Consider $z_n=(\frac{1}{\sqrt{2}}+i\frac{1}{\sqrt{2}})^n$.
Prove that $z_n$ does not converge.
My approach: it is easy to see that $z_n = e^{i\cdot n \frac{\pi}{4}}$. So $z_n$ is the eighth root of unity and is just rotating around the unit circle at angles of $\pi /4$. But I have to prove this using the $\epsilon$ definition of convergence of sequences.
I attack this problem the same way we approach proving that the harmonic sequence $(-1)^n$ does not converge.
I say that $z_n$ converges to some complex number $L$. Therefore, by the definition of convergence of sequences, $|z_n-L|<\epsilon$ for all $n>N$.
Let $\epsilon =1/2$.
Therefore, $|z_n-L|<1/2$, for $n>N$.
Let $n,m>N$ and $n=8k,m=8k+4$. This implies:
$|1-L|<1/2$ and $|1+L|<1/2$.
Now, is it okay for me to conclude that since the shortest distance between $1+0i$ and $-1+0i$ is $2$, no two discs with radius of at most $1/2$ that contain $1+0i$ and $-1+0i$ will ever intersect.
I think the final statement is kind of incomplete. How should I end this proof?
Because $z_{4n}=-1$ and $z_{8n}=1$
Let $\lim\limits_{n\rightarrow+\infty}z_n=L$.
Thus, by your work $$1>|1-L|+|1+L|\geq|1-L+1+L|=2,$$ which is contradiction.
Id est, our assuming was wrong.