1
$\begingroup$

$$ \large \displaystyle \int_0^\infty {\dfrac{e^{-2x} \tanh\frac{x}{2}}{x \cosh x}dx} = 2 \ln \frac{\pi}{2\sqrt{2}} $$

How to prove the above integral?

What I tried :

$\displaystyle I(s) = \int_0^\infty {\dfrac{e^{-sx} \tanh\frac{x}{2}}{x \cosh x}dx} $

$\displaystyle I'(s) = -\int_0^\infty {\dfrac{e^{-sx} \tanh\frac{x}{2}}{\cosh x}dx} $

$\displaystyle I'(s) = L [sech x](s+1) - L [sech \frac{x}{2}](s+\frac{1}{2}) $

where $L[f](s)$ is the Laplace transform of $f(x)$ as a function of $s$. But after this I don't follow much. Finding the laplace transform and then integrating doesn't seem a better approach.

$\endgroup$
  • 1
    $\begingroup$ In general, one can show that $$I(a)=\int_0^{\infty} \frac{e^{-a x}\tanh\frac{x}{2}}{x\cosh x}dx = 2 \ln\left(\large \frac{ \Gamma(a) \Gamma\left(\frac{a}{4}\right)^2}{2^{\frac{a+1}{2}} \Gamma\left(\frac{a}{2}\right)^3}\right)$$ $\endgroup$ – nospoon May 2 '16 at 18:32
2
$\begingroup$

We want to compute: $$ I =\int_{0}^{+\infty}\frac{1-e^{-x}}{x(e^x+1)(e^{2x}+1)}\,dx =\int_{0}^{+\infty}\left(\frac{1}{e^{2x}+1}-\frac{1}{e^{2x}+e^{x}}\right)\frac{dx}{x}$$ that is the same as computing $$ J = \int_{0}^{1}\frac{t^2(1-t)^2}{(1-t^4)\log t}\,dt.$$ Since $\frac{t^2(1-t)^2}{1-t^4}=\sum_{k\geq 1}\left(t^{4k-2}-2t^{4k-1}+t^{4k}\right)$ and $\int_{0}^{1}\frac{t^\alpha-1}{\log t}\,dt = \log(\alpha+1)$, $$ J = \sum_{k\geq 1}\log\left(\frac{(4k-1)(4k+1)}{(4k)^2}\right)=\log\prod_{k\geq 1}\left(1-\frac{1}{16k^2}\right) $$ where the value of the last infinite product can be recovered from: $$ \frac{\sin z}{z}=\prod_{k\geq 1}\left(1-\frac{z^2}{\pi^2 k^2}\right)$$ by setting $z=\frac{\pi}{4}$.

$\endgroup$
  • $\begingroup$ Is there no "straighter" approach? That substitution seems quite too random. $\endgroup$ – Kartik Sharma May 2 '16 at 14:45
  • 1
    $\begingroup$ To set $x=\log t$ in a integral that strongly depends on $e^x$ does not look so random to me, anyway you may also avoid that step and skip from the first centered formula to the third one by directly applying Frullani's theorem. $\endgroup$ – Jack D'Aurizio May 2 '16 at 14:54
  • $\begingroup$ Yes, now I get it. Thanks! $\endgroup$ – Kartik Sharma May 2 '16 at 15:19

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.