I have tried to use polar coordinates, which gives me:

$$\frac{\sin(r^2)}{1-\cos r}$$

And from here, I couldn't proceed no matter what I tried and then I tried the following identities In a bit of despair: The first is merely the fundamental limit, the second may be derived from it, I guess.

$$\sin x \approx x \quad \quad \quad \quad \quad \quad \quad \quad \cos x \approx 1-\frac{x^2}{2}$$

Which gives me:


And then:

$$\lim_{r\to0}\frac{2r^2}{r^2}=2 $$

The same result would also follow without the use of polar coordinates:

$$\frac{x^2+y^2}{\frac{x^2+y^2}{2}}= \frac{2(x^2+y^2)}{(x^2+y^2)}$$

And hence:

$$\lim_{(x,y)\to 0} \frac{2(x^2+y^2)}{(x^2+y^2)}=2$$

The result to both limits was confirmed by Wolfram Alpha and I see no other way to do it. Notice that it's highly "suspicious" that there is a square root inside the $\cos$ and there is a square in our identity for $\cos x$.

$$\Large\text{ Questions:}$$

  1. Is it correct? I feel as if I had done some forbidden move in there.
  2. How can I deduce $\cos x \approx 1-\frac{x^2}{2}$? I know it may follow from the fundamental limit, but I have tried a few things a nothing worked.
  • 2
    $\begingroup$ $\cos x=1-x^2/2+O(x^4)$ via the Maclaurin series. if you don't like that, $1-\cos r=2\sin^2(r/2)$. $\endgroup$ – Angina Seng Aug 31 '18 at 16:12
  • $\begingroup$ I think, since the converted fraction into polar coordinates doesn't depend to $\theta$, so the limit doesn't existt. $\endgroup$ – mrs Aug 31 '18 at 16:17
  • 1
    $\begingroup$ The square root is not suspicious, it is the square root that computes the modulus of $(x,y)$. The square (of the square root) under the sine is more artificial. $\endgroup$ – Yves Daoust Aug 31 '18 at 16:26
  • $\begingroup$ @BillyRubina Please recall that if the OP is solved you can evaluate to accept an answer among the given, more details here meta.stackexchange.com/questions/5234/… $\endgroup$ – user Sep 17 '18 at 20:31

We don't need polar coordinates indeed it suffices observe that $t=x^2+y^2 \to 0$ and then the limit becomes

$$\lim_{t\to 0}\frac{\sin(t)}{1-\cos \sqrt{t}}=\lim_{t\to 0}\frac{\sin(t)}{t}\frac{t}{1-\cos \sqrt{t}}=2$$

For the standard limit we can also use l'Hopital to obtain

$$\lim_{t\to 0}\frac{1-\cos t}{t^2}=\lim_{t\to 0}\frac{\sin t}{2t}=\frac12$$

or as an alternative

$$\frac{1-\cos t}{t^2}=\frac{1-(1-2\sin^2(t/2))}{t^2}=\frac12\frac{\sin^2(t/2))}{(t/2)^2}\to \frac12$$

Your first way by polar coordinates is not a wrong idea and you can conlude by standard limits as shown here. Note that if we want use series expansion we need to proceed as follow using little-o notation for the remainder (or also big-O notation if you prefer)

$$\frac{\sin(r^2)}{1-\cos r}=\frac{r^2+o(r^2)}{1-\left(1-\frac12r^2+o(r^2)\right)}=\frac{r^2+o(r^2)}{\frac12r^2+o(r^2)}=\frac{1+o(1)}{\frac12+o(1)}\to 2$$


I would write $$\frac{\sin(r^2)(1+\cos(r))}{(1-\cos(r))(1+\cos(r))}$$


The bidimensional limit

$$\lim_{(x,y)\to0}f\left(\sqrt{x^2+y^2}\right)$$ can very well be replaced by a unidimensional one



$$\lim_{r\to0}\frac{\sin r^2}{1-\cos r}=\lim_{r\to0}\frac{\sin r^2}{2\sin^2\dfrac r2}=2\lim_{r\to0}\frac{\sin r^2}{r^2}\left(\frac{\dfrac r2}{\sin\dfrac r2}\right)^2=2.$$

  • 1
    $\begingroup$ No Taylor, no L'Hospital, no Wolfram. $\endgroup$ – Yves Daoust Aug 31 '18 at 16:35

$\dfrac{\sin r^2}{1-\cos r} = \dfrac{\sin r^2}{r^2} \dfrac{r^2}{1-\cos r}$

Using L'Hospital's rule, $\dfrac{r^2}{1-\cos r} \to \dfrac{2r}{\sin r} \to 2$ as $r \to 0$.

So $\dfrac{\sin r^2}{1-\cos r} \to 2$ as $r \to 0$.


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.