This proof takes place in the complex space, and $C$ is a set of complex numbers. Also, let $\rightarrow$ denote the $\varepsilon$ definition of convergence. I have some experience with real analysis but this is my first attempt at complex analysis, I figure the proofs should be similar but I am not sure my proof is formal enough any suggestions or corrections are greatly appreciated! Also, I haven't learned any theorems so this will be a barebones proof using only definitions of open, closed and convergence.

Show that $C$ is closed if and only if $\{z_n\} \in C$ and $z_n \rightarrow z$ implies $z$ is in $C$.

I will prove the claim by contradiction.

Suppose $C$ is closed, consider a sequence $\{z_n\} \in C$ such that $z_n \rightarrow z$ for $z \notin C$. Thus, $z \in C^c.$ Since $z_n \rightarrow z$ then for every $\varepsilon$ there is an integer $N$ so that $|z_n - z| < \varepsilon$ for all $n \geq N$ such that all $\{z_n\} \in C$. Thus any disk of radius $\varepsilon$ centered around $z$ will contain some elements of $z_n$. So $z \in C^c$ and there is no $\varepsilon$ such that $D(z, \varepsilon) \subset C^c$, however, $C^c$ is open by definition. Thus, we have arrived at a contradiction.

Conversely, suppose $z_n \rightarrow z$ and $\{z_n\} \in C$ implies that $z \in C$.

Let $z_1 \in C^c$, I claim that there exists some $r > 0$ so that $D(z_1, r) \subset C^c$. Suppose not, then for any choice of $r$ $D(z_1, r) \cap C \neq \varnothing.$ In particular, consider $D(z_1, \frac{1}{n})$ for $n = 1, 2, 3 \dots$. Suppose $a_n \in D(z_1, \frac{1}{n}) \cap C$, then $a_n \in C$ and $|a_n - z_1| < \frac{1}{n}$, therefore $a_n \rightarrow z_1$ as $n \rightarrow \infty$. Then by the hypothesis $z_1 \in C$, however this is a contradiction. Therefore, there exists some $r > 0$ so that $D(r, z_1) \subset C^c$ for arbitrary $z_1 \in C^c$ and thus $C^c$ is open and $C$ is closed.

  • 3
    $\begingroup$ The proof is correct but the converse implication is missing. The if is an iff here. $\endgroup$ – Gribouillis Sep 2 '17 at 20:16
  • $\begingroup$ Yes haha I am still working on that side figured I'd see if I got this direction down first. Also, in the text the $C$ is meant to denote a set of complex numbers! $\endgroup$ – student_t Sep 2 '17 at 20:17
  • $\begingroup$ @Gribouillis Added the converse! $\endgroup$ – student_t Sep 3 '17 at 1:46
  • $\begingroup$ It is correct. Your proof works in any metric space, not just ${\mathbb C}$. Metric spaces are sequential spaces. $\endgroup$ – Gribouillis Sep 3 '17 at 5:46

The other direction is straightforward. Suppose $z \notin C$. Then consider the sequence of balls $B(z,{1 \over n})$. If an infinite number of these balls intersect $C$ then we have $z\in C$, hence at most a finite number of these balls intersect $C$. Hence there is some $n$ such that $B(z,{1 \over n}) \cap C = \emptyset$, and so $C^c$ is open.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.