1
$\begingroup$

I have come across many limit questions of the form $0^0$

For instance , here is one example

$$y=\lim_{x \to 0^+} (2\sin(\sqrt x) + \sqrt x\sin{1\over x})^x$$

In this example if I take logarithm on both sides , it becomes $$\ln y = \lim_{x \to 0^+} x\ln (2\sin(\sqrt x) + \sqrt x\sin{1\over x})$$

And then I make the argument that as $x \to 0$ ,$x$ reaches zero more dominantly than $\ln (2\sin(\sqrt x) + \sqrt x\sin{1\over x})$ reaches $-\infty$ . And so $\ln y = 0$ , thus , $y=1$

But I realised tht this the argument I make in almost all such cases , and it got to me thinking are all limits of the form $0^0$ eventually turn out to be 1 ?

So what I am looking for is :

(a) Various other ways of solving the above limit (as many different ways are possible) , but please avoid using L'hopital's rule as it would be too tedious to work with here.

(b) Counter examples , that is , limits of the form $0^0$ which are not equal to $1$

Thanks for your time !

Edit : I got the plenty counter examples , now I am only looking for other ways of solving the above limit

$\endgroup$
  • 2
    $\begingroup$ For any real number $a \neq 0$ we have $(\frac 1 n)^{\ln (1/a) /ln n} \to a$. $\endgroup$ – Kavi Rama Murthy Jul 2 at 6:00
  • 1
    $\begingroup$ No, the indeterminate form does not always come out to be $1$. And that is in particular true when you consider multi variable calculus, where more than not the answer won't be $1$ $\endgroup$ – imranfat Jul 2 at 6:02
2
$\begingroup$

$0^0$ we call "Uncertainty", because if we take different sequences with limits $0$ then we obtain different results: $$\begin{array} {} \left ( \frac{1}{n} \right )^{\frac{1}{n}}=e^{\frac{1}{n} \ln \frac{1}{n}} \to 1 \\ \left ( \frac{1}{n} \right )^{e^{-n^2}} = e^{\frac{1}{n} \ln e^{-n^2}} \to 0 \\ \left ( \frac{1}{n} \right )^{e^{-n}} = e^{\frac{1}{n} \ln e^{-n}} \to e^{-1} \end{array} $$ last line makes clear, that we can obtain any finite number $a>0$. One casuistic example we obtain if allow negative base (somebody forbid complex numbers?) $$\left ( -\frac{1}{n} \right )^{e^{-n^2}} = e^{-\frac{1}{n} \ln e^{-n^2}} \to +\infty$$

and last requirement $$\lim_{x \to 0^+} (2\sin(\sqrt x) + \sqrt x\sin{1\over x})^x = \lim_{x \to 0^+}e^{x \ln (2\sin(\sqrt x) + \sqrt x\sin{1\over x})}$$ now $$x \ln (2\sin(\sqrt x) + \sqrt x\sin{1\over x})=x \ln(2\sin(\sqrt x))+x\ln \left (1+\frac{\sqrt x\sin{1\over x}}{2\sin(\sqrt x)} \right ) \to0$$

| cite | improve this answer | |
$\endgroup$
2
$\begingroup$

$\lim\limits_{x \to 0} \left( 2x \right)^{\frac{3}{5 - \ln(4x)}} = \left[ 0^0 \right] = \lim\limits_{x \to 0} e^{\frac{3 \cdot \ln(2x)}{5 - \ln(4x)}} = \left[ \text{L'Hopital} \right] = \lim\limits_{x \to 0} e^{\frac{\frac{d}{dx}(3 \cdot \ln(2x))}{\frac{d}{dx}(5 - \ln(4x))}} = \lim\limits_{x \to 0} e^{\frac{\frac{3}{x}}{\frac{-1}{x}}} = \boxed{e^{-3}}$

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Yeah , I got the counter example ,thanks , but I am still looking for other ways of solving the above limit (that's why I am not green checking either of the two answers. $\endgroup$ – ARROW Jul 2 at 6:44
1
$\begingroup$

No, the limit of the form 0^0 not always comes out to be 1

Here is a counter example,

https://www.youtube.com/watch?v=89d5f8WUf1Y

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ This was indeed useful , but I am not giving this answer the green check mark because I am waiting for other answers as I am looking for other ways of solving the above limit also $\endgroup$ – ARROW Jul 2 at 6:12
1
$\begingroup$

Here is an approach which takes care of the oscillating term $\sin(1/x)$.

Clearly we have $|\sin(1/x)|\leq 1$ and therefore $$2\sin\sqrt{x}-\sqrt{x}\leq 2\sin\sqrt{x}+\sqrt {x} \sin\frac {1}{x}\leq 2\sin\sqrt{x}+\sqrt{x}$$ Raising each term to power $x$ we get an inequality of the form $g(x) \leq f(x) \leq h(x) $ where $f(x)$ is the function under limit in question. You can now show that $g(x) \to 1,h(x)\to 1$ so that by squeeze theorem $f(x) \to 1$.

We have $$g(x) =(2\sin\sqrt{x}-\sqrt{x})^x=x^{x/2}\left(2\cdot\frac{\sin\sqrt{x}}{\sqrt{x}}-1\right)^x$$ The expression $x^{x/2}$ tends to $1$ (prove this via the standard limit that $x\log x\to 0$) and the expression in large parentheses tends to $1$ so that the second factor also tends to $1$. It follows that $g(x) \to 1$ and in a similar manner $h(x) \to 1$.

| cite | improve this answer | |
$\endgroup$
1
$\begingroup$

Although some authors define $0^0=1$, does not mean that $$ \lim_{(x,y)\to(0,0)}x^y=1\tag1 $$ There are many paths to $(0,0)$ that do not lead to $1$. For example, if $x\gt0$, $$ x^{\frac\lambda{\log(x)}}=e^\lambda\tag2 $$ Taking the limit as $x\to0^+$ obviously gives $e^\lambda$.


However, for $0\le x\le\frac{\pi^2}{16}$, the concavity of $\sin(x)$ gives $\frac{2\sqrt2}\pi\sqrt{x}\le\sin\left(\sqrt{x}\right)$. Therefore, $$ \left(\frac{4\sqrt2}\pi-1\right)\sqrt{x}\le2\sin\left(\sqrt{x}\right)+\sqrt{x}\sin\left(\tfrac1x\right)\le3\sqrt{x}\tag3 $$ Furthermore, for $a\gt0$, $$ \begin{align} \lim_{x\to0^+}\left(a\sqrt{x}\right)^{\,x} &=\lim_{x\to0^+}\left(\sqrt2\,a\right)^x\lim_{x\to0^+}(x/2)^{x/2}\\[3pt] &=1\cdot1\tag4 \end{align} $$ The Squeeze Theorem, with $(3)$ and $(4)$, shows that $$ \lim_{x\to0^+}\left(2\sin\left(\sqrt{x}\right)+\sqrt{x}\sin\left(\tfrac1x\right)\right)^x=1\tag5 $$

| cite | improve this answer | |
$\endgroup$

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.