6
$\begingroup$

I am trying to prove that $ \lim_{n\rightarrow \infty}\left( \sqrt[n]{n}-1 \right)^{n}=0 $. My attempt is as follows.

Since for each $n\in \mathbb{N}$, $n>0$ by arithmetic geometric inequality $$ \sqrt[n]{n}=\sqrt[n]{n.1...1}=\le \dfrac{n+1+...+1}{n}=\dfrac{n+n-1}{n}=2-\dfrac{1}{n} $$.

Hence $$0<\sqrt[n]{n}-1\le 1-\dfrac{1}{n}<1.$$

So $ \lim_{n\rightarrow \infty}\left( \sqrt[n]{n}-1 \right)^{n}=0 $.

Is this correct ? If not how to show that $ \lim_{n\rightarrow \infty}\left( \sqrt[n]{n}-1 \right)^{n}=0 $ ? Please help me. Thanks.

$\endgroup$
  • $\begingroup$ For every $n\geqslant1$, $n<(3/2)^n$ hence $1\leqslant\sqrt[n]{n}<3/2$ hence the $n$th term of the sequence is between $0$ and $1/2^n$. (Note also that $\sqrt[n]{n}-1<1$ does not imply $(\sqrt[n]{n}-1)^n\to0$, for example $(1-1/n)^n\to1/e\ne0$.) $\endgroup$ – Did Jun 12 '15 at 17:18
  • $\begingroup$ @Did please could you help me to prove $ \lim_{n\rightarrow \infty}\left( \sqrt[n]{n}-1 \right)^{n}=0 $? $\endgroup$ – Aven Jun 12 '15 at 17:23
  • $\begingroup$ @Did argued that $(n^{1/n} - 1)^n \leq (3/2 - 1)^n \leq (1/2)^n$, which approaches 0 as $n$ goes to infinity. (This is a correct argument, while your original attempt was not.) More strongly, one may note that $log_2(n^{1/n}) = (1/n)\log_2 n$, which approaches 0, so that $n^{1/n} \rightarrow 1$. This also implies what you want to show. $\endgroup$ – Andy Drucker Jun 12 '15 at 17:27
  • $\begingroup$ @Aven "please could you help me to prove..." I just did. $\endgroup$ – Did Jun 12 '15 at 17:30
  • $\begingroup$ See also: Evaluating the limit $\lim_{n\to+\infty}(\sqrt[n]{n}-1)^n$. $\endgroup$ – Martin Sleziak Jan 12 '20 at 19:05
7
$\begingroup$

The inequality $$ \sqrt[n]{n} \leq \frac{1+\ldots+1+n}{n}= 2-\frac{1}{n}$$ just gives: $$ \limsup_{n\to +\infty}\left(\sqrt[n]{n}-1\right)^n \leq\frac{1}{e} $$ but if you consider the slightly improved inequality: $$ \sqrt[n]{n}\leq\frac{1+\ldots+1+\sqrt{n}+\sqrt{n}}{n}=1-\frac{2}{n}+\frac{2}{\sqrt{n}}$$ you easily get: $$ \lim_{n\to +\infty}\left(\sqrt[n]{n}-1\right)^n = 0 $$ as wanted.


For a sharp inequality, you need to write $n$ as a product of $n$ numbers, quite close one each other: that is not difficult to achieve. Since: $$ n=1\cdot\prod_{k=1}^{n-1}\frac{k+1}{k} $$ by the AM-GM inequality it follows that: $$ \sqrt[n]{n}\leq 1+\frac{H_{n-1}}{n} $$ that is essentially optimal since $e^x\geq 1+x$ implies: $$ \sqrt[n]{n} \geq 1+\frac{\log n}{n}.$$

$\endgroup$
  • $\begingroup$ Thanks for this incredible explanation! $ \lim_{n\rightarrow \infty}\left( \sqrt[n]{n}-1 \right)^{n}=0 $ because of the rule: "the limit of the products is the product of the limits" which in this case would be $0*0*0*...*0$ n times? $\endgroup$ – GniruT Nov 20 '15 at 12:08
2
$\begingroup$

Not exactly rigorous, but you can make it rigorous without too much effort, I think: It suffices to show that

$$ \lim_{n \to \infty} \sqrt[n]{n} < 2 $$

But this is trivial, since $2^n \gg n$. Hence the limit follows. (In fact, $\lim_{n \to \infty} \sqrt[n]{n} = 1$.)

ETA: If the $2$ makes you nervous, you can substitute any value $q, 1 < q < 2$. @Did chose $q = 3/2$ in the comments.

$\endgroup$
1
$\begingroup$

We can also approach this using brute force via application of L'Hospital's Rule.

$$\lim_{n\to \infty}(n^{1/n}-1)^n=\exp{\lim_{n\to \infty}n\log(n^{1/n}-1)}$$

Now, let's look at the interior limit.

$$\lim_{n\to \infty}n\log(n^{1/n}-1)=\lim_{n\to \infty}\frac{\log(n^{1/n}-1)}{n^{-1}}=-\lim_{n\to \infty}n^2\frac{1-n^{-2}\log n}{n^{1/n}-1}=-\infty$$

Inasmuch as $\lim_{x\to -\infty}e^x=0$, we obtain the expected result.

$\endgroup$

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.