Clopen Sets

Sets that are both open and closed — clopen. In R\mathbb{R} only ∅\varnothing and R\mathbb{R} qualify (proved in full), but change the ambient space and the same set's fate changes: [0,1][0,1] is merely closed in R\mathbb{R}, yet clopen inside [0,1]∪[2,3][0,1]\cup[2,3].

A clopen in X  ⟺  A open in X and closed in XA \text{ clopen in } X \iff A \text{ open in } X \text{ and closed in } X
[0,1]=([0,1]∪[2,3])∩(−0.5,1.5)=([0,1]∪[2,3])∩[−0.5,1.5][0,1] = \bigl([0,1]\cup[2,3]\bigr) \cap (-0.5, 1.5) = \bigl([0,1]\cup[2,3]\bigr) \cap [-0.5, 1.5]

Examples & Exercises

Let A⊆RA \subseteq \mathbb{R} be both open and closed (clopen), with ∅≠A≠R\emptyset \ne A \ne \mathbb{R}. Derive a contradiction, proving completely that the only clopen subsets of R\mathbb{R} are ∅\emptyset and R\mathbb{R} itself.

clopen in R  ⟹  A=∅ or A=R\text{clopen in } \mathbb{R} \implies A = \emptyset \text{ or } A = \mathbb{R}
X=RX = \mathbb{R}, A=(0,1)A = (0,1): open but not closed — no nontrivial clopen set exists in R\mathbb{R}.
Show proof
  1. If AA is clopen then so is its complement AcA^c: AA open makes AcA^c closed, and AA closed makes AcA^c open. Since A≠RA \ne \mathbb{R} we have Ac≠∅A^c \ne \emptyset, and since A≠∅A \ne \emptyset, AA itself is nonempty.
  2. Hence R=A∪Ac\mathbb{R} = A \cup A^c is a decomposition into two disjoint nonempty open sets — a separation of R\mathbb{R}. Now we show R\mathbb{R} is connected. Suppose R=U∪V\mathbb{R} = U \cup V with U,VU, V disjoint nonempty open; pick a∈Ua \in U, b∈Vb \in V (renaming so a<ba < b).
  3. Put c=sup⁡(U∩[a,b])c = \sup(U \cap [a,b]) (nonempty, containing aa; bounded above by bb). UU is not only open but also closed, since U=R∖VU = \mathbb{R} \setminus V. Now cc lies in the closure of U∩[a,b]U \cap [a,b], so c∈U‾=Uc \in \overline{U} = U. We have c≤bc \le b, and c≠bc \ne b (as b∈Vb \in V and the sets are disjoint), so c<bc < b. Since UU is open, (c−ε,c+ε)⊆U(c-\varepsilon, c+\varepsilon) \subseteq U for some ε>0\varepsilon > 0; but then c+ε/2∈[a,b]c + \varepsilon/2 \in [a,b] exceeds cc while belonging to U∩[a,b]U \cap [a,b] — contradicting c=sup⁡c = \sup. Hence no separation of R\mathbb{R} exists.
  4. This contradicts the separation from the second step, so the assumed nontrivial clopen set AA cannot exist: the only clopen subsets of R\mathbb{R} are ∅\emptyset and R\mathbb{R}. The scene's A=(0,1)A = (0,1) is open but not closed, dodging this theorem — closing it would require adding the missing limit points 0,10, 1, which would break openness.

For the subspace X=[0,1]∪[2,3]⊆RX = [0,1] \cup [2,3] \subseteq \mathbb{R} and A=[0,1]A = [0,1], prove completely that AA is clopen in XX but not open in R\mathbb{R}.

[0,1]=X∩(−1/2, 3/2)=X∩[−1/2, 3/2][0,1] = X \cap (-1/2,\, 3/2) = X \cap [-1/2,\, 3/2]
x=1x = 1, ε=0.5\varepsilon = 0.5: even at the boundary point 11, (1−0.5,1+0.5)∩X=(0.5,1]⊆A(1-0.5, 1+0.5) \cap X = (0.5, 1] \subseteq A, so the ball stays in AA relative to XX — openness in XX made visible.
Show proof
  1. Openness in XX: check [0,1]=X∩(−1/2, 3/2)[0,1] = X \cap (-1/2,\, 3/2). The inclusion ⊆\subseteq is clear; conversely a point of X∩(−1/2, 3/2)X \cap (-1/2,\, 3/2) lies in [0,1][0,1] or in [2,3]∩(−1/2, 3/2)[2,3] \cap (-1/2,\, 3/2), and the latter is empty, so it lies in [0,1][0,1]. Since (−1/2, 3/2)(-1/2,\, 3/2) is open in R\mathbb{R}, [0,1][0,1] is open in XX.
  2. Closedness in XX: the same argument gives [0,1]=X∩[−1/2, 3/2][0,1] = X \cap [-1/2,\, 3/2] (since [2,3]∩[−1/2, 3/2]=∅[2,3] \cap [-1/2,\, 3/2] = \emptyset). As [−1/2, 3/2][-1/2,\, 3/2] is closed in R\mathbb{R}, [0,1][0,1] is closed in XX. Hence [0,1][0,1] is clopen in XX.
  3. Not open in R\mathbb{R}: 1∈[0,1]1 \in [0,1], yet every ball (1−ε,1+ε)(1-\varepsilon, 1+\varepsilon) contains 1+ε/2∉[0,1]1 + \varepsilon/2 \notin [0,1]. So the same set [0,1][0,1] is clopen in XX while in R\mathbb{R} it is merely closed, not open — clopenness is relative to the ambient space.

Let XX be a topological space and A⊆XA \subseteq X clopen. Prove completely that if ∅≠A≠X\emptyset \ne A \ne X then XX is disconnected.

X=A∪Ac,A∩Ac=∅,A, Ac≠∅ open in XX = A \cup A^c, \qquad A \cap A^c = \emptyset, \qquad A,\, A^c \ne \emptyset \text{ open in } X
X=(0,1)∪(1,2)X = (0,1) \cup (1,2), A=(0,1)A = (0,1): X=A∪AcX = A \cup A^c is a separation into disjoint nonempty opens — showing XX is disconnected.
Show proof
  1. If AA is clopen then so is its complement Ac=X∖AA^c = X \setminus A: complementation swaps openness and closedness (AA open makes AcA^c closed; AA closed makes AcA^c open).
  2. Since ∅≠A≠X\emptyset \ne A \ne X, both AA and AcA^c are nonempty. Moreover X=A∪AcX = A \cup A^c, A∩Ac=∅A \cap A^c = \emptyset, and both are open in XX — a separation of XX. Hence XX is disconnected.
  3. The converse holds too: if X=U∪VX = U \cup V is a separation then U=X∖VU = X \setminus V is also closed, hence clopen. So "a nontrivial clopen set exists" is equivalent to "disconnected". Problem 1 above is the special case X=RX = \mathbb{R} — since R\mathbb{R} is connected, no nontrivial clopen set can exist there.

4. Topology