()


@
Applied General Topology

c© Universidad Politécnica de Valencia

Volume 14, no. 1, 2013

pp. 97-113

F-door spaces and F-submaximal spaces

Lobna Dridi, Sami Lazaar and Tarek Turki ∗

Abstract

Submaximal spaces and door spaces play an enigmatic role in topology.
In this paper, reinforcing this role, we are concerned with reaching two
main goals:
The first one is to characterize topological spaces X such that F(X) is
a submaximal space (resp., door space) for some covariant functor F
from the category Top to itself. T0, ρ and FH functors are completely
studied.
Secondly, our interest is directed towards the characterization of maps
f given by a flow (X, f) in the category Set, such that (X, P(f)) is
submaximal (resp., door) where P(f) is a topology on X whose closed
sets are exactly the f-invariant sets.

2010 MSC: 54B30, 54D10, 54F65, 46M15.

Keywords: categories, functors, door spaces, submaximal spaces, primal
spaces.

1. Introduction

Among the oldest separation axioms in topology there are some famous ones
as T0, T1 and T2. The T0−, T1− and T2−reflections of a topological space have
long been of interest to categorical topologists. The first systematic treatment
of separation axioms is due to Urysohn [34].

More detailed discussion was given by Freudenthal and Van Est [35]. The
first separation axiom between T0 and T1 was introduced by J.W. T. Youngs
[39] who encountered it in the study of locally connected spaces.

In 1962, C. E. Aull and W. J. Thron were interested in separation axioms
between T0 and T1-spaces (see [1]).

∗Corresponding author: Sami Lazaar.



98 L. Dridi, S. Lazaar and T. Turki

In 2004, Karim Belaid et al in [4] gave some new separation axioms using
the theory of categories and functors in the goal of studying Wallman compact-
ification.

Definition 1.1. Let i, j be two integers such that 0 ≤ i < j ≤ 2. Let us
denote by Ti the functor from Top to Top which takes each topological space
X to its Ti-reflection (the universal Ti-space associated with X). A topological
space X is said to be T(i,j)-space if Ti(X) is a Tj-space (thus we have three
new types of separation axioms; namely, T(0,1), T(0,2) and T(1,2)).

Definition 1.2. Let C be a category and F, G two (covariant) functors from
C to itself.

(1) An object X of C is said to be a T(F,G)-object if G(F(X)) is isomorphic
with F(X).

(2) Let P be a topological property on the objects of C. An object X of
C is said to be a T(F,P )-object if F(X) satisfies the property P .

One year later, H-P. Künzi and T. A. Richmond generalized the study of [4]
using the Ti−ordered reflections (i ∈ {0, 1, 2}) of a partially ordered topological
space (X, τ, ≤) and characterized ordered topological spaces whose T0-ordered
reflection is T1-ordered (see [28]).

On the other hand, recall that a subset A of a topological space X is locally
closed if A is open in its closure in X, or equivalently is the intersection of an
open subset and a closed subset of X. The study of locally closed sets deals to
important results in topology. An investigation is made in certain aspects of
the most discrete case, where every subset is locally closed.

Definition 1.3 ([8, Definition 1.1]). A space X is called submaximal if every
subset of X is locally closed.

One of the reasons to consider submaximal spaces is provided by the theory
of maximal spaces. (A topological space X is said to be maximal if and only
if for any point x ∈ X, {x} is not open).

In [5], the authors give some characterizations of submaximal spaces.

Theorem 1.4. [5, Theorem 3.1] Let X be a topological space. Then the fol-
lowing statements are equivalent:

(i) X is submaximal;
(ii) S \ S is closed, for each S ⊆ X;
(iii) S \ S is closed and discrete, for each S ⊆ X.

Furthermore, let a door space be a space in which every subset is either open
or closed. Clearly, every door space is submaximal.

Recently, some authors (see [2]) have been interested in topological spaces
X that have a compactification noted K(X) which is a door space (resp.,
submaximal space).

In this paper we mainly expose results concerning the following question:
how can we characterize topological spaces X such that F(X) is a submaximal



F -door spaces and F -submaximal spaces 99

space (resp., door space) where F is a covariant functor from the category Top
to itself? Recall the standard notion of reflective subcategory A of B that
is, a full subcategory such that the embedding A −→ B has a left adjoint
F : B −→ A (called reflection). Further, recall that for all i = 0, 1, 2, 3, 3.1

2
the subcategory Topi of Ti-spaces is reflective in Top, the category of all
topological spaces.

The Tychonoff (resp., functionally Hausdorff) reflection of X will be denoted
by ρ(X) (resp., FH(X)).

Specifically, we are interested in T0, ρ and FH functors.
In the first section of this paper, we characterize T0-door spaces and T0-

submaximal spaces.
The second section is devoted to the characterization of ρ-door (resp., FH-

door) spaces and ρ-submaximal (resp., FH-submaximal) spaces.
In section three, given a flow (X, f) in Set we characterize maps f such that

(X, P(f)) is submaximal (resp., door).

2. T0-door and T0-submaximal spaces

First, let us recall the T0-reflection of a topological space. Let X be a
topological space. We define the binary relation ∼ on X by x ∼ y if and only if

{x} = {y}. Then ∼ is an equivalence relation on X and the resulting quotient
space T0(X) := X/ ∼ is the T0-reflection of X.

The canonical surjection µX : X −→ T0(X) is a quasihomeomorphism.
(A continuous map q : X −→ Y is said to be a quasihomeomorphism if
U 7−→ q−1(U) (resp., C 7−→ q−1(C)) defines a bijection O(Y ) −→ O(X)
(resp., F(Y ) −→ F(X)), where O(X) (resp., F(X)) is the collection of all
open subsets (resp., closed subsets) of X [21]).

Let us give some straightforward remarks about quasihomeomorphisms.

Remarks 2.1.

(1) If f : X −→ Y , g : Y −→ Z are continuous maps such that two of the
three maps f, g, g ◦ f are a quasihomeomorphisms, then so is the third
one.

(2) Let q : X −→ Y be a quasihomeomorphism. Then, according to [4,
Lemma 2.7], the following properties hold.
(a) If X is a T0-space, then q is one-one.
(b) If Y is a TD-space, then q is onto.
(c) If Y is a TD-space and X is a T0-space, then q is a homeomorphism.
(d) If X is sober and Y is a T0-space, then q is a homeomorphism.

Now, we introduce some new notations.

Notations 2.2. Let X be a topological space, a ∈ X and A ⊆ X. We denote
by:

(1) d0(a) := {x ∈ X : {x} = {a}}.
(2) d0(A) := ∪[d0(a); a ∈ A].



100 L. Dridi, S. Lazaar and T. Turki

The following remarks follow immediately.

Remarks 2.3. Let X be a topological space and let A be a subset of X. Then
the following properties hold:

(i) d0(A) = µ
−1
X (µX(A)).

(ii) d0(∅) = ∅, d0(X) = X, d0(∪[Ai : i ∈ I]) = ∪[d0(Ai) : i ∈ I] and
d0(d0(A)) = d0(A). Consequently, d0 is a Kuratoweski closure.

(iii) A ⊆ d0(A) ⊆ A and consequently d0(A) = A.
(iv) In particular if A is open (resp., closed ), then d0(A) = A. Indeed,

µX is an onto quasihomeomorphism and thus by [15, Lemma 1.1], it is
an open map (resp., a closed map) and µ−1

X
(µX(A)) = A for any open

(resp., closed) subset A of X.

Thus, a characterization of T0-spaces and symmetric spaces in term of d0
will be useful.

Proposition 2.4. Let X be a topological space. Then the following statements
are equivalent:

(i) X is a T0-space;
(ii) For any subset A of X, d0(A) = A;
(iii) For any a ∈ X, d0(a) = {a}.

Proof. (i) =⇒ (ii) If X is a T0-space, then µX is an homeomorphism and thus
d0(A) = µ

−1
X

(µX(A)) = A.
(ii) =⇒ (iii) Straightforward.

(iii) =⇒ (i) {x} = {y} implies that x ∈ d0(y) = {y} and thus x = y.
�

Recall that a symmetric space is a space in which for any x, y ∈ X, we have

x ∈ {y} =⇒ y ∈ {x}. This notion is introduced by N. A. Shanin in [30] and
rediscovered by A. S. Davis in [12]. It is also studied by K. Belaid, O. Echi and
S. Lazaar in [4] and called T(0,1)-spaces.

Proposition 2.5. Let X be a topological space. Then the following statements
are equivalent:

(i) X is a T(0,1)-space;

(ii) For any a ∈ X, d0(a) = {a}.

Proof. (i) =⇒ (ii) Let X be a T(0,1)-space and a ∈ X, then:

d0(a) = {x ∈ X : {x} = {a}} = {x ∈ X : x ∈ {a}} = {a}.

(ii) =⇒ (i) x ∈ {y} implies that x ∈ d0(y) and thus {x} = {y}, therefore

y ∈ {x}. �

Proposition 2.6. Let X be a topological space. Then the following statements
are equivalent:

(i) X is an Alexandroff T(0,1)-space;

(ii) For any subset A of X, d0(A) = A.



F -door spaces and F -submaximal spaces 101

Proof. (i) =⇒ (ii) For any subset A of X, d0(A) = ∪[d0(a) : a ∈ A]. Now,

since X is T(0,1), we get d0(A) = ∪[{a} : a ∈ A] which is closed because X is
an Alexandroff space. Finally, Remarks 2.3 (iii) does the job.

(ii) =⇒ (i) Clearly, X is a T(0,1)-space.
Now, let {Fi : i ∈ I} be a family of closed subsets of X, then:

∪[Fi : i ∈ I] = ∪[Fi : i ∈ I] = ∪[d0(Fi) : i ∈ I] = d0(∪[Fi : i ∈ I]) = ∪[Fi : i ∈ I].

�

Example 2.7. Let X be an infinite set equipped with the co-finite topology.
Clearly X is a T1-space and thus a T(0,1)-space. Then for any a ∈ X we have
d0(a) = {a}.
Now, let m ∈ X and A = X \ {m}. It is easily seen that d0(A) = A 6= A = X.
Therefore a T(0,1)-space need not to be an Alexandroff T(0,1)-space.

Now, we introduce the following definition.

Definition 2.8. Let X be a topological space. X is called a T0-door space if
its T0-reflection is a door space.

Remark 2.9. Since every door space is a T0-space, then every door space is a
T0-door space. The converse does not hold.

Indeed, given a set X = {0, 1} such that {0} = {1}, we can easily see that
T0(X) is a one point space and thus a door space.

However {0} is not open and not closed in X.

The following result gives answer about the question mentioned in the in-
troduction concerning door spaces.

Theorem 2.10. Let X be a topological space. Then the following statements
are equivalent:

(i) X is a T0-door space;
(ii) For any subset A of X, d0(A) is either open or closed.

Proof. (i) =⇒ (ii) Let A be a subset of X. Since X is a T0-door space, then
µX(A) is either open or closed and consequently d0(A) = µ

−1
X (µX(A)) is either

open or closed.
(ii) =⇒ (i) Let µX(A) be a subset of T0(X), where A ⊆ X. Then, d0(A) =

µ−1
X

(µX(A)) is either open or closed in X and thus µX(A) is either open or
closed in T0(X). Therefore, X is a T0-door space. �

The following result is an immediate consequence of Theorem 2.10 and
Proposition 2.6.

Corollary 2.11. Every T(0,1) Alexandroff space is a T0-door space.

Examples 2.12. (1) A T0-door space need not to be a T(0,1) space.
For this let X be a Sierpinski pace {0, 1}. Then X is a T0-space which is

not T1 and thus X is not T(0,1).



102 L. Dridi, S. Lazaar and T. Turki

Clearly, T0(X) = X is a door space.
(2) A T0-door space need not to be an Alexandroff space.
It is sufficient to choose a door space which is not Alexandroff.
For this, let X be an infinite set and m ∈ X. Equip X with the topology

whose closed sets are all subsets of X containing m or all finite subsets of X.
Hence, if we consider a subset A of X, then two cases arise. If m ∈ A, then A
is closed. If not m ∈ X\A and consequently A is open, so X is a door space.
However X\{m} = ∪[{x} : x 6= m] is a union of closed subsets of X which is
not closed.

Now, the same study will be devoted to submaximal spaces. That’s why we
introduce the following definition.

Definition 2.13. Let X be a topological space. X is called a T0-submaximal
space if its T0-reflection is a submaximal space.

Remark 2.14. Since every submaximal space is T0, then every submaximal
space is a T0-submaximal space. The converse does not hold. The example in
Remark 2.9 does the job.

Theorem 2.15. Let X be a topological space. Then the following statements
are equivalent:

(i) X is a T0-submaximal space;
(ii) For any subset A of X, we have: A dense =⇒ d0(A) open;

(iii) For any subset A of X, d0(A)\d0(A) is a closed set of X.

Proof. We need a Lemma:

Lemma 2.16. Let f : X −→ Y be a quasihomeomorphism. Then the following
statements are equivalent:

(i) f is onto;

(ii) For any subset A of Y, we have f−1(A) = f−1(A).

Proof of the Lemma:
(i) =⇒ (ii) Clearly, f−1(A) ⊇ f−1(A) for any subset A of Y .
Conversely, let x be in f−1(A) and U be an open subset of X containing x.

Since f is a quasihomeomorphism, then there exists an open subset V of Y such
that U = f−1(V ). Now, f(x) ∈ V ∩ A and consequently V ∩ A 6= ∅. Consider
a point y = f(y′) in V ∩ A (since f is onto). Clearly, y′ ∈ f−1(V ) ∩ f−1(A) =

U ∩ f−1(A) and thus U ∩ f−1(A) is not empty which implies that x ∈ f−1(A).

(ii) =⇒ (i) Let y be in Y , by (ii), f−1({y}) = f−1({y}). Since f is a

quasihomeomorphism, f−1({y}) is not empty and consequently f−1({y}) is
not empty too, which implies that f is onto.

Proof of the Theorem:
(i) =⇒ (ii) Let A ⊆ X such that A = X. By Remarks 2.3 (iii), d0(A) = X

and thus µ−1
X

(µX(A)) = X.



F -door spaces and F -submaximal spaces 103

Now, according to Lemma 2.16, µ−1
X

(µX(A)) = X, which implies that

µX(A) = T0(X). Since T0(X) is a submaximal space, we get µX(A) open
and consequently d0(A) = µ

−1
X (µX(A)) is an open set of X.

(i) =⇒ (iii) Let A be a subset of X, then

d0(A)\d0(A) = µ
−1
X

(µX(A))\µ
−1
X

(µX(A))

= µ−1X (µX(A))\µ
−1
X (µX(A))

= µ−1X (µX(A)\ µX(A)).

Now, since X is a T0-submaximal space, then µX(A)\ µX(A) is a closed sub-

set of T0(X) and thus µ
−1
X (µX(A)\ µX(A)) is a closed subset of X. Therefore,

d0(A)\d0(A) is closed.
(ii) =⇒ (i) Let A ⊆ X such that µX(A) is a dense subset of T0(X), that

is, µX(A) = T0(X), then µ
−1
X (µX(A)) = X. Now, according to Lemma 2.16,

µ−1
X

(µX(A)) = X which means that d0(A) = X and thus A = X. By (ii),
d0(A) is open and finally µX(A) is open.

(iii) =⇒ (i) Let A be a subset of X such that d0(A)\d0(A) is closed, then

µ−1X (µX(A)\ µX(A)) is a closed subset of X and thus µX(A)\ µX(A) is a closed
subset of T0(X). Therefore, X is a T0-submaximal space. �

3. ρ-door and ρ-submaximal spaces

Let X be a topological space, F a subset of X and x ∈ X. x and F are said
to be completely separated if there exists a continuous map f : X −→ R such
that f(x) = 0 and f(F) = {1}. Now, two distinct points x and y in X are
called completely separated if x and {y} are completely separated.

A space X is said to be completely regular if every closed subset F of X
is completely separated from any point x not in F . Recall that a topological
space X is called a T1-space if each singleton of X is closed. A completely
regular T1-space is called a Tychonoff space [33].

A functionally Hausdorff space is a topological space in which any two
distinct points of this space are completely separated. Remark here that a
Tychonoff space is a functionally Hausdorff space and consequently a Hausdorff
space (T2-space).

Now, for a given topological space X, we define the equivalence relation ∼
on X by x ∼ y if and only if f(x) = f(y) for all f ∈ C(X) (where C(X)
designates the family of all continuous maps from X to R). Let us denote by
X/ ∼ the set of equivalence classes and let ρX : X −→ X/ ∼ be the canonical
surjection map assigning to each point of X its equivalence class. Since every f
in C(X) is constant on each equivalence class, we can define ρ(f) : X/ ∼−→ R
by ρ(f)(ρX(x)) = f(x). One may illustrate this situation by the following
commutative diagram.



104 L. Dridi, S. Lazaar and T. Turki

X

▽

ρ
X

// X/ ∼

ρ(f)
}}z
zz
z
zz
z
z

R
��

f

?
?
?
?
?
?
?
?

Now, equip X/ ∼ with the topology whose closed sets are of the form
∩[ρ(fα)

−1(Fα) : α ∈ I], where fα : X −→ R (resp., Fα) is a continuous
map (resp., a closed subset of R). It is well known that, with this topology,
X/ ∼ is a Tychonoff space (see for instance [36]) and its denoted by ρ(X).

The construction of ρ(X) satisfies some categorical properties:
For each Tychonoff space Y and each continuous map f : X −→ Y , there

exists a unique continuous map f̃ : ρ(X) −→ Y such that f̃ ◦ ρX = f. We will
say that ρ(X) is the ρ-reflection (or Tychonoff-reflection) of X.

From the above properties, it is clear that ρ is a covariant functor from
the category of topological spaces Top into the full subcategory Tych of Top
whose objects are Tychonoff spaces.

On the other hand, the quotient space X/ ∼ which is denoted by FH(X) is
a functionally Hausdorff space.

The construction FH(X) satisfies some categorical properties:
For each functionally Hausdorff space Y and each continuous map f : X −→

Y , there exists a unique continuous map f̃ : FH(X) −→ Y such that f̃◦ρX = f.
We will say that FH(X) is the functionally Hausdorff-reflection of X (or the
FH-reflection of X).

Consequently, it is clear that FH is a covariant functor from the category
of topological spaces Top into the full subcategory FunHaus of Top whose
objects are functionally Hausdorff spaces.

Notations 3.1. Let X be a topological space, a ∈ X and A a subset of X. We
denote by:

(1) dρ(a) := ∩[f
−1(f({a})) : f ∈ C(X)].

(2) dρ(A) := ∪[dρ(a) : a ∈ A].

The following results are immediate.

Proposition 3.2. Let X be a topological space, a ∈ X and A a subset of X.
Then:

(1) dρ(A) = ρ
−1
X (ρX(A)).

(2) dρ(a) is a closed subset of X.
(3) A ⊆ dρ(A) ⊆ ∩[f

−1(f(A)) : f ∈ C(X)].
(4) ∀f ∈ C(X), f(A) = f(dρ(A)).

Now, we give a characterization of functionally Hausdorff spaces in term of
dρ.



F -door spaces and F -submaximal spaces 105

Proposition 3.3. Let X be a topological space. Then the following statements
are equivalent:

(i) X is a functionally Hausdorff space;
(ii) For any subset A of X, dρ(A) = A;
(iii) For any a ∈ X, dρ(a) = {a}.

Proof. (i) =⇒ (ii) If X is a functionally Hausdorff space, then FH(X) = X
and µX is equal to 1X and thus dρ(A) = A.

(ii) =⇒ (iii) Straightforward.
(iii) =⇒ (i) First, remark that dρ(a) = {a} means that for any x ∈ X such

that x 6= a, there exists a continuous map f : X −→ R such that f(x) 6= f(a)
and thus X is a functionally Hausdorff space. �

Using Defintion 1.2 for the functor FH, one may define an other separation
axiom: A space X is called T(0,FH) if its T0-reflection is functionally Hausdorff.

The following result characterize when ρX : X −→ FH(X) is a quasihome-
omorphism.

Proposition 3.4. Let X be a topological space. Then the following statements
are equivalent:

(a) X is a T(0,FH)-space;
(b) The canonical surjection ρX : X −→ FH(X) is a quasihomeomor-

phism.

Proof. (a) =⇒ (b) Since X is a T(0,FH)-space, then T0(X) is a functionally
Hausdorff space and consequently there exists a unique continuous map f :
FH(X) −→ T0(X) making commutative the following diagram

X

▽

ρ
X

// FH(X)

f
yytt
tt
tt
tt
t

T0(X)
""

µX

EEEEEEEEE

That is f ◦ ρX = µX. On the other hand, since FH(X) is a T0-space, there
is a unique continuous map g : T0(X) −→ FH(X) such that g ◦ µX = ρX.
Now, combining the previous equalities we get easily f ◦g = 1T0(X) and g ◦f =
1FH(X) which means that f and g are homeomorphisms and finally ρX is a
quasihomeomorphism.

(b) =⇒ (a) Consider the following commutative diagram

X

�

ρ
X

// FH(X)

1

T0(X)
��

µX

T0(ρX )
// T0(FH(X)) = FH(X)



106 L. Dridi, S. Lazaar and T. Turki

Clearly, T0(ρX) is a quasihomeomorphism between a T0-space and a func-
tionally Hausdorff space. Now, since FH(X) is a TD-space, then according to
Remarks 2.1 (2.c), T0(ρX) is a homeomorphism which implies that T0(X) is
a functionally Hausdorff space. �

Now, we give a characterization of T(0,FH)-spaces in term of dρ.

Proposition 3.5. Let X be a topological space. Then the following statements
are equivalent:

(i) X is a T(0,FH)-space;

(ii) For any a ∈ X, dρ(a) = dρ({a}) = {a}.

Proof. (i) =⇒ (ii) Clearly, dρ(a) is a closed subset of X containing a and thus

{a} ⊂ dρ(a).
Conversely, let x ∈ dρ(a), then f(x) = f(a) for any f ∈ C(X). Now, suppose

that µX(x) 6= µX(a), then since T0(X) is a functionally Hausdorff space, there
exists a continuous map g from T0(X) to R satisfying g(µX(x)) 6= g(µX(a))
and thus g◦µX is a continuous map of C(X) separating x and y, contradiction.

Finally, µX(x) = µX(a), that is, {x} = {a} and consequently x ∈ {a}.
On the other hand, since X is a T(0,FH)-space, then by Proposition 3.4

ρX : X −→ FH(X) is an onto quasihomeomorphism and thus by [15, Lemma

1.1] dρ({a}) = ρ
−1
X

(ρX({a})) = {a}.
(ii) =⇒ (i) Let µX(x) and µX(a) be two distinct points in T0(X), that is,

{a} 6= {x}. Then x /∈ {a} or a /∈ {x} which means that x /∈ dρ(a) or a /∈ dρ(x)
and consequently there exists a continuous map f from X to R separating a

and x. Now, by universality of T0, let f̃ be the unique continuous map from

T0(X) to R such that f̃ ◦ µX = f. Clearly, f̃ is a continuous map separating
µX(x) and µX(a). �

Proposition 3.6. Let X be a topological space. Then the following statements
are equivalent:

(i) X is an Alexandroff T(0,FH)-space;

(ii) For any subset A of X, dρ(A) = A.

Proof. The same proof as in Proposition 2.6. �

Example 3.7. Let R be the real line equipped with usual topology. Clearly R
is a T(0,F H)-space which is not an Alexandroff space. Hence, dρ(a) = {a} = {a}

for any a ∈ R but dρ(Q) = Q 6= Q = R, where Q is the set of rational numbers.

Let us introduce the following definition.

Definition 3.8. Let X be a topological space. X is called a ρ-door (resp.,
FH-door) space if its ρ-reflection (resp., FH-reflection) is a door space.

By the same way as in Theorem 2.10, the following result gives immediately.



F -door spaces and F -submaximal spaces 107

Theorem 3.9. Let X be a topological space. Then the following statements
are equivalent:

(i) X is an FH-door space;
(ii) For any subset A of X, dρ(A) is either open or closed.

Before giving a characterization of ρ-door spaces, let us recall an interesting
result which characterizes Tychonoff spaces in term of zero-sets (resp., cozero-
sets). Let X be a topological space and A ⊆ X. A is called a zero-set if there
exists f ∈ C(X) such that A = f−1({0}). The complement of a zero-set is
called a cozero-set.

Proposition 3.10 ([36, Proposition 1.7]). A space is Tychonoff if and only if
the family of zero-sets of the space is a base for the closed sets (or equivalently,
the family of cozero-sets of the space is a base for the open sets).

Let us state a useful remark.

Remark 3.11. A closed (resp., open) subset of ρ(X) is of the form ∩[ρ(f)−1({0}) :
f ∈ H] (resp., ∪[ρ(f)−1(R⋆) : f ∈ H]) , where H is a collection of continuous
maps f : X −→ R.

Indeed, ρ(X) is a Tychonoff space, then the collection {g−1{0}; g : ρ(X) −→
R continuous} (resp., {g−1(R⋆); g : ρ(X) −→ R continuous}) is a basis of closed
(resp., open) subsets of ρ(X).

According to the universal property of ρ(X), each continuous map g :
ρ(X) −→ R may be written as g = ρ(f) with f = g ◦ ρX.

Theorem 3.12. Let X be a topological space. Then the following statements
are equivalent:

(i) X is a ρ-door space;
(ii) For any subset A of X, dρ(A) is either an intersection of zero-sets or

a union of cozero-sets of X.

Proof. (i) =⇒ (ii) Let A be a subset of X. Since ρ(X) is a door space, then
ρX(A) ⊆ ρ(X) is either open or closed.

• If ρX(A) is closed, then it is equal to
⋂
[ρ(fi)

−1({0}) : i ∈ I] (where
{fi : i ∈ I} is a family of continuous maps from X to R) and consequently
ρ
−1
X

(ρX(A)) =
⋂
[ρ−1

X
(ρ(fi)

−1({0})) : i ∈ I] =
⋂
[f−1i ({0}) : i ∈ I]. There-

fore, dρ(A) is an intersection of zero-sets of X.
• If ρX(A) is open, then it is equal to

⋃
[ρ(gi)

−1(R⋆) : i ∈ J] (where {gi :
i ∈ J} is a family of continuous maps from X to R) and thus ρ−1X (ρX(A)) =⋃
[g−1i (R

⋆) : i ∈ J]. Therefore, dρ(A) is a union of cozero-sets of X.
(ii) =⇒ (i) Conversely, let ρX(A) ⊆ ρ(X), where A is a subset of X.
• If dρ(A) = ρ

−1
X (ρX(A)) is an intersection of zero-sets of X, then let {fi :

i ∈ I} be a family of continuous maps from X to R such that ρ−1X (ρX(A)) =⋂
[ρ−1

X
(ρ(fi)

−1({0})) : i ∈ I]. Now, since ρX is onto, then:



108 L. Dridi, S. Lazaar and T. Turki

ρX(A) = ρX(
⋂
[f−1i ({0}) : i ∈ I])

= ρX(
⋂
[ρ−1X (ρ(fi)

−1({0}) : i ∈ I])
= ρX(ρ

−1
X (

⋂
[(ρ(fi)

−1({0}) : i ∈ I]))
=

⋂
[ρ(fi)

−1({0}) : i ∈ I]

Consequently, ρX(A) is a closed subset of ρX.
• If dρ(A) = ρ

−1
X (ρX(A)) is a union of cozero-sets of X, then let {gi : i ∈

J} be a family of continuous maps from X to R such that ρ−1
X

(ρX(A)) =⋃
[ρ−1X (ρ(gi)

−1(R⋆)) : i ∈ J].
It is clearly seen, by the same way as in the first case, that ρX(A) =⋃
[ρ(gi)

−1(R⋆) : i ∈ J] and thus ρX(A) is an open subset of ρ(X).
Finally, ρX(A) is either open or closed for every subset A of X which means

that ρ(X) is a door space. �

Definition 3.13. Let X be a topological space. X is said to be a ρ-submaximal
(resp., FH-submaximal) space if its ρ-reflection (resp., FH-reflection) is sub-
maximal.

Now, in order to characterize ρ-submaximal spaces and FH-submaximal
spaces, we introduce the following definitions:

Definition 3.14. Let X be a topological space.

(1) A subset V of X is called a functionally open subset of X (for short
F-open) if and only if dρ(V ) is open in X.

(2) A subset V of X is called a functionally dense subset of X (for short
F-dense) if and only if for any F-open subset W of X, dρ(V ) meets
dρ(W).

(3) A nonempty subset V of X is said to be a ρ-dense if g(V ) 6= {0} for
every nonzero continuous map g from X to R.

Remarks 3.15. (1) V is an F-open subset of X if and only if ρX(V ) is an open
subset of FH(X).

(2) Clearly, a dense subset is a ρ-dense subset. The converse does not hold.
Indeed, let X := {0, 1} be the Sierpinski space, then it is easily seen that ρ(X)
is a one point space, that is, any continuous map f from X to R is constant.
Hence, any nonempty subset A of X is ρ-dense. Now, to conclude choose
A = {1}.

(3) Every F-dense subset of X is ρ-dense. Indeed, let U be an open subset of
ρ (X) and A an F-dense subset of X, that is ρX(A) dense in FH(X). Since U =

∪[ρ (f)
−1

(R∗) : f ∈ H], where H is a collection of continuous maps f : X −→

R and ρ−1X (U) = ρ
−1
X (∪[ρ (f)

−1
(R∗) : f ∈ H]) = ∪[ρ−1X

(
ρ (f)

−1
(R∗)

)
:

f ∈ H] = ∪[f−1 (R∗) : f ∈ H], then U is an open subset of FH(X). Thus
ρX(A) ∩ U 6= ∅. Therefore A is ρ-dense.

(4) An F-dense subset of X is not necessary dense. Indeed, let X = {0, 1}
be the Sierpinski space. Clearly {1} is F-dense but not dense.



F -door spaces and F -submaximal spaces 109

Proposition 3.16. Let X be a topological space and A a subset of X. Then
the following statements are equivalent:

(i) A is a ρ-dense subset of X;
(ii) ρX(A) is a dense subset of ρ(X).

Proof. (i) =⇒ (ii) Let A be a ρ-dense subset of X. Then for any nonzero
continuous map g from X to R, we have A∩g−1(R⋆) 6= ∅. So, let a be in A such
that g(a) 6= 0, then ρ(g)(ρX(a)) = g(a) 6= 0 and thus ρX(A)∩ρ(g)

−1(R⋆) 6= ∅.
Now, by Remark 3.11, ρX(A) meets every nonempty open subset of ρ(X).

(ii) =⇒ (i) Let A be a subset of X. Since ρX(A) is dense, then for any
nonzero continuous map g from X to R, we have ρX(A) ∩ ρ(g)

−1(R⋆) 6= ∅
which means that there exists a ∈ A satisfying ρ(g)(ρX(a)) 6= 0 or equivalently
g(a) 6= 0. Therefore, A is a ρ-dense subset of X. �

We are now in a position to give the characterization of ρ-submaximal
spaces.

Theorem 3.17. Let X be a topological space. Then the following statements
are equivalent:

(i) X is a ρ-submaximal space;
(ii) For any subset A of X, we have: A ρ-dense =⇒ dρ(A) is a union of

cozero-sets of X.

Proof. (i) =⇒ (ii) Let A be a ρ-dense subset of X. According to Proposition
3.16, ρX(A) is a dense subset of ρ(X). Since X is a ρ-submaximal space, then
ρX(A) is an open subset of ρ(X) and thus ρX(A) =

⋃
[ρ(f)−1(R⋆) : f ∈ H]

(where H is a subfamily of C(X)). So that ρ−1
X

(ρX(A)) =
⋃
[ρ−1

X
(ρ(f)−1(R⋆)) :

f ∈ H]. Therefore, dρ(A) =
⋃
[f−1(R⋆) : f ∈ H] is a union of cozero-sets of

X.
(ii) =⇒ (i) Conversely, let A be a subset of X such that ρX(A) = ρ(X).

Then, by Proposition 3.16, A is a ρ-dense subset of X and consequently dρ(A)
is a union of cozero-sets of X. Hence there exists a subfamily {fi : i ∈ I} of
C(X) satisfying ρ−1

X
(ρX(A)) =

⋃
[f−1i (R

⋆) : i ∈ I]. Then:

ρX(A) = ρX(
⋃
[f−1i (R

⋆) : i ∈ I])
= ρX(

⋃
[ρ−1X (ρ(fi)

−1(R⋆)) : i ∈ I])
= ρX(ρ

−1
X

(
⋃
[ρ(fi)

−1(R⋆) : i ∈ I]))
=

⋃
[ρ(fi)

−1(R⋆) : i ∈ I]

Finally, ρX(A) is an open subset of ρ(X). �

Theorem 3.18. Let X be a topological space. Then the following statements
are equivalent:

(i) X is FH-submaximal;
(ii) For any F-dense subset A of X, dρ(A) is open.

Proof. (i) =⇒ (ii)
Let A be an F-dense subset of X. First, let us show that ρX(A) is a dense

subset of FH(X). Indeed, consider ρX(U) an open subset of FH(X). Then



110 L. Dridi, S. Lazaar and T. Turki

dρ(U) = ρ
−1
X

(ρX(U)) is open in X and consequently U is an F-open subset of
X. Since A is F-dense, dρ(U) ∩ dρ(A) 6= ∅ and thus ρX(U) ∩ ρX(A) 6= ∅.

Now, since X is FH-submaximal, then ρX(A) is open in FH(X) and dρ(A) =

ρ
−1
X

(ρX(A)) is open in X.
(ii) =⇒ (i)
Let ρX(A) be a dense subset of FH(X), where A is a subset of X, and V

an F-open subset of X, that is dρ(V ) is open in X and thus ρX(V ) is open
in FH(X). Since ρX(A) is dense in FH(X), then ρX(V ) ∩ ρX(A) 6= ∅. Thus
ρ
−1
X (ρX(V )) ∩ ρ

−1
X (ρX(A)) 6= ∅. Hence dρ(V ) ∩ dρ(A) 6= ∅. Therefore, A is

an F-dense subset of X.
Now, by (ii), dρ(A) is open in X and consequently ρX(A) is open in FH(X).

Therefore, FH(X) is a submaximal space. �

4. Alexandroff Topology

According to Kennisson, a flow in a category C is a couple (X, f), where X
is an object of C and f : X −→ X is a morphism, called the iterator (see [25]
and [26]).

Now, let (X, f) be a flow in the category Set. In [16], the author define the
topology P(f) on X with closed sets are exactly those A which are f-invariant
(i.e., f(A) ⊆ A). It is clearly seen that for any subset A of X, the topological
closure A is exactly ∪[fn(A) : n ∈ N]. In particular for any point x ∈ X,

{x} = Of (x) = {f
n(x) : n ∈ N} called the orbit of x by f. One can see

easily that the family {Vf (x) : x ∈ X} is a basis of open sets of P(f), where
Vf (x) := {y ∈ X : f

n(y) = x, for some n in N}.
Clearly, P(f) is an Alexandroff topology on X.
Characterizing maps f such that (X, P(f)) is submaximal, which is one of

our main goals, is given by the following result.

Proposition 4.1. Let (X, f) be a flow in Set. Then the following statements
are equivalent:

(i) (X, P(f)) is a submaximal space;
(ii) f2 = f.

Proof. (i) =⇒ (ii) let x ∈ X. Two cases arise.
• If f(x) = x, then f2(x) = f(x).

• If f(x) 6= x, then x ∈ {f(x)}c and thus f(x) ∈ {x} ⊆ {f(x)}c, con-

sequently {f(x)}c = X. Now, since (X, P(f)) is a submaximal space, then
{f(x)}c is open, equivalently {f(x)} is closed and finally f2(x) = f(x).

(ii) =⇒ (i) Let A be a dense subset of X. Since f2 = f, then any point in
f(X) is closed and thus, since the topology is principal, every subset of f(X)
is closed. In particular every subset of f(A) is closed. On the other hand
A = A ∪ f(A) = X, then Ac is closed (Ac ⊆ f(A)), so A is open. �



F -door spaces and F -submaximal spaces 111

Example 4.2. Consider the map f: N −→ N
n 7−→ n + 1

where N is the set of all

natural numbers including 0.
It is clearly seen that f2 6= f. Now, consider the topological space (N, P(f))
and set A = 2N. A is a dense subset of (N, P(f)) which is not open since for
each n ∈ N \ {0}, we have 2n − 1 ∈ Vf (2n).

Before giving a characterization of maps f such that (X, P(f)) is door, let
us recall that a point x ∈ X is called a fixed point if f(x) = x and we denote
by Fix(f) the family of all fixed points of X.

Proposition 4.3. Let (X, f) be a flow in Set. Then the following statements
are equivalent:

(i) (X, P(f)) is a door space;
(ii) |f(Fix(f)c)| ≤ 1.

Proof. (i) =⇒ (ii) Suppose that |f(Fix(f)c)| ≥ 2. Then there exist two dis-
tinct points x and y in Fix(f)c such that f(x) 6= f(y). Set A = {x, f(y)}.
Clearly A is neither closed (f(x) /∈ A) nor open (f(y) /∈ Ac).

(ii) =⇒ (i)
• If |f(Fix(f)c)| = 0, then Fix(f) = X and thus (X, P(f)) is the discrete

topology which is door.
• If |f(Fix(f)c)| = 1. Let x0 such that f(Fix(f)

c) = {x0} and let us show
that for any point x in X distinct from x0, {x} is open.

Indeed, set A = f−1({x})\{x}. Assume that A 6= ∅ and let y ∈ X such that
y 6= x and f(y) = x. Then y ∈ Fix(f)c and thus f(y) = x0, contradiction.
Hence A is empty and consequently for any point x ∈ X distinct from x0,
Vf (x) = {x} is open. Now, consider a subset C of X, then it is open if x0 /∈ C
and it is closed when it contains x0. Finally (X, P(f)) is a door space. �

Example 4.4. Let Z be the set of all integers and f: Z −→ Z
n 7−→ |n|

where |n|

denotes the absolute value of the integer n.
Then, we have Fix(f) = N and thus |f(Fix(f)c)| = |N \ {0}| > 1.
Now, consider the topological space (Z, P(f)) and set A = {−1, 2} . Clearly A
is neither closed (A = {−1, 1, 2}) nor open (Vf (2) = {2, −2}).

Acknowledgements. The authors gratefully acknowledge helpful corrections,
comments, and suggestions of the referee which reinforce the presentation of
our paper.



112 L. Dridi, S. Lazaar and T. Turki

References

[1] C. E. Aull and W. J. Thron, Separation axioms between T0 and T1, Indag. Math. 24
(1962), 26–37.

[2] K. Belaid, L. Dridi and O. Echi, Submaximal and door compactifications, Topology
Appl. 158 (2011), 1969–1975.

[3] K. Belaid, O. Echi and R. Gargouri, A-spectral spaces, Topology Appl. 138 (2004),
315 − 322.

[4] K. Belaid, O. Echi and S. Lazaar, T(α,β)-spaces and the Wallman compactification, Int.

J. Math. Math. Sc. 68 (2004), 3717–3735.
[5] G. Bezhanishvili, L. Esakia and D. Gabelaia, Some results on modal axiomatization

and definability for topological spaces, Studia Logica. 81 (2005), 325–355.
[6] E. Bouacida, O. Echi, G. Picavet and E. Salhi, An extension theorem for sober spaces

and the Goldman topology, Int. J. Math. Math. Sc. 2003, no. 51 (2003), 3217–3239.
[7] N. Bourbaki, Eléments de mathématiques, topologie générale, chapitres 1 à 4, 1990.
[8] N. Bourbaki, Topologie générale, chapitres , 3rd ed., Actualités Scientifiqueqs et indus-

trielles 1142 (Hermann, Paris, 1961).
[9] C. Cassidy, M. Hebert and J. M. Kelly, Reflective subcategories, localization and factor-

ization systems, J. Austral. Math. Soc (Ser. A) 38 (1985), 387–429.
[10] Y.S. Cho, On reflective subcategories, Kyungpook Math. J. 18 (1978), 143–146.
[11] J. Cincura, Closed structures on reflective subcategories of the category of topological

spaces, Topology Appl. 37 (1990), 237–247.
[12] A. S. Davis, Indexed systems of neighborhoods for general topological spaces, Am. Math.

Mon. 68 (1961), 886–894.
[13] C. Dorsett, Characterizations of spaces using T0-identification spaces, Kyungpook

Math. J. 17 (1977), 175–179.
[14] D. Drake and W. J. Thron, On the representations of an abstract lattice as the family

of closed sets on a topological space, Trans. Amer. Math. Soc. 120, no. 2 (1965), 57–71.
[15] O. Echi, Quasi-homeomorphisms, Goldspectral spaces and Jacspectral spaces, Boll.

Unione Mat. Ital. Sez. B Artic. Ric. Mat. (8) 6, no. 2 (2003), 489–507.
[16] O. Echi, The category of flows of Set and Top, Topology Appl. 159 (2012), 2357–2366.
[17] O. Echi, R. Gargouri and S. Lazaar, On the Hochster dual of topological space, Topology

Proc. 32 (2008), 153–166.
[18] O. Echi and S. Lazaar, Reflective subcategories, Tychonoff spaces, and spectral spaces,

Topology Proc. 34 (2009), 307–319.
[19] O. Echi and S. Lazaar, Universal spaces, Tychonoff and spectral spaces, Math. Proc. R.

Ir. Acad. 109 (2009), 35–48.
[20] P. D. Finch, On the lattice-equivalence of topological spaces, J. Austral. Math. Soc. 6

(1966), 495–511.
[21] A. Grothendieck and J. Dieudonné, Eléments de géométrie algébrique I: le langage des

schémas, Inst. Hautes Etudes Sci. Publ. Math. no. 4, 1960.
[22] A. Grothendieck and J. Dieudonné, Eléments de géométrie algébrique, Die Grundlehren

der mathematischen Wissenschaften, vol. 166, Springer-Verlag, New York, 1971.
[23] M. Hochster, Prime ideal structure in commutative rings, Trans. Amer. Math. Soc. 142

(1969), 43–60.
[24] J. L. Kelly, General topology, D. Van Nostrand Company, Inc., Princeton, NJ, 1955.
[25] J. F. Kennisson, The cyclic spectrum of a boolean flow, Theory Appl. Categ. 10 (2002),

392–409.
[26] J. F. Kennisson, Spectra of finitely generated boolean flows, Theory Appl. Categ. 16

(2006), 434–459.
[27] H-P. A. Künzi, A. E. Mccluskey and T. A. Richmond, Ordered separation axioms and the

Wallman ordered compactification, Pupl. Math. Debrecen 73, no. 3-4 (2008), 361–377.
[28] H-P. A. Künzi and T. A. Richmond, Ti-ordered reflections, Appl. Gen. Topol. 6, no. 2

(2005), 207–216.



F -door spaces and F -submaximal spaces 113

[29] M. W. Mislove, Topology, domain theory and theoretical computer science, Topology
Appl. 89 (1998), 3–59.

[30] N. A. Shanin, On separation in topological spaces, C. R. (Dokl.) Acad. Sci. URSS, n.
Ser. 38 (1943), 110–113.

[31] M. H. Stone, Applications of boolean algebra to topology, Mat. Sb. 1 (1936), 765–772.
[32] W. J. Thron, Lattice-equivalence of topological spaces, Duke Math. J. 29 (1962), 671–

679.
[33] J. W. Tukey, Convergence and uniformity in topology, Annals of Mathematics Studies,

no. 2. Princeton University Press, 1940 Princeton, N. J.
[34] P. Uryshon, Uber the mächtigkeit der zusammenhägenden Mengen, Math. Ann. 94

(1925), 262–295.
[35] W. T. Van Est and H. Freudenthal, Trennung durch stetige funktionen in topologischen

Räumen, Indag. Math. 13 (1951), 359–368.
[36] R. C. Walker, The Stone-Cech compactification, Ergebnisse der Mathamatik Band 83.
[37] Y-M. Wong, lattice-invariant properties of topological spaces, Proc. Amer. Math. Soc.

26, no. 1 (1970), 206–208.
[38] K. W. Yip, Quasi-homeomorphisms and lattice-equivalences of topological spaces, J.

Austral. Math. Soc. 14 (1972), 41–44.
[39] J. W. T. Youngs, A note on separation axioms and their application in the theory of a

locally connected topological space, Bull. Amer. Math. Soc. 49 (1943), 383–385.

(Received May 2012 – Accepted January 2013)

Lobna Dridi (lobna dridi 2006@yahoo.fr)
Department of Mathematics, Tunis Preparatory Engineering Institute. Uni-
versity of Tunis. 1089 Tunis, Tunisia.

Sami Lazaar (salazaar72@yahoo.fr)
Department of Mathematics, Faculty of Sciences of Tunis. University Tunis-El
Manar. “Campus Universitaire” 2092 Tunis, Tunisia.

Tarek Turki (tarek turki@hotmail.com)
Department of Mathematics, Faculty of Sciences of Tunis. University Tunis-El
Manar. “Campus Universitaire” 2092 Tunis, Tunisia.


	F-door spaces and F-submaximal spaces. By L. Dridi, S. Lazaar and T. Turki