# What is $\lim\limits_{z \to 0} |z \cdot \sin(\frac{1}{z})|$ for $z \in \mathbb C$?

What is $\lim\limits_{z \to 0} |z \cdot \sin(\frac{1}{z})|$ for $z \in \mathbb C^*$? I need it to determine the type of the singularity at $z = 0$.

-
What kind of singularity does $\sin$ have at $\infty$? – Antonio Vargas Jan 18 '13 at 17:28
It is easy to compute the Laurent series of your function at $0$, so that is how to tell what kind of singularity it has there. – GEdgar Jan 18 '13 at 17:39
@GEdgar That's what I did and found that the singularity at 0 must be essential. But when I tried to evaluate to this limit, I found that it had a result $< \infty$, when it should be undefined. – StringerBell Jan 18 '13 at 19:20
@AntonioVargas $\lim\limits_{z \to \infty} \sin(z)$ converges, but I thought since I'm multiplying with $z$ as $z$ approaches $0$, it wouldn't matter (which - judging from the answers below - was wrong; this counts for $x \in \mathbb R$ but not in $\mathbb C$). – StringerBell Jan 18 '13 at 19:24

We have $a_n=\frac{1}{n} \to 0$ as $n\to \infty$ and $$\lim_{n \to \infty}\Big|a_n\sin\Big(\frac{1}{a_n}\Big)\Big|=\lim_{n \to \infty}\frac{|\sin n|}{n}=0,$$ but $$\lim_{n \to \infty}\Big|ia_n\sin\Big(\frac{1}{ia_n}\Big)\Big|=\lim_{n \to \infty}\frac{e^n-e^{-n}}{2n}=\infty.$$ Therefore $\lim_{z \to 0}|z\sin(z^{-1})|$ does not exist.

Notice for every $k \in \mathbb{N}$ the limit $\lim_{z\to 0}|z\sin(z^{-1})|$ does not exist. In fact $$\lim_{n \to \infty}\Big|a_n^{k+1}\sin\Big(\frac{1}{a_n}\Big)\Big|=\lim_{n \to \infty}\frac{|\sin n|}{n^{k+1}}=0,$$ but

$$\lim_{n \to \infty}\Big|(ia_n)^{k+1}\sin\Big(\frac{1}{ia_n}\Big)\Big|=\lim_{n \to \infty}\frac{e^n-e^{-n}}{2n^{k+1}}=\infty.$$ It follows that $z=0$ is neither a pole nor a removable singularity. Hence $z=0$ is an essential singularity.

-
Can I apply the same arguments to $\lim\limits_{z \to \pi/2} \tan^2(z)$ for $z \in \mathbb C$? This limit would also be undefined. – StringerBell Jan 18 '13 at 21:11
Yes, and you'll find that the limit $\lim_{z \to \pi/2}(z-\pi/2)^k\tan^2x$ exists for $k=4$, i.e. $z=\pi/2$ is a pole of order $4$ for $\tan^2$. – Mercy King Jan 18 '13 at 21:34
This, I don't understand - but I have opened a new question for that (in case your interested: math.stackexchange.com/questions/281647/…) thanks already! – StringerBell Jan 18 '13 at 21:58
sorry, the order is $2$, not $4$! – Mercy King Jan 18 '13 at 22:33

It is undefined. Consider the function $g(z)=\frac{\sin{z}}{z}$. This can be made an entire function by setting $g(0)=0$. But then you are looking for $\lim_{z\to\infty} g(z)$. Any non-polynomial entire function must have an essential singularity at $\infty$, so this limit does not exist.

Alternatively:

For $x$ real, $\lim_{x\to\infty} g(x)$ is clearly zero, since $|\sin x|\leq 1$.

But when $z=ix$ is imaginary, then $g(z)=\frac{e^x-e^{-x}}{2x}$, which has infinite limit as $x\to\infty$.

-
Why must any non-constant entire function have an essential singularity at $\infty$? Because the "value" $\infty$ is technically not "allowed" for functions? – StringerBell Jan 18 '13 at 19:15
Hmmm, $f(z)=z$ is a non-constant entire function and it clearly does not have an essential singularity at infinity, does it? – Nils Matthes Jan 18 '13 at 19:26
@NilsMatthes Sorry, you are correct, it only has an essential singularity of the function is not a polynomial. – Thomas Andrews Jan 18 '13 at 19:31
@StringerBell I was slightly wrong, it is only true for non-polynomial entire functions that the singularity is essential. But check the Wikipedia page for Entire function: en.wikipedia.org/wiki/Entire_function – Thomas Andrews Jan 18 '13 at 19:33
@ThomasAndrews: I think you mean more generally rational function, not polynomial. – Nils Matthes Jan 18 '13 at 19:33