). Moreover, we
determine the structure of K0[S] being right (left) self-injective when K0[S] has a unity. As their applications,
we determine some sufficient and necessary conditions for the algebra of an IC abundant semigroup
(a primitively semisimple semigroup; a primitive abundant semigroup) over a field to be semisimple.

Keywords: (IC) abundant semigroup, regular semigroup, semigroup algebra, left (right) self-injective algebra

MSC 2010: Primary 20M25, Secondary 16S36, 16D50, 16L36

1 Introduction
Recall that an algebra (possibly without unity) R is right self-injective if R is an injective right R-module.
Dually, left self-injective algebra is defined. Equivalently, then R is right self-injective if and only if the
right R1-module R satisfies the Baer condition, where R1 is the standard extension of R to an algebra with
unity (see [1, Chap. 1]). In this case, R has a left unity. (Left; right) Self-injective algebras are known as the
generalizations of Frobenius algebras. These classes of algebras play an important role and have become a
central topic in algebras.
For group algebras, it is well known that the group algebra K[G] of a group G over the field K is right
self-injective if and only if the group G is finite; if and only if K[G] is Frobenius (for detail, see [2, Theorem
3.2.8]). Along this direction, self-injective and Frobenius semigroup algebras of finite semigroups have
been investigated by many authors (for references, see [3,4]). In particular, it was proved that in some
cases, the finiteness of the semigroup is a necessary condition for the semigroup algebra to be right
(respectively, left) self-injective; for example, the semigroup is an inverse semigroup, a countable
semigroup, a regular semigroup, etc. (cf. [3–10]). So, Okniński raised a problem: does that K[S] is a right
(respectively, left) self-injective imply that S is finite? (see [11, Problem 6, p. 328]).
A semigroup S is called right principal projective (rpp), if for any a ∈ S, the right principal ideal aS1, regarded
as a right S1-system, is projective. We can dually define left principal projective semigroup (lpp semigroup). As in
[12], an abundant semigroup is defined as a semigroup being both rpp and lpp. Moreover, El Qallali and
Fountain [13] defined idempotent-connected (IC) abundant semigroups. Indeed, IC abundant semigroups


* Corresponding author: Xiaojiang Guo, College of Mathematics and Information Science, Jiangxi Normal University,
Nanchang, 330022, China, e-mail: xjguo@jxnu.edu.cn
Junying Guo: College of Science and Technology, Jiangxi Normal University, Nanchang, 330022, China,
e-mail: 651945171@qq.com

Open Access. © 2020 Junying Guo and Xiaojiang Guo, published by De Gruyter. This work is licensed under the Creative
Commons Attribution 4.0 Public License.
334  Junying Guo and Xiaojiang Guo

become a large class of semigroups including the class of regular semigroups and that of cancellative monoids
as its proper subclasses. These three classes of semigroups have relationships as follows:

{Regular semigroups} ⊂ {IC abundant semigroups} ⊂ {Abundant semigroups}.

As known, rpp semigroups come from rpp ring. In precise, a ring R is (lpp; rpp) pp if and only if the
multiplicative semigroup of R is (lpp; rpp) pp.
Recently, Guo and Shum [14] proved that the semigroup K[S] of an ample semigroup S is right self-
injective; if and only if K[S] is left self-injective; if and only if K[S] is quasi-Frobenius; if and only if K[S] is
Frobenius; if and only if S is a finite inverse semigroup. These results show that the “distance” between
the class of finite inverse semigroups and that of ample semigroups is the right (left) self-injectivity of
semigroup algebras. So-called an ample semigroup is an IC abundant semigroup whose set of regular
elements forms an inverse subsemigroup. The class of ample semigroups contains properly the class of
inverse semigroups. Indeed, for the self-injectivity, the symmetry of  ⁎ −  ⁎ in semigroups need not be so
important. For semigroup algebras of finite ample semigroups, see ref. [15]. Guo and Guo [16] pointed out
that the mentioned results as above in [14] is valid when the ample semigroup is weakened into a strict RA
semigroup or a strict LA semigroup, especially, a right (left) ample monoid.
By inspiring the result of Okniński in [7]: for a regular semigroup S, if K[S] is right (left) self-injective
then S is finite, we have a natural problem: whether the Okniński problem is valid for IC abundant
semigroups? This is the main aim of this paper. It is worthy to record here that K[S] is right (respectively,
left) self-injective if and only if so is K0[S] (see [11, p. 188]). We shall prove the following result:

Theorem. Let K be a field and S be in one of the following cases:

(a) primitive abundant semigroups;
(b) IC abundant semigroups; and
(c) primitively semisimple semigroups.
If K0[S] is right (left) self-injective, then S is a finite regular semigroup.

As its applications, we determine when the algebra of IC abundant semigroups (respectively,

primitively semisimple semigroups; primitive abundant semigroups) is semismiple (Theorem 6.2).

2 Preliminaries
Throughout this paper, we shall use the notions and notations of the monographs of Okniński [11] and
Kelarev [17]. For semigroups, the readers can be referred to the textbooks of Clifford [18] and Howie [19].
Let S be a semigroup; we denote the set of idempotents of S as E(S), and the semigroup obtained from S by
adjoining an identity if S does not have one by S1.

2.1 IC abundant semigroups

The Green’s relations on a semigroup are well known; see for example [19, Chapter II]. As generalizations
of Green’s  - and  -relations, we have  ⁎ - and  ⁎-relations defined by

a  ⁎b if (ax = ay ⇔ bx = by for all x , y ∈ S 1),

a  ⁎b if (xa = ya ⇔ xb = yb for all x , y ∈ S 1).

 Self-injectivity of semigroup algebras  335

It is well known that  ⁎ is a right congruence and  ⁎ is a left congruence. In general,  ⊆  ⁎ and
 ⊆  ⁎. And, if a, b are regular, then a  ( ) b if and only if a  ⁎ ( ⁎) b. For the relations  ⁎ and  ⁎, the
reader can refer to [12].
A left ideal L of S is a left *-ideal of S if L = ⊔x ∈ L Lx⁎ , where Lx⁎ is the  ⁎ -class of S containing x. Dually,
right *-ideal is defined. Moreover, an ideal of S is a *-ideal of S if it is both a left *-ideal and a right *-ideal.
For a ∈ S, we denote by J*(a) the smallest *-ideal of S containing a. Following Fountain [12], we define
 ⁎ =  ⁎ ⊓  ⁎ and  ⁎ =  ⁎ ∨  ⁎ . Also, we define: for a, b ∈ S

a  ⁎b if J ⁎ (a) = J ⁎ (b).

Evidently,  ⁎ is an equivalence on S. It is verified that  ⁎ ⊆  ⁎ . Denote

I ⁎ (a) = {x ∈ J ⁎ (a): (a , x ) ∉  ⁎}.

It is well known that I*(a) is a *-ideal of S. And, it is clear that J ⁎ (a) = Ja⁎ ⊔ I ⁎ (a) where Ja⁎ is the  ⁎-class
of S containing a. If  is one of  ⁎,  ⁎,  ⁎ and  ⁎ , we shall use Ka to denote the  -class of S
containing a.
The following observation is used in the sequel.

Observation. (*): If S is a semigroup with zero θ, then

Hθ⁎ = Lθ⁎ = Rθ⁎ = Dθ⁎ = Jθ⁎ = {θ}.

Indeed, let a ∈ S.
(i) If a  ⁎θ , then by θθ = θ, we get θ = aθ = a, whereby Lθ⁎ = {θ}. Dually, we have Rθ⁎ = {θ}. Hence Hθ⁎ = {θ}
since Hθ⁎ = Lθ⁎ ⊓ Rθ⁎ .
(ii) Suppose that a  ⁎θ . Notice that  ⁎ is the smallest equivalence containing  ⁎ and  ⁎ . Then by [19,
Proposition 5.14, p. 28], there are x1, x2,…,x2n−1 ∈ S such that

(a , x1) ∈  ⁎, (x1, x2) ∈  ⁎, …, (x2n − 1, θ ) ∈  ⁎.

Now by the foregoing proof, x2n−1 = θ, and so x2n−2 = θ,…,x1 = θ, thus, a = θ. Therefore, Dθ⁎ = {θ}.
(iii) Assume that a  ⁎θ . By definition, J*(a) = J*(θ). But by the foregoing proof, {θ} is a *-ideal of S, so
J*(θ) = {θ}. It follows that a = θ since a ∈ J*(a). Thus, Jθ⁎ = {θ}.
A semigroup S is abundant if each  ⁎ -class and  ⁎ -class of S contains at least one idempotent.
Moreover, an abundant semigroup S is idempotent-connected (for short, IC) if for any a ∈ S, there exist
idempotents e, f satisfying the conditions:
(i) e  ⁎a  ⁎f ;
(ii) for each x ∈ 〈e〉, there exists y ∈ 〈f〉 such that xa = ay, where for g ∈ E(S), 〈g〉 is the subsemigroup of S
generated by the set E(gSg).
Regular semigroups are IC abundant semigroups [13]. Also, an abundant semigroup is adequate if all
of its idempotents commute; that is, all of its idempotents form a semilattice. It is proved by El Qallali and
Fountain in ref. [13] that ample semigroups are adequate IC abundant semigroups, and vice versa.

Lemma 2.1. (i) [20, Lemma 3.5] Let S be an abundant semigroup and U a *-ideal of S. For any a, b ∈ S/U,
(a , b) ∈  ⁎S ( ⁎S) if and only if (a , b) ∈  ⁎S / U ( ⁎S / U );

(ii) [20, Lemma 2.2] Let S be an (IC) abundant semigroup and I a *-ideal of S. Then the Rees quotient S/I is
(IC) abundant.

Given an abundant semigroup S, we define: for any a, b ∈ S,

336  Junying Guo and Xiaojiang Guo

a≤b if there exist e, f ∈ E (S ) such that a = eb = bf .

It is verified that ≤ is a partial order on S; see for example [21]. It is pointed out that if a ≤ b, then b ∈ E(S)
whenever a ∈ E(S) [22]. A nonzero idempotent e of S is primitive if for any f ∈ S, f ≤ e can imply that f = e or
f = θ if S has zero θ. And, S is primitive if each nonzero idempotent of S is primitive.

Lemma 2.2. Let S be an IC abundant semigroup and e, f be primitive idempotents of S.

(i) J*(e) is a primitive abundant subsemigroup of S.
(ii) J ⁎ (e) = J ⁎ ( f ) or J ⁎ (e)⋅J ⁎ (f ) = {θ} if S has zero θ .

Proof. (i). By definition,

 ⁎S ⊓ ( J ⁎ (e) × J ⁎ (e)) ⊆  ⁎J ⁎ (e) and  ⁎S ⊓ (J ⁎ (e) × J ⁎ (e)) ⊆  ⁎J ⁎ (e) .

It follows that J*(e) is an abundant subsemigroup of S. For any nonzero idempotent f, g ∈ J*(e) and
f ≤ g, since f ∈ J*(e) and by [21, Lemma 3.7], there exists a nonzero idempotent g ∈ Df⁎ such that g ≤ e. By e
is primitive, it follows from Observation (*) that g = e. So that f ∈ De⁎; similarly, g ∈ De⁎. Thus, f , g ∈ De⁎.
Now by [21, Lemma 4.3], f = g, since e is primitive so that e has no infinite chains under ≤. This shows that
any nonzero idempotent of J*(e) is not comparable for ≤. Therefore, any nonzero idempotent in J*(e) is
primitive. So, J*(e) is primitive.
(ii) Suppose that J*(e) ≠ J*( f ). For any a ∈ J*(e), b ∈ J*( f ), it is clear that ab ∈ J ⁎ (e) ⊓ J ⁎ (f ). If ab ≠ θ, then
since ab ∈ J*(e) and by [21, Lemma 3.7], we have h ∈ E(S) such that h ≤ e and h ⁎ab. From Observation (*),
it follows that h ≠ θ. But e is primitive, so e = h. It follows that e  ⁎ab. Thus, e ∈J*( f ), and J*(e) ⊆ J*( f ); and
similarly, J*( f ) ⊆ J*(e). Therefore, J*(e) = J*( f ), contrary to our hypothesis. Consequently, ab = θ and J*
(e)·J*( f ) = θ. □

2.2 Primitive abundant semigroups

Let I, Λ be nonempty sets and let Γ be a nonempty set indexing partitions P(I) = {Iα:α ∈ Γ}, P(Λ) = {Λα:α ∈ Γ}
of I and Λ, respectively. For each pair (α,β) ∈ Γ × Γ, let Mαβ be a set such that for each α ∈ Γ, Tα := Mαα is a
monoid and for α ≠ β, either Mαβ = ∅ or Mαβ is a (Tα,Tβ)-bisystem. Let 0 be a symbol not in any Mαβ. By the
(α,β)-block of an I × Λ matrix we mean those (i,λ) positions with i ∈ Iα, λ ∈ Λβ. The (α,α)-blocks are called
the diagonal blocks of the matrix. Following the usual convention, we use (a)iλ to denote the I × Λ matrix
with entry a in the (i,λ) position and zeros elsewhere, and denote by 0 the I × Λ matrix all of whose entries
are 0. Let P = (pλi) be an Λ × I sandwich matrix where a non-zero entry in the (α,β)-block is a member of
Mαβ. Suppose that the following conditions are satisfied:
(M) For all α, β, γ ∈ Γ, if Mαβ, Mαγ are both non-empty, then Mαγ is non-empty and there is a (Tα,Tβ)-
homomorphism ϕαβγ: Mαβ ⊗ Mβγ → Mαγ such that if α = β or β = γ, then ϕαβγ is a canonical isomorphism
such that the square

idαβ ⊗ ϕβγδ
Mαβ ⊗ Mβγ ⊗ Mγδ ⟶ Mαβ ⊗ Mβδ
ϕαβ ⊗ idβδ ↓ ↓ϕαβδ
Mαγ ⊗ Mγδ ⟶ Mαδ
ϕαγδ

is commutative, where idαβ is the identity mapping on Mαβ.

(C) (In what follows, we simply denote (a ⊗ b)ϕαβγ by ab, for a ∈ Mαβ, b ∈ Mαγ). If a, a1, a2 ∈ Mαβ, b, b1,
b2 ∈ Mαγ, then ab1 = ab2 implies b1 = b2; a1b = a2b implies a1 = a2. Clearly, each Tα is a cancellative monoid.
(U) For each a ∈ Γ and each λ ∈ Λα (i ∈ Iα), there is a member i of Iα (λ of Λα) such that pλi is a unit of Tα.
 Self-injectivity of semigroup algebras  337

(R) If Mαβ, Mβα are both non-empty where α ≠ β, then aba ≠ a for all a ∈ Mαβ, b ∈ Mβα.
Now, let

S = {(a)iλ : a ∈ Mαβ , i ∈ Iα , λ ∈ Λβ , α, β ∈ Γ } ⊔ {0}

and A, B ∈ S. If A = 0 or B = 0, then APB = 0. Assume that A = (a)iλ and B = (b)jμ are non-zero. If pλj = 0,
then (apλj)b = 0 = a(pλjb) so that APB = 0. Assume that pλj ≠ 0 and let (i, λ) ∈ Iα × Λβ, (j, μ) ∈ Iγ × Λδ. Then,
a ∈ Mαβ, b ∈ Mγδ and pλj ∈ Mβγ so that by Condition (M),we see that (apλj)b = a(pλjb) is a well-defined
member of Mαδ and (apλjb)iμ ∈ S. Thus, we have a product ∘ defined on S by A ∘ B = APB. We can easily
check that (S,∘) is an abundant semigroup which is primitive, and called the PA blocked Rees matrix
semigroup with the sandwich matrix P. For the sake of convenience, we denote this semigroup by
 (Mαβ ; I , Λ, Γ ; P ). It was pointed out by Fountain that a semigroup is a primitive abundant semigroup if
and only if it is isomorphic to some  (Mαβ ; I , Λ, Γ ; P ) [12]. Moreover, by [12, Proposition 2.4 (7)], the
number of nonzero regular  -classes of  (Mαβ ; I , Λ, Γ ; P ) is equal to |Γ|.
Let Q = (V, E) be a quiver (a directed graph) with set V of vertex and set E of edges. Then, a vertex α is
a source of Q if no edges end at α; α is a sink if no edges begin at α. Assume that  :=  (Mαβ ; I , Λ, Γ ; P )
is a primitive abundant semigroup. From  , we construct a quiver Q ( ) whose set of vertex is Γ and in
which there is an edge beginning at α and ending at β if α ≠ β and Mαβ ≠ ∅. By Condition (M), it is easy to
see that if Mαβ and Mβγ are both not empty, then Mαγ must not be empty. So, in Q ( ), there is a path
beginning at α and ending at β if and only if Mαβ ≠ ∅.

Lemma 2.3. Let  (Mαβ ; I , Λ, Γ ; P ) satisfy the following conditions:

(F1) |Tα| < ∞ for all α; that is, Tα is a finite group;
(F2) |Γ| = n.
Then
(i) Q ( ) is acyclic;
(ii) Q ( ) has sources and sinks;
(iii) The vertex set Γ of Q ( ) can be labeled Γ = {1, 2,…, n} in such a way that (i,j) ∈ E implies i < j.

Proof. We first prove Claim (*): For any α, β ∈ Γ such that α ≠ β, at most one of Mαβ and Mβα is not empty. If
not, take some u ∈ Mαβ, some v ∈ Mβα, some k ∈ Iα, and some λ ∈ Λβ. By the definition of blocked Rees matrix
semigroup and Condition (U), there exist μ ∈ Λβ, l ∈ Iβ such that the entry pμl of the sandwich matrix P is in
Tβ. By (upμlv)kλ = (u)kμ∘(v)lλ, we have upμlv ∈ MαβTβMβα ⊆ Tα. But Tα is a group, so there exists a ∈ Tα such that
upμlva = 1α. It follows that u·pμlva·u = 1αu = u, contrary to Condition (R). We prove Claim (*).
Let us consider the quiver Q ( ). By Claim (*), we know that Q ( ) is acyclic and by [23, Lemma,
p. 142], Q ( ) has sources and sinks; by [23, Corollary, p. 143], we have (iii). □

3 Primitive abundant semigroup algebras

In this section, we determine when a primitive abundant semigroup algebra is right (left) self-injective.
We first recall some known results.

Lemma 3.1. Let S be a semigroup and K a field. If K0[S] is a right self-injective K-algebra, then
(i) [7,8] There exist ideals Si, i = 0, 1,…, n, such that θ = S0 ⊏ S1 ⊏⋯⊏ Sn = S and the Rees quotients Si/Si+1
are completely 0-simple or T-nilpotent.
(ii) [11, Lemmas 9 and 10, p. 192] S satisfies the descending chain condition on principal left ideals and has
no infinite subgroups.
(iii) [11, Lemma 1, p. 187] K0[S] has a left identity.
338  Junying Guo and Xiaojiang Guo

Lemma 3.2. Let be a right self-injective K-algebra. Then

(i) [11, Lemma 3, p. 188] Let J be an ideal of . If J2 = 0 and /J is finite dimensional, then is finite
dimensional.
(ii) [11, Lemma 1, pp. 187–188] ann ℓ (annr (
)) ⊆
+ ann ℓ ( ) for any finitely generated left ideal
of
, where annr (X )(ann ℓ (X )) stands for the right (left) annihilators of a subset X of in .

Lemma 3.3. Let  =  (Mαβ ; I , Λ, Γ ; P ) be a primitive abundant semigroup and K be a field. If K0 [ ] is

right self-injective, then
(i) |I| < ∞ and |Γ | < ∞.
(ii) Tα is a finite group for any α.

Proof. (i) By Lemma 3.1, K0 [ ] has a left identity ε. Denote

Iε = {i ∈ I : (a)iλ ∈ supp (ε)} and Λε = {λ ∈ Λ: (b)iλ ∈ supp (ε)}.

For any (u)iλ ∈  , since ε(u)iλ = (u)iλ, there is (a)kμ ∈ supp(ε) such that (a)kμ(u)iλ = (u)iλ. It follows that
k = i. So, i ∈ Iε and i ⊆ Iε. Thus, |I| < ∞ since |Iε| < ∞. Note that, by definition, ⊔α ∈ Γ Iα is a partition of I. We
can observe that |Γ| ≤ |I| < ∞.

(ii) Let x ∈ Tα and j ∈ Iα. By Condition (U), there exists ξ ∈ Λα such that the entry pξj of the sandwich matrix
P is a unit of Tα. Because

 (x )jξ ⊇  (x )2jξ =  (xpξj x )jξ ⊇⋯ (x )njξ =  ((xpξj )n − 1 x ) jξ ⊇ ⋯,

there exists n such that

 ((xpξj )n − 1 x ) jξ =  (x )njξ =  (x )njξ+ 1 =  (xpξj )n x ) jξ

by Lemma 3.1. Thus, there exists (v )kγ ∈  such that ((xpξj)n−1x)jξ = (v)kγ((xpξj)nx)jξ. It follows that

((xpξj )n − 1 x ) jξ = (pξj−1 ) jξ ((xpξj )n − 1 x ) jξ

= (pξj−1 ) jξ (v )kγ ((xpξj )n x ) jξ
= (pξj−1 pξk vpγj (xpξj )n x ) jξ

and (xpξj )n − 1 x = pξj−1 pξk vpγj (xpξj )n x . Therefore, by Condition (C), 1α = pξj−1 pξk vpγj ⋅xpξj since Tα is a cancellative
monoid and

x , (xpξj )n − 1 , (pξj−1 pξk vpγj pξj−1 ) ∈ Tα.

So, xpξj is a unit in Tα, whereby x is invertible in Tα since pξj is a unit in Tα. Consequently, Tα is a group
and by Lemma 3.1, Tα is finite. □

Lemma 3.4. Let  =  (Mαβ ; I , Λ, Γ ; P ) be a primitive abundant semigroup and K be a field. If K0 [ ] is

right self-injective, then |Mαβ| < ∞ for any α, β ∈ Γ with α ≠ β.

Proof. We first verify

Fact (†). If β is a sink of Q ( ) and Mαβ ≠ ∅, then |Mαβ| = |Tβ| < ∞.

Pick a ∈ Mαβ. We shall prove that Mαβ = aTβ. If not, there exists b ∈ Mαβ/aTβ. Obviously, bTβ ⊆ Mαβ and
aTβ ⊓ bTβ = ∅ since Tβ is a group. Take some i0 ∈ Iα and set
 Self-injectivity of semigroup algebras  339

= {(x )i0 λ : λ ∈ Λβ , x ∈ aTβ}, and  = {(y )i0 μ : μ ∈ Λβ , y ∈ bTβ}.

Again by β is a sink of Q ( ), we know that Mβγ ≠ ∅ for any γ ≠ β. By definition, we have that
pλj ∈ Mβγ ⊔ {0} for all j ∈ Iγ, so that

Fact (‡). The entry pλj of P is equal to 0 whenever λ ∈ Λβ, j ∈ Iγ.

By Fact (‡), a routine check shows that  := ⊔  ⊔ {0} is a right ideal of  . Moreover, K0 [] is a right
ideal of K0 [ ]. By Fact (‡), a routine computation shows that the mapping φ is defined by the linear span
of the mapping:

 x if x ∈ ;
x↦
 0 if x ∈  ⊔ {0}

is a homomorphism of K0 [] into K0 [ ]. But K0 [ ] is injective, so by Baer condition, there exists

z ∈ K0 [ ] such that

φ (x ) = zx for all x ∈ K0 []. (1)

By Condition (U), there exists η ∈ Λα such that pηi0 (the entry of the sandwich matrix P) is a unit of Tα. It is
easy to see that
(a) (pηi−10 )i η x = x for any x ∈ ⊔ .
0
(b) (pηi−10 )i η z (pηi−10 )i η = ∑w ∈ J kw (w )i0 η k where J ⊆ Tα .
0 0

For any (a)i0 λ ∈ , since, by Eq. (1),

(a)i0 λ = (pηi−1 )i η ∘ (a)i λ = (pηi−1 )i η ∘ φ (a)i λ

0 0
0 0 0
0

= (pηi−1 )i η z (a)i λ = (pηi−1 )i η z (pηi−1 )i η ∘ (a)i λ

0 0
0 0 0 0 0
0

  (2)
=  w∑∈J kw (w)i η ∘ (a)i λ
0 0

 
= ∑ w ( ηi )i λ
k wp a 0 0
w∈J

and, by Condition (C), w1 pηi0 a ≠ w2 pηi0 a for any w1, w2 ∈ Tα and w1 ≠ w2, we get that wpηi0 a = a for any
w ∈ J. So, w = pηi−10 for any w ∈ J. Again by Eq. (2), (a)i0 λ = (∑w ∈ J kw )⋅(a)i0 λ and ∑w ∈ J kw = 1 (the unity of K).
Therefore, (pηi−10 )i η z (pηi−10 )i = (pηi−10 )i η. Let y ∈  . Then, by Eq. (1),
0 0η 0

0 = (pηi−10 )i ∘ φ (y ) = (pηi−10 )i ∘ zy = (pηi−10 )i η z (pηi−10 )i ∘ y = (pηi−10 )i ∘y=y

0η 0η 0 0η 0η

and  = 0, contrary to the definition of  . Thus, Mαβ = aTβ and we prove Fact (†).
Now, let Mπζ ≠ ∅. If ζ is not a sink of Q ( ), then as Q ( ) is acyclic and has only finite vertices, there
exists a path ζ = α0 → α2 → ⋯ → αn such that αn is a sink of Q ( ). So, by Condition (M),

Mπζ Mα0 α1 Mα1 α2⋯Mαn −1 αn ⊆ Mπζ Mα0 αn ⊆ Mπαn .

For d ∈ Mα0 αn , since Mπζ d ⊆ Mπαn , we have |Mπζ d| ≤ |Mπαn |. By Fact (†), |Mπαn | < ∞ and |Mπζd| < ∞. But,
by Condition (C), |Mπζ| = |Mπζd|, now |Mπζ| < ∞. This proves the lemma. □

Lemma 3.5. Let  =  (Mαβ ; I , Λ, Γ ; P ) be a primitive abundant semigroup and K a field. If K0 [ ] is right
self-injective, then  is finite.
340  Junying Guo and Xiaojiang Guo

Proof. Let ε be a left identity, and Iε and Λε have the same meaning as in Lemma 3.3. Set = K0 [ ]⋅ε
and  = {x − x⋅ε: x ∈ K0 [ ]}. It is easy to know that
(a) K0 [ ] = ⊕  (regarded as K-vector spaces);
(b)  2 = 0;
(c)  is an ideal of K0 [ ].

So, dimK (K0 [ ]/ ) = dimK ( ). Take X = {(a)iλ ∈  : i ∈ Iε , λ ∈ Λε }. Notice that |Iε| < ∞ and
|Λε| < ∞. By Lemma 3.4, we get |X| < ∞, it follows that dimK () < ∞ where  is a K-space with a basis
X. On the other hand, for any u ∈ , we know that εuε = u, and further for any (v)kl ∈ supp(u), there exist
(x)iλ, (y)jμ ∈ supp(ε), (z)mn ∈ supp(u) such that (v)kl = (x)iλ(z)mn(y)jμ = (xpλmzpnjy)iμ. Hence, k = i ∈ Iε and
l = μ ∈ Λε. Thus, (v)kl ∈ X and so u ∈  . It follows that ⊆  . This means that is a subspace of  .
Therefore, dimK ( ) ≤ dimK () and is finite dimensional. Now by Lemma 3.2 (i), K0 [ ] is finite
dimensional. Consequently,  is a finite semigroup. □

Theorem 3.6. Let  =  (Mαβ ; I , Λ, Γ ; P ) be a primitive abundant semigroup and K a field. If K0 [ ] is

right self-injective, then  is a finite primitive regular semigroup.

Proof. We need to only verify that  is a regular semigroup. By Lemmas 2.3 and 3.3, we may assume that
(a) Γ = {1, 2,…,n};
(b) For any 1 ≤ i, j ≤ n and i ≠ j, whenever Mij ≠ ∅, we have i < j;
(c) Ti is a finite group, for any 1 ≤ i ≤ n. Moreover, we let |Ii| = pi and |Λi| = qi for any i. By computation,

 M p ×q (K [T1])
1 1
M p1 × q2 (K [M12]) ⋯ M p1 × qn (K [M1n]) 
K0 [ ] =  0 M p2 × q2 (K [M22]) ⋯ M p2 × qn (K [M2n]),
 ⋯ ⋯ ⋯ ⋯ 
 0 0 ⋯ M pn × qn (K [Tn]) 

where M pi × qj (K [Mij]) is the set consisting of all pi × qj matrices over K[Mij], and

 P11 P12 ⋯ P1n 


P =  0 P22 ⋯ P2n ,
⋯ ⋯ ⋯ ⋯
 
0 0 ⋯ Pnn 

where Pij ∈ Mqi × pj (Mij ⊔ {0}). By multiplying with  , we easily know that the contracted semigroup algebra
K0 [ ] is an algebra with the usual matrix addition and the multiplication is defined by: for A, B ∈ K0 [ ]

A ∘ B = APB,

where the right side is the usual matrix multiplication of A, P and B.

Let

 ε1 ε12 ⋯ ε1n 
0
ε = ⋯ ε2 ⋯ ε2n 
⋯ ⋯ ⋯
0 εn 
 0 ⋯

be a left identity of K0 [ ]. Since,

 Self-injectivity of semigroup algebras  341

0 0 ⋯ 0  ε1 ε12 ⋯ ε1n   0 0 ⋯ 0 
0 0 ⋯ 0  0 ε2 ⋯ ε2n  ∘  0 0 ⋯ 0 
⋯ ⋯ ⋯ ⋯ = ⋯ ⋯ ⋯ ⋯  ⋯ ⋯ ⋯ ⋯
  0  
0 0 ⋯ A  0 ⋯ εn   0 0 ⋯ A 
 ε1 ε12 ⋯ ε1n   0 0 ⋯ 0 
0
= ⋯ ε2 ⋯ ε2n  P  0 0 ⋯ 0  (3)
⋯ ⋯ ⋯  ⋯ ⋯ ⋯ ⋯
0  
 0 ⋯ εn   0 0 ⋯ A 
⁎ ⁎ ⋯ ⁎ 
= ⋯0 ⁎ ⋯ ⁎ 
⋯ ⋯ ⋯ 
0 εn Pnn A
 0 ⋯

we have

εn Pnn A = A where A ∈ M pn × qn (K [Mnn]), (4)

especially, εnPnnεn = εn ≠ 0. So,

0 0 ⋯ A1  0 0 ⋯ A1 
0 0 ⋯ A2  ∘ ε =  0 0 ⋯ A2  Pε
⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯
   
0 0 ⋯ εn  0 0 ⋯ εn 
0 0 ⋯ A1 Pnn εn 

= 0 0 ⋯ A2 Pnn εn 
⋯ ⋯ ⋯ ⋯ 
 
0 0 ⋯ εn Pnn εn 
0 0 ⋯ A1 Pnn εn 

= 0 0 ⋯ A2 Pnn εn 
⋯ ⋯ ⋯ ⋯ 
 
0 0 ⋯ εn 

and

0 0 ⋯ A1 
0 0 ⋯ A2  ∉ ann (K [ ]). (5)
⋯ ⋯ ⋯ ⋯ ℓ 0
 
0 0 ⋯ εn 

We next verify that

(*) All of Min with 1 ≤ i ≤ n − 1 are empty sets.

We assume, on the contrary, that not all Min with 1 ≤ i ≤ n − 1 are empty sets. Without loss of generality,
assume that M1n ≠ ∅ and u ∈ M1n. By a routine check,

0 0 ⋯ M p1 × qn (K [M1n]) 
n :=  0 0 ⋯ M p2 × qn (K [M2n])
⋯ ⋯ ⋯ ⋯ 
0 0 ⋯ 0 
is a left ideal of K0 [ ] generated by the finite set
342  Junying Guo and Xiaojiang Guo

0 0 ⋯ M p1 × qn (M1n ⊔ {0}) 
0
 0 ⋯ M p2 × qn (M2n ⊔ {0}) 
⋯ ⋯ ⋯ ⋯ 
0 0 ⋯ M pn −1 × qn (Mn − 1, n ⊔ {0})
 
0 0 ⋯ 0 

(since all Mij are finite). If

0 0 ⋯ U   X1 X12 ⋯ X1n 
0 0 ⋯ 0 0 X2 ⋯ X2n  = 0,
⋯

⋯ ⋯ ⋯ ∘ ⋯
 ⋯ ⋯ ⋯
0   0 
0 0 ⋯ 0 ⋯ Xn 

i.e.,

0 0 ⋯ UPnn Xn   0 0 ⋯ U   X1 X12 ⋯ X1n 

0 0 ⋯ 0  = 0 0 ⋯ 0 0 X2 ⋯ X2n  = 0,
⋯ ⋯ ⋯ ⋯  ⋯ ⋯ ⋯ ⋯ P ⋯
 ⋯ ⋯ ⋯
  
0 0 ⋯ 0  0 0 ⋯ 0   0 0 ⋯ Xn 

then UPnnXn = 0. It is not difficult to know that Pnn Xn ∈ Mqn × qn (K [Tn]). Let U = (u)11 and u ∈ M12. Let

 a11 a12 ⋯ a1qn


Pnn Xn =  21
a a22 ⋯ a2q2
.
a⋯ ⋯ ⋯ ⋯ 
 q1n
aqn 2 ⋯ aqn qn 

We have

ua11 ua12 ⋯ ua1qn


0 = UPnn Xn =  0 0 ⋯ 0 
⋯ ⋯ ⋯ ⋯ 
 0 0 ⋯ 0 

and ua1j = 0 for any 1 ≤ j ≤ qn. Now let J = supp(a1j) and a1j = ∑vl ∈ J rl vl (rl ∈ K ). Note that by Condition (C)
ux ≠ uy for any x ≠ y ∈ supp(a1j). So, ∑l ∈ J rl (uvl) = ua1j = 0 implies that rl = 0. Thus, a1j = 0. If U = (u)i1, then
by applying a similar argument as above, we may obtain that aij = 0 for any 1 ≤ i, j ≤ qn. So, PnnXn = 0.
Now, by Eq. (4), Xn = εnPnnXn = 0. It follows that

 M p ×q (K [T1])
1 1
M p1 × q2 (K [M12]) ⋯ M p1 × qn (K [M1n]) 
 0 M p2 × q2 (K [M22]) ⋯ M p2 × qn (K [M2n]) 
annr (n) ⊆  ⋯ ⋯ ⋯ ⋯ 
 0 0 ⋯ M pn −1 × qn (K [Mn − 1, n])
 0 0 ⋯ 0 
and a routine check shows the reverse inclusion. Thus,

 M p ×q (K [T1])
1 1
M p1 × q2 (K [M12]) ⋯ M p1 × qn (K [M1n]) 
 0 M p2 × q2 (K [M22]) ⋯ M p2 × qn (K [M2n]) 
annr (n) =  ⋯ ⋯ ⋯ ⋯ . (6)
 0 0 ⋯ M pn −1 × qn (K [Mn − 1, n])
 0 0 ⋯ 0 
 Self-injectivity of semigroup algebras  343

On the other hand, by computation and Lemma 3.2, we have

0 0 ⋯ M p1 × qn (K [M1n]) 
0 0 ⋯ M p2 × qn (K [M2n])  ⊆ ann (ann ( )) ⊆  + ann (K [ ]). (7)
r n n 0
⋯ 
ℓ ℓ
⋯ ⋯ ⋯
0 0 ⋯ M pn × qn (K [Mnn])

Obviously,

0 0 ⋯ 0 0 0 ⋯ M p1 × qn (K [M1n]) 
0
Vn := ⋯ 0 ⋯ 0 0 0 ⋯ M p2 × qn (K [M2n]) .
⋯ ∈ ⋯ ⋯
0
⋯ ⋯
εn 
 ⋯ ⋯ 
 0 ⋯
0 0 ⋯ M pn × qn (K [Mnn])

By Eq. (5), Vn ∉ n + ann ℓ (K0 [ ]), contrary to Eq. (7). Thus, Min = ∅ for any i < n. Moreover, Pin = 0 in the
sandwich matrix P, for any i < n.
By applying a similar argument as above to the set

0 0 ⋯ M p1 × qn −1 (K [M1, n − 1]) 0
0 0 ⋯ M p2 × qn −1 (K [M2, n − 1]) 0
⋯
n − 1 := 
⋯ ⋯ ⋯ ⋯
0 0 ⋯ M pn −2 × qn −1 (K [Mn − 2, n − 1]) 0
0 0 ⋯ 0 0
0 0 ⋯ 0 0

and

0 0 ⋯ 0 0
0 0 ⋯ 0 0
Vn − 1 := ⋯ ⋯ ⋯ ⋯ ⋯,
0 0 ⋯ εn − 1 0
 
0 0 ⋯ 0 0

we can obtain that Mj,n−1 = ∅ for any 1 ≤ j < n − 1. Continuing this process, we can prove that Mji = ∅ for
j = 1, 2,…,i − 1, i = 2, 3,…,n − 2. Thus, Mji = ∅ for any i = j. Now by [12, Proposition 2.4 (7)],  is regular.
We complete the proof. □

4 Algebras of IC abundant semigroups

In this section, we shall research the self-injectivity of algebras of IC abundant semigroups.

Lemma 4.1. Let S be a semigroup and K a field. If K0[S] is right self-injective, then
(i) S has only finite regular  -classes.
(ii) S has a primitive idempotent.

Proof. (i) By Lemma 3.1, there exist ideals Si, i = 0, 1,…,n, of S such that θ = S0 ⊏ S1 ⊏⋯⊏ Sn = S and the
Rees quotients Si/Si+1 are completely 0-simple or T-nilpotent. Note that S = ⊔ni = 1 Si \ Si − 1. So, any regular
 -class of S must be in some Rees quotient Si/Si−1 being completely 0-simple. It follows that the number
of regular  -classes of S is smaller than n.
344  Junying Guo and Xiaojiang Guo

(ii) We claim: under ≤, S has a minimal nonzero idempotent e0; for, if no, S has a chain of nonzero
idempotents: e1 > e2 > ⋯ > en > ⋯, so ⋯Sen ⊂ ⋯ ⊂ Se2 ⊂ Se1, contrary to Lemma 3.1 (ii). It is not difficult to
know that e0 is primitive. □

Lemma 4.2. [11, Lemma 5, p. 189] Assume that K0[S] is right self-injective. Then there is no infinite sequence
of elements a1, a2,… of K0[S] such that the principal right algebra ideals generated by the ai are independent
and dimK(K0[S]ai) = ∞ for all i = 1, 2,….

Lemma 4.3. Assume that J =  (Mαβ , I , Λ; Γ ) is a primitive abundant ideal of an IC abundant semigroup S.
If K0[S] is right self-injective, then
(i) |I| < ∞;
(ii) K0[J] has a left identity;
(iii) all Tα = Mαα are finite groups.

Proof. We first prove that for any α ∈ Γ, |Iα| < ∞. Indeed, if Iα is infinite and choose elements i1, i2,… from Iα,
then by Condition (U), there exist elements λ1, λ2,… of Λα such that p λj ij (the entry of the sandwich matrix P)
is a unity of Tα, for j = 1, 2,…. By a routine check, every (p λ−j1ij )i λ is an idempotent in J. Moreover, we have
j j

      
(a) K0  (p λ−j1ij ) S  = K0  (p λ−j1ij ) ∘  (p λ−j1ij ) S   = K0  (p λ−j1ij ) J  ;
 ij λ j
 
ij λ j

ij λ j
  ij λ j

 −1   −1 
(b) K0  (p λ1 i1 )i λ J , K0  (p λ2 i2 )i λ J , ⋯ are independent;
 1 1
  2 2

(c) By computation,

 −1  −1
{ (a)iλ j
∈ J : a ∈ Tα, i ∈ Iα } ⊆ S (p λ−j1ij )
ij λ j
= S (p λ i )i λ  ∘ (p λ i )i λ
j j j j
= J (p λ−j1ij )
ij λ j
 
j j j j

  
and so is infinite. It follows that dimK K0 S (p λ−j1ii )   = ∞.
  
ij λ j

This is contrary to Lemma 4.2. Thus, |Iα| < ∞ for any α. On the other hand, by Lemma 4.1, S has finite
regular  -classes. So, J has finite regular  -classes. But, by [[12], Proposition 2.6 (6) and Proposition 4.1],
the number of regular  -classes of J is equal to |Γ| + 1, now |Γ| < ∞. Thus, |I| < ∞ since I = ⊔α ∈ Γ Iα .
For any i ∈ I, by Condition (U), there exists λi ∈ Λ such that p λi i (the entry of the sandwich matrix P) is
a unit of some Tα. So, (p λ−i1i )iλ is an idempotent of J and
i

K0 [J ] = ∑ K0 [(p λ−1i )iλ J] = ∑ K0 [(p λ−1i )iλ S] = i⊕∈I K0 [(p λ−1i )iλ S].
i i i i i i
i∈I i∈I

Now by [11, Lemma 1 (iv), pp. 187–188], K0[J] = eK0[S] for some e = e2 ∈ K0[J]. It follows that K0[J] has a left identity.
In addition, by the same reason as Lemma 3.3 (ii), we can prove (iii). We omit the detail. □
The following lemma is a key result to research the self-injective algebras of IC abundant semigroups,
which may be proved by revising the proof of Theorem 3.6. For the completeness, we give the proof.

Lemma 4.4. With notations in Lemma 4.3, if J is a proper ideal of S, then J is a regular subsemigroup of S.

Proof. Suppose that J is a proper ideal of S. By Lemmas 4.3 and 2.3, we may assume that
J =  (Mαβ ; I , Λ, Γ ; P ) in which

(i) Γ = {1, 2,…,n};

(ii) For any 1 ≤ i, j ≤ n and i ≠ j, whenever Mij ≠ ∅, we have i < j;
(iii) Ti is a finite group, for any 1 ≤ i ≤ n.
 Self-injectivity of semigroup algebras  345

Moreover, we let |Ii| = pi and |Λi| < qi for any i. We shall use the notations in the proof of Theorem 3.6. By (5),

0 0 ⋯ A1 
0 0 ⋯ A2  ∉ ann (K [J ]) (8)
⋯ ⋯ ⋯ ⋯ ℓ 0
 
0 0 ⋯ εn 

and of course, not in ann ℓ (K0 [S]). Notice that by (4)

0 0 ⋯ 0 0 0 ⋯ 0 0 0 ⋯ 0
0 0 ⋯ 0 0 0 ⋯ 0 0 0 ⋯ 0
⋯ ⋯ ⋯ ⋯ ∘ ⋯ ⋯ ⋯ ⋯ = ⋯ ⋯ ⋯ ⋯
0 εn   0 εn   0 εn 
 0 ⋯ 0 ⋯ 0 ⋯

and K0[J] is an ideal of K0[S], we have

0 0 ⋯ 0 0 0 ⋯ 0 0 0 ⋯ 0
0 0 ⋯ 0  K [S] =  0 0 ⋯ 0 0 0 ⋯ 0
⋯ ⋯ ⋯ ⋯ 0 ⋯ ⋯ ⋯ ⋯ ∘ ⋯ ⋯ ⋯ ⋯ K0 [S]
0 εn  0 εn   0 0 ⋯ εn 
 0 ⋯  0 ⋯
0 0 ⋯ 0
0 0 ⋯ 0
⊆ ⋯ ⋯ ⋯ ⋯ ∘ K0 [J ] (9)
0 εn 
 0 ⋯

0 0 ⋯ 0 
⊆ ⋯0 0
⋯
⋯
⋯ ⋯
0
.
0 
 0 ⋯ M pn × qn (K0 [Mnn])

We next prove that Min = ∅ for i = 1, 2,…,n − 1. Suppose, on the contrary, that not all of Min are empty
sets. Obviously,

0 0 ⋯ M p1 × qn (K [M1n]) 
n :=  0 0 ⋯ M p2 × qn (K [M2n]) ≠ 0.
⋯ ⋯ ⋯ ⋯ 
0 0 ⋯ 0 
For any X ∈ annr (n),

 0 0 ⋯ 0    0 0 ⋯ M p ×q (K [M1n])   0 0 ⋯
1 n 0 
 0
n ∘  ⋯ 0 ⋯ 0    0 0 ⋯ M p ×q (K [M2n]) P ⋯0 0 ⋯ 0 
⋯ X  ⊆ 2
 n
⋯ ⋅X
 0
⋯ ⋯
εn  
 ⋯ ⋯ ⋯ ⋯ ⋯ ⋯
 0 0 ⋯ εn 

 0 ⋯
 0 0 ⋯ 0   
 0 0 ⋯ M p ×q (K [M1n]) 
1 n

⊆  0 0 ⋯ M p ×q (K [M2n])⋅X = n⋅X = 0.
2 n

⋯ ⋯ ⋯ ⋯ 
0 0 ⋯ 0 
Hence,

0 0 ⋯ 0
0 0 ⋯ 0
⋯ X ∈ annr (n) ⊓ K0 [J ] = annr (n),
J
⋯ ⋯ ⋯
0 
 0 ⋯ εn 
346  Junying Guo and Xiaojiang Guo

where annrJ (n) is the right annihilators of n in K0[J]. From (6), it follows that

 M p ×q (K [T1])
1 1
M p1 × q2 (K [M12]) ⋯ M p1 × qn (K [M1n]) 
0 0 ⋯ 0
 0 M p2 × q2 (K [M22]) ⋯ M p2 × qn (K [M2n]) 
0 0 ⋯ 0
⋯ ⋯ ⋯ ⋯ X ∈  ⋯ ⋯ ⋯ ⋯ 
0 εn  ⋯ M pn −1 × qn (K [Mn − 1, n])
 0 ⋯  0 0
 0 0 ⋯ 0 
and so

0 0 ⋯ 0 0 0 ⋯ 0   0 0 ⋯ 0 
0 0 ⋯ 0 X =  0 0 ⋯ 0   0 0 ⋯ 0 
⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ∘  ⋯ ⋯ ⋯ ⋯ X 
0 εn  0 εn    0 εn  
 0 ⋯  0 ⋯ 0 ⋯
0 0 ⋯ 0   0 0 ⋯ 0 
0 0 ⋯ 0   0 0 ⋯ 0 
= ⋯ ⋯ ⋯ ⋯ P  ⋯ ⋯ ⋯ ⋯ X 
0 εn    0 εn  
 0 ⋯ 0 ⋯
 B11 B12 ⋯ B1n 
0 ⋯ 0 
⋯ B2n 
0
0 B
0
= ⋯ 0 ⋯ 0   ⋯ ⋯22 ⋯ ⋯ 
⋯ ⋯ ⋯  
0 ⋯ εn Pnn   0 0 ⋯ Bn − 1, n 
 0 
0 0 ⋯ 0 
= 0,

where,

 B11 B12 ⋯ B1n 

0 0 0 B ⋯ B2n 
0 ⋯
0 0 ⋯ 0  X =  ⋯ ⋯22 ⋯ ⋯ .
⋯ ⋯ ⋯ ⋯  
0 εn   0 0 ⋯ Bn − 1, n 
 0 ⋯ 
0 0 ⋯ 0 

Thus,

0 0 ⋯ 0
0 0 ⋯ 0
⋯ ⋯ ⋯ ⋯ ∈ ann ℓ (annr (n)).
0 εn 
 0 ⋯

Now, by Lemma 3.2 (2),

0 0 ⋯ 0
0 0 ⋯ 0
⋯ ⋯ ⋯ ⋯ ∈ n + ann ℓ (K0 [S])
0 εn 
 0 ⋯

it follows that ann ℓ (K0 [S]) has an element of the form:

0 0 ⋯ A1 
0 0 ⋯ A2 ,
⋯ ⋯ ⋯ ⋯
 
0 0 ⋯ εn 
 Self-injectivity of semigroup algebras  347

in contradiction to (5). We have now proved that Min = ∅ for i = 1, 2,…,n − 1.

By applying the similar arguments as above to

0 0 ⋯ M p1 × qn −1 (K [M1, n − 1]) 0
0 0 ⋯ M p2 × qn −1 (K [M2, n − 1]) 0
⋯
n − 1 := 
⋯ ⋯ ⋯ ⋯
0 0 ⋯ M pn −2 × qn −1 (K [Mn − 2, n − 1]) 0
0 0 ⋯ 0 0
0 0 ⋯ 0 0

and

0 0 ⋯ 0 0
0 0 ⋯ 0 0
⋯ ⋯ ⋯ ⋯ ⋯,
0 0 ⋯ εn − 1 0
 
0 0 ⋯ 0 0

we may verify that Mi,n−1 = ∅ for i = 1, 2,…,n − 2. Continuing this process, we can show that Mij = ∅ for j =
1, 2,…,i − 1, i = 2,…,n − 2. Thus, Mij = ∅ whenever i = j. By [12, Proposition 2.4 (7)], J is a regular
subsemigroup of S. □

Lemma 4.5. Let  be a proper ideal of an algebra . If

(i)  has a left identity e; and
(ii) is right self-injective,
then /  is right self-injective.

Proof. Assume that =  . As pointed out in the Introduction, by hypothesis that is right self-injective,
has a left identity and let ε be a left identity of . Then, = e ⊕ (ε − e) as right 1-modules. Thus,
(ε − e) is an injective right 1-module. Hence, /  ≅ (ε − e) and is an injective right 1-module.
Now, let J be a right ideal of ( / )1 and ϕ a ( / )1-module homomorphism of J into /  . Observe
that the 1-module and 1/  -module structures on /  coincide. Notice that 1/  ≅ ( / )1. If we
identity 1/  with ( / )1, then the inclusion mapping ι: J → ( / )1 is an injective 1 -module
homomorphism and ϕ also an 1-module homomorphism of J into /  . By /  is an injective right
1-module, there exists an 1 -module homomorphism φ of ( / )1 into /  such that ϕ = φι. On the
other hand, for any x +  ∈ 1/  , x +  = (1 + )⋅x , it follows that there is u +  ∈ /  such that
φ (x + ) = (u + )⋅x . From (u + )⋅x = (u + )(x + ), it follows that φ is indeed an ( / )1-module
homomorphism. Therefore, /  is a right injective ( / )1-module, when /  is right self-injective. □

Lemma 4.6. Let S be a semigroup and U an ideal. Then S is a regular semigroup if and only if U and the Rees
quotient S/U are both regular.

Proof. We only verify the sufficiency. To the end, we assume that U and S/U are both regular. For any a ∈
S, if a ∈ U, then a is regular in S; if a ∈ S/U, then as S/U is regular, there is b ∈ S/U such that a ∘ b ∘ a = a in
S/U, in this case, by the definition of Rees quotient, a ∘ b ∘ a = aba and so aba = a in S, it follows that a is
regular in S. However, a is regular in S. Thus, S is a regular semigroup. □

We arrive now at the main result of this section, which generalizes the main result of Okniński
on right (respectively, left) self-injective algebras of a regular semigroup (see [7, Theorem 2], which
answers affirmatively the Okniński’s problem mentioned in the Introduction for the IC abundant
semigroup case.
348  Junying Guo and Xiaojiang Guo

Theorem 4.7. Let S be an IC abundant semigroup and K a field. If K0[S] is right (respectively, left) self-
injective, then S is a finite regular semigroup. In this case, K0[S] is Artinian.

Proof. By Lemma 4.1, we pick a primitive idempotent f1 of S. By Lemma 2.2, S1 = J*(f1) is a primitive
abundant ideal of S.

If S = S1, then by Theorem 3.6, S is a finite regular semigroup.

Suppose that S1 is a proper ideal of S. Then by Lemma 4.4, S1 is a regular subsemigroup of S; and by
Lemmas 4.3 and 4.5, K0[S/S1] is right self-injective.

Case (i). If S/S1 is primitive, then by Theorem 3.6, S/S1 is regular and by Lemma 4.6, S is regular.

Case (ii). Assume that T1 := S/S1 is not primitive. By Lemma 4.6, S is regular if and only if T1 is regular. On
the other hand, by the definition of Green’s  -relation, it is not difficult to see that for any ideal I of S, Ja ⊆
I for all a ∈ I. This shows that Dr(T1) < Dr(S) where Dr(T) stands for the number of nonzero regular
 -classes of T. By applying the similar argument to T1, there exits a primitive abundant ideal S2 of T1
such that
(1) S2 is a regular semigroup;
(2) T2 = T1/S2 is an IC abundant semigroup (by Lemma 2.1);
(3) S is regular if and only if T2 is regular (by Lemma 4.6);
(4) K0[T2] is right self-injective; and
(5) Dr(T2) < Dr(T1).
This proceedings can continue only finite times since |Dr(S)| < ∞ (by Lemma 4.1). So, there exists a positive
integer r such that
(a) Tr is a primitive abundant semigroup;
(b) K0[(Tr)] is right self-injective;
(c) S is regular if and only if so is Tr.
Again by Theorem 3.6, Tr is a finite regular semigroup. Therefore, S is a regular semigroup. By the
result of [7], for a regular semigroup S, if K0[S] is right self-injective, then S is finite, we get that S is a finite
regular semigroup. We have finished the proof. □

5 Algebras of primitively semisimple semigroups

Following [13], we call an abundant semigroup S to be primitively semisimple if for all a ∈ S, the Rees
quotient J*(a)/I*(a) is primitive. Indeed, by Lemma 2.1, J*(a)/I*(a) is a primitive abundant semigroup. Of
course, S is a completely semisimple semigroup if and only if S is a primitively semisimple semigroup
being regular.
By the definition of Rees quotients, J*(a)/I*(a) is a semigroup whose lying set is Ja⁎ ⊔ {θ} and in which
the multiplication is defined as follows: for any x, y ∈ J*(a)/I*(a)

 xy if x , y , xy ∈ Ja⁎ ;
x⋅y = 
 θ otherwise,

where xy is the product of x and y in S. This shows that

(a) for any x ∈ Ja⁎, x is an idempotent in J*(a)/I*(a) if and only if so is x in S;
(b) for any f , g ∈ E (Ja⁎ ), f ≤ g in J*(a)/I*(a) if and only if f ≤ g in S.
 Self-injectivity of semigroup algebras  349

It follows that for an abunadnt semigroup S, J*(a)/I*(a) is primitive if and only if any two nonzero
idempotents in Ja⁎ are not comparable under ≤. Based on this argument, the following lemma is
immediate.

Lemma 5.1. Let S be an abundant semigroup. Then S is primitively semisimple if and only if any two nonzero
idempotents related by  ⁎ are not comparable under ≤.

Lemma 5.2. Let S be an abundant semigroup and U a *-ideal of S. If a, b ∈ S/U, then

(i) a  ⁎S b if and only if a  ⁎S / U b.
(ii) a  ⁎S b if and only if a  ⁎S / U b.

Proof. (i). Suppose that (a , b) ∈  ⁎S . Because  ⁎ is the smallest equivalence containing  ⁎ and  ⁎ , it
follows from [19, Proposition 5.14, p. 28] that there exist x1, x2,…,x2n−1 ∈ S such that

(a , x1) ∈  ⁎S , (x1, x2) ∈  ⁎S , ⋯, (x2n − 1, b) ∈  ⁎S .

By the definition of *-ideal, a ∉ U implies that x1 ∉ U, whereby x2 ∉ U,… and x2n−1 ∉ U. From Lemma 2.1 (i),
it follows that

(a , x1) ∈  ⁎S / U , (x1, x2) ∈  ⁎S / U , ⋯, (x2n − 1, b) ∈  ⁎S / U ;

that is, (a , b) ∈  ⁎S / U . By interchanging the roles of  ⁎S and  ⁎S / U , we may equally show well the sufficiency.
(ii). By the definition of the relation  ⁎ , a  ⁎b if and only if a ∈ J*(b) and b ∈ J*(a). Suppose now that
(a , b) ∈  ⁎S . Then, a ∈ JS⁎ (b) and b ∈ JS⁎ (a). It follows from [12, Lemma 1.7 (3)] and b ∈ J ⁎ (a) that there are
elements a0, a1,…,an ∈ S, x1,…,xn, y1,…,yn ∈ S1 such that a = a0, b = an and (ai , xi ai − 1 yi) ∈  ⁎S for i = 1,…,n.
Notice that U = ⊔x ∈ U Jx⁎ (by the definition of  ⁎ ) and  ⁎ ⊆  ⁎ . We observe that if (x , y ) ∈  ⁎S , then x ∈ U if
and only if y ∈ U. So, b ∉ U can imply that xnan−1yn ∉ U, whereby xn, yn, an−1 ∉ U, moreover by the same
reason, we may show that xn−1, yn−1, an−2,…,x1, y1, a0 = a ∉ U. Hence, a0, a1,…,an ∈ S/U, x1,…,xn, y1,…, yn ∈
(S/U)1 and (ai , xi ai − 1 yi) ∈  ⁎S / U for i = 1,…,n. Now, by [12, Lemma 1.7 (3)], b ∈ JS⁎/ U (a). Similarly, a ∈ JS⁎ (b)
can imply that a ∈ JS⁎/ U (b). Thus, (a , b) ∈  ⁎S / U . By interchanging the roles of  ⁎S and  ⁎S / U , we may
equally well show the sufficiency. □

Lemma 5.3. Let S be a primitively semismple semigroup and U a *-ideal of S. Then S/U is a primitively
semisimple semigroup.

Proof. For any idempotents e, f ∈ S\U, by the arguments before Lemma 5.1, we know that e ≤ f in the
semigroup S if and only if e ≤ f in the Rees quotient S/U. The remainder of the proof is immediate from
Lemmas 5.1 and 5.2. □

Until now, we does not know whether any primitively semisimple semigroup is an IC abundant
semigroup in the literature. But for algebras of primitively semisimple semigroups, we have the following
theorem.

Theorem 5.4. Let S be a primitively semisimple semigroup and K a field. If K0[S] is a right (resp. left) self-
injective algebra, then S is a finite regular semigroup; that is, S is a finite completely semisimple semigroup.

Proof. By Theorem 4.7, we need to only prove that S is regular since any regular semigroup is always an IC
abundant semigroup. On the set S /  ⁎ = {Ja⁎ : a ∈ S}, define

Ja⁎ ≤J Jb⁎ if J ⁎ (a) ⊆ J ⁎ (b).

350  Junying Guo and Xiaojiang Guo

It is a routine check that ≤J is a partial order on S /  ⁎ . (In what follows, we use Ja⁎ <J Jb⁎ to denote Ja⁎ ≤J Jb⁎
but Ja⁎ = Jb⁎.) Because S is abundant, we know that any nonzero  ⁎ -class of S contains at least one nonzero
regular  -class. It follows from Lemma 4.1 (i) that S /  ⁎ is a finite set. So, S /  ⁎ exists a minimum nonzero
 ⁎ -class Ja⁎0 under ≤J.

Lemma 5.5. I*(a0) = ∅ or I*(a0) = {θ} if S has zero θ.

Proof. Assume that I*(a0) ≠ ∅ and S has zero θ. For any x ∈ I*(a0), since I*(a0) is a *-ideal of S, we get J*(x)
⊆ I*(a0). But, by definition, I*(a0) ⊂ J*(a0), now J*(x) ⊆ I*(a0) ⊂ J*(a0). So then, Jx⁎ <J Ja⁎0 . By the minimality
of Ja⁎0 , Jx⁎ = θ so that x = 0 by Observation (*), whereby I*(a0) = {θ}. □

However, J*(a0) is isomorphic to J*(a0)/I*(a0), so then J*(a0) is a primitive abundant semigroup.

Case (i). If S = J*(a0), then by Theorem 3.6, S is a finite regular semigroup.

Case (ii). If S ≠ J*(a0), then by Lemma 4.4, S0 := J*(a0) is regular and by Lemmas 4.3 and 4.5, K0[S1] ≅
K0[S]/K0[S0] and is right self-injective where S1 = S/J*(a0). Also, Dr(S1) < Dr(S). By Lemma 5.3, S1 is a
primitively semisimple semigroup. To verify that S is regular, it suffices to show that S1 is regular by
Lemma 4.6. By applying the foregoing proof to K0[S1], there exits a primitively semisimple semigroup S2
such that
(i) S is regular if and only if S2 is regular;
(ii) K0[S2] is right self-injective; and
(iii) Dr(S2) < Dr(S1).
This proceedings can continue only finite times since |Dr(S)| < ∞ (by Lemma 4.1). So, there exists a positive
integer r such that
(a) Sr is a primitive abundant semigroup;
(b) S is regular if and only if Sr is regular; and
(c) K0[Sr] is right self-injective.
By Theorem 3.6, Sr is regular. Consequently, S is a regular semigroup.
However, S is a regular semigroup and further a finite regular semigroup by Theorem 4.7 and
hypothesis that K0[S] is right self-injective. □
Recall that a semigroup is said to be semisimple if all its principal factors are 0-simple. Obviously, any
regular semigroup is semisimple. By the proof of (ii) ⇒ (iii) in Theorem [11, Theorem 17, p. 196], each Gi in
Theorem [11, Theorem 17, p. 196] (iv) comes from the completely 0-simple semigroup Si/Si−1 such that
Si/Si−1 is isomorphic to  0 (Gi , Ii , Λi ; Pi), hence Gi is isomorphic to a maximum subgroup of Si/Si−1, so that
Gi may be chosen as a maximum subgroup of S, thus by [18, Theorem 2.20, p. 61], all Gi are just non-
isomorphic nontrivial maximum subgroups of S. By Theorems 3.6, 4.7 and 5.4, and [11, Theorem 17,
p. 196], the following theorem is immediate and extends the main result of Guo and Shum in [14].

Theorem 5.6. Let S be an IC abundant semigroup (respectively, a primitively semisimple semigroup; a

primitive abundant semigroup) and K a field. If K0[S] has an identity, then the following statements are
equivalent:
(i) K0[S] is a left self-injective algebra;
(ii) K0[S] is a right self-injective algebra;
(iii) K0[S] is a quasi-Frobenius algebra;
(iv) K0 [S] ≅ Mn1 (K [G1]) ⊕ Mn2 (K [G 2]) ⊕⋯⊕ Mnr (K [Gr]), where
(a) r ≥ 1, ni ≥ 1;
(b) all Gi are just all non-isomorphic nontrivial maximum subgroups Gi of S and are finite.
 Self-injectivity of semigroup algebras  351

The following example, due to Okniński [7], shows that not all of right self-injective algebras of IC
abundant semigroups have identities.

Example 5.7. Let S = {g,h} be the semigroup of left zeros, and  the field of rational numbers. Obviously,
S is a regular semigroup and of course an IC abundant semigroup. Consider the algebra  [S] =  0 [S] and
the standard extension  [S]1 of  [S] to a  -algebra with unity. It may be shown that for any left ideal I of
 [S]1, any homomorphism of left  [S]1-modules I →  [S], extends to a homomorphism of  [S]1-modules
I  [S]1 →  [S]. Moreover, by computing the right ideals of  [S]1, one can easily check that  [S]1 satisfies
Baer’s condition. Hence,  [S] satisfies Baer’s condition as  [S]1-module, which means that  [S] is left
self-injective. It is easy to see that  [S] has no left identities.

Remark 5.8. Let S be a semisimple semigroup. If K0[S] is right (left) self-injective, then by [11, Theorem 14, p. 194],
S is finite. Note that any finite 0-simple semigroup is a completely 0-simple semigroup (for completely 0-simple
semigroups, see [19, p. 60]). So, all principal factors of S are completely 0-simple semigroups, and hence S is
regular; that is, S is a completely semisimple semigroup. Based on this view, Theorem 5.6 is indeed a
generalization of [11, Theorem 14, p. 194] while Theorem 4.7 is a generalization of [11, Theorem 17, p. 196].

6 An application
Ji [24], and Ji and Luo [25] researched the semisimplicity of orthodox semigroup algebras. We next
consider the semisimplicity of algebras of IC abundant semigroups. Obviously, any semisimple algebra is
right (respectively, left) self-injective. So, the following is an immediate consequence of Theorem 4.7.

Proposition 6.1. Let S be an IC abundant semigroup and K a field. If K[S] is semisimple, then S is a finite
regular semigroup.

The following theorem gives a sufficient and necessary condition for an algebra of IC abundant
semigroup to be semisimple.

Theorem 6.2. Let S be an IC abundant semigroup (respectively, a primitively semisimple semigroup; a primitive
abundant semigroup) and K a field. Then K[S] is semisimple if and only if K0 [S] ≅ Mn1 (K [G1]) ⊕ Mn2 (K [G 2])
⊕⋯⊕ Mnr (K [Gr]) where
(i) r ≥ 1, ni ≥ 1 for i = 1, 2,…,r;
(ii) each Gi is a maximal subgroup of S and each K[Gi] is semisimple.

Proof. We need to only verify the necessity. Assume that K[S] is semisimple. Then, K[S] is a right self-
injective algebra with unity. The rest of proof follows from Theorem 5.6. □

Based on Theorems 3.6, 4.7, and 5.4, we have

Corollary 6.3. Let S be an IC abundant semigroup (respectively, a primitively semisimple semigroup; a

primitive abundant semigroup) and K a field. Then K[S] is semisimple if and only if
(i) K0[S] has an unity;
(ii) K0[S] is left (resp. right) self-injective;
(iii) for any maximum subgroup G of S, K[G] is semisimple.

Notice that any right (left) self-injective algebra has a left (right) identity. By Corollary 5.3, we have
the following.
352  Junying Guo and Xiaojiang Guo

Corollary 6.4. Let S be an IC abundant semigroup (respectively, a primitively semisimple semigroup; a

primitive abundant semigroup) and K a field. Then K[S] is semisimple if and only if
(i) K0[S] is right self-injective;
(ii) K0[S] is left self-injective;
(iii) for any maximum subgroup G of S, K[G] is semisimple.

Acknowledgements: This research is jointly supported by the National Natural Science Foundation of
China (grant: 11761034; 11361027; 11661042); the Natural Science Foundation of Jiangxi Province (grant:
20161BAB201018) and the Science Foundation of the Education Department of Jiangxi Province, China
(grant: GJJ14251).

References
[1] C. Faith, Lectures on injective modules and quotient rings, Lectures Notes in Mathematics No: 49, Springer-Verlag, Berlin,
Heidelberg, New York, 1967.
[2] D. S. Passman, The Algebraic Structures of Group Algebras, 2nd ed., Robert E. Krieger Publishing, Melbourne, 1985.
[3] I. B. Kozuhov, Self-injective semigroup rings of inverse semigroups, Izv. Vyss. Uceb. Zaved. 2 (1981), 46–51. (In Russian).
[4] H. Saito, Semigroup rings construction of Frobenius extensions, J. Reine Angew. Math. 324 (1981), 211–220.
[5] J. Lawrence, A countable self-injective ring is quasi-Frobenius, Proc. Amer. Math. Soc. 65 (1977), 217–220.
[6] J. Okniński, When is the semigroup rings perfect? Proc. Amer. Math. Soc., 89 (1983), 49–51.
[7] J. Okniński, On self-injective semigroup rings, Arch. Math. 43 (1984), 407–411.
[8] J. Okniński, On regular semigroup rings, Proc. Roy. Soc. Edinb. 99A (1984), 145–151.
[9] R. Wenger, Some semigroups having quasi-Frobenius algebras I, Proc. London Math. Soc. 18 (1968), 484–494.
[10] R. Wenger, Some semigroups having quasi-Frobenius algebras II, Canadian J. Math. 21 (1969), 615–624.
[11] J. Okniński, Semigroup Algebras, Monographs and textbooks in pure and applied mathematics, Marcel Dekker, Inc.,
New York, Basel, Hong Kong, 1991.
[12] J. B. Fountain, Abundant semigroups, Proc. London Math. Soc. s3-44 (1982), no. 1, 103–129.
[13] A. El Qallali and J. B. Fountain, Idempotent-connected abundant semigroups, Proc. Roy. Soc. Edinb. 91A (1981), 91–99.
[14] X. J. Guo and K. P. Shum, Ample semigroups and Frobenius algebras, Semigroup Forum 91 (2015), 213–223.
[15] X. J. Guo and L. Chen, Semigroup algebras of finite ample semigroups, Proc. Roy. Soc. Edinb. 142A (2012), 1–19.
[16] J. Y. Guo and X. J. Guo, Algebras of right ample semigroups, Open Math. 16 (2018), 842–861.
[17] A. V. Kelarev, Ring Constructions and Applications, World Scientific, New Jersey, 2002.
[18] A. H. Clifford and G. B. Preston, The Algebraic Theory of Semigroups Vol. 1, Mathematical Surveys No. 7, American
Mathematical Society, Providence, RI, USA, 1961.
[19] J. M. Howie, An Introduction to Semigroup Theory, Academic Press, London, 1976.
[20] J. B. Fountain and V. Gould, Endomorphisms of relatively free algebras with weak exchange properties, Algebra
Universalis 51 (2004), 257–285.
[21] X. J. Guo and Y. F. Luo, The naturally partial orders on abundant semigroups, Adv. Math. (China) 34 (2005), 297–308.
[22] M. V. Lawson, The natural partial order on an abundant semigroup, Proc. Edinb. Math. Soc. 30 (1987), 169–186.
[23] R. S. Pierce, Associative Algebras, Springer-Verlag, World Publishing Corporation, New York Heidelberg Berlin, Beijing,
China, 1986.
[24] Y. Ji,  -unipotent semigroup algebras, Comm. Algebra 46 (2018), 740–755.
[25] Y. Ji and Y. Luo, Semiprimitivity of orthodox semigroup algebras, Comm. Algebra 44 (2016), 5149–5162.