So I know that $\sum_{n\in\mathbb{N}}1/n$ diverges and $\sum_{n\in\mathbb{N}}1/n^2$ converges. What about the series $\sum_{n=2}^\infty1/n(\log(n))$? I'm pretty confident that it diverges but is there a quick justification?
-
Another one: https://math.stackexchange.com/q/2223275/42969. – Martin R Nov 20 '18 at 20:53
3 Answers
Diverges by Integral test, since $$\displaystyle \int_2^\infty \dfrac{dt}{t \ln(t)} = \int_2^\infty \dfrac{d}{dt} \ln(\ln(t))$$ diverges.

- 448,999
HINT
By Cauchy condensation test we can consider the condensed series
$$\sum_{n=2}^\infty \frac{1}{n\log(n)} \rightarrow\sum_{n=2}^\infty \frac{2^n}{2^n\log(2^n)}$$
and the first converges if and only if the second one converges.
By the same test we can also determine the convergence for the more general case
$$\sum_{n=2}^\infty \frac{1}{n^a\log^b(n)}$$
known as Bertrand Series.

- 154,566
I'll see if I can come up with a discrete version of the integral test.
$\begin{array}\\ \ln(\ln(t+1))-\ln(\ln(t)) &=\ln\left(\dfrac{\ln(t+1)}{\ln(t))}\right)\\ &=\ln\left(\dfrac{\ln(t)+(\ln(t+1)-\ln(t))}{\ln(t))}\right)\\ &=\ln\left(1+\dfrac{\ln(t+1)-\ln(t)}{\ln(t))}\right)\\ &=\ln\left(1+\dfrac{\ln(\frac{t+1}{t})}{\ln(t))}\right)\\ &=\ln\left(1+\dfrac{\ln(1+\frac{1}{t})}{\ln(t))}\right)\\ &<\ln\left(1+\dfrac{\frac{1}{t}}{\ln(t))}\right) \qquad\text{since } \ln(1+x) < x \text{ for } 0 < x\\ &=\ln\left(1+\dfrac{1}{t\ln(t))}\right)\\ &<\dfrac{1}{t\ln(t))}\\ \end{array} $
so
$\begin{array}\\ \sum_{k=2}^n \dfrac{1}{k\ln(k))} &>\sum_{k=2}^n (\ln(\ln(k+1))-\ln(\ln(k)))\\ &=\ln(\ln(n+1))-\ln(\ln(2)))\\ &\to \infty\\ \end{array} $
Yep - that works.
By iterating this, we can show that $\sum_n \dfrac1{n\ln(n)\ln(\ln(n))\ln(\ln(\ln(n)))...} $ diverges.
All that is needed is $\ln(1+x) < x$.

- 107,799
-
1Similar approach here: https://math.stackexchange.com/a/2223354/42969. – Martin R Nov 20 '18 at 20:58
-