6.Closed set and Limit points

we treat notations of closed set, closure of set and limit point. These lead naturally to consideration of a certain axiom for topological spaces called the Hausdorff axiom.

Closed sets

A subset AA of a topological space XX is said to be closed if the set X−A is open.

ex1) the subset [a,b][a,b] of R is closed because its complement

R−[a,b]=(+∞,a) ∪ (b,∞)

is open. similarly [a,+∞) is closed, because ints complement (−∞,a) is open. these facts justify our use of terms closed interval and closed ray. the subset [a,b)[a,b) of R is neither open nor closed.

ex2) In the plane R2\mathbb R^2, the set

x×yx0andy0{x \times y \, | \, x \ge 0 \; \text{and} \; y\ge 0 }

is closed, because its complement is the union of the two sets

()×RandR×(,0)(-\infty) \times \mathbb R \;\text{and}\; \mathbb R \times (-\infty,0)

each of which is a product of open sets of and is open in R2\mathbb R^2.

ex3) In the finite complement topology on a set XX, the closed sets consist of XX itself and all finite subsets of XX.

ex4) In the discrete topology on the set XX, every set is open; it follows that every set is closed as well.

ex5) consider the following subset of the real line

Y=[0,1](2,3)Y=[0,1]\cup(2,3)

in the subspace topology. In this space, the set [0,1][0,1] is open. similarly (2,3)(2,3) is open as a subset of YY; it's even open in R\mathbb R. since [0,1][0,1] and (2,3)(2,3) are complements in YY of each other, we conclude that both are closed as subset of YY.

Theorem 17.1 Let XX be a topological spaces. then the following conditions hold

  1. XX and ϕ\phi are closed
  2. Arbitrary intersections of closed sets are closed
  3. Finite unions of closed sets are closed

proof.

  1. ϕ\phi and XX are closed because they are the complements of the open sets XX and ϕ\phi, respectively.

  2. Given a collection of closed sets {Aα}αj\{A_\alpha\}_{\alpha \in j}, Applying DeMorgan's law
    Xαj{Aα}=αj(XA)X - \bigcap_{\alpha \in j}\{A_\alpha\}=\bigcup_{\alpha \in j}(X-A)
    since the sets X−A_α are open by definition, the right side represents an arbitrary union of open sets, and is open. therefore Aα\cap A_\alphais closed.

  3. similiarly if AiA_i is closed for i=1, ⋯,n, consider the equation
    Xi=1nAi=i=1n(xAi)X-\bigcup_{i=1}^{n}A_i= \bigcap_{i=1}^n(x-A_i)
    The set on the right side is a finite intersection of open sets and is open. hence Ai\cup A_i is closed.

If YY is a subspace of XX, we say that a set AA is closed in YY if AA is a subset of YY and AA is closed in the subspace topology of YY.

@Fig 17.1 @Fig 17.2

Theorem 17.2 Let YY be a subspace of XX. then a set AA is closed in YY if and only if it equals the intersection of a closed set of XX with YY.

proof. Assume that A=CYA=C \cup Y, where CC is closed in XX (see Fig 17.1) then X−C is open in XX, so that (XC)Y(X-C) \cup Y is open in YY, by definition of the subspace topology, but (XC)Y=(XY)(CY)=YA(X-C) \cap Y=(X \cap Y)-(C \cap Y)=Y-A. hence YAY-A is open in YY. so that AA is closed in YY. conversely, assume that AA is closed in YY(Fig 17.2) then Y−A is open in Y, so by definition it equals the intersection of an open set U of X with Y. the set X−U is closed in X, and

YA=UYY-A=U\cap Y

A=Y(UY)=(XY)(UY)=(XU)YA=Y-(U\cap Y)=(X\cap Y)-(U\cap Y)=(X-U)\cap Y

so that AA equals the intersection of a closed set of XX with YY, as desired.

Theorem 17.3 Let YY be a subspace of XX. If AA is closed in YY and YY is closed in XX, then A is closed in XX.

proof. since AA is closed in YY, by thm 17.2, A=CYA=C\cup Y, where CC is closed in XX. so that A is closed in XX. because YY is closed in XX, by closed condition 2.

Closure and Interior of a set

Given a subset AA of a topological space XX, the interior of AA is defined as the unions of all open set contained in AA. and the closure of AA is defined as the intersections of all closed set containing AA. The interior of AA is denoted by IntAInt\, A and the closure of AA is denoted by ClAClA by A\overline A. obvisously IntAIntA is an open set and A\overline A is a closed set; furthermore,

IntAAAIntA\subset A \subset \overline A

if AA is open, A=IntAA=Int\,A; while AA is closed, A=AA=\overline A.

Theorem 17.4 Let YY be a subspace of XX; Let AA be a subset of YY; Let A\overline A denoted by a closure of AA in XX. then a closure of AA in YY equals AY\overline A \cap Y.

proof. Let B\overline B denoted by a closure of AA in YY. the set A\overline A is closed in XX. so AY\overline A \cap Y is closed in YY by Thm 17.2. since AY\overline A \cap Y contains AA, and since by definition, B\overline B equals the intersections of all closed set containing AA, we must have B(AY)\overline B \subset (\overline A \cap Y).
on the other hand, we know that B\overline B is closed in YY. hence by thm 17.2, B=CY\overline B = C \cap Y for some set CC is closed in XX. then CC is a closed set of XX containing AA; because A\overline A is the intersections of all such closed sets we conclude AC\overline A \subset C then (AY)(CY)=B(\overline A \cap Y ) \subset ( C \cap Y ) = B.

The definition of the closure doesn't give us a convenient way for finding the closures of a specific set.
First let us introduce some convenient terminology. we shall say that a set AA intersects a set BB in the intersection ABA\cap B is not empty.

Theorem 17.5 Let AA be a subset of the topoloical space XX.
(a) then xAx \in \overline A if and only if every open set UU containing xx intersects AA.
(b) supposing the topology of XX is given B\mathcal B by a basis, then xAx \in \overline A if and only if every basis element BB containing xx intersects AA.

proof.consider the statement in (a), It's a statement form PQP \Leftrightarrow Q. Let us transform each implication to its contrapositive, thereby obtaining the logically equivalent statement ¬P¬Q\neg P \Leftrightarrow \neg Q written out, it's following.
xAx \notin \overline A \Leftrightarrow there exists an open set UU containing xx that doesn't intersect AA.
In this form, our theorem is easy to prove.
If xx is not in A\overline A, the set U=XAU = X - \overline A is an open set containg xx that doesn't intersect AA.
conversely, if there exists an open set UU containing xx which doesn't intersect AA, then XUX-U is a closed set containing AA. by definition of the closure A\overline A, the set XUX-U must contain AA; therefore xx cannot be in A\overline A.
Statement (b) follows readily, If every open set containing xx intersects AA, so does every basis element containing xx, because BB is open. conversely if every basis element containing xx intersects AA, so does every open set VV containing xx because VV contains a basis element that contains xx.

we shorten the statement "UU is an open set containing xx" to the phrase
    "UU is a neighborhood of xx"
using this terminology, one can write the first half of the preceding theorem as follows.

If AA is a subset of topological space XX, then xAx \in \overline A if and only iff every neighborhood of xx intersects AA.

Example Let XX be the real line R\mathbb R. If A=(0,1]A=(0,1] for every neighborhood of 0 intersects AA, while every point outside [0,1][0,1] has a neighborhood disjoint from AA. similar arguments apply to the following subset of XX:
If B={1/nnZ+}B=\{1/n\,|\,n \in \mathbb Z_+\}, then B={0}B\overline B = \{0\}\cup B. If C={0}(1,2)C=\{0\}\cup (1,2) then C={0}[1,2]\overline C = \{0\} \cup [1,2]. If Q\mathbb Q is the set of ational numbers then Q=R\mathbb {\overline Q} = \mathbb R, Z+=Z+\overline {\mathbb Z}_+ = \mathbb Z_+. if R+\mathbb R_+ is the set of positive reals, then the closure of R+\mathbb R_+ is R+{0}\mathbb R_+ \cup \{0\}.

Example Consider the subspace Y=(0,1]Y=(0,1] of the real line R\mathbb R. the set A=(0,12)A=(0, \frac{1}{2}) is a subset of YY; its closure in R\mathbb R is the set [0,12][0,\frac{1}{2}] and its closure in YY is the set [0,12]Y=(0,12][0,\frac{1}{2}]\cap Y = (0,\frac{1}{2}].

Limit Points

If AA is a subset of the topological space XX and if xx is a point of XX, we say that xx is a limit point of AA if every neighborhood of xx intersects AA in some point other than xx itself. say differently,
xx is a limit point of AA if it belongs to the closure of A{x}A-\{x\}.

Example Consider the real line in R\mathbb R, If A=(0,1]A=(0,1], then the point 0 is a limit point of AA and so is the point 12\frac{1}{2} In fact every point of the interval [0,1][0,1] is a limit point of AA.
    If B={1nnZ+}B=\{ \frac{1}{n} \,|\, n \in \mathbb Z_+\}, then 0 is the only limit point of BB. if C={0}(1,2)C=\{0\} \cup (1,2) then the limit point of CC are the points of the interval [1,2][1,2]. a limit point of Q\mathbb Q is every point of R\mathbb R. a limit point of Z+\mathbb Z_+ is no point of R\mathbb R. a limit point of R+\mathbb R_+ is a point of {0}R+\{0\} \cup \mathbb R_+.

Theorem 17.6 Let AA be a subset of the topological space XX; Let AA' be the set of all limit points of AA then A=AA\overline A = A \cup A'.

proof. If xx is in AA', every neighborhood of xx intersects AA. therefore, by thm 17.5 xx belongs to A\overline A hence AAA' \subset \overline A, since by definition AAA \subset \overline A, AAAA\cup A' \subset \overline A.
To demonstrate the reverse inclusion, we let xx be a point of A\overline A and show that xAAx \in A \subset A'. if xx happens to lie in AA, it is trivial that xAAx \in A \cup A'; suppose that xx does not lie in AA. since xAx \in \overline A, we know that every neighborhood UU of xx intersects AA; because xAx \notin A, the set UU must intersect AA in a point from different from xx; then xx is a limit point , xAx \in A', so that xAAx \in A \cup A'.

Corollary 17.7 A subset of a topological space is closed if and only if it contains all its limit points.

proof. the set AA is closed if and only if A=AA=\overline A and the latter holds if and only if AAA'\subset A.

Hausdorff spaces

Definition A topological space XX is called Hausdorff space if for each pair x1x_1, x2x_2 of distinct points of XX, there exists neighborhood U1U_1 and U2U_2 of x1x_1, x2x_2, respectively, that are disjoint.

Theorem 17.8 Every finite point set in Hausdorff space XX is closed.

proof. It suffices to show that every one-point {x0}\{x_0\} is closed. if xx of XX is different from x0x_0, then xx and x0x_0 has disjoint neighborhood UU and VV, respectively. since UU does not intersect {x0}\{x_0\}, the point xx cannot belong to the closure of {x0}\{x_0\}. as a result, the closure of the {x0}\{x_0\} is itself, so that it is closed.

The condition that finite point sets be closed is in fact weaker than Hausdorff condition. for example,the real line R\mathbb R in the finite complement topology is not a Hausdorff space, but it is a space in which finite sets are closed.
the condition that finite sets be closed has been given a name of tis own: its called T1T_1 axiom.

Theorem 17.9 Let XX be a space satisfying the T1T_1 axiom; Let AA be a subset of XX. then the point xx is a limit point of AA. if and only if every neighborhood of xx contains infinitely many points of AA.

proof. If every neighborhood of xx intersects AA in infinitely many points, it intersects AA in some points other than xx itself so that xx is a limit point of AA.
conversely, suppose that xx is a limit point of AA, and suppose some neighborhood UU of x intersects AA in only finitely many points, then UU also intersects A{x}A-\{x\} in finitely points; let {x1,,xn}\{x_1,\, \ldots,\, x_n \} be the points of U(A{x})U \cap (A-\{x\}). the set X{x1,,xn}X-\{x_1,\, \ldots,\, x_n \} is open, since the finite point set {x1,,xn}\{x_1,\, \ldots,\, x_n \} is closed.

U(X{x1,,xm})U \cap (X-\{x_1,\,\ldots,\, x_m\})

is a neighborhood of xx that intersects AA. thist contradicts assumption that xx is a limit point of AA.

Theorem 17.10 If XX is a hausdorff space, then a sequence of points of XX converges to at most one point of XX.

proof. Suppose that xnx_n is a sequence of points that converges to xx. if yxy \neq x, let UU and VV be disjoint neighborhood of xx and yy, respectively. since UU contains xnx_n for all but finitley many values of n, the set VV cannot. therefore, xnx_n cannot converge to yy.

Definition If the sequence xnx_n of points of the Hausdorff space XX converges to the point xx of XX, we often write xnxx_n \to x, and we say that is the limit of the sequence xnx_n.

Theorem 17.11 (1)Every simply ordered set is a Hausdorff space in the order topology. (2)the product of two Hausdorff space is a Hausdorff space. (3)A subspace of a Hausdorff space is a Hausdorff space.

proof. (1) If x1<x2x_1 < x_2, x1x2x_1\, x_2 is points of the ordered space XX. (,x1),(x2,)(-\infty,\,x1),\,(x_2,\,\infty) is elements of the ordered topology, that are disjoint.
(2) Let XX and YY be a Hausdorff spaces. there exists U1U_1 and U2U_2 that contains x1x_1 and x2x_2, respectively. which are disjoint the the topology of XX. similiary there exists V1V_1 and V2V_2 in the space YY. x1×y1x_1 \times y_1, x2×y2x_2 \times y_2 belong to U1×V1U_1\times V_1 and U2×V2U_2 \times V_2, respectively. U1×V1U_1\times V_1 and U2×V2U_2 \times V_2 are disjoint.
(3) let AA be a subset of the Hausdorff space XX. there exists U1U_1 and U2U_2, which are open and disjoint, that contains x1x_1 and x2x_2, respectively. x1x_1 is different than x2x_2. and AA contains x1x_1 and x2x_2. then AU1A\cap U_1 and AU2A \cap U_2 is open in the subspace topology of AA, that are disjoint.