On closures in semitopological inverse semigroups with continuous inversion

We study the closures of subgroups, semilattices and different kinds of semigroup extensions in semitopological inverse semigroups with continuous inversion. In particularly we show that a topological group G is H-closed in the class of semitopological inverse semigroups with continuous inversion if...

Повний опис

Збережено в:
Бібліографічні деталі
Дата:2014
Автор: Gutik, O.
Формат: Стаття
Мова:English
Опубліковано: Інститут прикладної математики і механіки НАН України 2014
Назва видання:Algebra and Discrete Mathematics
Онлайн доступ:http://dspace.nbuv.gov.ua/handle/123456789/153347
Теги: Додати тег
Немає тегів, Будьте першим, хто поставить тег для цього запису!
Назва журналу:Digital Library of Periodicals of National Academy of Sciences of Ukraine
Цитувати:On closures in semitopological inverse semigroups with continuous inversion / O. Gutik // Algebra and Discrete Mathematics. — 2014. — Vol. 18, № 1. — С. 59–85. — Бібліогр.: 33 назв. — англ.

Репозитарії

Digital Library of Periodicals of National Academy of Sciences of Ukraine
id irk-123456789-153347
record_format dspace
spelling irk-123456789-1533472019-06-15T01:30:35Z On closures in semitopological inverse semigroups with continuous inversion Gutik, O. We study the closures of subgroups, semilattices and different kinds of semigroup extensions in semitopological inverse semigroups with continuous inversion. In particularly we show that a topological group G is H-closed in the class of semitopological inverse semigroups with continuous inversion if and only if G is compact, a Hausdorff linearly ordered topological semilattice E is H-closed in the class of semitopological semilattices if and only if E is H-closed in the class of topological semilattices, and a topological Brandt λ⁰-extension of S is (absolutely) H-closed in the class of semitopological inverse semigroups with continuous inversion if and only if so is S. Also, we construct an example of an H-closed non-absolutely H-closed semitopological semilattice in the class of semitopological semilattices. 2014 Article On closures in semitopological inverse semigroups with continuous inversion / O. Gutik // Algebra and Discrete Mathematics. — 2014. — Vol. 18, № 1. — С. 59–85. — Бібліогр.: 33 назв. — англ. 1726-3255 2010 MSC:22A05, 22A15, 22A26; 20M18, 20M15. http://dspace.nbuv.gov.ua/handle/123456789/153347 en Algebra and Discrete Mathematics Інститут прикладної математики і механіки НАН України
institution Digital Library of Periodicals of National Academy of Sciences of Ukraine
collection DSpace DC
language English
description We study the closures of subgroups, semilattices and different kinds of semigroup extensions in semitopological inverse semigroups with continuous inversion. In particularly we show that a topological group G is H-closed in the class of semitopological inverse semigroups with continuous inversion if and only if G is compact, a Hausdorff linearly ordered topological semilattice E is H-closed in the class of semitopological semilattices if and only if E is H-closed in the class of topological semilattices, and a topological Brandt λ⁰-extension of S is (absolutely) H-closed in the class of semitopological inverse semigroups with continuous inversion if and only if so is S. Also, we construct an example of an H-closed non-absolutely H-closed semitopological semilattice in the class of semitopological semilattices.
format Article
author Gutik, O.
spellingShingle Gutik, O.
On closures in semitopological inverse semigroups with continuous inversion
Algebra and Discrete Mathematics
author_facet Gutik, O.
author_sort Gutik, O.
title On closures in semitopological inverse semigroups with continuous inversion
title_short On closures in semitopological inverse semigroups with continuous inversion
title_full On closures in semitopological inverse semigroups with continuous inversion
title_fullStr On closures in semitopological inverse semigroups with continuous inversion
title_full_unstemmed On closures in semitopological inverse semigroups with continuous inversion
title_sort on closures in semitopological inverse semigroups with continuous inversion
publisher Інститут прикладної математики і механіки НАН України
publishDate 2014
url http://dspace.nbuv.gov.ua/handle/123456789/153347
citation_txt On closures in semitopological inverse semigroups with continuous inversion / O. Gutik // Algebra and Discrete Mathematics. — 2014. — Vol. 18, № 1. — С. 59–85. — Бібліогр.: 33 назв. — англ.
series Algebra and Discrete Mathematics
work_keys_str_mv AT gutiko onclosuresinsemitopologicalinversesemigroupswithcontinuousinversion
first_indexed 2025-07-14T04:33:58Z
last_indexed 2025-07-14T04:33:58Z
_version_ 1837595509178499072
fulltext Algebra and Discrete Mathematics RESEARCH ARTICLE Volume 18 (2014). Number 1, pp. 59 – 85 © Journal “Algebra and Discrete Mathematics” On closures in semitopological inverse semigroups with continuous inversion Oleg Gutik Communicated by M. Ya. Komarnytskyj Abstract. We study the closures of subgroups, semilattices and different kinds of semigroup extensions in semitopological inverse semigroups with continuous inversion. In particularly we show that a topological group G is H-closed in the class of semitopological inverse semigroups with continuous inversion if and only if G is compact, a Hausdorff linearly ordered topological semilattice E is H-closed in the class of semitopological semilattices if and only if E is H-closed in the class of topological semilattices, and a topological Brandt λ0-extension of S is (absolutely) H-closed in the class of semitopological inverse semigroups with continuous inversion if and only if so is S. Also, we construct an example of an H-closed non-absolutely H-closed semitopological semilattice in the class of semitopological semilattices. 1. Introduction and preliminaries We shall follow the terminology of [2, 8, 12,27,30]. A subset A of an infinite set X is called cofinite in X if X \A is finite. Given a semigroup S, we shall denote the set of idempotents of S by E(S). A semilattice is a commutative semigroup of idempotents. For 2010 MSC: 22A05, 22A15, 22A26; 20M18, 20M15. Key words and phrases: semigroup, semitopological semigroup, topological Brandt λ 0-extension, inverse semigroup, quasitopological group, topological group, semilattice, closure, H-closed, absolutely H-closed. 60 On closures in semitopological inverse semigroups a semilattice E the semilattice operation on E determines the partial order 6 on E: e 6 f if and only if ef = fe = e. This order is called natural. An element e of a partially ordered set X is called minimal if f 6 e implies f = e for f ∈ X. An idempotent e of a semigroup S without zero (with zero 0S) is called primitive if e is a minimal element in E(S) (in (E(S)) \ {0S}). A maximal chain of a semilattice E is a chain which is properly contained in no other chain of E. The Axiom of Choice implies the existence of maximal chains in any partially ordered set. A semigroup S with the adjoined unit [zero] will be denoted by S1 [S0] (cf. [8]). Next, we shall denote the unit (identity) and the zero of a semigroup S by 1S and 0S , respectively. Given a subset A of a semigroup S, we shall denote by A∗ = A \ {0S} and |A| = the cardinality of A. A semigroup S is called inverse if for any x ∈ S there exists a unique y ∈ S such that xyx = x and yxy = y. Such an element y is called inverse of x and it is denoted by x−1. If h : S → T is a homomorphism (or a map) from a semigroup S into a semigroup T and if s ∈ S, then we denote the image of s under h by (s)h. A semigroup homomorphism h : S → T is called annihilating if (s)h = (t)h for all s, t ∈ S. Let S be a semigroup with zero and λ a cardinal > 1. We define the semigroup operation on the set Bλ(S) = (λ× S × λ) ∪ {0} as follows: (α, a, β) · (γ, b, δ) = { (α, ab, δ), if β = γ; 0, if β 6= γ, and (α, a, β)·0 = 0·(α, a, β) = 0·0 = 0, for all α, β, γ, δ ∈ λ and a, b ∈ S. If S = S1 then the semigroup Bλ(S) is called the Brandt λ-extension of the semigroup S [13]. Obviously, if S has zero then J = {0} ∪ {(α, 0S , β) | 0S is the zero of S} is an ideal of Bλ(S). We put B0 λ(S) = Bλ(S)/J and the semigroup B0 λ(S) is called the Brandt λ0-extension of the semigroup S with zero [19]. Next, if A ⊆ S then we shall denote Aαβ = {(α, s, β) | s ∈ A} if A does not contain zero, and Aα,β = {(α, s, β) | s ∈ A \ {0}} ∪ {0} if 0 ∈ A, for α, β ∈ λ. We shall denote the semigroup of λ×λ-matrix units by Bλ and the subsemigroup of λ×λ-matrix units of the Brandt λ0-extension of a monoid S with zero by B0 λ(1). We always consider the Brandt λ0-extension only O. Gutik 61 of a monoid with zero. Obviously, for any monoid S with zero we have B0 1(S) = S. Note that every Brandt λ-extension of a group G is isomor- phic to the Brandt λ0-extension of the group G0 with adjoined zero. The Brandt λ0-extension of the group with adjoined zero is called a Brandt semigroup [8, 27]. A semigroup S is a Brandt semigroup if and only if S is a completely 0-simple inverse semigroup [7, 25] (cf. also [27, Theo- rem II.3.5]). We also observe that the semigroup Bλ of λ× λ-matrix units is isomorphic to the Brandt λ0-extension of the two-element monoid with zero S = {1S , 0S} and the trivial semigroup S (i. e. S is a singleton set) is isomorphic to the Brandt λ0-extension of S for every cardinal λ > 1. Let {Sι : ι ∈ I} be a disjoint family of semigroups with zero such that 0ι is zero in Sι for any ι ∈ I. We put S = {0} ∪ ⋃ {S∗ ι : ι ∈ I}, where 0 /∈ ⋃ {S∗ ι : ι ∈ I}, and define a semigroup operation “ · ” on S in the following way s · t = { st, if st ∈ S∗ ι for some ι ∈ I; 0, otherwise. The semigroup S with the operation “ · ” is called an orthogonal sum of the semigroups {Sι : ι ∈ I} and in this case we shall write S = ∑ ι∈I Sι. A non-trivial inverse semigroup is called a primitive inverse semi- group if all its non-zero idempotents are primitive [27]. A semigroup S is a primitive inverse semigroup if and only if S is an orthogonal sum of Brandt semigroups [27, Theorem II.4.3]. In this paper all topological spaces are Hausdorff. If Y is a subspace of a topological space X and A ⊆ Y , then by clY (A) we denote the topological closure of A in Y . A (semi)topological semigroup is a Hausdorff topological space with a (separately) continuous semigroup operation. A topological semigroup which is an inverse semigroup is called an inverse topological semigroup. A topological inverse semigroup is an inverse topological semigroup with continuous inversion. We observe that the inversion on a (semi)topological inverse semigroup is a homeomorphism (see [10, Proposition II.1]). A semitopological group is a Hausdorff topological space with a separately continuous group operation. A semitopological group with continuous inversion is a quasitopological group. A paratopological group is called a group with a continuous group operation. A paratopological group with continuous inversion is a topological group. Let STSG0 be a class of semitopological semigroups. A semigroup S ∈ STSG0 is called H-closed in STSG0, if S is a closed subsemigroup 62 On closures in semitopological inverse semigroups of any topological semigroup T ∈ STSG0 which contains S both as a subsemigroup and as a topological space. The H-closed topological semigroups were introduced by Stepp in [31], and there they were called maximal semigroups. A semitopological semigroup S ∈ STSG0 is called absolutely H-closed in the class STSG0, if any continuous homomorphic image of S into T ∈ STSG0 is H-closed in STSG0. An algebraic semigroup S is called: • algebraically complete in STSG0, if S with any Hausdorff topology τ such that (S, τ) ∈ STSG0 is H-closed in STSG0; • algebraically h-complete in STSG0, if S with discrete topology d is absolutely H-closed in STSG0 and (S, d) ∈ STSG0. Absolutely H-closed topological semigroups and algebraically h-complete semigroups were introduced by Stepp in [32], and there they were called absolutely maximal and algebraic maximal, respectively. Recall [1], a topological group G is called absolutely closed if G is a closed subgroup of any topological group which containsG as a subgroup. In our terminology such topological groups are called H-closed in the class of topological groups. In [28] Raikov proved that a topological group G is absolutely closed if and only if it is Raikov complete, i.e. G is complete with respect to the two-sided uniformity. A topological group G is called h-complete if for every continuous homomorphism h : G → H the subgroup f(G) of H is closed [9]. In our terminology such topological groups are called absolutely H-closed in the class of topological groups. The h-completeness is preserved under taking products and closed central subgroups [9]. H-closed paratopological and topological groups in the class of paratopological groups studied in [29]. In [32] Stepp studied H-closed topological semilattice in the class of topological semigroups. There he proved that an algebraic semilattice E is algebraically h-complete in the class of topological semilattices if and only if every chain in E is finite. In [23] Gutik and Repovš established the closure of a linearly ordered topological semilattice in a topological semilattice. They proved the criterium ofH-closedness of a linearly ordered topological semilattice in the class of topological semilattices and showed that every H-closed topological semilattice is absolutely H-closed in the class of topological semilattices. Also, such semilattices studied in [6, 14]. In [3] the structure of closures of the discrete semilattices (N,min) and (N,max) is described. Here the authors constructed an example of an H- closed topological semilattice in the class of topological semilattices which O. Gutik 63 is not absolutely H-closed in the class of topological semilattices. The constructed example gives a negative answer on Question 17 from [32]. Definition 1.1 ([19]). Let STSG0 be a class of semitopological semi- groups. Let λ > 1 be a cardinal and (S, τ) ∈ STSG0. Let τB be a topology on B0 λ(S) such that a) ( B0 λ(S), τB ) ∈ STSG0; b) the topological subspace (Sα,α, τB|Sα,α ) is naturally homeomorphic to (S, τ) for some α ∈ λ. Then ( B0 λ(S), τB ) is called a topological Brandt λ0-extension of (S, τ) in STSG0. In the paper [24] Gutik and Repovš established homomorphisms of the Brandt λ0-extensions of monoids with zeros. They also described a category whose objects are ingredients in the constructions of the Brandt λ0-extensions of monoids with zeros. Here they introduced fi- nite, compact topological Brandt λ0-extensions of topological semigroups and countably compact topological Brandt λ0-extensions of topological inverse semigroups in the class of topological inverse semigroups, and established the structure of such extensions and non-trivial continuous homomorphisms between such topological Brandt λ0-extensions of topo- logical monoids with zero. There they also described a category whose objects are ingredients in the constructions of finite (compact, count- ably compact) topological Brandt λ0-extensions of topological monoids with zeros. These investigations were continued in [20–22], where estab- lished countably compact topological Brandt λ0-extensions of topological monoids with zeros and pseudocompact topological Brandt λ0-extensions of semitopological monoids with zeros their corresponding categories. In the papers [4,15,16,19,26] where studies H-closed and absolutely H-closed topological Brandt λ0-extensions of topological semigroups in the class of topological semigroups. In Section 2 we study the closure of a quasitopological group in a semi- topological inverse semigroup with continuous inversion. In particularly we show that a topological group G is H-closed in the class of semitopo- logical inverse semigroups with continuous inversion if and only if G is compact. Section 3 is devoted to the closure of a semitopological semilattice in a semitopological inverse semigroup with continuous inversion. We show that a Hausdorff linearly ordered topological semilattice E is H-closed 64 On closures in semitopological inverse semigroups in the class of semitopological semilattices if and only if E is H-closed in the class of topological semilattices. Also, we construct an example of an H-closed semitopological semilattice in the class of semitopological semilattices which is not absolutelyH-closed in the class of semitopological semilattices. In Section 4 we show that a topological Brandt λ0-extension of S is (absolutely) H-closed in the class of semitopological inverse semigroups with continuous inversion if and only if so is S. Also, we study the preserving of (absolute) H-closedness in the class of semitopological inverse semigroups with continuous inversion by orthogonal sums. 2. On the closure of a quasitopological group in a semi- topological inverse semigroup with continuous inver- sion Proposition 2.1. Every left topological inverse semigroup with continu- ous inversion is semitopological semigroup. Proof. We write an arbitrary right translation ρa : S → S : x 7→ xa of a left topological inverse semigroup S with continuous inversion inv : S → S on three steps in the following way: ρa(x) = xa = ( a−1x−1 ) −1 = (inv ◦λa−1 ◦ inv) (x). This implies the continuity of right translations i S. It is well known that the closure of an inverse subsemigroup of a topo- logical inverse semigroup is again a topological inverse semigroup (see: [10, Proposition II.1]). The following proposition extends this result to semitopological inverse semigroups with continuous inversion. Proposition 2.2. The closure of an inverse subsemigroup T in a semi- topological inverse semigroup S with continuous inversion is an inverse semigroup. Proof. By Proposition 1.8(ii) from [30, Chapter I, Proposition 1.8(ii)] the closure clS(T ) of T in a semitopological semigroup S is a semitopo- logical semigroup. Then the continuity of the inversion inv : S → S and Theorem 1.4.1 from [11] imply that inv(clS(T )) ⊆ clS(inv(T )) = clS(T ) and hence we get that inv(clS(T )) = clS(T ). This implies that clS(T ) is an inverse subsemigroup of S. O. Gutik 65 We observe that the statement of Proposition 2.2 is not true in the case of inverse topological semigroup. It is complete to consider the set R+ = [0,+∞) of non-negative real numbers with usual topology and usual multiplication of real numbers. This implies that in Proposition 2.2 the condition that S has continuous inversion is essential. In a compact topological semigroup the closure of a subgroup is a topological subgroup (see: [5, Vol. 1, Theorems 1.11 and 1.13]). Also, since for a topological inverse semigroup S the map f : S → S : x → xx−1 is continuous, the maximal subgroup of S is closed, and hence the closure of a subgroup of a topological inverse semigroup is a subgroup. The previous observation implies that this is not true in the general case of topological semigroups. Also, the following example shows that the closure of a subgroup in a semitopological inverse semigroup with continuous inversion is not a subgroup. Example 2.3. Let Z be the discrete additive group of integers. We put A(Z) is the one point Alexandroff compactification of the space Z with the remainder ∞. We extend the semigroup operation from Z onto A(Z) in the following way: n+ ∞ = ∞ + n = ∞ + ∞ = ∞, for every n ∈ Z. It is well known that A(Z) with such defined operation is a semitopological inverse semigroup with continuous inversion and Z is not a closed subgroup of A(Z) [30]. A quasitopological group G is called precompact if for every open neighbourhood U of the neutral element of G there exists a finite subset F of G such that UF = G [2]. The following proposition gives examples quasitopological groups which are non-closed subgroups of some semitopological inverse semigroups with continuous inversion. Proposition 2.4. For every non-precompact regular quasitopological group (G, τ) there exists a regular semitopological inverse semigroup with continuous inversion which contains (G, τ) as a non-closed subgroup. Proof. Since the quasitopological group (G, τ) is non-precompact there exists an open neighbourhood U of the neutral element e of the group G such that FU 6= G and UF 6= G for every finite subset F in G. Let Be be a base of the topology τ at the neutral element e of (G, τ). Since the inversion is continuous in (G, τ), without loss of generality we may 66 On closures in semitopological inverse semigroups assume that all elements of the family Be are symmetric, i.e., V = V −1 for every V ∈ Be. We put BU = {V ∈ Be : clG(V ) ⊆ U} . Since the quasitopological group (G, τ) is not precompact we have that FV 6= G and V F 6= G for every V ∈ BU and for every finite subset F in G. By G0 we denote the group G with a joined zero 0. Now, we put P0 = {Wg,V = {0} ∪G \ clG(gV ) : V ∈ BU , g ∈ G} ∪ {WV,g = {0} ∪G \ clG(V g) : V ∈ BU , g ∈ G} and τ ∪ P0 is a subbase of a topology τ0 on G0. Since (G, τ) a quasitopological group, it is sufficient to show that the semigroup operation on (G0, τ0) is separately continuous in the following two cases: h · 0 = 0 and 0 · h = 0, for h ∈ G. Then for arbitrary subbase neighbourhoods Wg1,V1 , . . . ,Wgn,Vn and WV1,g1 , . . . ,WVn,gn we have that h · (Wg1,V1 ∩ · · · ∩Wgn,Vn ) ⊆ Whg1,V1 ∩ · · · ∩Whgn,Vn and (WV1,g1 ∩ · · · ∩WVn,gn ) · h ⊆ WV1,g1h ∩ · · · ∩WVn,gnh. Also, since translations in the quasitopological group (G, τ) are home- omorphisms, for every open subbase neighbourhood V ∈ BU of the neutral element of G and every g ∈ G we have that (Wg,V )−1 ⊆ WV −1,g−1 . Therefore (G0, τ0) is a quasitopological inverse semigroup with continuous inversion. Now for every open subbase neighbourhoods V1, V2 ∈ BU of the neutral element of G such that clG(V1) ⊆ V2 and every g ∈ G the following conditions holds: clG(Wg,V2 ) ⊆ Wg,V1 and clG(WV2,g) ⊆ WV1,g. Hence we get that the topological space (G0, τ0) is regular. Theorem 2.5. A topological group G is H-closed in the class of semi- topological inverse semigroups with continuous inversion if and only if G is compact. O. Gutik 67 Proof. The implication (⇐) is trivial. (⇒) Let a topological group G be H-closed in the class of semitopo- logical inverse semigroups with continuous inversion. Suppose to the contrary: the space G is not compact. Then G is H-closed in the class of topological groups and hence it is Rǎıkov complete. If G is precompact then by Theorem 3.7.15 of [2], G is compact. Hence the topological group G is not precompact. This contradicts Proposition 2.4. The obtained contradiction implies the statement of our theorem. Theorem 2.5 implies the following two corollaries: Corollary 2.6. A topological group G is absolutely H-closed in the class of semitopological inverse semigroups with continuous inversion if and only if G is compact. Corollary 2.7. A topological group G is H-closed in the class of semi- topological semigroups if and only if G is compact. The following example shows that there exists a non-compact qua- sitopological group with adjoined zero which H-closed in the class of semitopological inverse semigroups with continuous inversion. Example 2.8. Let R be the additive group of real numbers with usual topology. We put G is the direct quare of R with the product topology. It is well known that G is a topological group. Let G0 be the group G with the adjoined zero 0. We define the topology τ on G0 in the following way. For every non-zero element x of G0 the base of the topology τ at x coincides with base of the product topology at x in G. For every (x0, y0) ∈ R2 and every ε > 0 we denote by Oε(x0, y0) = { (x, y) ∈ R2 : √ (x− x0)2 + (y − y0)2 6 ε } the usual closed ε-ball with the center at the point (x0, y0). We denote A(x0, y0) = { (x0, y) ∈ R2 : y ∈ R } ∪ { (x, y0) ∈ R2 : x ∈ R } and Uε(x0, y0) = G0 \ (Oε(x0, y0) ∪A(x0, y0)) . Now we put P(0) = { Uε(x, y) : (x, y) ∈ R2, ε > 0 } and P(0) ∪ BG is a subbase of the topology τ on G0, where BG is a base of the topology of the topological group G. Simple verifications show that (G0, τ) is 68 On closures in semitopological inverse semigroups a Hausdorff semitopological inverse semigroup with continuous inversion and (G0, τ) is not a regular space. Then for any finitely many points (x1, y1), . . . , (xn, yn) ∈ R2 and finitely many ε1, . . . , εn > 0 the following conditions hold: (a) Oε1 (x1, y1) ∪ · · · ∪ Oεn(xn, yn) is a compact subset of the space (G0, τ); (b) clG0(Uε1 (x1, y1)∩· · ·∩Uεn(xn, yn))∪Oε1 (x1, y1)∪· · ·∪Oεn(xn, yn)= G0. This implies that (G0, τ) is an H-closed topological space and hence the semigroup (G0, τ) is H-closed in the class of semitopological inverse semigroups with continuous inversion. 3. On the closure of a semilattice in a semitopological inverse semigroup with continuous inversion It is well known that the subset of idempotent E(S) of a topological semigroup S is a closed subset of S (see: [5, Vol. 1, Theorem 1.5]). We ob- serve that for semitopological semigroups this statement does not hold [30]. Amassing, but the subset of all idempotent E(S) of a semitopological inverse semigroup S with continuous inversion is a closed subset of S. Proposition 3.1. The subset of idempotents E(S) of a semitopological inverse semigroup S with continuous inversion is a closed subset of S. Proof. First we observe that for any topological space X and any con- tinuous map f : X → X the set Fix(f) of fixed point of f is closed subset of X (see: [5, Vol. 1, Theorem 1.4] or [11, Theorem 1.5.4]). Since e−1 = e for every idempotent e ∈ S, the continuity of inversion implies that E(S) ⊆ Fix(inv). Let be x ∈ S such that x ∈ Fix(inv). Since S is an inverse semigroup we obtain that xx = xx−1 ∈ E(S) and hence Fix(inv) ⊆ E(S). This completes the proof of the proposition. Proposition 3.1 implies the following Corollary 3.2. The closure of a subsemilattice in a semitopological inverse semigroup S with continuous inversion is a subsemilattice of S. Since the closure of a subsemilattice in a Hausdorff topological semi- group is again a topological semilattice, an (absolutely) H-closed topo- logical semilattice in the class of topological semilattices is (absolutely) O. Gutik 69 H-closed in the class of topological semigroups [16]. In [32] Stepp proved that an algebraic semilattice E is algebraically h-complete in the class of topological semilattices if and only if every chain in E is finite. The following example shows that for every infinite cardinal λ there exists an algebraically h-complete semilattice E(λ) in the class of topological semilattices of cardinality λ such that E(λ) with the discrete topology is not H-closed in the class of semitopological semigroups. Example 3.3. Let λ be any infinite cardinal. We fix an arbitrary a0 ∈ λ and define the semigroup operation on λ by the formula: xy = { x, if x = y; a0, if x 6= y. The cardinal λ with so defined semigroup operation we denote by E(λ). It is obvious that E(λ) is a semilattice such that a0 is zero of E(λ) and any two distinct non-zero elements of E(λ) are incomparable with respect to the natural partial order on E(λ). Let be a /∈ E(λ). We extend the semigroup operation from E(λ) onto S = E(λ) ∪ {a} in the following way: aa = ax = xa = a0, for any x ∈ E(λ). It is obvious that S with so defined operation is not a semilattice. We define a topology τ on S in the following way. Fix an arbitrary sequence of distinct points {xn : n ∈ N} from E(λ) and put Un(a) = {a} ∪ {xi : i > n}. Put all elements of the set E(λ) are isolated points of the space (S, τ) and the family B(a) = {Un(a) : n ∈ N} is a base of the topology τ at the point a ∈ S. Simple verifications show that (S, τ) is a metrizable 0-dimensional semitopological semigroup and E(λ) is a dense subsemilattice of (S, τ). Also, we observe that by Theorem 9 from [32] the semilattice E(λ) is algebraically h-complete in the class of topological semilattices. Remark 3.4. We observe that for every infinite cardinal λ and every Hausdorff topology τ on E(λ) such that (E(λ), τ) is a semitopological semilattice we have that all non-zero idempotents of (E(λ), τ) are isolated points and moreover (E(λ), τ) is a topological semilattice. Also, a simple modification of the proof in the Example 3.3 shows that a semitopological semilattice (E(λ), τ) is H-closed in the class of semitopological semigroups if and only if the space (E(λ), τ) is compact. Suppose that E is a Hausdorff semitopological semilattice. If L is a maximal chain in E, then by Proposition IV-1.13 of [12] we have that 70 On closures in semitopological inverse semigroups L = ⋂ e∈L(↑e ∪ ↓e) is a closed subset of E and hence we proved the following proposition: Proposition 3.5. The closure of a linearly ordered subsemilattice of a Hausdorff semitopological semilattice E is a linearly ordered subsemilat- tice of E. It is well known that the natural partial order on a Hausdorff semi- topological semilattice is semiclosed (see [12, Proposition IV-1.13]). Also, by Lemma 3 of [33] a semiclosed linear order is closed, and hence every linearly ordered set with a closed order admits the structure of a Hausdorff topological semilattice. This implies the following proposition: Proposition 3.6. Every linearly ordered Hausdorff semitopological semi- lattice is a topological semilattice. Propositions 3.5 and 3.6 imply Theorem 3.7. A Hausdorff linearly ordered topological semilattice E is H-closed in the class of semitopological semilattices if and only if E is H-closed in the class of topological semilattices. Theorem 3.7 and results obtained in the paper [23] imply Corollar- ies 3.8—3.12. A linearly ordered semilattice E is called complete if every non-empty subset of S has inf and sup. Corollary 3.8. A linearly ordered semitopological semilattice E is H- closed in the class of semitopological semilattices if and only if the following conditions hold: (i) E is complete; (ii) x = supA for A = ↓A \ {x} implies x ∈ clE A, whenever A 6= ∅; and (iii) x = inf B for B = ↑B \ {x} implies x ∈ clE B, whenever B 6= ∅ Corollary 3.9. Every linearly ordered H-closed semitopological semilat- tice in the class of semitopological semilattices is absolutely H-closed in the class of semitopological semilattices. Corollary 3.10. Every linearly ordered H-closed semitopological semi- lattice in the class of semitopological semilattices contains maximal and minimal idempotents. O. Gutik 71 Corollary 3.11. Let E be a linearly ordered H-closed semitopological semilattice in the class of semitopological semilattices and e ∈ E. Then ↑e and ↓e are (absolutely) H-closed topological semilattices in the class of semitopological semilattices. Corollary 3.12. Every linearly ordered semitopological semilattice is a dense subsemilattice of an H-closed semitopological semilattice in the class of semitopological semilattices. Remark 3.13. Theorem 3.7, Example 7 and Proposition 8 from [23] imply that there exists a countable linearly ordered σ-compact 0-dimensional scattered locally compact metrizable topological semilattice which does not embeds into any compact Hausdorff semitopological semilattice. At the finish of this section we construct an H-closed semitopologi- cal semilattice in the class of semitopological semilattices which is not absolutely H-closed in the class of semitopological semilattices. A filter F on a set X is called free if ⋂ F = ∅. Example 3.14 ([3]). Let N denote the set of positive integers. For each free filter F on N consider the topological space NF = N ∪ {F} in which all points x ∈ N are isolated while the sets F ∪ {F}, F ∈ F, form a neighbourhood base at the unique non-isolated point F. The semilattice operation min of N extends to a continuous semilat- tice operation min on NF such that min{n,F} = min{F, n} = n and min{F,F} = F for all n ∈ N. By NF,min we shall denote the topological space NF with the semilattice operation min. Simple verifications show that NF,min is a topological semilattice. Then by Theorem 2(i) of [3] the topological semilattice NF,min is H-closed in the class of topological semilattices and hence by Theorem 3.7 it is H-closed in the class of semitopological semilattices. Later by E2 = {0, 1} we denote the discrete topological semilattice with the semilattice operation min. Theorem 3.15. Let F be a free filter on N and F ∈ F be a set with infinite complement N \ F . Then the closed subsemilattice E = ( NF,min × {0} ) ∪ ((N \ F ) × {1}) of the direct product NF,min ×E2 is H- closed not absolutely H-closed in the class of semitopological semilattices. 72 On closures in semitopological inverse semigroups Proof. The definition of the topological semilattice NF,min ×E2 implies that E is a closed subsemilattice of NF,min × E2. Suppose the contrary: the topological semilattice E is not H-closed in the class of semitopological semilattices. Since the closure of a sub- semilattice in a semitopological semilattice is a semilattice (see [30, Chap- ter I, Proposition 1.8(ii)]) we conclude that there exists a semitopological semilattice S which contains E as a dense subsemilattice and S \ E 6= ∅. We fix an arbitrary a ∈ S \E. Then for every open neighbourhood U(a) of the point a in S we have that the set U(a)∩E is infinite. By Theorem 2(i) of [3] and Theorem 3.7, the subspace NF,min × {0} of E with the induced semilattice operation from E is an H-closed in the class of semitopological semilattices. Therefore there exists an open neighbourhood U(a) of the point a in S such that U(a) ∩ E ⊆ (N \ F ) × {1} and hence the set U(a) ∩ ((N \ F ) × {1}) is infinite. Since the subset NF,min × {0} is an ideal of E, the H-closedness of NF,min × {0} in the class of semitopological semilattices implies that NF,min ×{0} is a closed ideal in S and hence we have that x ·a ∈ NF,min × {0} for every x ∈ NF,min × {0}. Since for every open neighbourhood U(a) of the point a in S the set U(a) ∩ ((N\F ) × {1}) is infinite the semilattice operation in E implies that for every x ∈ (NF,min × {0}) \ {(F, 0)} the set x · U(a) is infinite and hence we have that x · a /∈ N × {0} = (NF,min × {0}) \ {(F, 0)}. Therefore we obtain that x · a = (F, 0). Now, since in NF,min the sets F∪{F}, F ∈ F, form a neighbourhood base at the unique non-isolated point F, we conclude that x ·U(a) * (F ∪{F})×{0}, which contradicts the separate continuity of the semilattice operation on S. Hence we get that S \ E = ∅. This implies that the topological semilattice E is H-closed in the class of semitopological semilattices. Now, by Theorem 3 of [3] the topological semilattice E is not absolutely H-closed in the class of topological semilattices, and hence E is not absolutely H-closed in the class of semitopological semilattices. Remark 3.16. Corollary 3.2 implies that the topological semilattice E determined in Theorem 3.15 is an example a topological inverse semi- group which is H-closed but is not absolutely H-closed in the class of semitopological semigroups with continuous inversion. Remark 3.17. Proposition 3.6 and Theorem 3.7 imply that Theorem 2 of [3] describes all H-closed semilattices in the class of semitopological semilattices which contain the discrete semilattice (N,min) or the discrete semilattice (N,max) as a dense subsemilattice. O. Gutik 73 4. On the closure of topological Brandt λ-extensions in a semitopological inverse semigroup with continuous inversion In this section we study the preserving of H-closedness and absolute H-closedness by topological Brandt λ0-extensions and orthogonal sums of semitopological semigroups. Theorem 4.1. Let S be a Hausdorff semitopological inverse monoid with zero and continuous inversion. Then the following conditions are equivalent: (i) S is absolutely H-closed in the class of semitopological inverse semigroups with continuous inversion; (ii) there exists a cardinal λ > 2 such that every topological Brandt λ0- extension of S is absolutely H-closed in the class of semitopological inverse semigroups with continuous inversion; (iii) for each cardinal λ > 2 every topological Brandt λ0-extension of S is absolutely H-closed in the class of semitopological inverse semigroups with continuous inversion. Proof. (i) ⇒ (iii). Suppose that the semigroup S is absolutely H-closed in the class of semitopological inverse semigroups with continuous inversion. We fix an arbitrary cardinal λ > 2. Let B0 λ(S) be a topological Brandt λ0-extension of S in the class of semitopological inverse semigroups with continuous inversion, T be a semitopological inverse semigroup with continuous inversion and h : B0 λ(S) → T be a continuous homomorphism. First we observe that by Proposition 2.3 of [24], either h is an annihi- lating homomorphism or the image (B0 λ(S))h is isomorphic to the Brandt λ0-extension B0 λ((Sα,α)h) of the semigroup (Sα,α)h for some α ∈ λ. If h is an annihilating homomorphism then (Sα,α)h is a singleton, and therefore we have that (Sα,α)h is a closed subset of T . Hence, later we assume that h is a non-annihilating homomorphism. Next we show that for any γ, δ ∈ λ the set (Sγ,δ)h is closed in the space T . By Definition 1.1 there exists α ∈ λ such that (Sα,α)h is a closed subset of T . We define the maps ϕh, ψh : T → T by the formulae (x)ϕh = (α, 1S , γ)h·(x)h·(δ, 1S , α)h and (x)ψh = (γ, 1S , α)h·(x)h·(α, 1S , δ)h. Then the maps ϕh and ψh are continuous because left and right translations in T and homomorphism h : B0 λ(S) → T are continuous maps. Thus, the full preimage A = ((Sα,α)h)ϕ−1 h is a closed subset of T . Then the restriction 74 On closures in semitopological inverse semigroups map (ϕh ◦ψh)|A : A → (Sγ,δ)h is a retraction, and therefore the set (Sγ,δ)h is a retract of A. This implies that (Sγ,δ)h is a closed subset of T . Suppose to the contrary that (B0 λ(S))h is not a closed subsemigroup of T . By Lemma II.1.10 of [27], (B0 λ(S))h is an inverse subsemigroup of T . Since by Proposition 2.2 the closure of an inverse subsemigroup (B0 λ(S))h in a semitopological inverse semigroup T with continuous inversion is an inverse semigroup, without loss of generality we may assume that (B0 λ(S))h is a dense proper inverse subsemigroup of T . We fix an arbitrary x ∈ clT ((B0 λ(S))h) \ (B0 λ(S))h. Then only one of the following cases holds: a) x is an idempotent of the semigroup T ; b) x is a non-idempotent element of T . Suppose that case a) holds. By the previous part of the proof we have that every open neighbourhood U(x) of the point x in the topological space T intersects infinitely sets of the form (Sα,β)h,α, β ∈ λ. By Proposition 2.3 of [24], (B0 λ(S))h is isomorphic to the Brandt λ0-extension B0 λ((Sα,α)h) of the semigroup (Sα,α)h for some α ∈ λ, and since (B0 λ(S))h is a dense subsemigroup of semitopological semigroup T , the zero 0 of the semigroup (B0 λ(S))h is zero of T (see [20, Lemma 23]). Then the semigroup operation of B0 λ((Sα,α)h) implies that either 0 ∈ (α, e, α)h · U(x) or 0 ∈ U(x) · (α, e, α)h for every non-zero idempotent (α, e, α) of B0 λ(S), e ∈ E(S), α ∈ λ. Now by the Hausdorffness of the space T and the separate continuity of the semigroup operation of T we have that either (α, e, α)h · x = 0 or x · (α, e, α)h = 0 for every non-zero idempotent (α, e, α) of B0 λ(S), e ∈ E(S), α ∈ λ. Since in an inverse semigroup any two idempotents commute we conclude that (α, e, α)h · x = x · (α, e, α)h = 0 for every non-zero idempotent (α, e, α) of the semigroup B0 λ(S). We fix an arbitrary non-zero element (α, s, β)h of the semigroup (B0 λ(S))h, where α, β ∈ λ and s ∈ S∗. Then by the previous part of the proof we obtain that x · (α, s, β)h = x · (α, ss−1s, β)h = x · ((α, ss−1, α)(α, s, β))h = x · (α, ss−1, α)h · (α, s, β)h = 0 · (α, s, β)h = 0 and (α, s, β)h · x = (α, ss−1s, β)h · x = ((α, s, β)(β, s−1s, β))h · x = (α, s, β)h · (β, s−1s, β)h · x = (α, s, β)h · 0 = 0. O. Gutik 75 This implies that for every open neighbourhood U(x) of the point x in the space T we have that 0 ∈ x · U(x) and 0 ∈ U(x) · x. Then by Hausdorffness of the space T and the separate continuity of the semigroup operation in T we get that x · x = 0, and hence x = 0. This implies that E(T ) = E((B0 λ(S))h). Suppose that case b) holds. If xx−1 = 0, then x = xx−1x = 0 · x = 0, and similarly if x−1x = 0, then x = xx−1x = x · 0 = 0. This implies that xx−1, x−1x ∈ E((B0 λ(S))h) \ {0}. Then by Lemma I.7.10 of [27] there exist idempotents (α, e, α), (β, f, β) ∈ B0 λ(S) such that xx−1 = (α, e, α)h and x−1x = (β, f, β)h, where e, f ∈ (E(S))∗ and α, β ∈ λ. Then we have that x · (β, f, β)h = (α, e, α)h · x = x. Since x ∈ clT ((B0 λ(S))h) \ (B0 λ(S))h, every open neigh- bourhood U(x) of the point x in the space T intersects infinitely many sets (Sγ,δ)h, γ, δ ∈ λ, and hence we obtain that either U(x) · (β, f, β)h ∋ 0 or (α, e, α)h ·U(x) ∋ 0. Then the Hausdorffness of the space T and the sepa- rate continuity of the semigroup operation on T imply that x·(β, f, β)h = 0 or (α, e, α)h·x = 0. If x·(β, f, β)h = 0 then x = x·xx−1 = x·(β, f, β)h = 0 and if (α, e, α)h · x = 0 then x = xx−1x = (α, e, α)h · x = 0. All these two cases imply that x = 0, and hence we get that T = (B0 λ(S))h, which completes the proof of our theorem. The implication (iii) ⇒ (ii) is trivial. (ii) ⇒ (i). Suppose to the contrary: there exists semigroup S such that S is not absolutely H-closed semigroup S in the class of semitopological inverse semigroups with continuous inversion and condition (ii) holds for S. Then there exists a semitopological inverse semigroup T with continuous inversion and continuous homomorphism h : S → T such that (S)h is non-closed subset of T . Now, by Proposition 2.2 without loss of generality we may assume that (S)h is a proper dense inverse subsemigroup of T . Next, for the cardinal λ we define topologies τB T and τB S on Brandt λ0-extensions B0 λ(T ) and B0 λ(S), respectively, in the following way. We put BT (α,t,β) = {(U(t))α,β : 0 /∈ U(t) ∈ BT (t)} and Bs (α,s,β) = {(U(s))α,β : 0 /∈ U(s) ∈ BS(s)} are bases of topologies τB T and τB S at non-zero elements (α, t, β) ∈ B0 λ(T ) and (α, s, β) ∈ B0 λ(S), respectively, α, β ∈ λ, where BT (t) and BS(s) are bases of topologies of spaces T and S at non-zero elements t ∈ T and s ∈ S, respectively. Also, if BT (0T ) and BS(0S) are bases at zeros 0T ∈ T 76 On closures in semitopological inverse semigroups and 0S ∈ S then we define BT 0 = { {0} ∪ ⋃ α,β∈λ (U(0T ))∗ α,β : U(0T ) ∈ BT (0T ) } and BS 0 = { {0} ∪ ⋃ α,β∈λ (U(0S))∗ α,β : U(0S) ∈ BS(0S) } to be the bases of topologies τB T and τB S at zeros 0 ∈ B0 λ(T ) and 0 ∈ B0 λ(S), respectively. Simple verifications show that if T and S are semitopological in- verse semigroups with continuous inversion, then so are (B0 λ(T ), τB T ) and (B0 λ(S), τB S ). Also the continuity of homomorphism h : S → T implies that the map hB : B0 λ(S) → B0 λ(T ) defined by the formulae (α, s, β)hB = { (α, (s)h, β), if (s)h 6= 0T ; 0, otherwise, s ∈ S∗, α, β ∈ λ, and (0)hB = 0 is continuous. Also, by Theorem 3.10 of [24] so defines map hB : B0 λ(S) → B0 λ(T ) is a homomorphism. The definition of the topology τB T on B0 λ(T ) implies that the homomorphic image (B0 λ(S))hB is a dense proper subsemigroup of the semitopological inverse semigroup (B0 λ(T ), τB T ) with continuous inversion, which contra- dicts to statement (ii). The obtained contradiction implies the requested implication. Now, if we put h is a topological isomorphic embedding of semitopo- logical semigroups with continuous inversions in the proof of Theorem 4.1, then we get the proof of the following theorem: Theorem 4.2. Let S be a Hausdorff semitopological inverse monoid with zero and continuous inversion. Then the following conditions are equivalent: (i) S is H-closed in the class of semitopological inverse semigroups with continuous inversion; (ii) there exists a cardinal λ > 2 such that every topological Brandt λ0-extension of S is H-closed in the class of semitopological inverse semigroups with continuous inversion; (iii) for each cardinal λ > 2 every topological Brandt λ0-extension of S is H-closed in the class of semitopological inverse semigroups with continuous inversion. O. Gutik 77 Theorem 4.1 implies Corollary 4.3 which generalizes Corollary 20 from [17]. Corollary 4.3. For any cardinal λ > 2 the semigroup of λ×λ-units Bλ is algebraically h-complete in the class of semitopological inverse semigroups with continuous inversion. Also, Theorems 4.1 and 4.2 imply the following corollary: Corollary 4.4. For an inverse monoid S with zero the following condi- tions are equivalent: (i) S is algebraically complete (algebraically h-complete) in the class of semitopological inverse semigroups with continuous inversion; (ii) there exists a cardinal λ > 2 such the Brandt λ0-extension of S is algebraically complete (algebraically h-complete) in the class of semitopological inverse semigroups with continuous inversion; (iii) for each cardinal λ > 2 the Brandt λ0-extension of S is algebraically complete (algebraically h-complete) in the class of semitopological inverse semigroups with continuous inversion. Theorems 4.5, 4.6 and 4.7 give a method of the construction of abso- lutely H-closed and H-closed semigroups in the class of semitopological inverse semigroups with continuous inversion. Theorem 4.5. Let S = ⋃ α∈A Sα be a semitopological inverse semigroup with continuous inversion such that (i) Sα is an absolutely H-closed semigroup in the class of semitopological inverse semigroups with continuous inversion for any α ∈ A; and (ii) there exists an ideal T of S which is absolutely H-closed in the class of semitopological inverse semigroups with continuous inversion such that Sα · Sβ ⊆ T for all α 6= β, α, β ∈ A. Then S is an absolutely H-closed semigroup in the class of semitopological inverse semigroups with continuous inversion. Proof. Suppose to the contrary that there exists a semitopological inverse semigroup K with continuous inversion and continuous homomorphism h : S → K such that the image (S)h is not a closed subsemigroup of K. By Lemma II.1.10 of [27], (S)h is an inverse subsemigroup of K. Since by Proposition 2.2 the closure clK((S)h) of an inverse subsemigroup (S)h in 78 On closures in semitopological inverse semigroups a semitopological inverse semigroup K with continuous inversion is an inverse semigroup, without loss of generality we may assume that (S)h is a dense proper inverse subsemigroup of K. We observe that the assumption of the theorem states that T is an ideal of S. This implies that (T )h is an ideal in (S)h. Then by Proposi- tion I.1.8(iii) of [30] the closure of an ideal of a semitopological semigroup is again an ideal, and hence we get that (T )h is a closed ideal of the semigroup K. We fix an arbitrary x ∈ K \ (S)h. Then only one of the following cases holds: a) x is an idempotent of the semigroup K; b) x is a non-idempotent element of K. First we show that x · y, y · x ∈ (T )h for every y ∈ (S)h. We fix an arbitrary open neighbourhood U(x) of the point x in the space K. Since U(x) intersects infinitely many subsemigroups of K from the family {(Sα)h : α ∈ A} we conclude that U(x) · y ∩ (T )h 6= ∅ and y · U(x) ∩ (T )h 6= ∅ for every y ∈ (S)h. Then the separate continuity of the semigroup operation in K implies that any open neighbourhoods W (x · y) and W (y · x) of the points x · y and y · x in K, respectively, intersect the ideal (T )h. This implies that x · y, y · x ∈ clK((T )h). Since the ideal (T )h is closed in K we conclude that x · y, y · x ∈ (T )h. Suppose that case a) holds. Then there exists an open neighbourhood U(x) of the point x in the space K such that U(x) ∩ (T )h = ∅ and the neighbourhood U(x) intersects infinitely many semigroups from the family {(Sα)h : α ∈ A}. By the separate continuity of the semigroup operation in K we have that for every open neighbourhood U(x) of the point x in K such that U(x) ∩ (T )h = ∅ there exists an open neighbourhood V (x) of x in K such that x·V (x) ⊆ U(x) and V (x)·x ⊆ U(x). Now, the previous part of proof implies that x · V (x) ∩ (T )h 6= ∅ and V (x) · x∩ (T )h 6= ∅, which contradict the assumption U(x) ∩ (T )h = ∅. The obtained contradiction implies that E((S)h) = E(K). Suppose that case b) holds. Then there exist idempotents e and f in (S)h such that xx−1 = e and x−1x = f . We observe that e, f /∈ (T )h. Indeed, if e ∈ (T )h or f ∈ (T )h, then we have that x = xx−1x = ex ∈ (T )h and x = xx−1x = xf ∈ (T )h, because (T )h is an ideal of the semigroup K. Since x ∈ clK((S)h), every open neighbourhood of the point x in K intersects infinitely many O. Gutik 79 semigroups from the family {(Sα)h : α ∈ A}, and hence we get that (U(x) · f) ∩ (T )h 6= ∅ and (e · U(x)) ∩ (T )h 6= ∅. Then the Hausdorffness of K and the separate continuity of the semigroup operation in K imply that x = xx−1x = x · f = e · x ∈ (T )h. This contradicts the assumption that x 6= (T )h. The obtained contradiction implies the statement of our theorem. The proof of Theorem 4.6 is similar to the proof of Theorem 4.5. Theorem 4.6. Let S = ⋃ α∈A Sα be a semitopological inverse semigroup with continuous inversion such that (i) Sα is an H-closed semigroup in the class of semitopological inverse semigroups with continuous inversion for any α ∈ A; and (ii) there exists an ideal T of S which is H-closed in the class of semi- topological inverse semigroups with continuous inversion such that Sα · Sβ ⊆ T for all α 6= β, α, β ∈ A. Then S is an H-closed semigroup in the class of semitopological inverse semigroups with continuous inversion. Theorem 4.7. Let a semitopological semigroup S with continuous inver- sion be the orthogonal sum of a family {Sα : α ∈ I} of semitopological inverse semigroups with zeros. Then S is an (absolutely) H-closed semi- group in the class of semitopological inverse semigroups with continuous inversion if and only if so is any element of the family {Sα : α ∈ I}. Proof. First we observe that if S is a semitopological semigroup with con- tinuous inversion then so is every semigroup from the family {Sα : α ∈ I}. The implication (⇐) follows from Theorems 4.5 and 4.6. First we shall prove the implication (⇒) in the case of absolute H- closedness. Suppose to the contrary that there exists an absolute H-closed semi- group S in the class of semitopological inverse semigroups with continuous inversion which is an orthogonal sum of a family {Sα : α ∈ I} of semi- topological inverse semigroups and there exists a semigroup Sα0 in this family such that Sα0 is not absolute H-closed in the class of semitopo- logical inverse semigroups with continuous inversion. Then there exists a semitopological inverse semigroup K with continuous inversion and continuous homomorphism h : Sα0 → K such that the image (Sα0 )h is not 80 On closures in semitopological inverse semigroups a closed subsemigroup of K. By Lemma II.1.10 of [27], (Sα0 )h is an inverse subsemigroup of K. Since by Proposition 2.2 the closure clK((Sα0 )h) of an inverse subsemigroup (Sα0 )h in a semitopological inverse semigroup K with continuous inversion is an inverse semigroup, without loss of gener- ality we may assume that (Sα0 )h is a dense proper inverse subsemigroup of K. Also, the semigroup K has zero because (Sα0 )h contains zero. We define a map f : S → K by the formula (x)f = { 0K , if x ∈ S \ S∗ α0 ; (x)h, if x ∈ S∗ α0 , where 0K is zero of the semigroup K. Simple verifications show that so de- fined map f is a continuous homomorphism, but the image (S)f = (Sα0 )h is a dense proper subsemigroup of K. This contradicts the assumption that the semigroup S is absolutely H-closed semigroup in the class of semitopological inverse semigroups with continuous inversion. Now, we suppose that there exists an H-closed semigroup S in the class of semitopological inverse semigroups with continuous inversion which is an orthogonal sum of a family {Sα : α ∈ I} of semitopological inverse semigroups and there exists a semigroup Sα0 in this family such that Sα0 is not H-closed in the class of semitopological inverse semigroups with continuous inversion. Then there exists a semitopological inverse semigroup K with continuous inversion such that Sα0 is not a closed subsemigroup of K. Since by Proposition 2.2 the closure clK(Sα0 ) of an inverse subsemigroup Sα0 in a semitopological inverse semigroup K with continuous inversion is an inverse semigroup, without loss of generality we may assume that Sα0 is a dense proper inverse subsemigroup of K. Next, we put S′ be the orthogonal sum of the family {Sα :α ∈ I\{α0}} and the semigroup K. We determine a topology τ on S′ in the following way. First we observe if the orthogonal sum T = ∑ i∈J Tj is an inverse Hausdorff semitopological semigroup, then for every non-zero element t ∈ Tj ⊂ T there exists an open neighbourhood U(t) of t in T such that U(t) ⊆ T ∗ j . Indeed, for every open neighbourhood W (t) 6∋ 0 of t in T there exists an open neighbourhood U(t) of t in T such that tt−1 ·U(t) ⊆ W (t). The neighbourhood U(t) is requested. We put that the bases of topologies at any point s of S \ Sα0 and of S′ \K coincide in S and in S′, respectively. Also the bases at any point s of subspace K∗ ⊆ S′ coincide with the base at the point s of K∗. The following family determines the base of the topology τ at zero of the O. Gutik 81 semigroup S′: B0 =            U ⊆ S′ : there exist an element V of the base at zero of the topology of S and an element W of the base at zero of the topology of K such that U ∩ S′ \K = V ∩ S \ Sα0 , U ∩K = W and U ∩ Sα0 = W ∩ Sα0            . Simple verifications show that (S′, τ) is a Hausdorff semitopological inverse semigroup with continuous inversion and moreover S is a dense proper inverse subsemigroup of (S′, τ), which contradicts the assumption of our theorem. The obtained contradiction implies the statement of the theorem. Theorem 4.7 implies the following corollary: Corollary 4.8. A primitive Hausdorff semitopological inverse semigroup S is (absolutely) H-closed in the class of semitopological inverse semi- groups with continuous inversion if and only if so is every its maximal subgroup G with adjoined zero with an induced topology from S. Remark 4.9. We observe that the statements of Theorems 4.5, 4.6 and 4.7 hold for H-closed and absolute H-closed semitopological semilattices in the class of semitopological semilattices. Theorem 4.10. An infinite semitopological semigroup of λ× λ-matrix units Bλ id H-closed in the class of semitopological semigroups if and only if the space Bλ is compact. Proof. Implication (⇐) is trivial. (⇒). Suppose to the contrary that there exists a Hausdorff non- compact topology τB on the semigroup Bλ such that (Bλ, τB) is an H- closed semigroup in the class of semitopological semigroups. By Lemma 2 of [18] every non-zero element of Bλ is an isolated point in (Bλ, τB). Then there exists an infinite open-and-closed subset A ⊆ Bλ \ {0}. Then we have that at least one of the following cases holds: 1) there exist finitely many i1, . . . , in ∈ λ such that if (i, j) ∈ A then i ∈ {i1, . . . , in}; 2) there exist finitely many j1, . . . , jn ∈ λ such that if (i, j) ∈ A then i ∈ {j1, . . . , jn}; 3) cases 1) and 2) don’t hold. 82 On closures in semitopological inverse semigroups Suppose case 1) holds. Then there exists an element i0 ∈ {i1, . . . , in} such that the set {(i0, j) : j ∈ λ} ∩ A is infinite. We denote Ai0 = {(i0, j) ∈ Bλ : (i0, j) ∈ A}. It is obvious that Ai0 is infinite subset of the semigroup Bλ. By Lemma 2 of [18] every non-zero element of Bλ is an isolated point in (Bλ, τB) and hence Ai0 is an open-and- closed subset in the topological space (Bλ, τB). Since the left shift l(i0,i) : Bλ → Bλ : x 7→ (i0, i) · x is a continuous map for any i ∈ λ, Ai = {(i, j) ∈ Bλ : (i0, j) ∈ A} is an infinite open-and-closed subset in (Bλ, τB) for every i ∈ λ. This implies that the set Bλ \ {Ai1 ∪ · · · ∪Aik } is an open neighbourhood of the zero in (Bλ, τB) for every finite subset {i1, . . . , ik} ⊂ λ. Now. for every i ∈ λ we put ai /∈ Bλ. We extend the semigroup operation from Bλ onto the set S = Bλ ∪ {ai : i ∈ λ} in the following way: (i) ai · aj = ai · 0 = 0 · ai = 0 for all i, j ∈ λ; (ii) (s, p) · ai = { as, if p = i; 0, if p 6= i for all (s, p) ∈ Bλ \ {0} and i ∈ λ; (iii) ai · (s, p) = 0 for all (s, p) ∈ Bλ \ {0} and i ∈ λ. Simple verifications show that so defines binary operation on S is asso- ciative, and hence S is a semigroup. Next, we define a topology τS on the semigroup S in the following way. For every element x ∈ Bλ we put that bases of topologies τB and τS at the point x coincide. Also, for every i ∈ λ we put BS(ai) = {{ai} ∪ Ci : Ci is a cofinite subset of Ai} is a base of the topology τS at the point ai ∈ S. It is obvious that (S, τS) is a Hausdorff topological space. The separate continuity of the semigroup operation in (S, τS) follows from the cofinality of the set Ci in Ai for each i ∈ λ. Therefore we get that the semitopological semigroup (Bλ, τB) is a dense proper subsemigroup of (S, τS), which contradicts the assumption of the theorem. In case 2) the proof is similar. Suppose that cases 1) and 2) don’t hold. By induction we construct an infinite sequence {(xi, yi)}i∈N in Bλ in the following way. First we fix an arbitrary element (x, y) ∈ A and denote (x1, y1) = (x, y). Suppose that for some positive integer n we construct the finite sequence {(xi, yi)}i=1,...,n. Since the set A is infinite and cases 1) and 2) don’t hold, there exists (x, y) ∈ A such that x /∈ {y1, . . . , yn} and y /∈ {x1, . . . , xn}. Then we put (xn+1, yn+1) = (x, y). O. Gutik 83 Let a /∈ Bλ. We put T = Bλ ∪{a} and extend the semigroup operation from Bλ onto T in the following way: a · x = x · a = a · a = 0, for every x ∈ Bλ. Next, we define a topology τT on the semigroup T in the following way. For every element x ∈ Bλ we put that bases of topologies τB and τT at the point x coincide. Also, we put BT (a) = {{a} ∪ C : C is a cofinite subset of the set {(xi, yi) : i ∈ N}} is a base of the topology τT at the point a ∈ T . It is obvious that (T, τT ) is a Hausdorff topological space, the semigroup operation in (T, τT ) is separately continuous, and Bλ is a dense subsemigroup of (T, τT ). This contradicts the assumption of the theorem. The obtained contradictions imply the statement of our theorem. Remark 4.11. By Theorem 2 [18] for every infinite cardinal λ there exists a unique Hausdorff pseudocompact topology τc on the semigroup Bλ such that (Bλ, τc) is a semitopological semigroup. This topology is compact and it is described in Example 1 of [18]. Acknowledgements The author acknowledges Oleksandr Ravskyi for his comments and suggestions. References [1] A. D. Aleksandrov, On an extension of Hausdorff space to H-closed, Dokl. Akad. Nauk SSSR. 37 (1942), 138—141 (in Russian). [2] A. Arhangel’skii and M. Tkachenko, Topological Groups and Related Structures, Atlantis Press, 2008. [3] S. Bardyla and O. Gutik, On H-complete topological semilattices, Mat. Stud. 38:2 (2012), 118—123. [4] T. Berezovski, O. Gutik, and K. Pavlyk, Brandt extensions and primitive topological inverse semigroups, Int. J. Math. Math. Sci. 2010 (2010) Article ID 671401, 13 pages, doi:10.1155/2010/671401. [5] J. H. Carruth, J. A. Hildebrant, and R. J. Koch, The Theory of Topological Semigroups, Vols I and II, Marcell Dekker, Inc., New York and Basel, 1983 and 1986. [6] I. Chuchman and O. Gutik, On H-closed topological semigroups and semilattices, Algebra Discr. Math. no. 1 (2007), 13—23. 84 On closures in semitopological inverse semigroups [7] A. H. Clifford, Matrix representations of completely simple semigroups, Amer. J. Math. 64 (1942), 327—342. [8] A. H. Clifford and G. B. Preston, The Algebraic Theory of Semigroups, Vols. I and II, Amer. Math. Soc. Surveys 7, Providence, R.I., 1961 and 1967. [9] D. N. Dikranjan and V. V. Uspenskij, Categorically compact topological groups, J. Pure Appl. Algebra 126 (1998), 149—1168. [10] C. Eberhart and J. Selden, On the closure of the bicyclic semigroup, Trans. Amer. Math. Soc. 144 (1969), 115—126. [11] R. Engelking, General Topology, 2nd ed., Heldermann, Berlin, 1989. [12] G. Gierz, K. H. Hofmann, K. Keimel, J. D. Lawson, M. W. Mislove, and D. S. Scott, Continuous Lattices and Domains, Cambridge Univ. Press, Cambridge, 2003. [13] O. V. Gutik, On Howie semigroup, Mat. Metody Phis.-Mech. Polya. 42:4 (1999), 127—132 (in Ukrainian). [14] O. Gutik, D. Pagon and D. Repov̌s, On chains in H-closed topological pospaces, Order 27:1 (2010), 69—81. [15] O. V. Gutik and K. P. Pavlyk, H-closed topological semigroups and Brandt λ- extensions, Mat. Metody Phis.-Mech. Polya. 44:3 (2001), 20—28 (in Ukrainian). [16] O. Gutik and K. Pavlyk, Topological Brandt λ-extensions of absolutely H-closed topological inverse semigroups, Visn. L’viv. Univ., Ser. Mekh.-Mat. 61 (2003), 98—105. [17] O. V. Gutik and K. P. Pavlyk, On topological semigroups of matrix units, Semigroup Forum. 71:3 (2005), 389—400. [18] O. V. Gutik and K. P. Pavlyk, Topological semigroups of matrix units, Algebra Discr. Math. no. 3 (2005), 1—17. [19] O. V. Gutik and K. P. Pavlyk, On Brandt λ 0-extensions of semigroups with zero, Mat. Metody Phis.-Mech. Polya. 49:3 (2006), 26—40. [20] O. Gutik and K. Pavlyk, On pseudocompact topological Brandt λ 0-extensions of semitopological monoids, Topological Algebra and its Applications 1 (2013), 60—79. [21] O. Gutik, K. Pavlyk, and A. Reiter, Topological semigroups of matrix units and countably compact Brandt λ 0-extensions, Mat. Stud. 32:2 (2009), 115—131. [22] O. Gutik and O. Ravsky, On pseudocompact inverse primitive (semi)topological semigroups, Preprint (arXiv: 1310.4313). [23] O. Gutik and D. Repovš, On linearly ordered H-closed topological semilattices, Semigroup Forum 77:3 (2008), 474—481. [24] O. Gutik and D. Repovš, On Brandt λ 0-extensions of monoids with zero, Semigroup Forum 80:1 (2010), 8—32. [25] W. D. Munn, Matrix representations of semigroups, Proc. Cambridge Phil. Soc. 53 (1957), 5—12. [26] K. P. Pavlyk, Absolutely H-closed topological semigroups and Brandt λ-extensions, Appl. Problems Mech. Math. 2 (2004), 61—68 (in Ukrainian). [27] M. Petrich, Inverse Semigroups, John Wiley & Sons, New York, 1984. O. Gutik 85 [28] D. A. Raikov, On a completion of topological groups, Izv. Akad. Nauk SSSR. 10 (1946), 513—528. [29] O. Ravsky, On H-closed paratopological groups, Visn. L’viv. Univ., Ser. Mekh.-Mat. 61 (2003), 172—179. [30] W. Ruppert, Compact Semitopological Semigroups: An Intrinsic Theory, Lect. Notes Math., 1079, Springer, Berlin, 1984. [31] J. W. Stepp, A note on maximal locally compact semigroups, Proc. Amer. Math. Soc. 20 (1969), 251—253. [32] J. W. Stepp, Algebraic maximal semilattices, Pacific J. Math. 58:1 (1975), 243— 248. [33] L. E. Ward, Jr., Partially ordered topological spaces, Proc. Amer. Math. Soc. 5:1 (1954), 144—161. Contact information O. Gutik Faculty of Mechanics and Mathematics, National University of Lviv, Universytetska 1, Lviv, 79000, Ukraine E-Mail: o_gutik@franko.lviv.ua, ovgutik@yahoo.com Received by the editors: 17.09.2014 and in final form 17.09.2014.